Geo-Environment and Landscape Evolution II

Geo-Environment andLandscape Evolution II

Evolution, Monitoring, Simulation, Management andRemediation of the Geological Environment and

Landscape

WIT Press publishes leading books in Science and Technology.Visit our website for the current list of titles.

www.witpress.com

WITeLibraryHome of the Transactions of the Wessex Institute.

Papers presented at Geo-Environment and Landscape Evolution 2006 are archived in theWIT elibrary in volume 89 of WIT Transactions on Ecology and the Environment

(ISSN 1743-3541).The WIT electronic-library provides the international scientific community with

immediate and permanent access to individual papers presented at WIT conferences.http://library.witpress.com

Transactions Editor

Carlos Brebbia Wessex Institute of Technology

Ashurst Lodge, Ashurst Southampton SO40 7AA, UK Email: [email protected]

WIT Transactions on Ecology and the Environment

Editorial Board

Y N Abousleiman University of Oklahoma USA

A Aldama IMTA Mexico

D Almorza Gomar University of Cadiz Spain

A M Amer Cairo University Egypt

M Andretta Montecatini Italy

J M Baldasano Universitat Politecnica de Catalunya Spain

J G Bartzis Institute of Nuclear Technology Greece

A Bejan Duke University USA

J Boarder Cartref Consulting Systems UK

B Bobee Institut National de la Recherche Scientifique Canada

H Boileau ESIGEC France

C A Borrego University of Aveiro Portugal

A H-D Cheng University of Mississippi USA

C-L Chiu University of Pittsburgh USA

A Cieslak Technical University of Lodz Poland

W Czyczula Krakow University of Technology Poland

M da Conceicao Cunha University of Coimbra Portugal

M Davis Temple University USA

A B de Almeida Instituto Superior Tecnico Portugal

K Dorow Pacific Northwest National Laboratory USA

C Dowlen South Bank University UK

R Duffell University of Hertfordshire UK

J P du Plessis University of Stellenbosch South Africa

A Ebel University of Cologne Germany

D Elms University of Canterbury New Zealand

D M Elsom Oxford Brookes University UK

D Emmanouloudis Technological Educational Institute of Kavala Greece

J W Everett Rowan University USA

R A Falconer Cardiff University UK

D M Fraser University of Cape Town South Africa

G Gambolati Universita di Padova Italy

N Georgantzis Universitat Jaume I Spain

F Gomez Universidad Politecnica de Valencia Spain

K G Goulias Pennsylvania State University USA

W E Grant Texas A & M University USA

C Hanke Danish Technical University Denmark

A H Hendrickx Free University of Brussels Belgium

S Heslop University of Bristol UK

I Hideaki Nagoya University Japan

W F Huebner Southwest Research Institute USA

W Hutchinson Edith Cowan University Australia

D Kaliampakos National Technical University of Athens Greece

K L Katsifarakis Aristotle University of Thessaloniki Greece

H Kawashima The University of Tokyo Japan

B A Kazimee Washington State University USA

D Kirkland Nicholas Grimshaw & Partners Ltd UK

D Koga Saga University Japan

J G Kretzschmar VITO Belgium

B S Larsen Technical University of Denmark Denmark

A Lebedev Moscow State University Russia

D Lewis Mississippi State University USA

K-C Lin University of New Brunswick Canada

J W S Longhurst University of the West of England UK

T Lyons Murdoch University Australia

U Mander University of Tartu Estonia

N Marchettini University of Siena Italy

J D M Marsh Griffith University Australia

J F Martin-Duque Universidad Complutense Spain

K McManis University of New Orleans USA

C A Mitchell The University of Sydney Australia

M B Neace Mercer University USA

R Olsen Camp Dresser & McKee Inc. USA

R O'Neill Oak Ridge National Laboratory USA

K Onishi Ibaraki University Japan

J Park Seoul National University Korea

G Passerini Universita delle Marche Italy

B C Patten University of Georgia USA

M F Platzer Naval Postgraduate School USA

V Popov Wessex Institute of Technology UK

H Power University of Nottingham UK

M R I Purvis University of Portsmouth UK

Y A Pykh Russian Academy of Sciences Russia

A D Rey McGill University Canada

A C Rodrigues Universidade Nova de Lisboa Portugal

R Rosset Laboratoire d'Aerologie France

J L Rubio Centro de Investigaciones sobre Desertificacion Spain

S G Saad American University in Cairo Egypt

R San Jose Technical University of Madrid Spain

J J Sharp Memorial University of Newfoundland Canada

H Sozer Illinois Institute of Technology USA

I V Stangeeva St Petersburg University Russia

E Tiezzi University of Siena Italy

T Tirabassi Institute FISBAT-CNR Italy

S G Tushinski Moscow State University Russia

J-L Uso Universitat Jaume I Spain

R van Duin Delft University of Technology Netherlands

A Viguri Universitat Jaume I Spain

Y Villacampa Esteve Universidad de Alicante Spain

G Walters University of Exeter UK

SECOND INTERNATIONAL CONFERENCE ON EVOLUTION,MONITORING, SIMULATION, MANAGEMENT AND

REMEDIATION OF THE GEOLOGICAL ENVIRONMENT ANDLANDSCAPE

Geo-Environment & Landscape Evolution 2006

J. F. Martín-DuqueComplutense University, Spain

C. A. BrebbiaWessex Institute of Technology, UK

D. E. EmmanouloudisTechnological Educational Institute of Kavala, Greece

U. ManderUniversity of Tartu, Estonia

CONFERENCE CHAIRMEN

INTERNATIONAL SCIENTIFIC ADVISORY COMMITTEE

Organised byWessex Institute of Technology, UK

Complutense University, Spain

Sponsored byMunicipality of Rhodes, Greece

Technological Education Institute of Kavala, GreeceWIT Transactions on Ecology and The Environment

A. CecioniM. E. ContiW. S. Fyfe

A. E. GodfreyG. Lorenzini

C. MillerM. Noormets

J. Pedraza

This page intentionally left blank

Geo-Environment andLandscape Evolution II

Editors



D. E. EmmanouloudisTechnological Educational Institute of Kavala, Greece


Published by

WIT PressAshurst Lodge, Ashurst, Southampton, SO40 7AA, UKTel: 44 (0) 238 029 3223; Fax: 44 (0) 238 029 2853E-Mail: [email protected]://www.witpress.com

For USA, Canada and Mexico

Computational Mechanics Inc25 Bridge Street, Billerica, MA 01821, USATel: 978 667 5841; Fax: 978 667 7582E-Mail: [email protected]://www.witpress.com

British Library Cataloguing-in-Publication Data

A Catalogue record for this book is availablefrom the British Library

ISBN: 1-84564-168-XISSN: 1746-448X (print)ISSN: 1743-3541 (on-line)

The texts of the papers in this volume were set individually by the authors or under their supervision. Only minor corrections to the text may have been carried out by the publisher.

No responsibility is assumed by the Publisher, the Editors and Authors for any injuryand/or damage to persons or property as a matter of products liability, negligence orotherwise, or from any use or operation of any methods, products, instructions or ideascontained in the material herein.

© WIT Press 2006

Printed in Great Britain by Cambridge Printing.

All rights reserved. No part of this publication may be reproduced, stored in a retrievalsystem, or transmitted in any form or by any means, electronic, mechanical, photocopying,recording, or otherwise, without the prior written permission of the Publisher.



D. E. EmmanouloudisTechnological Educational Institute of Kavala,Greece


Preface

This book contains papers presented at the Second International Conference onMonitoring, Simulation, Management, and Remediation of the GeologicalEnvironment and Landscape, called Geo-Environment and Landscape Evolution,held in Rhodes, Greece, in June of 2006.

The success of the first Geo-Environment conference held in Segovia, Spain(2004), prompted the organizers to reconvene a new meeting in Rhodes in 2006. Thebasic reason for this second Geo-Environment Conference is the perception thatthe information that all the geosciences can bring to environmental management isstill being underutilized by politicians and developers. We strongly believe that athrough understanding of all the elements of natural systems and processes isnecessary for a proper use of the land. To get a broader perspective on this issue,other multidisciplinary studies in the fields of geoecology, earth surface processes,and landscape ecology are also included by the addition of ‘landscape evolution’ tothe Conference name.

In this regard, we believe that the group of geosciences is the most complete ofthe physical sciences. This is because to be able to practice them properly, onemust have a mastery of mathematics, physics, chemistry, and biology. In addition,they require an understanding of time, measured in terms of thousands and millionsof years. It is the tools provided by these areas of knowledge that a geoscientistbrings together to study the processes, rates and durations that form the landscapesand conditions that present opportunities and hazards to society. From a pragmaticpoint of view, environmental geo-science is simply basic earth sciences applied to aprocess or condition that society perceives important. On the other hand, therehave been innumerable geologic and landscape studies that started out as intellectualdelights, but later turned into valuable information for humanity’s use of the land ina sustainable manner.

Based on these principles, this book contains important contributions bygeologists, geographers, soil scientists, ecologists, engineers, planners, and otherprofessionals interested in the contribution of geo-information to sustainabledevelopment or suitable use of the land.

The editors of this book are grateful to all the authors for their outstandingcontributions. We especially would like to thank all the distinguished scientistswho supported this effort by serving on the International Scientific AdvisoryCommittee, by promoting the conference and reviewing the submitted abstracts

and papers. We thank the Complutense University of Madrid, Spain, for supportingthe conference, through the collaboration of its Faculty of Geology. We are alsoindebted to the Municipality of Rhodes and the Technical Institute of Kavala fortheir support as well as the University of Tartu in Estonia. The Wessex Institute ofTechnology and WIT Press coordinated all the work, organizing the meeting andproducing this book to their usual high standards.

The EditorsRhodes, 2006

Contents

Keynote address Towards a truly sustainable future and a high quality of life for all people on our planet: the role of the Earth Sciences W. S. Fyfe ..............................................................................................................1 Section 1: Environmental planning and management Ethics, geological risks, politics and society A. Cecioni & V. Pineda.........................................................................................7 Environmental impact assessment and environmental management plans: an example of an integrated process from the UK M. A. Broderick & B. Durning............................................................................15 Geological factors in monitoring and planning nature trails at tourist centres in northern Finland K. Lehtinen & P. Sarala......................................................................................25 The acoustical improvement plan as a process to re-establish acceptable acoustical conditions M. Clerico & G. Soffredini..................................................................................33 Estimation of the on-farm-costs of soil erosion in Sleman, Indonesia A. Möller & U. Ranke .........................................................................................43 Section 2: Environmental modelling and monitoring Large scale soil erosion modeling for a mountainous watershed P. Zhou, J. Nieminen, T. Tokola, O. Luukkanen & T. Oliver..............................55

Decreased deposition of sulphate and the responses in soilwater at Estonian integrated monitoring sites 1995–2004 J. Frey, K. Pajuste, K. Treier, Ü. Mander, P. Kask & T. Frey ...........................69 SAKWeb© – Spatial Autocorrelation and Kriging Web Service J. Negreiros & M. Painho...................................................................................79 Section 3: Environmental pollution and remediation After treatment of landfill leachate in peat filters M. Kõiv, M. Kriipsalu & Ü. Mander...................................................................93 Studies on the distribution of heavy metal Cd in contaminated soils of various particle sizes and removal efficiencies of heavy metal using acid washing S. F. Cheng, C. Y. Huang & L. S. Hsiao ...........................................................105 Mitigation of alkaline mine drainage in a natural wetland system J. Kelly, P. Champagne & F. Michel ................................................................115 A sequential aerated peat biofilter system for the treatment of landfill leachate P. Champagne & M. Khalekuzzaman ...............................................................125 Are crop and range lands being contaminated with cadmium and lead in sediments transported by wind from an adjacent contaminated shallow lake? H. O. Rubio, T. R. Saucedo, M. R. Bautista, K. Wood, C. Holguin & J. Jiménez .....................................................................................................135 The sorption characteristics of PAHs onto soils in the presence of synthetic and bio surfactant J.-H. Chang, Z. Qiang & C.-P. Huang .............................................................143 Carbon dioxide sequestration in coal: implications for CO2 disposal and CH4 displacement from coal seams M. Mirzaeian & P. J. Hall.................................................................................151 Section 4: Climatological processes Response of dissolved organic carbon in a shallow groundwater ecosystem to a simulated global warming experiment K. P. Wilson & D. D. Williams .........................................................................163

Regional analysis of climate and bioclimate change in South Italy A. Capra, P. Porto & B. Scicolone ...................................................................175 Section 5: Geo-environment in urban settings The urban geo-science model: an essential tool to support planning and sustainable development D. McC. Bridge, B. L. Morris & J. R. A. Giles .................................................187 Developing design-oriented strategies to combat regional scale climate change B. Stone, Jr ........................................................................................................195 Land subsidence evolution and controlling mechanisms near Mexico City: environmental planning and management M. A. Ortega-Guerrero .....................................................................................205 Stability of slopes of municipal solid waste landfills with co-disposal of biosolids M. Chopra, D. Reinhart, M. Vajirkar & B. Koodhathinkal ..............................215 Section 6: Geoecology Net photosynthetic O2 evolution and calcium precipitation in Chlamydomonas reinhardtii Y. Y. Wu, P. P. Li, B. L. Wang & C. Q. Liu.......................................................225 Dynamics of suspended sediment concentration and the impact on a lake-inhabiting bivalve (Corbicula japonica) in the Abashiri River basin, Hokkaido, northern Japan S. Yanai, Y. Nishihama & R. Tamura ................................................................231 Potential of poplar plantation for enhancing Polish farm sustainability A. Czarnecki & A. Lewandowska-Czarnecka ...................................................241 Section 7: Hydrological studies Modified quantitative estimation model of erosion and degradation in three mountainous watersheds D. Emmanouloudis & M. Kaikis .......................................................................253

Bathymetric curve (75 years old) validation, using the soil erosion transportation at Cuitzeo Lake Watershed J. Lafragua, A. Gutiérrez, A. Bahena, G. Moriel & S. Férnández....................263 Efficient watershed modeling using a multi-site weather generator for meteorological data M. Khalili, R. Leconte & F. Brissette................................................................273 Hydrological modelling for river basin management in a highly hydro-geological conditioned environment D. Guida, A. Longobardi & P. Villani ..............................................................283 Hydrological modelling of snow cover in the large upper Po river basin: winter 2004 results and validation with snow cover estimation from satellite D. Rabuffetti, A. Salandin & R. Cremonini.......................................................293 Section 8: Landscape analysis Spatial correlograms and landscape metrics as indicators of land use changes R. Aunap, E. Uuemaa, J. Roosaare & Ü. Mander ............................................305 The role of geosciences and landscape in the management of Natural Parks of Guadalajara (Central Spain): in search of suitable applications A. García-Quintana, M. P. Abad, M. Aguilar, L. Alcalá, I. Barrera, M. Cebrián, M. C. Fernández de Villalta, J. F. García-Hidalgo, R. Giménez, A. E. Godfrey, J. A. González-Martín, A. Lucía, J. F. Martín-Duque, M. Martín-Loeches, E. Quijada, J. M. Rodríguez-Borreguero, R. Ruiz López de la Cova & A. Solís .................317 Exploring the effect of demographic elements on the evaluation of the scenic beauty of various landforms – preliminary results A. Tsouchlaraki .................................................................................................329 An approach to the landscape analysis B. Badiani..........................................................................................................341 Evaluation and analysis criteria of the environmental risk factor of the anthropic perturbation in the infrastructure works G. Gecchele & G. Pizzo ....................................................................................349

Section 9: Natural hazards and risks M3 (Monitoring, Management and Mapping) – tool for the solution of the conflict: “people and environment” Z. Boukalová, V. Beneš & P. Kořán..................................................................361 Gas hazard: an often neglected natural risk in volcanic areas W. D’Alessandro ...............................................................................................369 Section 10: Remote sensing Identification of As-bearing minerals associated with mine wastes from former metalliferous mines in France using laboratory reflectance spectra V. Carrère .........................................................................................................381 Rapid mapping with remote sensing data during flooding 2005 in Switzerland by object-based methods: a case study Y. A. Buehler, T. W. Kellenberger, D. Small & K. I. Itten ................................391 Using spatial technology for analyzing disturbed areas and potential site selection in Chihuahua, Mexico V. M. Tena, A. C. Pinedo, A. H. Rubio, P. de L. G. Barragán, A. A. Pinedo, M. V. Hernandez & C. Velez .......................................................401 Section 11: Soil and rock properties The pedoecologial conditions of natural and opencast peat fields in Estonia M. Noormets, T. Köster, T. Tõnutare, K. Kauer, R. Kõlli, T. Paal & M. Oder .........................................................................................................413 Monitoring programme for underground rock characterization facility K. Lehto & J. Lahdenperä.................................................................................423 Geotechnical evaluation of Stabilized Dredged Material (SDM) from the New York/New Jersey Harbor A. Maher, A. Sarmad & M. Jafari.....................................................................431 The impact of soil quality on cocoa yield in Nigeria O. A. Amusan & F. O. Amusan .........................................................................443

Section 12: Vulnerability studies Influence of ground water extraction in the seismic hazard of Mexico City J. Avilés, L. E. Pérez-Rocha & H. R. Aguilar ...................................................457 Predicting favourable areas for landsliding through GIS modelling in Aparados da Serra (Brazil) A. J. Strieder, S. A. Buffon, T. F. P. de Quadros & H. R. Oliveira...................467 Author Index ...................................................................................................477

Towards a truly sustainable future and a high quality of life for all people on our planet: the role of the Earth Sciences

W. S. Fyfe Department of Earth Sciences, University of Western Ontario, Canada

Abstract

We are living in a new period of earth history. I was impressed by the writing of Vladimir Vernadsky, 1929, in his book The Biosphere. He said we were in a transition to a new geological era which he termed the psychozoic era. As he stated, mankind as a whole would become a new and powerful geological entity able to transform the planet. I very much liked the words of Sir Crispin Tickell in his British Association lecture of 1993 “I was recently asked if I was an optimist or pessimist. The best answer was given by someone else. He said that he had optimism of the intellect but pessimism of the will. In short, we have most of the means for coping with the problems we face, but are distinctly short on our readiness to use them. It is never easy to bring the long term into the short term. Our leaders, whether in politics or business, rarely have a time horizon of more than five years”. I have been fortunate that I have walked in over 60 nations. When you walk you see, smell, the problems and you meet people of all types. Keywords: the Psychozoic era, education needed for all people, diversity, energy and waste reduction, soil quality.

www.witpress.com, ISSN 1743-3541 (on-line)

© 2006 WIT PressWIT Transactions on Ecology and the Environment, Vol 89,

Geo-Environment and Landscape Evolution II 1

doi:10.2495/GEO060011

Keynote Address

1 What are the great world problems today?

When we look at many regions of our planet there is no question we face the problem of too many people. An excellent data source is provided by the publication by The Economist [1], pocket world in figures, which is published every two years. I am looking at the edition for 2003. For example, we see that the average female in Niger has eight children, Ethiopia 6.75. This can be compared with advanced nations like Spain 1.13 children, Italy 1.20, Sweden 1.29. But there is no doubt that related to such problems is quality of education for every male and female. For example, in Niger adult literacy is only 15.9%, compare Spain 98%, Italy 99%, Sweden 99%. I have noticed when I have visited some nations in Africa, most females do not go to school! It is interesting that the Scientific America (September 05), recently reviewed such problems including a section on the Population Peak. I was born on a farm in New Zealand. The main activity was sheep farming. One thing we all knew: Every year the climate, rainfall, can be different. If there is too little rain, there is too little food and the animals die. This is also true for humans. There is also no doubt today that our climate is changing - Human activities have changed the chemistry of our atmosphere and the convection patterns in the oceans and atmosphere. There is a fundamental feature of all fluid convecting systems. As the temperature increases, so do chaos and the unreliability of predictions. Today we see new deserts, new floods, etc. etc. As the recent UN meeting in Montreal, Canada, showed, even our leaders in politics are beginning to accept the reality of climate change and the technologies causing such changes. Our life support systems include climate, air to breathe, water quality and quantity, and quality food. At this time millions of people suffer from serious malnutrition as in the nations of Africa, India and many more. As the famous French scientist and philosopher, Montaigne, said in 1581, “diversity is security”. There is no doubt that good nutrition requires great diversity. Food diversity again depends on climate, water and soil quality. Soil quality is not well appreciated by many. In many of the most fertile regions of the world the soil quality and quantity is related to recent tectonic events, including volcanism and mountain building. Good soils need complex mineral components, quantity and diversity. Recent studies in the NW of the U.S.A. show that crop yields increase following many volcanic eruptions. Work in many nations has shown that the use of pesticides, etc., can decrease quality and increase health problems. As they say in the U.K., such chemicals are BIOCIDES. As with all animal species, we require a very wide range of chemical elements in our food. We have studied soils of laterite soils, and water chemistry in many parts of the Amazon region. Temperature is warm, rainfall can be extreme. The key nutrients in the soils can be washed away leaving soils not productive and not producing quality food. The great trees live because of



2 Geo-Environment and Landscape Evolution II

complex root systems and dust. We have shown that some parts of the Amazon River water systems are very deficient in mineral components. For a general discussion, see Van Straaten, 2002 [5], Brown and Wolf, 1984 [1]. Without our present energy technologies we would not have our present population. Imagine Canada, where I now live, if there was no electricity, natural gas or oil. One of the most urgent problems on our planet is to improve energy technologies. The largest consumers of energy at the present time are: the U.S.A., China, and Russia. Major components include sources from coal, oil, nuclear, hydro, etc. All lead to environmental problems. Given our present knowledge, can we produce adequate clean energy for all nations? The first problem we must consider is how to reduce waste by intelligent construction technologies, and use of fossil fuels. I have noticed the potential use of underground construction in some nations. Do we all need a big automobile? Work in many nations has shown how we can use natural energy resources such as the Sun, Wind, tides, and beneath our feet, geothermal energy. As the temperature rises about 30o C every km beneath the surface, there is vast potential for geothermal heating. Even more potential is available where we have recent active volcanism. A few nations have made use of such resources. And there is active volcanism in certain rock types, the hot fluids are rich in hydrogen, a clean gas energy resource (see Fyfe, 1999) [2]. We must improve our management of waste products. In this respect, ‘Europe leads the world! First reduce the use of materials which cannot be recycled and reused, as with plastic bags, etc. In most nations that use nuclear power, there is a problem as to where do we put nuclear waste where it will be safe for millions of years (Fyfe, 1999) [2]. But we also have many other examples of waste problems, as with materials which involve elements such as arsenic in mine wastes. Recent work we have done on the volcanic basalt rocks of Hawaii show that these rocks will absorb carbon dioxide. When we burn coal and oil, why not put the gas products underground? I have discussed this problem with many people in China and India. Do we design cities intelligently? Again, in this respect Europe leads the world. As I mentioned above, storage facilities, etc. can be built underground. Don’t waste the surface! To live in a city, do you need a car? My daughter lived in London, U.K., for several years. As she once said, a car is a nuisance. Public transport is excellent. Can you walk to work, etc. etc.? I have been in many cities in Europe. You do not need a car. Many hotels are designed to save energy and water needs. It is possible! We all use massive quantities of materials derived from our mining industries. How much iron do we need? Do we recycle what we have used? Do you need a gold ring? How much rock was mined to get that gold? Was it in your nation or imported? Most gold ores are rich in arsenic and there are many examples of deadly arsenic pollution related to gold mining. The same is true for many of the materials used in the fertilizer industries of our agri-technology.




2 Concluding thoughts

I always remember discussions during a meeting of the International Union of Geological Sciences in Florence, Italy. There was agreement that the two basic sciences are astronomy and earth science, from which all others are derived. Also, there was concern about how few scientists were working on the urgent world problems. As was said, too many scientists spend their life finishing their Ph.D. research projects. World data are clear. The quality of life in nations is related to the quality of education for all people, all ages. And another feature of education is clear. Numerous studies of animal life including our relatives, the apes, show that in general the females are more intelligent than males. There are many reasons why this should be. When we examine nations and their problems, such as AIDS, there is a clear relation between problems like this and female education. We also need more females in governments as with the situation in N. Europe. I recently attended a wonderful conference in England. It was based on the need for education for all. We had about 10,000 young people of all ages, and about 1,000 school teachers from all Europe. The discussions were wonderful (see Moody et al, 2000) [3]. We need new systems, new groups of experts to integrate knowledge and the planning of our world for future generations. Such groups must include scientists, sociologists, economists, politicians and, particularly, citizens of both sexes. As Sir Crispin Tickell stressed, we must plan for now and for future generations. That is why the focus of this meeting on Landscape Evaluation is so important. My final question to all experts at this meeting. When you leave the planet, is it better for all people than when you arrived? My answer is NO, it is not better and a major cause is overpopulation in many nations related to a poor quality of education for all people. We must improve our systems and plan for the next generations.

References

[1] Brown, L.R. and Wolf, E.C. Soil erosion and quiet crisis in the world economy, WorldWatch paper 60. The Economist 2003. Pocket world in figures. 236 pp. 1984

[2] Fyfe, W.S. Clean energy for 10- billion humans in the 21st century - Is it possible? Coal Geology 40:85-90, 1999.

[3] Moody, R.J.J. et al, Earth Alert, the past, present and future of our planet. Geologists Association, London, 149 pp. 2000.

[4] Crossroads for Planet Earth, Scientific American, Special Issue, September 2005.

[5] Van Straaten, P. Rocks for crops. Department of Land Resources, University of Guelph, Canada. 338 pp, 2002.

[6] Vernadsky, V. The Biosphere. Synergetic Press Inc. 1986. 82 pp. 1929.




Section 1 Environmental planning

and management

Ethics, geological risks, politics and society

A. Cecioni & V. Pineda Departamento Ciencias de la Tierra, Universidad de Concepción, Chile

Abstract

Since humans live in communities, the moral regulation of behavior has been necessary for the collective well-being. From the antiquity, the Greek philosophers discussed moral behavior, which led to the later development of ethics. Some geological processes, such as earthquakes, tsunami, landslides, volcanism, and geological structures such as faults, are potentially hazardous for the inhabitants. Some geological processes can affect entire cities and active faults can damage buildings constructed on them. In general, the urban and urban development areas are sustained on numerous components. Among these components can be pointed out the economic one that regulates the urban already consolidated zone and the areas of urban expansion; the information provided by geologists about the possible occurrence of geologically hazardous or risky processes; and the political authorities that must take the difficult decisions based on the economic component of the private investors, the geologically dangerous processes and the commitment of the State to provide inexpensive accommodation to low social-economic people and the responsibility of granting safety to the society. Analysis and questions about professional ethics, investors’ ethics and political ethics of the State, taking as a goal the information that must be provided to the citizenship and to the civil protection, are discussed. Keywords: ethic, geological risks, land planning.

1 Introduction

Ethics (gr. êthos) is the support of the principles of the human behavior, called also morality (lat. moralis).

Ethics is the natural norm of the human conduct. Since the peoples live in community, the moral regulation of the conduct has been necessary to obtain the collective well-being.




doi:10.2495/GEO060021

Among the classic civilizations, the Egyptian developed the ethics combined with the religion; in classic China culture Confucio’s doctrine was accepted as a morality code. From the VIth century B.C., the Greek philosophers, thought about the moral conduct, developing the ethics as a way of personal life related to the respect for the social community.

Socrates did not partake the ideas of the sophists. His philosophical position it is represented in the dialogues of his disciple Plato. The Socratic vision can be summarized in the following concepts: the virtue is knowledge; the people will be virtuous if it has instruction and culture, and the vice (evil) it is the fruit of the ignorance. According to Plato, the human soul is composed by three components: the intellect, the determination (“the want”) and the emotion. The virtue of the intellect is the knowledge of the intentions of the life; the determination has the virtue of the value in the aptitude for acting; and the virtue of emotion it is the prudence. Plato thinks that the real reality is the purification of all the passions, which originates knowledge, temperance, justice and virtue [1] talents that he reiterate in his works. Plato has an absolute conception of the “good” and of the virtues of the ethical and moral attitudes.

Christianity extended the application of the moral values to the whole society, including the slaves. This conception of equality in agree to the ethics and to the morality, is represented in several references and concepts, such as loves your neighbor as to you itself [2].

It is interesting to observe the masterly classification of the faults that Dante Alighieri describes in the Chapter of the Hell of the Divine Comedy [3]. He places the more condemnable attitudes between 7th and 9th circle of the Hell; the violent ones against God, against the nature and against the society imprisoned in the seventh circle; and the traitors to their benefactors condemned to the ninth circle. Then, according to Dante, we have that the treason and the lack of ethics and of morality, are located in the maximum circles of punishment.

For need of living together, there were formed groups of persons that constitute the society. The city represents the symbol of the social collectivity. From the first cultures and men’s groups, the city constitutes a living space in which the society feels protected. In the city there develop the economic, cultural and family activities. Consequently, to be able to coexist, the persons must act with ethics and be mutually respectful. Leaders and scientists are not exempt from the ethical duty towards the society.

The society must be regulated in the ethics and morality conditions. Institutions and laws were created. But the laws become anachronistic, since they are applied to a way of life and of attitudes that happen in a certain temporary and circumstantial episode. If this social situation is overcome, the law remains obsolete. It is very difficult to be able to establish laws on events that might happen in certain future circumstances, in conditions that we do not know. The serious problem is not to legislate and not to take decisions related to situations that we know, that can happen and that are hazardous for the safety of the society.

In this respect, no person has the right to damage other one, understanding the damage in all the possible forms. The duties and the rights exist. One of the




principal duties is the honest and complete information, since giving only a part of the truth is a falsehood. One of the principal rights is the attribution of living safely, and this is an obligation of the State.

Politics, science and society must constitute a harmonic whole founded on the ethics.

2 Geological hazards and some related question: Concepción area as a pattern

In Concepción (South-Central Chile) area, that could be a case of other similar regions of other countries, there occur constructions that are founded on faults, in zones of flood for tsunami and/or that can be affected by landslides. This situation causes a decrease of the quality of life and a dangerousness situation for the inhabitants.

It is thus convenient to design and apply procedures for determining responsibilities. At this point, we must consider some questions related to (A) all ready constructed areas, and (B) use of expansion urban areas. (A) For already constructed areas, some questions are: Who is the responsible to notify to the inhabitants of a building that it is

constructed over a fault that could damage the building in case of an earthquake?

Who is the responsible to notify to the inhabitants that the coastal zone in which they live or work is dangerous for tsunami hazard?

Who is the responsible to notify to the inhabitants that the area where they live or work could be destroyed by a tsunami?

Who is the responsible to notify to the tourist skiers that the volcano where they are is not monitored?

(B) For expansion urban areas: Who is the responsible to approve the use of a certain hazardous areas for

general social benefit? Who is the responsible to consent the construction of a building on a fault? Knowing the dangerousness of some area, who is responsible to accept the

proposal of an investor, which have, obviously, an economic interest to develop constructions?

Who is the responsible to request information of hazard natural processes, vulnerability and risk of some areas?

In the area of Concepción there are active faults, some zones with high seismic risk (Fig. 1), landslide risk (Fig. 2) and tsunami risk [4] (Fig. 3).

3 Some answers

Answers to these questions could be very complicate for some ones. But applying ethics, the solution is simple.




(A)

(B)

Figure 1: (A) Main faults in Concepción city. (B) Zones of major seismic dangerousness, considering faulting, soil, underground waters and other geological factors.

B

B

B

B




(A)

(B)

(C)

Figure 2: (A) Landslide in zones of housings. (B) Slide in mountain highway. (C) Crown of new landslides (retrogresives) – hazardous zone.




Figure 3: Zone of flood by tsunami indicated with light gray colour, extending across the whole city of Talcahuano.

The occupation of an investor is to make investments to develop social activities, for which it is normal to expect economic benefits. Scientists must don’t have conflict of interests and its role is to study neatly the territory, analyze in an objective way the hazardousness and associate risks, prepare maps of vulnerability and to warn the authorities about these dangers. The politicians and managers of the State, as the scientists, must not have created interests and they has the obligation to be informed about the geological and natural risks, in general, before granting the permissions of urban development.

Consequently, the administrator of the State must warn the population about the dangers and risks of geological processes that could occur in the areas where they live and work. Additionally, the administrators must evaluate the convenience of authorize investor to occupies territories for some social activities. On the other hand, the administrators of the State must request studies of vulnerability of geological risks that allow to carry out a suitable managing of the territory and, in the case that some dangerous zone should be occupied for urbanism, industry and tourism, among others, they will have to request and finance a appropriate monitoring system for the geological hazardous processes that affect this territorial space.

4 Thoughts and conclusions

In several countries there do not exist procedure that force to the State to effect studies of vulnerability, risk and monitoring of geological hazardous processes.




The State provides ambulances and hospitals to care patients who have accidents; it provides policemen to protect the citizens of the delinquency; it provides firemen to attack fires and to carry out rescues. Then, why the State does not place in the same plane the geological hazardous processes and provides to the society of a strategy of monitoring intended to save lives?

We are strongly sure that an international organization must demand to the States the application of studies on geological hazardous processes and to carry out an ethical planning of the territory, in order that the society could feel surer of living in an environment in which these processes are monitored.

References

[1] Platón. Diálogos Socráticos. Traducción de Patricio de Azcárate. Editorial Océano de México. S.A. 365 p.

[2] Sagrada Biblia. Biblioteca de Autores Cristanos, de la Editorial Católica, S.A. Madrid 1970. Cuarta Edición. San Lucas (10:27). 1377 p.

[3] Dante Alighieri. La Divina Comedia. Adapt. Francisco José Fernández Defez - México, D.F.: Editorial Selector, S.A. de C.V., 2004. 96 p.

[4] Servicio Hidrográfico y Oceanográfico de la Armada de Chile. Bahía de San Vicente. Carta de Inundación por tsunami, escala 1:10.000. Armada de Chile, SHOA, 2000.




Environmental impact assessment and environmental management plans: an example of an integrated process from the UK

M. A. Broderick1 & B. Durning2 1Halcrow Group Ltd, UK 2Oxford Brookes University, UK

Abstract

Environmental impact assessment (EIA) is a mature process implemented around the globe to identify significant impacts from development and provide mitigation measures to reduce these impacts. Increasingly in the UK the process is being supplemented through the integration of an environmental management plan (EMP) into the resulting environmental statement. The EMP specifically aims to manage the impacts during the construction phase of the development. This paper presents an example of practice from the UK in this integrated process for the installation of a high pressure natural gas pipeline through open countryside. It demonstrates the added benefit that the integrated process provides in managing and reducing environmental impacts from the development. Keywords: environmental assessment, environmental management, construction, pipeline.

1 Introduction

The environmental impact assessment (EIA) process is a well established method used around the globe in the identification and mitigation of the impact of developments on the environment. It consists of a series of studies and discussions which are designed to: • identify which legislation is relevant to the proposals (screening) • assess the scope of the project (scoping+consultations)




doi:10.2495/GEO060031

• identify the nature of the existing environment (baseline) • obtain stakeholders views on proposals (consultation) • identify the impacts of the proposals and predict the likely magnitude and

significance of those impacts on the environment (environmental assessment) • allow the formulation of mitigation measures (mitigation). The outcome of the process is the production of an environmental statement (ES). Increasingly in the UK the EIA process for developments at varying scales is being supplemented through the integration of a voluntary environmental management plan (EMP) into the resulting ES. The purpose of the EMP is to serve as an operational manual for implementing appropriate environmental controls and monitoring procedures within the construction phase of the proposed development. It sets out to ensure that the construction of the works complies with relevant environmental legislation, licence conditions and accepted good practice and that measures to mitigate impacts discussed in the project ES are implemented. In this paper we present a case study from the UK of the construction of a high pressure natural gas pipeline, in central southern England, which was subject to the EIA process with a fully integrated EMP (Environmental Resources Management [1]). It is based on the experiences of one of the authors who at the time was employed by consultants Environmental Resources Management (ERM) and undertook the EIA and implementation to the EMP. We aim to demonstrate how the combined process ensures that environmental management procedures during the construction phase limit environmental impacts and ensure that good quality restoration of the pipeline route limits environmental damage, to the extent that eight years later there is very little evidence in the landscape for the presence of the pipeline.

2 Background to case study

The pipeline, from Aylesbury in Buckinghamshire to Chalgrove in Oxfordshire, a distance of approximately 26km, was constructed by National Grid (called Transco at the time of pipeline construction in 1999). National Grid is responsible for the operation of the national gas distribution system in Britain known as the National Transmission System (NTS). The NTS transports gas at high pressure along a network of pipelines from the gas production terminals to major gas users (“40 power stations, a small number of large industrial consumers”) and a series of “Local Distribution Zones” from which the gas is then distributed at a lower pressure to consumers (National Grid [2]). The case study pipeline was constructed to provide additional capacity for the NTS in order to satisfy an increase in demand for gas in the south of England. It supports an existing pipeline which runs parallel and continues past Chalgrove.

3 EMP as part EIA process

The requirement for screening of developments which are possibly subject to the environmental assessment (EA) process has been a requirement in UK law since




the late 1980s (although the legislation was revised in the late 1990s [3]). Although not all developments are required to go through the process the requirement for high pressure natural gas pipelines falling within certain criteria (length, diameter) to go through the EA process is enacted in specific gas pipeline legislation [4, 5]. There is no requirement in UK or EU legislation for an EMP to be part of the EIA process. However, it is part of National Grid corporate procedures that an EMP be included in the contract documents on which the commercial terms for the construction contract are negotiated.

3.1 The EMP process

The broad purpose of the EMP is to: • provide a mechanism for ensuring that measures to mitigate potentially

adverse environmental impacts are implemented • ensure that standards of good construction practice are adopted throughout

the construction of the pipeline • provide a framework for mitigating impacts that may be unforeseen or

unidentified until construction is underway To be successful an EMP should evolve throughout the life of the project. For this project the EMP was issued for consultation to various stakeholders and was refined as additional information, design changes or comment from stakeholders becomes available. An EMP can therefore also provide assurance to stakeholders that their requirements, with respect to environmental performance, will be met. Although the EMP detailed the mechanisms through which the issues outlined above were to be addressed during construction of the pipeline and the responsibilities for meeting them, it was the contractor who was required to provide method statements of the details of the actions to be taken, in order to implement each aspect of the EMP. The method statements had to demonstrate how compliance with the requirements of the EMP were to be achieved, and specify the names of the individual people who will be charged with achieving and monitoring compliance

3.2 Auditing/monitoring during construction

The EMP also provided a framework for compliance auditing and monitoring to ensure that its aims are being met. As the EMP formed part of the commercial contract for the contractor during the construction of the pipeline, National Grid required that inspections and audits were undertaken to ensure that the plan was being implemented. In addition to any audits the contractor may undertake, National Grid also commissioned their consultants to undertake periodic site audits. A checklist pro forma was used which covered the environmental issues addressed in the ES and the EMP. Where problems were identified corrective actions were required to be undertaken. These could include further direct mitigation, changes to procedures or additional training.




4 Aylesbury-Chalgrove pipeline

For natural gas pipelines the EIA process can be divided into a series of successive components which mirror the overall engineering design and construction process and also the general approach adopted to EIA for other developments (see Table 1).

Table 1: Overview of environmental, design and construction of pipelines.

EIA process for natural gas pipelines

Pipeline design and construction phases

General approach to EIA required in UK legislation

Establish need Pre-feasibility Screening Route corridor information study

Feasibility Scoping

EIA (baseline data collation, impact prediction) and production of ES

Conceptual design EIA (baseline data collation, impact prediction) and production of ES

Production of EMP Detailed design, commissioning and construction

Consenting from Department of Trade & Industry

4.1 Route corridor information study

The pipeline route was to be from Aylesbury in Buckinghamshire to Chalgrove in Oxfordshire, a distance of approximately 26km. The desk based route corridor information study was carried out at the pipeline feasibility study stage and constraint/overlay maps were produced at a scale of 1:50 000, based on:

• geological/ground conditions • distances • archaeological sensitivities • ecological sensitivities • numbers of roads, rivers, railway, hedgerow crossings • two route corridors 1km wide

The route corridors passed completely through open countryside, in an area known as the Vale of Aylesbury. The topography along the route was gently rolling, although some areas of high ground lay on the fringes of the route. It also passed through a river valley system (River Thame). The majority of the route corridors was underlain by clay (Jurassic and Cretaceous) with some minor areas of Cretaceous sandstone. Very little superficial deposits occurred along the route and those that were encountered were clay deposits and some floodplain gravels. The key environmental issues identified from the route corridor study which would require detailed evaluation were: archaeology; ecology and land take.




Secondary issues which would require consideration, were impact on the landscape and drainage as the land use along the route was agriculture.

4.2 EIA and production of ES

The next stage in the process was the undertaking of the EIA, which involved detailed baseline surveys and the collation of the information into an ES where impacts were identified and mitigation measures proposed. Information concerning the project itself, including background, project schedule, construction techniques, restoration and operation were also included. Important information and areas of particular concern were identified and plotted on constraint/overlay maps at a scale of 1:10 000. A preferred route corridor was brought forward for EIA based on distance crossing and avoidance of sensitive areas.

4.3 Production of an EMP

The final stage involved the production of the EMP, detailing all environmental constraints along the final route, and the mitigation measures to be taken. It also included detailed restoration practices and highlighted areas where aftercare was necessary, and the nature of the care required. Areas of particular concern were identified and plotted on constraint/overlay maps at a scale of 1:2500. The EMP identified eleven activities which may give rise to potential impacts during the construction of the pipeline and for which mitigation measures were required (ERM [1]). These are reproduced in Table 2.

Table 2: Proposed mitigation to environmental impacts contained within EMP.

Activity Potential impacts Proposed mitigation Pipeline construction (physical disruption from clearance of the working width, pipe stores and temporary working areas)

Damage to significant ecological, archaeological sites and species

Will be avoided by: re-routing; restriction of working width; bore underneath site; rescue dig for archaeology; translocate rare plants; minimise hedgerow removal and avoid trees; carefully reinstate topsoil and habitat

Fuel storage Leakage/ spillage may give rise to contamination affecting: abstraction downstream; ground-waters; ecology of surface waters

Site stores located >50m from watercourses. Bunded (110% capacity) design with impermeable liners for stores and refuelling point will be used. Use drip trays wherever possible. Provide local first response absorbents, booms etc. Training will be given to all staff. Inspect the works frequently.




Table 2: Continued.

Activity Potential impacts Proposed mitigation Prepare and exercise oil spill

contingency plans. Carry out regular inspections and maintenance of plant.

Machinery operations

Smoke and fumes Proper maintenance will be maintained

Noise Sitting (pumps, generators etc) away from dwellings. Provide adequate silencing. Switch off when not needed

Dust Restrict vehicle speeds. Spray in dry weather

Construction of river crossings

Fisheries (migration and spawning)

Schedule construction activities to avoid sensitive times or minimise sedimentation effects.

Stimulation of bank/bed erosion

Design adequate emplacements and protection measures.

Clearance and activities within the Right of Way

Sediment run off to watercourses, ponds and lakes

Provision of cut-off drainage and settlement ponds.

Effects on fish (asphyxiation; indirect effects on mitigation and spawning)

Discharge of pump outlet to soil surface/crops to promote seepage (subject to agreement).

Waste management

Contamination of soils and water from wastes

Effective containment and management of wastes.

External pipe cleaning

Contamination of soils and water by blast grits

Use mechanical brushing in lieu of grit blasting.

Machinery movement (trafficking)

Topsoil compaction Strip and segregate topsoil. Break up panned sections

Hydrotesting Disruption from abstraction and disposal of large volumes of water

Plan and agree abstraction and discharge points, rates and contingency arrangements.

Pollution from additives Minimise/avoid use of dyes, corrosion inhibitors, oxygen scavengers

Erosion from failure under test

Reinstate to pre-erosion conditions

Disruption to field drains

Water-logging/crop losses Insert header drains Reconnect severed drains

Following completion of the ES and EMP the document was sent to the relevant Government department for approval, as required under relevant




legislation. They undertook a consultation process with statutory bodies and regulators before granting approval for the pipeline.

Figure 1: Normal sequence for pipeline construction.

Figure 2: Pipeline route at Stage 4 – excavator is passing over a road and between a gap where hedging has been removed along the field boundary.

4.4 Construction of pipeline and auditing EMP

Barnett and Jordin [6] provide a useful guide to the pipeline construction process adapted for the case study. A “spread” method is employed for the construction which “involves several groups of workers and equipment who collectively

1. Receiving material 2. Setting out 3. Pre-construction land

drainage 4. Right of way and topsoil

stripping 5. Stringing 6. Welding 7. Excavation/trenching 8. Ditching/lowering and lay 9. Tie-ins 10. Bedding and covering pipe 11. Backfilling 12. Reinstatement 13. Post construction land drains14. Final trim




conduct the various stages of the construction operation. Each group completes an activity which picks up where the last one left off, advancing the construction process a step at a time and leaving it ready for the next step to begin”. Of note is that fact that construction work is limited to a “seasonal window which extends from March/April to October during which time the weather is more predictable and ground conditions are more favourable”. The normal construction sequence is given in Figure 1 (taken from Barnett and Jordin [6]) Auditing during construction of the potentially impacting activities listed in table 2 was carried out regularly by consultants using the methodology referred to in section 3.2.

Figure 3: Area of restored hedging and field (taken January 2006 from a road looking along the restored pipeline route).

4.5 Post construction follow-up

Due to the national importance of the NTS there is post construction follow up of this development: aerial surveys are regularly employed by National Grid to monitor the route and agriculture liaison officers and land agents maintain regular contact with land owners. In order to assess whether there was any evidence of degradation of the landscape caused by either the pipeline installation or poor restoration a number of points along the pipeline were visited by the authors in January 2006. The points included a road crossing (similar to that in Figure 2), a footpath and a stream crossing. The location of the pipeline was determined by identifying its position on the ES map and locating the National Grid marker post in the nearest road. In all cases very little evidence was found of a legacy of the pipeline. In one field the route could be identified by a darker green swath of grass, suggesting that restoration had improved the




land quality. The most obvious evidence was in the restored hedging (see example in Figure 3) where the extent of growth is not yet equivalent to that of the hedging removed.

5 Discussion

EMPs provide a critical link between EIA and project implementation. In effect EMPs comprise the operational response plan that implements the mitigation and monitoring programs for the project. The execution of an EMP is increasingly becoming conditional to project approval or licensing, and/or to project financing (Equator Principles [7]). The preparation of an EMP acceptable to all stakeholders is therefore a key part of the project development process. Preparing an effective EMP requires a balance between what is desirable, what is affordable, and what can be implemented. In particular it requires: • all stakeholders to have a common understanding of the objectives of the

EMP and particularly to understand the link between the EMP and any approvals or conditions that may be applied to the project on its implementation

• project owners/proponents to have an understanding of the requirements of relevant permitting processes applicable to the EMP and/or be familiar with the needs of specific, relevant financing agencies

• EMP costs to be clearly defined and understood by all parties • provision to be made for sustainability in implementation of the EMP,

particularly in post construction monitoring of impacts. EMP provides a concrete reassurance that construction/operational impacts identified in the EIA are addressed and mitigated during construction/operation. However, the absence in legislation of a requirement for EMPs and follow up environmental audits is a weakness, a fact which is increasingly being recognised (Morrison-Saunders and Arts [8]). EIA legislation could be strengthened and made more credible if EMPs were mandatory, incorporating environmental auditing during construction/operation.

References

[1] Environmental Resources Management (1998) Aylesbury to Chalgrove Gas Pipeline: Environmental Review. Report prepared for Transco, UK.

[2] National Grid http://www.nationalgrid.com/uk/Gas/About/ How+Gas+is+Delivered/

[3] Town and Country Planning (Environmental Impact Assessment) (England and Wales) Regulations (1999) SI 293

[4] Gas Act (1995) c 45 [5] Public Gas Transporter Pipe-line Works (Environmental Impact

Assessment) Regulations (1999). SI 1672 [6] Barnett, J. and Jordin, M, (1998) Pipelines – a worm’s eye view. Transco,

Ambergate, UK [7] Equator Principles http://www.equator-principles.com/principles.shtml




[8] Morrison-Saunders, A. and Arts, J. (2005) Learning from experience: emerging trends in environmental impact assessment follow up. Impact Assessment and Project Appraisal 23 (3) 170-174




Geological factors in monitoring and planning nature trails at tourist centres in northern Finland

K. Lehtinen & P. Sarala Geological Survey of Finland

Abstract Increasing tourism in northern Finland produces challenges for tourist centres, their land use plans and sustainable development. Increasing activities in areas that are sensitive to environmental changes need solutions for sustainable land use. Equipment to minimize environmental effects, and the planning of ecologically, culturally, and visually sustainable built up areas at popular tourist centres on the Ylläs and Levi fells are investigated in the LANDSCAPE LAB project. The project is partly financed by the EU LIFE Environment and the Geological Survey of Finland is involved as a partner. Geological factors such as the quality and composition of bedrock, the maturity of the matrix of surficial sediments, and geomorphology affect the resistance to erosion in different geoenvironments. In Fennoscandia, the glacial erosion has been intensive and the terrain is composed of eroded hill slopes and glacial landforms. Traditionally, in Finland, erosion resistance of nature trails has focused on vegetation and trampling resistance. The erosion rate is studied mainly by measuring the width and depth of the path and the amount of exposed roots and stones. Geological factors are not studied for nature trails, but geology and geological factors are the basic elements affecting resistance to erosion and should be included in land use planning. Careful planning and monitoring are the keys to creating visually impressive and geologically sustainable nature trails. Keywords: erosion resistance, geological factors, nature trail, land use.

1 Introduction

Increasing tourism causes pressure on the land use planning and sustainable development of tourist destinations. Northern Finland tourist destinations have




doi:10.2495/GEO060041

several thousands of kilometres of outdoor routes. These routes are situated in areas of different status like protected nature reserves, nature parks, state owned recreation areas and privately owned land. Many outdoor routes have multiple uses like, for example, nature trails for hikers, skiing and snow mobile routes and dog sled routes. Outdoor routes that are used as nature trails can be classified in four groups: 1. Old, naturally formed summer routes, which are historical travelling routes; 2. Old, naturally formed routes that were developed as nature trails after the

construction of guiding services; 3. Natural types of routes or other older routes, which have become nature

trails when taken into multiple, round-the-year use because of increased tourism (for example some winter routes taken into nature trail use);

4. Other new outdoor routes, originally made for different purpose, but are afterwards used as nature trails (for example constructed skiing routes, which are used as nature trail at summer).

Nature trails at tourist destinations are situated in places that have something special to see. Speciality is often based on the geology of the area. Geological features like block fields, gorges, lateral channels, esker chains and hanging bogs are major components of fell landscape (Fig. 1). These features are the key to creating the perceptions of hikers and other tourists that these areas are unique. These geoenvironments are sensitive to erosion due to their geological properties. Furthermore, tourist destinations in northern Finland (Fig. 2) are situate in a severe climatic zone with four seasons; a short growing season in summer, frost heaving and frost weathering problems caused by the cold winter, erosion impacts of melt waters during spring and autumn rain. Thus, nature trails become sensitive to impact of increasing amounts of visitors. Nature trails used at spring and summer, are particularly sensitive to erosion problems.

Figure 1: Scenery of highland on the Pallas fell.




Figure 2: The location of study and target areas in northern Finland.

The effects of nature tourism on nature trails have traditionally studied in Finland using the trampling resistance of vegetation [1, 2, 3]. The erosion status of nature trails is also estimated by measuring the width and depth of the path and by the amount of exposed roots and stones. Several visitor studies were made [4, 5] concerning visitor counting and impact of the amount and the type of visitors at nature trails. Geological factors on the resistance to erosion in relation to natural trails have earlier been studied for example in Sweden and North America [6, 7] but not in Finland. Finnish studies concerning construction work and road building have been done, but the circumstances and objectives of these studies were different. However, geology is the basic factor affecting both vegetation and erosion resistance properties of the areas. For planning, monitoring and conservation of nature trails, geological factors and the geological history of the area must be studied. Things to investigate are the bedrock quality and composition, geomorphology, Quaternary deposits and their properties like maturity and composition of the matrix of surficial sediments. Geological factors in land use and nature trail planning are studied by the Geological Survey of Finland as a part of the LANDSCAPE LAB project. The aim of the study is to investigate how and at what rate geological factors are affecting the resistance to erosion in different geological environments at tourist destinations in northern Finland. The objective is to develop a system to predict erosion effects and to develop a tool for nature trail planning and land use. Based on field and laboratory studies and visual examination, a classification of erosion resistance of different landform and soil types will be created and the equipment for monitoring erosion will be investigated. Furthermore, one other aim is to make recommendations for the planning of new nature trails and for suitable cover material for nature trails in different environments.




2 The Project LANDSCAPE LAB

The project LANDSCAPE LAB (2004-2007) ‘Tourist Destinations as Landscape Laboratories – Tools for Sustainable Tourism’ is an EU LIFE Environment supported project situated in Finnish Lapland. The beneficiary is the University of Lapland and there are nine partner organizations and two municipalities involved in the project. The main target of the project is to determine solutions to sustainable land use, and to plan ecologically, culturally and visually sustainable built-up areas, where disadvantages caused by tourism, would be minimized. The main research areas are the fells of Ylläs and Levi, popular tourist centres in western Lapland. The LANDSCAPE LAB Project consists of five tasks of which one concerns dissemination and management. The others are:

• The LABLAND task, in which the ecologically, culturally and visually sound urban structures at tourist destinations are studied. It makes landscape analyses in which the focus is on aspects like geology, landscape structure, history of land use, visual landscape, soundscape and landscape experiences.

• The LABECO task, in which the extent and types of the environmental impacts are studied to determine bio-indicators suitable for monitoring environmental effects for tourist destinations.

• The LABSOC task, where functional and social structures and activities of community are studied.

• The LABPLANT task, where usable and hardy plant species for sustainable restoration needs and landscaping in northern areas are sought. The task will select hardy plant species, develop methods of plant propagation and produce plant material for planting.

By combining all the data produced in the project, recommendations for use in the planning of tourist areas and implementation of environmental care will be presented.

3 Study of geological factors

3.1 Geological environment

The study area is situated in northern Finland and consists of nine geologically variable target areas (Fig. 2). The target areas have variations in bedrock quality and composition, geomorphology, glaciation history, Quaternary deposits and deepness of surficial sediments. The bedrock in Finland is part of Fennoscandian Shield, which is the oldest part in Europe (1.8-3.5 Ga). The bedrock consists mainly of quartzite, amphibolites, granulites and granites. In the study area, the rock types are quite resistant to erosion and are nowadays seen as fell areas at Finnish Lapland. The bedrock in the topographic depressions consists mainly of schist and greenstones [8, 9].




Finland was repeatedly covered by continental ice sheets over the last two and half million of years. Finland’s position in the central area of the ice sheet and the variation in glacial dynamics have formed various Quaternary deposits and glacioerosional features [10]. Different kind of till beddings, moraine formations and glaciofluvial deposits were formed [11, 12, 13]. These features control land use and have an influence, for example, on the location of settlements and the formation of passages.

3.2 Field and laboratory studies

Field studies started at summer 2005. The status of erosion on nature trails was mapped by measuring the width and depth of the path and the amount of exposed stones and roots. In luxuriant areas, where soil is enriched of nutrients, vegetation is worn and roots are often exposed (Fig. 3). In barren areas, stones are usually exposed and impeding walking (Fig. 3). The erosion rate caused by approximately the same amount of visitors was not equal in different geoenvironments. Different kind of soils and Quaternary deposits are the key to studying erosion on nature trails. Mire and areas where bog formation is on-going, and deposits like dunes and deflation areas (Fig. 4), are particularly sensitive to erosion. Nature trails in these areas need protective structures like duckboards and stairs to prevent erosion problems.

Figure 3: Exposed roots on nature trail at Oulanka (left); nature trail on moraine slope at Ylläs (right). Angular quartzite stones are exposed when fine material of moraine have been eroded.

One of the objectives is to study how suitable different cover materials are in different kinds of environments and how resistant they are against, for example,




melt water erosion (Fig. 5) and their stability in variable slope steepness. The aim is to make recommendations for the use of cover material on nature trails in different areas. For this purpose, present cover materials were mapped and samples were taken at target areas. In some places, nature trails sensitive to erosion are protected with covering materials like bark chip, sawdust, gravel, mineral aggregate and stone ash (industrial product) to prevent erosion or to conserve the path. Preliminary results show that the bark chip and sawdust seem to be problematic, because they tend to stay wet for a long time after rain and melting of snow. Sawdust also is not an aesthetic material on a nature trail. Gravel and stone ash do not have a water content problem, but coarse gravel seems to be unstable on slopes (Fig. 5). In some places, stone ash seems to be a suitable covering material, but its use is not economic due to high transport distances and costs.

Figure 4: Deflation area on nature trail at Saariselkä.

The objective is also to define indicators for monitoring geological factors on nature trails and for estimating erosion potential in planning new trails. 3 to 6 sites on nature trails have been chosen from every target areas for soil sampling. The samples are taken from the surface layer (depth ca. 30 cm) and analyzed to determine physical characteristics like matrix composition and maturity, weathering intensity, absorbing properties and frost heaving properties. To develop a useful method for studying soil properties and their variations in the field, geophysical methods will be tested. Electrical conductivity will be measured at all the sites on the nature trail and beside the trails in early spring and autumn to study adsorption, permeability and compression properties.




Figure 5: Nature trail at the Ylläs fell (left). Study areas are covered by snow

about six months of the year. During spring, melt waters cause erosion problems on nature trails. Some trails act as drainage channels, particularly during spring when the soil does not absorb water because the ground is still partly frozen; example of erosion problems on nature trail caused by unsuitable cover material (right).

4 Conclusions

Significance of this study is to investigate measurable geological factors affecting the resistance to erosion on glaciated terrain. As a result, equipment and recommendations will be created for planning, monitoring and conserving nature trails at tourist destinations in northern Finland. Preliminary results show that geological factors seem to have a significant impact on resistance to erosion on nature trails. The composition and quality of bedrock, depth of glacial overburden and grain size, stone content and quality of surficial sediment have a direct impact on absorption and permeability of soil. Stone content and quality of surficial sediments also affects the rate of wearing of fine-grained material at nature trails. However, soil and bedrock properties affect on vegetation and its luxuriance. To protect and conserve badly eroded or heavily loaded nature trails, it is important to find naturally looking but durable cover material.

References

[1] Tolvanen, A., Rämet, J., Siikamäki, P., Törn, A. & Orell, M., Research on ecological and social sustainability on nature tourism in northern Finland, Working Paper of the Finnish Forest Research Institute 2, 2004.




[2] Jämbäck, J., Aspects of ecological capacity: trampling tolerance and disturbance, Saarinen, J. & Järviluoma, J. (eds.). Luonto virkistys- ja matkailu-ympäristönä. Metsäntutkimuslaitoksen tiedonantoja 619, pp. 143-163, 1996.

[3] Hoogster, M., The effect of trampling on vegetation at four cottages in Torne Lapland, northern Sweden. Report from Kevo Subarctic Research Station 19, pp. 25-34, 1984.

[4] Erkkonen, J., Jokimäki, J., Saarinen, J., Tuulentie, S. & Virtanen, E. (eds.). Policies, Methods and Tools for Visitor Management. Proceedings of The Second International Conference on Monitoring and Management of Visitor Flows in Recreational and Protected areas, June 16-20.2004, Rovaniemi, Finland. Working Papers of the Finnish Forest Research Institute. 2004.

[5] Erkkonen, J. & Sievänen, T., Standardisation of Visitor Surveys – Experiences from Finland. Arnberger, A., Brandenburg, C. & Muhar, A. (eds.). Monitoring and Management of Visitor Flows in Recreational and Protected Areas, Conference Proceedings p. 252–257, 2002.

[6] Bryan, R., The Influence of soil properties on degradation of mountain hiking trails at Grövelsjön. Geografiska annaler 59 A, pp. 49-65, 1977.

[7] Hammit, W., Cole, D., Wildland recreation, ecology and management. John Wiley & Sons, 1998.

[8] Mikkola, E. Muonio-Sodankylä-Tuntsajoki. General Geological Map of Finland 1:400 000, Explanation to the Map of Rocks, sheets B7, C7, D7 (with an English summary). Helsinki: Geological Survey of Finland, 1941.

[9] Lehtonen, M., Airo, M-L., Eilu, P., Hanski, E., Kortelainen, V., Lanne, E., Manninen, T., Rastas, P., Räisänen, J., Virransalo, P., Kittilän vihreä-kivialueen geologia, Lapin vulkaniittiprojektin loppuraportti, (summary in English), Geological Survey of Finland, Report of Investigation 140, 1998.

[10] Sarala, P., Glacial morphology and dynamics with till geochemical exploration in the ribbed moraine area of Peräpohjola, Finnish Lapland. PhD thesis. Geological Survey of Finland, 2005.

[11] Hirvas, H., Pleistocene stratigraphy of Finnish Lapland. Geological Survey of Finland, Bulletin 354, 1991.

[12] Kujansuu, R., On the deglaciation of western Finnish Lapland. Bulletin de la Comission Geologique de Finlande 232, 1967.

[13] Johansson, P, Kujansuu, R. (eds.), Pohjois Suomen maaperä: maaperäkarttojen 1:400 000 selitys (Quaternary deposits of Northern Finland – Explanation to the maps of Quaternary deposits 1:400 000) (summary in English), Espoo: Geologian tutkimuskeskus. 2005.




The acoustical improvement plan as a process to re-establish acceptable acoustical conditions

M. Clerico & G. Soffredini

Politecnico di Torino, DITAG, Torino, Italy

Abstract

The word “Acoustical Improvement Plan” indicates a set of provisions, related to land management and suitable to reach the targets defined in the planning, with particular reference to acoustical standards satisfaction. The acoustical improvement plan could be interpreted as a loosening of the most critical nodes checked by the comparison between the noise mapping and the acoustical characterisation of a territory, but this interpretation could not reply to the most diffused question of acoustical quality. This condition derives by an approach founded on a multiplicity of actions and provisions, able to implement new logics in decisional processes that determines the territory planning and manages the transformation, with a particular attention to environmental noise problems. The purpose of this paper is to describe the plan identity, which isn’t represented by a specific planning action, but it invests and interests in particular actions of all politics of planning and territorial management, involving therefore the necessity to coordinate and to interact with the main instruments of territorial management. The acoustical improvement plan will not be the design of the intervention aimed to restore the sound levels limits, but a process, structured as a set of provisions and principles of urban planning and government of the territory, with the purpose to re-establish acceptable acoustic conditions for the critical zones, but also in order to prevent eventual future suffering. Keywords: noise pollution, acoustical improvement plan, environmental planning, land management.




doi:10.2495/GEO060051

1 The characteristics due to a healing plan

The word “Acoustical Improvement Plan” indicates a set of provisions, related to land management and suitable to reach the targets defined in the planning, with particular reference to acoustical standards satisfaction. A healing plan could be made of different measures, as administrative, legislative and through intervention to reduce environmental noise. The accordance of these procedures indicates a healing plan as a project that requires an interaction and coordination among the most important instrument of environmental management. The most strategic interaction could be with the functions asked to plan and manage the traffic and related infrastructures. The plan identity is not recognized in a specific action of a specific project , but involves intentions and actions of the whole politics of the territorial management [1]. This need of coordination, does not remain an internal request of Local Administrator, but becomes essential when other Subjects have to prepare and project a healing plan for the acoustical and environmental improvement. The healing plan is not the intervention project to take the acoustical value into law limits, but a whole of coordinate intervention for the progressive improvement. The interventions have to be different as type, time and related to specific part of territory or specific sources. The plan aimed to these objectives is not a project describing works, but a process, fixing structure and administrative conditions that determine a progressive acoustical improvement [2].

2 The contents of the Municipal Acoustic Improvement Plan (PRAC)

An acoustical improvement plan (PRAC) is the result of the comparison between the territorial PCA, “acoustic classification plan” (municipal action mandatory for the Italian law LQ 447/95) and the relevant survey: it must represent a solution of the most critical issues and the recovery actions have to answer to an extensive acoustic quality request which can only be the outcome of an integrated approach leading to the implementation of new specific strategies focused on the environmental noise and aimed to the territory management and evolution. The improvement plan will therefore be based on a wide actions range not only finalized to the mitigation of specific limits overcoming, but mainly intended as a coherent project where at the same time mitigations of the most critical situations, urban development plans, territorial government and administration will act to restore an acceptable acoustic situation and to prevent future problems. The correct approach to the acoustical improvement plan is therefore to put in place a set of effective activities, extending the application range from the reduction of the environmental noise intensity from fixed an mobile sources, to the mitigation of the acoustic impact on the affected people and to the




optimization of the passive protection performances of the buildings where the human activities are carried out, starting from the most sensitive situations. In this logic the technical and design actions to be developed in an improvement plan can be targeted to:

• noise emission reduction (source matrix reduction) • receivers exposure reduction (propagation path countermeasures) • buildings passive protection improvement (where sensitive activities are

located) [1].

3 The applied interaction between noise and land management

The PRAC is a dynamic tool in the acoustic pollution management and control, which includes design and direct mitigation capabilities always based on administrative actions, territorial (town planning and acoustics) and mobility management, through mandatory carry-back on actuation regulations [3]. Therefore administration, legislation and regulations aspects represent the core of the plan finalized to containment and mitigation of already compromised situations: such a scope requires interactions and coordination with the other territorial management tools and with current regulations. Even if the PRAC is not intended to produce administrative measures or urban development plans, this instrument must be used as an integration tool to include the acoustical impacts analysis on the territory, showed into “acoustic classification plan” PCA, and finalize proposals for future development. This point can be developed by the identification of procedures, operative tools and processes for the monitoring of acoustic effects related to different options, or through the comparison of acoustical effects of alternative scenarios. This can for example applied to the development of the City (PRGC) and Mobility plans and to the definition of the acoustical optimization criteria, or define references to be used in administrative measures finalized to the traffic management (such as Limited Traffic Zone (ZTL), heavy vehicles traffic, speed limitation along specific city areas). Due to the important role the territorial government plays especially on mobility, the contribution of the improvement plan also accounting for environmental acoustic quality can be regarded as a significant issue both in terms of recovery and prevention. Also the normative aspects can provide an essential contribution, in addition to the administrative ones, with specific reference to the prevention of further problems (i.e. the European Directive of 1996 about IPPC “Integrated Pollution Prevention and Control”, including noise pollution problems). The normative horizon of the improvement plan is deployed through a coherent program involving the local administration government, also including integrative instructions for building codes, health and safety issues and municipal police regulations about “noisy activities”.




In fact, as far as the building codes are concerned, the building codes refer to the passive acoustic requirements to be specified as a function of use, construction characteristics, infrastructural situation, and acoustic impact documentation to be attached to the authorization request. The integration of the operative technical specifications into the city plan plays a major role in territory government, through regulations finalized to the implementation of the acoustical classification as a tool in urban and building evolution [2].

Table 1: Scheme for the applied interaction between noise and land managements.

Administration Planning Actions

Laws and Codes

Operative Actions

Reduction of noise emissions in the environment

• City Plan (PRGC) • Acoustic

Classification Plan (PCA)

• Mobility City Planning (PUT)

• Infrastructural PRAC

• Private PRAC • IPPC

European, National and Local laws

• Impact Noise Assessment

• ZTL • Transport

Noise Mapping

Mitigation of impact on exposed people

• City Plan (PRGC) • Strategic Noise

Mapping • Noise Exposure

Assessment

European, National and Local laws

• Noise Conditions Assessment

• Private PRAC

Improvement of passive protection performances on receivers

• Building Codes National laws and codes

• Passive Noise Building Assessment

4 The application of the improvement plans: case histories

The national Italian legislation (LQ 447/95 and actuation decrees) establishes the need of municipal improvement plans and private company improvement plans for specific sources as transportation infrastructures and production sites under specified conditions: such conditions are identified when pre-existing problems are to be recovered. In particular, for the private company and transportation infrastructure improvement plans (DM 29/11/2000) [6] the intervention is required when the emission limits, defined by PCA, are overcome.




Differences are addressed as a function of the operational areas: • Municipalities: the acoustic improvement plan represents a complete

process, involving different competencies, regarding the whole city asset, town planning, administration, production and transportation infrastructures.

• Transportation infrastructures and production sites: in case of environmental acoustical pollution generated by a source are committed to restore the pre-existing compatibility acting on the source (when sustainability is not achievable) with specific containment actions.

As far as the municipal plans are concerned, the authority intervention is mandatory when the warning limits are overcome (LQ 447/95, art. 2, comma 1, lett. G) which in long term perspective coincides the emission limits of the PCA or in case of borderline connection between urban areas having quality levels with more than 5 dB (A) discrepancy. A case history of the last condition is showed in the Figure 1, where a discrepancy is caused by an industrial plan into the residential zone. Figure 1: Example of the urban area where PRAC is mandatory: case of

borderline connection between urban areas having in PCA quality levels with more than 5 dB (A) discrepancy, caused by an industrial plan into the residential zone.

The plans are also managed by different authorities who are required to deliver a fully exhaustive result homogeneous across the whole territory. From this point of view, it is easier to check the need and the validity of a production site improvement plan compared with the plan for a transportation infrastructure where is far more difficult to define the right counterpart authority: at municipal level the local government is competent for the whole territory except the pertinence bands of the transportation infrastructures. As mentioned, the municipal improvement plan process can be activated by the administrations in case of need (limits overcoming or acoustically non compatible boundary conditions) but also when an acoustic quality target is to be




achieved. This characteristic is in line with and anticipates the UE regulation, which supports the need to preserve the good acoustical areas. Indeed, the authority intervention is discretional when the local administration wants to reach a fixed quality level, even if the warning thresholds are not overcome or he needs to co-ordinate and to manage different private or/and infrastructural acoustical improvement plans (i.e. Figure 2).

Figure 2: Case history of a discretional PRAC of an urban area: the residential part of the town does not need any improvement actions, but the industrial and the railway noise sources require specific acoustical improvement plans.

The Italian legislation addresses the general PRAC contents and the activation procedures and criteria are still undefined. The PRAC is by definition a dynamic instrument whose effectiveness depends on the checking and updating work to be carried out according with a pre-defined timing, to allow the efficient revision of the achieved results (and their validity) and of course to monitor the on going activities. The main open points refer to inter-functional relationship with the transportation infrastructures and private companies and the correlation mechanisms with the other town planning and territory management tools.

Industrial improvement plan Railway improvement plan

Acoustic classification plan




The Italian law deals with the transport noise pollution [4], allowing in the pertinence band of roads and railways (until 100 – 250 m of wildness from the centre of the infrastructure line) emission values higher than safety noise limits as defined in the PCA. Therefore, from the legal point of view, if the noise does not exceed the legal infrastructure noise limits, the PRAC is not mandatory, also in case of strong noise exposure and pollution. However, if the objective is to reach the safety conditions of people exposure, by respecting the noise limit values as indicated by WHO, the PRAC is needed and is a precondition in any situation of noise pollution, even if the legal limits are formally respected. In figure 3 is showed a case history of an urban area where one national road, one highway and one railway generate, in particular in the night period, the noise values incompatible with the people safety conditions, without exceeding the Italian law limits for transport noise. In this case PRAC is not legally mandatory, but is the precondition to reach the required improvement in terms of people noise exposure.

Figure 3: A case history of an urban area where one national road, one highway and one railway generate, in particular in the night period, the noise values incompatible with the people safety conditions, without exceeding the Italian law limits for transport noise. In this case PRAC is not legally mandatory, but is the precondition to reach the required improvement in terms of people noise exposure.

City Plan Acoustical

Classification Plan

Road, highway and railway noise

pertinence bands




5 How to integrate the national and UE action plan levels?

It is evident that, in urban areas strongly affected by the transportation noise pollution (i.e. figure 4), no effective short/medium term prevention are actually achievable and therefore it becomes mandatory to manage the territory development together with the infrastructures to and obtain at least long term improvements in population exposure conditions. This approach had been adopted by UE requiring jointed actions from the national communities.

Figure 4: A case of an urban areas strongly affected by the transportation noise pollution where it becomes mandatory to manage the territory development together with the transport actions.

The tool for acoustic pollution and the associated problems management is defined in the action plan, which must be developed to account for urban areas and for the transportation infrastructures, with a timing plan depending on the dimensions of the involved area. The UE standards do not contain a detailed specification of a plan, but underline the priorities whose identification depends on limits overcoming. The application of such a tool is related with the strategic mapping which is the knowledge instrument intended to define the acoustic situation. The relationship between the National and the European tools is realized with the acknowledgment of the DIR 2002/49/CE [5] through the legislative decree 19 AGOSTO 2005, N.194 [6]. Such a decree defines the competencies and the procedures for the elaboration of the acoustical mapping and of the acoustical strategic maps, as fundamental instruments to define the existing acoustic situation. The Acoustical Strategic Maps is a map finalized to the definition of the global exposition to noise in a certain area due to various noise sources, or to define the general forecast for such an area.

Road, highway and railway noise pertinence band ( ____60dB(A); ____ 70dB(A))




The Acoustic Mapping (which is required for transportation infrastructures) consists in the representation of data describing an existing or forecasted noise situation in an area, generated by a specific source, as a function of an acoustic parameter showing the overcoming of the applicable limits together with the number of exposed people or the number of houses exposed to a determined level. From the above definitions it is clear that the acoustical strategic map represents the noise emissions while the acoustic map describes the acoustical emission of specific sources. The acoustic mapping and the strategic acoustical maps finally represent data describing an existing noise situation and the relevant exposure: both types can be presented as graphs or diagrams. The directive application field appears different from the Italian laws: if, according the national law, it is necessary to design healing plans for all the Municipalities (consequent to the application of the acoustic classification of the territory) in the European normative action plans are demanded for the big agglomerates and for the transport infrastructures. Differences are minimal for the infrastructures, for which a study of the acoustical emission (also in term of acoustical map) and a healing plan is required: the European directive doesn’t include the road infrastructures < 3000 vehicles/year and the rail infrastructures < 30.000 passages/year. There are important differences between European and Italian law considering about the environment of life, because the Italian law includes all the towns in the necessity of the healing plans, but are included in the European Directive only the agglomeration having population in excess of 100.000 persons and a population density that the Member State considers it to be an urbanised area. The distribution of the Italian population, divided in little or medium dimension town, should implicate a reduced impact for the European norm, with many realities excluded by an intervention of acoustical healing. In spite of what above stated, the action plans defined by UE should constitute the guide lines about the transport noise analysis for what refers to the acoustical improvement plan (PRAC) and represent an important integration of the land management in all urban areas.

6 Conclusions

The aim of the European Directive and of the Italian law shall be to preventing and reducing environmental noise where necessary and to preserving environmental quality where it is good. The analysis of the environmental noise and its representation in term of maps, the evaluation of the population noise exposure are the novelty in the Italian regulation. Because the objective is to reach the safety conditions of people exposure, by respecting the noise limit values as indicated by WHO, the acoustical




improvement plan (PRAC) is needed and precondition in any situation of noise pollution, even if there is the formal respect of the legal limits. The healing plan, beyond the content of the single measure, express the theme of the acoustical quality in the local administration’s culture and it represents a systematic instrument of political addresses and concrete actions, to support the choices of territory management and its activities. In fact, a territorial planning integrated with the acoustical management instruments, through a correct urbanite planning, should permit to eliminate from the beginning the critical situations. The healing plan is not an instrument to improve compromise situations or a set of passive protection interventions on the buildings. These solutions are obviously necessaries where an incorrect urbanite planning, in acoustical terms, has generated situations solvable only with a rilocalization, not easily realizable.

References

[1] ANPA Agenzia Nazionale Per La Protezione Dell’Ambiente “Linee guida per l’elaborazione di piani comunali di risanamento acustico”, a cura di – ANPA , APPA Bolzano, APPA Trento, ARPA Emilia Romagna, ARPA Liguria, ARPA Valle d’Aosta, ARPA Veneto, ARPA Toscana, Regione Lombardia, 1998

[2] Alberto Muratori “Piani di risanamento acustico: dimensione amministrativa, pianificatoria e normativa”, Convegno Nazionale “I piani di risanamento delle aree urbane”, Modena 22 – 23 febbraio 1999, in Atti pp 3- 17

[3] Jacopo Fogola, Rosario Romano “Piani d’azione e piani di risanamento acustico”, Convegno Nazionale “La direttiva 2002/49/CE: Determinazione e gestione del rumore ambientale e suo impatto sulla legislazione italiana”Pisa, 18 novembre 2004, in Atti pp. 71-81.

[4] Decreto del Ministero dell’Ambiente 29/11/2000 “Criteri per la predisposizione, da parte delle società e degli enti gestori dei servizi pubblici di trasporto o delle relative infrastrutture, dei piani degli interventi di contenimento e abbattimento del rumore”

[5] Direttiva 2002/49/CE del Parlamento Europeo e del Consiglio relativo alla “Determinazione e gestione del rumore ambientale”

[6] Decreto legislativo 19 agosto 2005, n.194 “Attuazione della Direttiva 2002/49/CE relativa alla determinazione e alla gestione del rumore ambientale”




Estimation of the on-farm-costs of soil erosion in Sleman, Indonesia

A. Möller & U. Ranke Federal Institute for Geosciences and Natural Resources (BGR), Hannover, Germany

Abstract

Soils are non-renewable resources. World-wide in many regions a sustainable use of soils is endangered through anthropogenic accelerated soil erosion. From the economic point of view erosion protection is the transfer soil use potential into the future. However, in developing countries such as Indonesia usually only short term profit counts and consequently soil resources suffer from accelerated exploitation. Advising farmers that soil erosion protection measures not only ensure a prolonged agricultural potential for the future, but can also include economic benefits is a promising attempt to promote the use of soil erosion protection measures. The prerequisite for the decision process on a farm level is the possibility to estimate the costs of soil erosion. In Indonesia mostly reliable data are missing, and also no expensive surveys can be accomplished. Therefore, less data intensive methods such as the “replacement cost” or the “productivity change” method were used to estimate the “on-farm-costs” of soil erosion in Sleman on Java. The “replacement cost” method resulted in clearly higher costs compared to the “productivity change” method. This is due to an over-estimation of the costs by the “replacement cost” method. The use of both methods comparing costs and benefits of soil protection measures indicate similar decision guidelines. However, more information is necessary on the additional effects of soil conservation and political constraints to be a base for sound decision-making on a farm level, but making information available on the benefits of conservation measures helps farmers in their decision process to invest in soil conservation. Beyond this, soil erosion is also a societal problem, including external costs making up a large portion of the economic effects of soil erosion. Keywords: soil erosion, on-farm costs, replacement cost method, productivity change.




doi:10.2495/GEO060061

1 Introduction

Land conversion within the developing world is occurring at an unprecedented rate. Expansion of subsistence farming practices in the form of field crop agriculture and pasture within rural areas is contributing significantly to ecological alteration in many tropical countries (Uhl et al. [1]; Landa et al. [2]; Lubchenco [3]).

Soil erosion is thereby a major environment threat for the sustainability and productive capacity of agriculture. During the last 40 years, nearly one third of the world’s arable land has been lost by erosion and continues to be lost at a rate of more than 10000 million hectares per year (Pimentel et al. [4]).

In Indonesia more than 50% is mountainous and consequently highly vulnerable towards soil erosion. On Java about 23.7 million hectares are stated to be “critical” land concerning soil erosion.

Erosion adversely affects soil quality and productivity by reducing nutrients, infiltration rates, water-holding capacity, organic matter, soil biota, and soil depth. Several studies have shown a reduction of soil productivity in the long term between 2 and 70% for many soils (Wolman [5]). The main reduction of soil productivity in the short term is thereby due to the loss of soil nutrients and water availability. In the long term the loss of soil depth, water holding capacity and organic matter can contribute largely to the loss of soil productivity.

While it is widely accepted that erosion lowers agricultural productivity, there is little agreement on exactly how productivity is related to erosion or on the quantitative impact of erosion on yields (Magrath and Arens [6]). Erosion involves changes in the availability and relative concentration of nutrients for plant growth and changes in the soil structure which influences root growth and affects the availability of water.

On the other hand, soils get differently affected by erosion based on their individual fertility. They range form soils whose natural fertility is accumulated, along with the soil organic matter, in the top few centimetres only, to soils being fertile throughout the whole profile. Furthermore, different crops get differently affected by a potential loss of soil productivity. Demanding crops may react with high yield reductions, while non-demanding crops like Alfalfa may only be little affected.

In measuring the on-site costs of soil erosion the main objective is usually to estimate the present value of net income lost through excessive (i.e. sub-optimal) soil erosion. According to Barbier [7], to be an economic cost, the onsite costs of soil erosion must be an opportunity cost, which is defined as the value of a forgone alternative like the investment in soil conservation. Because soil conservation is not costless, the on site cost of soil erosion must be the loss in the long-run net profitability of the farming system not investing in soil conservation, providing of course that such an investment is an economically worthwhile alternative. The on-site costs of soil erosion are than the difference between the net returns of the farming system with soil conservation and the net returns with erosion (Barbier [7]).




Although the methodology seems to be the best choice to estimate the on-site costs of soil erosion, it has often proven to be very difficult to implement empirically. It is not easy to determine an economically viable alternative conservation investment to current erosive practices. Particularly, in developing regions such as Southeast Asia, with diverse and heterogeneous small-scale farming systems, the data constraints are often enormous, whereas simplifying assumptions and generalizations may be misleading.

Therefore, the On-site costs of soil erosion were estimated in this study based on alternative empirical models determining changes in soil productivity, or the costs to replace the lost nutrients, water, eroded topsoil or organic matter (Kim and Dixon [8]; Magrath and Arens [6]; Dixon et al. [9]; Gunatilake and Vieth [10]; Krausse et al. [11]).

These approaches may by less reliable or even second-best from an economic perspective, but they were the only implementable choices based on the data available. Especially the data required for the replacement cost approach are easier to generate in developing countries.

2 Site description

The Kabupaten Sleman is situated in the northern part of the province of Yogyakarta (Figure 1). It is located at the southern flank of the Merapi volcano starting almost from see level in the south to the top of the volcano at an altitude of 2986 m.

The Climate in Sleman is humid tropical with a distinct dry season from Mai until October. The average annual rainfall rages between less than 1500 mm and 3300 mm. Heavy rainfalls with more than 100 mm per day or within a period of three days are common. These storms are a major driving force of heavy soil erosion and can trigger Lahars (mud flows) at the upper slopes of Mt. Merapi.

Figure 1: Topographic map of the Province of Yogyakarta on Java.




Topographically Sleman can be divided into two major areas, the upper and medium slopes of the Mt. Merapi and the foothill area below. The upper slope represents the young Merapi cone with devastated to bare land and some thin gullies. The medium slopes comprise a complex of old and young Merapi products forming deep valleys, mainly in pyroclastic deposits. Smooth slopes and shallow valleys characterize the foothill area consisting of fluvial deposits.

The major soils according to US soil classification in the area are Andisols and Mollisols at the upper slopes of the Mt. Merapi and Inceptisols at the medium slopes and the floodplains of Mt. Merapi. Some Entisols and Vertisols can be found in the mountainous areas in the south of Sleman.

Land use in Sleman is stamped by a vertical zonal distribution according to the increasing slope of the Merapi volcano. At lower slopes from about 0 to 5% paddy fields dominate the mostly agricultural used area. From about 5 to 10% slope mainly rain feed agriculture of vegetables and cash crops can be found. Weather at slopes greater than 10 % the land use is dominated by forest, shrubs and grassland.

3 Estimating soil erosion costs

To estimate the economic significance of soil erosion its physical dimension has to be determined, and linked and valued to changes in crop production and farming systems.

In this study soil erosion was determined using the Universal Soil Loss Equation (USLE). The model is widely used in science to estimate soil erosion at multiple scales (Renard et al. [12]; Turnage et al. [13]), and also commonly used with some adoptions under tropical conditions (Millward and Mersey [14]; Wiriosudarmo and Bisri [15]; El-Swaify [16]).

3.1 Replacement cost method

The so-called replacement cost approach estimates the forgone input which is necessary to overcome the negative effects of soil erosion (Kim and Dixon [8]; Dixon et al. [9]; Gunatilake and Vieth [10]; Krausse et al. [11]).

Usually only the fertilizer replacement as major costs is considered. Thus, the replacement costs can be seen as the costs to replace the lost nutrients and the additional energy, maintenance and labour work to apply the extra fertilizer to the fields. In the ith land use of an area it can be presented as eqn (1):

∑ ++−= + iriljijtti CCPNSSRC )( )1( (1)

kjni ...1,...1 == where:

iRC is the replacement cost of nutrients in ith category of land use, Rp/ha

)1( +− tt SS ist he soil loss from time t to t+1, t/ha




ijN is the quantity of jth nutrient in ith land use type, kg/t

jP is the price of jth nutrient, Rp/kg

ilC is the cost of labor in spreading fertilizer, Rp/ha

irC is the cost of repair and maintenance of damages due to soil erosion Rp/ha

Included in the calculation of the replacement costs are the major nutrient Nitrogen, Phosphorous and Potassium. Nutrient levels of eroded soil are usually not available. Therefore it is assumed that the nutrient level in the eroded soil and the farm soil are the same. Based on the assumption the amount of nutrients lost is calculated using representative soil nutrient analyses from the farm soils.

3.2 Productivity change method

The change in productivity approach determines the difference in crop yields with and without erosion, multiplied by the unit price of the crop, and less the variable costs of the production (Magrath and Arens [6]; Gunatilake and Vieth [10]). Although this seems straightforward and simple, in practice the quantification of the effect on crop yield losses is conceptually difficult. In this study the approach of Magrath and Arens [6] introduced for Java was used to quantify the changes in productivity.

They assumed that if output falls farmers adjust variable inputs in production to yield declines and that fixed costs remain fixed. Percentage productivity declines are denominated base on the response of sensitive and less sensitive crops. The result of this procedure is a linear decline in profits as productivity falls. To account for possible adjustments in cropping systems, farm budgets for a variety of representative dry land cropping systems across Java were constructed, and then used to estimate the effects of the yield losses from erosion on net farm incomes. This was done comprehensively for a single year.

They have estimated an average yield reduction on Java between 4–7%, depending on soil type and crops planted. Using their results the costs for reduced yields were calculated by relating the lost yields to the average cross margin of the agricultural production in the region, based on prices of important cash crops and vegetables in the District Yogyakarta in 2002.

4 Results and discussion

4.1 Soil erosion

On most of the cultivated area in Sleman bench terraces or at least raised bed terraces are used. This is reflected by low erosion rates < 5 t ha-1 yr-1 (Figure 2). At the upper part of the Merapi volcano at areas with steep slopes and badly maintained or no terraces erosion rates exceed by fare a sustainable level. Here erosion rates with more than 100 t ha-1 yr-1 can be found at cultivated areas.




Figure 2: Map of the average annual soil erosion in the District of Sleman.

Areas with particularly high erosion rates at the top of the volcano with up to 500 t ha-1 yr-1 are not cultivated, but consist of loose pyroclastic deposits triggering debris flows during heavy rains in the rainy season.

Similar erosion rates were found in other studies carried out on Java. Magrath and Arens [6] has estimated for steep slopes on Java erosion rates up to 500 t ha-1 yr-1 with an average of 123 t ha-1 yr-1 for “Tegal” land use (rain fed agriculture) and 87 t ha-1 yr-1 for degraded forests. Kusumandri and Mitchell [17] found soil erosion rates for the Citarik watershed on West Java to be about 100 t ha-1 yr-1.

Although at most of the agricultural land in Sleman already some soil conservation is practiced, some areas at the upper slopes of the volcano and at some hilly parts in the south have significant soil erosion problems with more than 15 t ha-1 yr-1. The area affected is 2737 ha with an average soil erosion rate of 30 t ha-1 yr-1 (Figure 3). These areas took center stage in the estimation of the costs of soil erosion in Sleman.

4.2 Soil erosion costs

The costs to replace lost nutrients in these agricultural areas were estimated to be 14100 Rp t-1 soil, taking into account an average soil nutrient content of N = 1.1 kg t-1, P = 0.8 kg t-1, and K = 3.7 kg t-1 and prices of common fertilizers in 2003




(Urea = 1000 Rp; TSP = 1000 Rp; KCl = 1200 Rp). Thus, the replacement cost with an average soil erosion rate of 30 t ha-1 yr-1 is 423000 Rp ha-1 yr-1. The additional costs for energy, maintenance and labour work were estimated based on results from literature and expert experience to be 85000 Rp ha-1 yr-1 or about 20% of the replacement costs. Thus, the total costs were estimated to be 508000 Rp ha-1 yr-1. However, the approach overestimates soil erosion costs, based on the conceptual assumption estimating the difference between erosion and “zero” erosion and the assumption that all nutrients lost would be available for plants in the long term, which is in reality in agriculture not realizable. On the other hand other effects like the loss of organic matter or water holding capacity are not considering.

On the bases of the change in productivity approach of Magrath and Arens [6] the average productivity loss at agricultural areas with significant erosion was calculated to be 160000 Rp ha-1 yr-1. This seems to underestimate the actual costs of soil erosion. However, Magrath and Arens [6] “capitalized” the one year cost of erosion by a factor of 10 to obtain a total present value of current and future losses, assuming that one year loss in net income recurs over each successive year. On the other hand considering only plant available nutrients lost (N = 1.1 kg t-1; P = 0.45 kg t-1; K = 0.8 kg t-1) within the replacement cost approach the average costs are comparable, with 250000 Rp ha-

1 yr-1 (Figure 3). These costs are equal to ≈ 17 % of the average farmers net income per ha agricultural land.

A similar order of magnitude of soil erosion costs was reported by Krausse et al. [11] and Gunatilake and Vieth [10] for agricultural soils in New Zealand and Sri Lankan high land soils, respectively. Krausse et al. [11] estimated the actual costs for soils suffering significant erosion in New Zealand to range between Aus$ 8 and Aus$ 25 ha-1, with an average erosion rate of 10 t ha-1 yr-1. Considering the average erosion rate of eroded soils in Sleman (30 t ha-1 yr-1) this would approximate between 135000 Rp and 420000 Rp ha-1 yr-1. Gunatilake and Vieth [10] estimated slightly higher soil erosion costs depending on the type of crop and the erosion rate with e.g. ≈ 200000 Rp (21-25 USD) for paddy fields and ≈ 600000 Rp (68 USD) for market gardens.

On the other hand to minimise soil erosion rates soil conservation measures are necessary, which are not cost less. Adiningsih and Karama [18] estimated the additional annual costs for bench terraces and raised bed terraces compared to conventional farming practices in East Java to be 113 USD (≈ 1 million Rp) and 56 USD (500000 Rp), respectively. Thus, the costs for terracing are not covered by the benefits from reduced erosion alone. Other potential benefits like the possibility to intensify/change the agricultural production coming along with conservation measures have to be considered as well. Quantifying these benefits is very difficult, but point based studies comparing the net income of farms with and without conservation measures indicate that adequate soil conservation measurements can be economically worthwhile (Adiningsih and Karama [18]; Posthumus and De Graaff [19]). Adiningsih and Karama [18] showed that the net income of the farmers was by a multiple higher after changing to an integrated farming system with terraces.




0.125 - 0.250.25 - 0.50.5 - 0.750.75 - 1.01.0 - 1.51.5 - 2.0over 2.0

Figure 3: Map of the On-farm cost of erosion (at agricultural areas with significant soil erosion [million Rp ha-1a-1]).

Adventitiously, other decision factors like the availability of credits or the fact that in many countries soil conservation is not reflected in land prices makes it difficult or impossible for a farmer to decide if it is worthwhile to invest in a certain soil conservation measure. Nevertheless, information available on the costs of soil erosion and on possible benefits is an important economic factor to help farmers in their decision to invest in soil conservations measure.

Besides, not reflected in farmer’s decision-making are off-site or external costs of soil erosion, but they play an important part of the economic impact of soil erosion. In many studies off-site cost are estimated to be higher than on-site costs. About these costs, which are not reflected by the markets prices, the society has to be concerned and against the background that decisions on pure economic basis usually only consider the next maximum 50 years, but sustainability of soil resources is a matter of the next centuries, soil conservation has to be a general goal for the society and can’t be shouldered by the farmers alone. However, it is not easy to design appropriate policies to include the off-site or external costs into the decision-process of soil conservation on farm level.




Acknowledgements

The project was funded by the Federal Minister for Economic Cooperation and Development, Germany. The Authors acknowledge the contribution of the Directorate General of Geology and Mineral Resources of Indonesia to the project.

References

[1] Uhl, C.J., Carl, C., Clark, H., Herrera, R., Ecosystem recovery in Amazon Caatinga forest after cutting, cutting and burning, and bulldozer clearing techniques. Oikos 38(3), pp. 313–320, 1982.

[2] Landa, R., Meave, J., Carabias, J., Environmental deterioration in rural Mexico: an examination of the concept. Ecological Applications 7(1), pp. 316–329, 1997.

[3] Lubchenco, J., Entering the century of the environment: a new social contract for science. Science 27(5350), pp. 491–497, 1998.

[4] Pimentel, D., Harvy, C., Resosudarmo, P., Sinclair, K., Kurz, D., McNair, M., Crist, S., Shpritz, L., Fitton, L., Saffouri, R., Blair, R., Environmental and Economic Costs of soil Erosion and Conservation Benefits. Science 267(5126), pp. 1117-1123, 1995.

[5] Wolman, M.G., Soil erosion and crop production: A worldwide perspective. Soil erosion and crop productivity, eds. R.F. Follett & B.A. Stewart, ASA, CSSA & SSSA: Madison, pp. 10-22, 1985.

[6] Magrath, W. & Arens, P., “The Costs of Soil Erosion on Java: A Natural Resource Accounting Approach”, Environment Department Working Paper No. 18, Washington D.C.: The World Bank, 1989.

[7] Barbier, E.B., The Economics of Soil Erosion: Theory, Methodology and Examples. Special Papers, Fifth Biannual Workshop on Economy and Environment in Southeast Asia, Singapore, 1995.

[8] Kim, S.H., Dixon, J.A., Economic valuation of environmental quality aspects of upland agricultural projects in Korea. In: Doxon, J.A., Hufschmidt, M.M. (eds.). Validation techniques for the environment: A case study workbook. Baltimore: Johns Hopkins University, 1986.

[9] Dixon, J.A., Scura, L.F., Carpenter, R.A., Sherman, P.B., Economic analysis of the environmental impacts. London, Earthscan Publication Ltd, 1994.

[10] Gunatilake, H.M., Vieth, G.R., Estimation of On-site Cost of Soil Erosion: A Comparison of Replacement and Productivity Change Methods. Journal of soil and water conservation 55(2), pp. 197-204, 2000.

[11] Krausse, M., Eastwood, C., Alexander, R.R., Muddied waters Estimating the national economic cost of soil erosion and sedimentation in New Zealand. Manaaki Whenua landcare research, Palmerston North, 2001.

[12] Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., Yoder, D.C., Predicting soil erosion by water: a guide to conservation planning with the




Revised Universal Soil Loss Equation RUSLE. Handbook No. 703. US Department of Agriculture, 404 pp., 1997.

[13] Turnage, K.M., Lee, S.Y., Foss, J.E., Kim, K.H., Larsen, I.L., Comparison of soil erosion and deposition rates using radiocesium, RUSLE, and buried soils in dolines in East Tennessee. Environmental Geology, 29, pp. 1-9, 1997.

[14] Millward, A.A., Mersey, J.E., Adapting the RUSLE to model soil erosion potential in a mountainous tropical watershed. Catena, 38, pp. 109-129, 1999.

[15] Wiriosudarmo, S., Bisri, D.H., Use of environmental geological information to plan erosion control in the upper part of the Citanduy catchment, West Java province, Indonesia. Sixth Regional Congress on Geology, Mineral and Hydrocarbon Resources of Southeast Asia, 1987.

[16] El-Swaify, S.A,. Susceptibilities of certain Tropical Soils to Erosion by Water. In: Greenland, D.J. and Lal, R. Soil Conservation and Management in the Humid Tropics. John Wiley & Sons, New York, 1977.

[17] Kusumandri, A., Mitchell, B., Soil erosion and sediment yield in forest and agroforestry areas in West Java, Indonesia. Journal of Soil and Water Conservation, 52(4), pp. 376-380, 1997.

[18] Adiningsih, J.S., Karama, A.S., A sustainable upland farming system for Indonesia. http://www.fftc.agnet.org/library/abstract /eb354b.html, 1992.

[19] Posthumus, H., De Graff, J., Cost-benefit analysis of bench terraces, a case study in Peru. Land Degradation & Development, 16(1), pp. 1-11, 2004.




Section 2 Environmental modelling

and monitoring

Large scale soil erosion modeling for a mountainous watershed

P. Zhou1, J. Nieminen2, T. Tokola 2, O. Luukkanen1 & T. Oliver3 1Viikki Tropical Resources Institute, University of Helsinki, Finland 2

3University of Barcelona, Spain

Abstract

Soil erosion control requires a quantitative evaluation of potential soil erosion on a specific site. The Revised Universal Soil Loss Equation (RUSLE), Remote Sensing (RS), and Geographic Information System (GIS) were used to model soil erosion intensity for soil conservation and vegetation rehabilitation in an Upper Min River (UMR) watershed, which is in the Upper Yangtze River basin. Data used in this study to generate the soil loss were Landsat Enhanced Thematic Mapper (ETM) images, Digitized Elevation Model (DEM), soil erodibility, rainfall erosivity, and inventory data. The non-parametric k-nearest neighbor (k-NN) method was used to produce the cover management map by integrating the ETM images and vegetation coverage data measured in the 625 sample plots. The root mean square errors and significance of biases at pixel level were evaluated in order to find optimal parameters. Four raster maps have been produced for the soil erodibility, rainfall erosivity, slope length and steepness, and cover management factor, and the map with different soil loss risks has been produced for soil erosion potential. The result can be beneficial to the erosion control and ecological restoration in the degraded mountainous watershed. Keywords: soil erosion, RUSLE, DEM, k-NN method, Upper Min River Watershed.

1 Introduction

Soil erosion is a worldwide environmental problem that degrades soil productivity and water quality, causes sedimentation and increases the




doi:10.2495/GEO060071

University of Helsinki, Finland Department of Forest Resource Management,

probability of floods. The 1998 flood in the Upper Yangtze raised public attention to the problems of soil erosion and sedimentation. Soil loss control calls for a quantitative evaluation of potential soil erosion on a specific site. Scientists have studied different methods to assess soil erosion loss by water, for instance, universal soil loss equation (USLE) [1], revised soil loss equation (RUSLE) [2], the 137Cs technique [3], and WEPP Hillslope model [4]. Among them, the USLE and RUSLE provided a convenient tool for soil loss evaluation by taking the climate, geographical terrain, conservation support practice, soil, and vegetation into considerations. The RUSLE was developed to incorporate the considerable amount of erosion information and to address specifically the application of the USLE to land uses other than agriculture [2]. The model can be used to any geographic region by modifying its factors. The factors are rainfall runoff erosivity factor, soil erodibility factor, slope length and steepness factor, cover management factor, and support practice factor [1, 2]. An important reference manual for applying the USLE to disturbed forestlands is “A guide for predicting sheet and rill erosion on forest land” [5]. The USLE and RUSLE have been used widely in evaluating the soil erosion risk in watershed and highland [6, 7, 8]. Slope length and steepness factor, which reflects the terrain on a given site, can be computed from the digital elevation model (DEM) [9, 10]. Rainfall and runoff erosivity factor was calculated based on the storm events and rainfall data in many studies [6, 7, 8, 11]. However, in mountainous watershed, orographic effects caused by mountainous terrain can result in a significant positive correlation between precipitation and elevation [12, 13, 14]. In the UMR watershed, precipitation tends to increase with an increase in elevation because of the orographic effect of mountainous terrain and the foehn effect [15]. In our study area, we examined the relationship between elevation and precipitation for 38 stations, and selected cokriging as a method for estimating average annual precipitation of the whole watershed. Cover management C-factor in the soil loss equation was defined as the ratio of soil loss from land cropped under specified conditions from the corresponding loss from clean tilled, continuous fallow [1]. However, in large scale UMR watershed, where are not mainly covered by agricultural lands, the cover management factor is not only affected by the agricultural crops. Ma et al. used the proposition of vegetation reflectance in pixel end members to calculate C factor [7]. In our study, we used the non-parametric k-nearest neighbour (k-NN) multi-source estimation method to estimate coverage data and produce the coverage map by integrating the satellite images and field data with optimal parameters. The k-NN method has been widely used in a variety of forest estimation and biomass mapping applications over the years [16, 17, 18, 19], and therefore, can be applied in vegetation cover estimation. Lu et al. [20] explored the relationships between the soil erosion and land use and land cover distribution, they found that most climax and mature forests are in low erosion risk areas, while agroforestry and pasture are usually associated with medium to high risk areas. A good plant cover is generally capable of preventing surface erosion, and reducing landslides as well. Removal of vegetation can greatly increase runoff and soil erosion particularly in




mountainous areas [21]. Soil erosion control especially calls for the forest restoration or rehabilitation to reduce the erosion loss and improve soil stability. The questions we are going to answer in the paper are: (1) How to model the soil erosion loss in this mountainous watershed? and (2) How much is the soil erosion risk in the area?

2 Study area

The Upper Yangtze River Basin is a mountainous region, which has an area of 1.04×106 km2, a mean annual runoff discharge of 4.35×108 m3, a mean sediment yield of 5.17×108 t and a population of 1.4×108 [22]. The basin is one of the most severely eroded areas in China. Water erosion results in both on-site soil degradation and off-site problems related to downstream sedimentation [23]. The Upper Min River, which is one of the most important tributaries of the Upper Yangtze River, is 341 km long with a drainage area of 23,037 km2. The watershed is located in Sichuan Province, South West China. The area is governed by the southeast and southwest monsoons. The complex topography, with elevations ranging from 900 m to 5 700 m, results in steep gradients of rainfall. The Upper Min river watershed has been divided into five ecozones: the Sub-tropical (1300–2200 m), Temperate (2200–2600 m), Sub-alpine (2600–3200 m), Boreal (3200–3600 m) and Arctic zone (3600–5700 m) [24]. At present, the forest cover is around 21% of the whole watershed area. Our 625 inventory plots were randomly placed in the middle and upper reaches of the UMR watershed, between 31º-34º N, 103º-104º E, with an area of about 7 400 km2, see fig. 1. The vegetation ranges from subtropical evergreen broadleaved forest to the alpine meadows.

3 Method

3.1 Model structure

The soil loss (A) due to water erosion per unit area per year (Mg ha-1yr-1) was quantified using RUSLE by the following equation:

PCLSKRA ××××= (1) where A is the average soil loss due to water erosion, R the rainfall and runoff erosivity factor (MJ mm ha-1h-1yr-1), K the soil erodibility factor (Mg h MJ-1

mm-1), L the slope length factor, S the slope steepness factor, C the cover and management practice factor, and P the support practice.

3.2 Data and processing

3.2.1 Rainfall and runoff erosivity factor (R) R is the long term annual average of the product of event rainfall kinetic energy (E) in MJ ha-1 and the maximum rainfall intensity in 30 minutes (I30) in mm h-1.




The R values were correlated with annual precipitation [1, 25]. We used the following equation to calculate the R factor, which has been adopted for application in the RUSLE model [2]:

2006661.00334.0 aa PPR +−= (2) where R is rainfall and runoff factor (MJ mm ha-1h-1 yr-1), and aP the measured annual precipitation in mm. The average annual precipitation (APP) and elevation data from 38 meteorological stations in the research area were obtained to check the correlation between precipitation and elevation. The APP surface was interpolated with a multivariate geostatistic cokriging model [26]. The R factor surface was then calculated by eqn. (2) from the APP surface using the raster calculation in spatial analyst.

Figure 1: Upper Yangtze River watershed and sample plots in the research area.

( • )showed the 625 sample plots, ( ) showed the drainage net work, ( ) showed the Upper Yangtze River basin, and ( ) showed the detail research area.




3.2.2 Slope length and slope steepness factors (LS) L factor and S factor, which reflect the topographic erosion susceptibility on a given site, were computed together from the digital elevation model (DEM). The DEM used is based on a digital topographic map, with 100-m elevation contour lines and stream data. In order to achieve a geomorphological realistic surface, it was interpolated to a 25-m cellsize grid with the Topogrid algorithm [27] which generates a hydrologically correct grid DEM using contour lines and stream data. The slope was calculated using the maximum downhill direction method, in which the slope value for each raster cell is obtained from the angle formed between the cell itself and the lowest neighboring cell. The flow direction was calculated with the D∞ (infinite directions) method developed by Tarboton [28], by which dispersed or rilled flow is estimated for each cell from the slopes to the lower neighboring cells. In flat areas of the DEM (no lower neighboring cells), the method approached by [29] was used to calculate flow direction. Flow accumulation, the number of cells contributing with its flow to each cell was calculated from the flow direction raster. The DEM sinks filling the slope angle, the flow direction and the flow accumulation were calculated by Taudem, an ArcGIS 9.0 extension developed by Tarboton. For this project, an approach developed by Moore and Burch [9, 10] was used to compute LS factor:

4.1××= SLLS (3)

mcsfaL )13.22/( ×= (4)

where LS is computed slope length (L) and slope steepness (S) factor, fa the flow accumulation (contributing area or upslope area), cs the cellsize, and m the slope-length exponent, as explained in the equation:

)1/( ββ +=m (6) where β is the ratio of rill to the interrill erosion for conditions when the soil is moderately susceptible to both, and is computed byθ with the following equation [30]:

[ ]56.0)(sin0.3/)0896.0/(sin 8.0 +×= θθβ (7) Whereθ is slope angle in degree. Table 1 shows the values for m computed from eqn (6) and (7), and applied to eqn (4) to calculate a raster map for L factor. S is calculated by the following equations:

03.0sin10 +×= θS If slope < 9 percent (8) 05.0sin8.16 −×= θS If slope ≥ 9 percent (9)




Table 1: Calculation of m value from angle.

θ angle in degree m value θ <1° m=0.2 1°≤ θ <2° m=0.3 2°≤ θ <3° m=0.4 3°≤ θ <6° m=0.45 6°≤ θ <10° m=0.55 10°≤ θ <26° m=0.65 26°≤ θ m=0.7

3.2.3 Soil erodibility factor (K) K factor is soil erodibility factor, which represents both susceptibility of soil to erosion and the rate of runoff. Specifically, the k factor is a function of particle size distribution, organic matter content, structure and permeability [1, 2, 11]. K was calculated using the equation recommended by Wischmeier and Smith [1]:

1317.0))3(025.0)2(0325.0)12(10)(1.2( 614.12 ×−×+−×+−×××+×= − stpeomsavfvfK (10) Where K is the soil erodibility factor (Mg h MJ-1mm-1), vf the percentage of very fine sand plus silt, sa the percentage of sand, om the percentage of organic matter, pe the permeability class, and st the structure class. In this study area, eqn (10) was used to calculate K value of each soil type. K values and the map of soil type were used to produce the raster map of K factor.

3.2.4 Cover management factor (C) Vegetation cover at three levels (canopy cover, under canopy cover, surface cover) was recorded from 625 sample plots. The canopy cover was measured by densiometer. The k nearest neighbor (k-NN) method was used to produce the canopy cover map and total vegetation cover map by integrating Landsat ETM+ image information and collected vegetation coverage. In this study we used two consecutive Landsat ETM+ scenes (WRS2 130/037 & 130/038, 10th July 2002 [31]) for the cover factor estimation. A set of parameters was chosen for the k-NN method in predicting the vegetation coverage map. The parameters were the image features, the weight for each band, the distance, the number of nearest neighbors the value of k, and the geographical reference area from which the nearest field plots are selected. The leave-one-out cross validation method is applied to calculate the root mean square errors (RMSE) and the average biases of predictions at the single pixel level for different combination of k-NN estimation. The RMSE and biases were used as a measure of reliability of the continuous variables. The cover management factor (C) then was calculated from vegetation coverage data using the equation recommended by Renard et al. [2].

)03048.0exp(1 HFCC c ×−×−= (11)




where CC is the canopy cover subfactor range from 0 to 1, cF is fraction of land surface covered by canopy, and H (m) is distance that raindrops fall after striking the canopy. The equation was used to calculate both the canopy cover subfactor and under canopy cover subfactor. For the former, we used the average weighted tree height 8.341m; and for the latter, we used the estimated shrub and grass height 0.5 m.

3.2.5 Support practice factor (P) P values range from 0 to 1, value 0 represents a very good manmade erosion resistance facility, and value 1 represents no manmade resistance erosion facility. In the study area, there are some agricultural support practices [7]. However, most of the farmlands in the study area are small and self-managed lands, and the spatial resolution of ETM+ imageries is 30 m, so it is impossible to distinguish the practices in a large scale watershed from the available data. We use P = 1 for all the lands.

4 Results

The historical precipitation data and station elevations from 38 meteorological stations were obtained to estimate the average annual precipitation (AAP) over the entire watershed. The AAP showed a significant (p < 0.01) correlation of r = 0.74 with the station elevation. A multivariate cokriging interpolation method was used in the analysis since it takes into consideration the elevation which significantly affects precipitation. Root mean square errors (RMSE) were calculated to investigate the estimation accuracy. The RMSE by cokriging estimation was 86.88 mm, which was reduced by 28.2% to kriging estimation (121.2 mm). The estimated AAP was used for calculation of rainfall and runoff erosivity R-factor in ArcGIS. The R factor varied from 1288 to 3342 MJ mm ha-

1h-1 yr-1 (fig. 2). The watershed occupied a raster grid space of 7700 rows by 2736 columns, and elevations ranged from 1261 m to 5537 m. Approximately 94.7 percent of the watershed has slopes steeper than 9 percent. Slope angles ranged from 0 to 77.2 degrees with a mean of 25.9 degrees and standard deviation of 12.0 degrees. As a result of applying Taudem, flow accumulation ranged from 1 to 15496180 m with 98% less than 1000 m, and slope length factor ranged from 0 to 3398, with a mean of 85.9 and 99.2% less than 120. The slope steepness S factor varied between 0.03 to 15.88 with a mean of 6.7 and a standard deviation of 3.1. The canopy cover map and total vegetation cover map were produced using k-NN method. The root mean square errors (RMSE) and the average biases of predictions at the single pixel level were evaluated for each combination of parameters. The value of k (8), the distance (55 km), the bands (1, 2, 3, 4, 5, 7) and their optimal weights were chosen when RMSE and bias were minimal. The C factor was calculated from the produced raster maps by using eqn (5). The cover management C factor ranged from 0.015 to 0.892 (fig. 2).




Figure 2: Factors for calculation of soil loss potential and the map of classified erosion risk.

The soil erodibility K factor was between 0.036 and 0.043 Mg h MJ-1mm-1 (fig. 2). The estimated soil loss for the research area varied from 325 to 83240 Mg ha-1 per year. According to the soil loss amount and field inventory result, we divided them into four ordinal classes: extreme risk (> 10000), high risk (3000 - 10000), moderate risk (1000 - 3000), low risk (< 10000) and No data (fig. 2). No




data values were assigned in two circumstances: firstly, excluded data with flow accumulation values higher than 1200 m (1.7% of the cells), which are coincident with the main stream paths; and secondly, excluded data with LS factor values higher than 1600 (0.0023% of the cells), which happened only in isolated cells with extremely high slopes and contributing areas. Totally 0.7% of the cells had No data value. Table 2 showed the area and proportion of each of the soil erosion potential categories. More than half of the watershed (58.1%) showed moderate, high or extremely high erosion risks. Table 2: Derivation of the ordinal categories of soil erosion potential and the

area and proportion of each category.

Numeric range (Mg ha-1yr-1)

Erosion potential Area (ha) Proportion (%)

0 - 1000 low 297007.7 40.2 1000 - 3000 moderate 306909.9 41.6 3000 - 10000 high 120109 16.3 >10000 extreme 9202.3 1.2 no data 4896.8 0.7

5 Discussion

In the UMR watershed, the average annual precipitation was positively correlated with elevation (r = 0.74, p < 0.01), which supports similar findings in Algarve (Portugal), southern Nevada and southeastern of California (r = 0.75, p < 0.05) [12, 13, 14]. The effect of elevation on precipitation can be used to improve the geostatistical interpolation. The RMSE by cokriging estimation was reduced by 28.2% to kriging estimation in the UMR watershed, and 54% reduction has been reported in Nevada and southeastern of California [13]. Slope calculations made with a maximum downhill method conserved the variability and the maximum slope values. This method produced no underestimation, since no averaging was used. Flow directions calculated by D∞ (infinite directions) improved significantly the water flow modelling, by allowing dispersed flow to be modelled over the surface. This method calculated the flow direction from the lowest continuous neighbouring cells and fractionated the water flow between them, simulating dispersed water flow and generating natural looking flow maps. Other studies mostly use the D8 approach method, by O'Callaghan and Mark [32]. However, D8 method produces unrealistically rilled water flow with lots of straight lines in flow accumulation maps, because it can only produce 8 different flow directions, to one of the neigbouring cells (cardinal or diagonal direction). The K values of the soils in our study area ranged from 0.036 to 0.043 Mg h MJ-1mm-1. Compared to the K values of tested soils in USLE (0.03 - 0.69 tons acre hr/hundreds of acre ft-ton in) [1], which are from 0.004 to 0.09 Mg h MJ-1mm-1, the soils in the UMR watershed have the moderate erodibility.




To find a suitable C-factor, a canopy cover and total vegetation surface cover were calculated using the k-NN technique for their estimation. Several methods were applied for topographic normalization of the imagery [33]. However, the elevation changes in the study area were so great (1254 m - 5527 m) that not all of the shadowing effects were removed from the imageries. To maximize spectral variability bands 1-5 and 7 were included in the analysis. The calibration of the k-NN parameters was performed as outlined by several articles dealing with forest estimation using k-NN methods [27, 34, 35]. The RMSE and significance of biases at sample plot pixel level were evaluated in order to choose the most optimal parameters, such as numbers of k, distance, and bands weights. The value of k (8) was chosen when the variation in bias and RMSE were minimal. The built in cross-validation method of bias and error estimation was applied in all calculations. Morgan [36] argues that 10 Mgha-1yr-1 is an appropriate boundary measure of soil loss over which agriculturists should be concerned. This was identified as the separation of the low and moderate categories in RUSLE [2]. Soil loss in highland conditions in Kenya ranged from 30 to 666 Mg ha-1yr-1 [8], while the calculated soil loss (325 - 83240 Mg ha-1yr-1) in our research area is much higher. The complex terrain, with elevations from 1261 m to 5537 m, and slopes from 0 to 77.2 degree could be one reason to compute so high soil loss potential. Van Remortel et al [37] argue that erosion model can be used to derive patterns of erosion, but not necessarily the actual loss of erosion, because of the limitations of the methods used to derive some component factor values. Millward and Mersey [6] found that relative comparisons of soil loss among land areas are more critical than assessing the absolute soil loss in a particular cell. A visual interpretation and validation of the resulting erosion risk map was performed for all the sample clusters. The sites were given a subjective risk scale ranging from No risk – Low – Moderate – High – Extreme bases upon the general site characteristics. The high or extreme high erosion risks mostly occurred on the downhill gullies with long proceeding slope lengths (red areas in the map). Some considerations should be given to the vulnerable areas, where the landslides or mudslides could happen easily according to the soil loss potential. A good plant cover is generally capable of preventing surface erosion [21]. The cover management factor with a range from 0.015 to 0.892 indicated that the loss of soil erosion can be greatly reduced by a higher vegetation cover. For the large scale soil conservation, little work can be done to reduce rainfall and runoff erosivity, soil erodibility, slope length and slope steepness, so vegetation restoration and support practice would be the way to reduce the soil loss risk. We have estimated the error from ETM+ images to the canopy cover map and total vegetation coverage map, and calculated the interpolation error of average annual precipitation surface. However, the model is still subjected to errors due to the limitations of the methods to estimate some component factor values, and the lack of possibilities on quantitatively verifying the actual erosivity from our sites. Problems seemed to be mostly concentrated in areas with thick canopy coverage and a high measured ground cover percentage. The erosivity




discrepancies in the forests caused by the modelled rainfall influence from the tree canopies on the ground without taking properly into account the forest floors ground layer. Problematic areas also include rill and valleys where the calculated risk values seemed relatively high when compared with field experiences. This may be explained by the methods used in calculating the LS factor. On the other hand hill tops within the Arctic zone (3600-5400m) were somewhat overestimated since the ETM+ cover of these regions was highly cloud covered. The results could be improved if a cloud free fully topographically normalized image was available, and the C-factor calculation model would take into consideration under canopy ground coverage.

References

[1] Wischmeier, W.H. & Smith, D.D. Predicting rainfall erosion losses – a guide to conservation planning. U.S. Dept. of Agriculture, Agric. Handbook No. 537, 58p, 1978.

[2] Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K, & Yoder, D.C. Predicting Soil Erosion by Water: A Guide to Conservation Planning with the Revised Soil Loss Equation (RUSLE). U.S. Dept. of Agriculture, Agric. Handbook No. 703, 404p, 1997.

[3] Zhang X., Zhang Y., Wen A., & Feng M., Assessment of soil losses on cultivated land by using the 137Cs technique in the Upper Yangtze River Basin of China. Soil and Tillage Research, 69 (1-2), pp. 99-106, 2003.

[4] Grønsten H.A. & Lundekvam H. Prediction of surface runoff and soil loss in southeastern Norway using the WEPP Hillslope model. Soil and Tillage Research, 85 (1-2), pp.186-199, 2006.

[5] Dissmeyer, G.E. & Foster, G.R. A guide for predicting sheet and rill erosion on forest land. Technical Publication SA-TP-11. USDA-Forest Service-State and Private Forestry-Southeastern Area. 40p. (Technical report), 1980.

[6] Millward, A.A. & Mersey, J.E. Adapting the RUSLE to model soil erosion potential in a mountainous tropical watershed. Catena, 38, pp. 109-129, 1999.

[7] Ma, J.W., Xue, Y., Ma, C.F. & Wang, Z.G. A data fusion approach for soil erosion monitoring in the Upper Yangtze River Basin of China based on the Universal Soil Loss Equation (USLE) model. International Journal of Remote Sensing, 24 (23), pp. 4777-4789, 2003.

[8] Angima, S.D., Stott, D.E., O’Neill, M.K., Ong, C.K. & Weesies, G.A. Soil erosion prediction using RUSLE for central Kenyan highland conditions. Agriculture, Ecosystems and Environment, 97, pp. 295-308, 2003.

[9] Moore, I.D. & Burch, G. Physical Basis of the Length-Slope Factor in the Universal Soil Loss Equation. Soil Science Society of America Journal, 50, pp. 1294-1298, 1986.

[10] Moore, I.D. & Burch, G. Modelling erosion and deposition: topographic effects. Transactions of ASAE, 29 (6), pp. 1624-1630, 1640, 1986.




[11] Renard, K.G., Foster, G.R., Yoder, D.C. & McCool, D.K. RUSLE revisited: Status, questions, answers, and the future. Journal of Soil and Water Conservation, 49 (3), pp. 213-220, 1994.

[12] Hevesi, J.A., Istok, J.D. & Flint, A.L. Precipitation estimation in mountainous terrain using multivariate geostatistics: I. Structural analysis. Journal of Applied Meteorology, 31, pp.661-676, 1992.

[13] Hevesi, J.A., Flint, A.L. & Istok, J.D. Precipitation estimation in mountainous terrain using multivariate geostatistics: II. Isohuetal maps. Journal of Applied Meteorology, 31, pp. 677-688, 1992.

[14] Goovaerts, P. Using elevation to aid the geostatistical mapping of rainfall erosivity. Catena, 34, pp. 227-242, 1999.

[15] Ma, K.M., Fu, B.J. Liu, S.L. Guan, W.B. Liu, G.H. Lü, Y.H. & Anand, M. Multiple-scale soil moisture distribution and its implications for ecosystem restoration in an Arid River valley, China. Land Degradation & Development, 15, pp. 75-85, 2004.

[16] Tokola, T. The influence of ground truth data location on growing stock volume estimation in Landsat TM-based forest inventory in Eastern Finland. Remote Sensing of Environment, 74, pp. 422-431, 2000.

[17] Franco-Lopez, H., Ek, A.R. & Bauer, M.E., Estimation and mapping of forest stand density, volume and cover type using the k-Nearest Neighbors method. Remote Sensing of Environment, 77 (3), pp.251-274, 2001.

[18] Katila, M. & Tomppo, E. Selecting estimation parameters for the Finnish multisource National Inventory. Remote Sensing of Environment, 76, pp. 16-32, 2001.

[19] McCool, D.K., Foster, G.R., Mutchler, C.K., & Meyer, L.D. Revised slope length factor for the Universal Soil Loss Equation. Transactions of the ASAE, 32, pp. 1571-1576, 1989.

[20] Lu, D., Li, G., Valladares, G.S. & Batistella, M. Mapping soil erosion risk in Rondônia, Brazilian Amazonia: using RUSLE, remote sensing and GIS. Land Degradation & Development, 15 (5), pp. 499-512, 2004.

[21] Gurevitch, J., Scheiner, S.M. & Fox, G.A. The Ecology of Plants. Sinauer Associates, Inc., Publishers, U.S.A. 523p., 2002.

[22] Zhang, X.B. & Wen, A.B. Current changes of sediment yields in the upper Yangtze River and its two biggest tributaries, China. Global and Planetary Change, 41 (3-4), 221-227, 2004.

[23] Zhang, X.B, Zhang, Y.Y, Wen, A.B. & Feng, M.Y. Assessment of soil losses on cultivated land by using the 137Cs technique in the Upper Yangtze River Basin of China. Soil and Tillage Research, 69 (12), pp. 99-106, 2003.

[24] Editorial Board of Sichuan Vegetation, Sichuan Vegetation. People’s Publishing House of Sichuan, Chengdu, 463p. (in Chinese), 1980.

[25] Renard, K.G. & Freidmund, J.R. Using monthly precipitation data to estimate the R-factor in the RUSLE. Journal of Hydrology, 157, pp. 287-306, 1994.

[26] David, M. Geostatistical Ore Reserve Estimation. Elsevier Scientific Publishing Company, 364p., 1977.




[27] Hutchinson, M.F. A new method for gridding elevation and stream line data with automatic removal of pits. Journal of Hydrology, 106, pp.211-232, 1989.

[28] Tarboton, D.G. Anew method for the determination of flow directions and contributing areas in grid digital elevation models. Water Resources Research, 33 (2), pp. 309-319, 1997.

[29] Garbrecht, J. & Martz, L. W. The Assignment of Drainage Direction Over Flat Surfaces in Raster Digital Elevation Models. Journal of Hydrology, 193, pp. 204-213, 1997.

[30] McRoberts, R. E., Nelson, M.D. & Wendt, D.G. Stratified estimation of forest area using satellite imagery, inventory data, and the k-Nearest Neighbors technique. Remote Sensing of Environment, 82, pp. 457-468, 2002.

[31] GLCF (Global Land Cover Facility) - www.landcover.org as a source of: - U.S. Geological Survey. 10th July 2002, Landsat ETM+, Scene, WRS-2 Path 130, Row 037, Orthorectified Geocover, Sioux Falls, South Dakota: USGS. - U.S. Geological Survey. 10th July 2002, Landsat ETM+, Scene, WRS-2 Path 130, Row 038, Orthorectified Geocover, Sioux Falls, South Dakota: USGS

[32] O'Callaghan, J. F. & Mark, D. M. The extraction of drainage networks from digital elevation data. Computer Vision, Graphics and Image Processing, 28, pp. 328-344, 1984.

[33] Labrecque S., Fournier R.A., Luther J.E. & Piercey, D. A comparison of four methods to map biomass from Landsat-TM and inventory data in western Newfoundland. Forest Ecology and Management, in press, 2006.

[34] Tokola, T., Pitkanen, J., Partinen, S., & Muinonen, E. Point accuracy of a non-parametric method in estimation of forest characteristics with different satellite materials. International Journal of Remote Sensing, 17 (12), pp. 2333-2351, 1996.

[35] Tokola, T., Sarkeala, J. & Van der Linden, M. Use of topographic correction in Landsat TM-based forest interpretation in Nepal. International Journal of Remote Sensing, 22 (4), pp. 551-563, 2001.

[36] Morgan, R.P.C. Soil erosion and conservation. Addison-Wesley Longman, Edinburgh, 198p. 1995.

[37] Van Remortel, R., Hamilton, M. & Hickey, R. Estimating the LS factor for RUSLE through iterative slope length processing of digital elevation data. Cartography, 30 (1), pp. 27-35, 2001.




Decreased deposition of sulphate and

J. Frey1, K. Pajuste2, K. Treier1, Ü. Mander1, P. Kask3 & T. Frey3 1Institute of Geography, University of Tartu, Estonia 2Estonian Environmental Research Centre, Tallinn, Estonia 3Estonian University of Life Sciences, Tartu, Estonia

Abstract

This paper presents the results of trend analyses of a ten-year study of deposition (throughfall+stemflow) and soilwater in two pine stands and a spruce stand (ICP IM areas). The reported decreasing trends of deposited sulphate and cations at the easterly located Saarejärve integrated monitoring area are in good agreement with the decline of SO2 and fly ash emissions in Estonia. At Vilsandi pine stand, which has the westernmost location and is under a marine influence, the decrease of sulphur deposition was comparable with that of the eastern Saarejärve pine stand. However, total base cation load in Vilsandi pine stand remained unchanged resulting in a decrease of throughfall acidity. At Saarejärv coniferous stands deposition of base cations decreased more than that of acid anions causing an increase in K leaching from canopies. A good accordance between sulphate decline in deposition and topsoil water was accompanied by base cation decline in soilwater under organic horizon at both IM areas. In podzolized soil at Saarejärve the decline of SO4 and base cations resulted in increased levels of total soluble Al. Keywords: deposition, throughfall, stemflow, pine and spruce stand, soilwater, sulphate, cations, aluminium, trend analyses.

1 Introduction

Predominant sources of SO2 and particle emission in Estonia are four large oil-shale-based thermal power plants and chemical industries in the Northeastern region. Oil shale mining and combustion accounted for about 81% of total




doi:10.2495/GEO060081

the responses in soilwater at Estonian integrated monitoring sites 1995–2004

harmful emissions up till 2005. Estonian oil shale as a fuel is characterized by high ash share (45-50%) and moderate sulphur content (1.4-1.8%). A peculiarity of Estonian oil shale is its chlorine content (0.75%) combined with organic matter. Oil-shale-centred energetics and economy have led to specific deposition patterns. The high amount of alkaline dust emitted contributed to a considerable neutralization of acid pollutants already in the atmosphere, and even resulted in alkalinization and damage of sensitive ombrotrophic raised bogs [1]. A rapid decrease in emissions took place in 1990s due to reshaping of local economy after restoration of independent statehood in 1991. During the study period (1995-2004), the quantities of emitted SO2 and fly ash decreased by 40% and 65%, respectively, mostly due to a decline in annually combusted oil shale amounts (from about 24 megatons in 1994 to 13 megatons in 2004) i.e. due to a decrease in production of electric energy in the power plants.

The decrease of emissions led to a decline in concentrations of both anions (SO4 and Cl) and cations (Ca, Mg, Na and K) in bulk precipitation at all Estonian monitoring stations during 1994-2001 [2].

The first long-term national intensive ecosystem monitoring programme in Estonia: International Co-operative Programme on Integrated Monitoring (ICP IM) under LTRAP Convention was initiated by the Nordic Council of Ministers in 1994 [3]. In the framework of ICP IM, an intensive monitoring site at Lake Saare forested catchment area in eastern Estonia and a biomonitoring site in the westernmost Island Vilsandi were established. The IM monitoring sites represent areas of boreal coniferous forest receiving background loads of air pollution and deposition. Although the two sites are not regionally representative, they provide an opportunity to monitor changes in air pollution and its impacts on the environment of coniferous stands at the western border of Estonia (Vilsandi), where higher concentrations of SO2 are measured from southern and south-western air mass transport direction [4], and in eastern Estonia (Lake Saare area), which is more affected by local sources of air pollution from the region of oil shale industry in North-Eastern Estonia.

The current paper presents results of a ten-year study of throughfall, stemflow and soilwater in two pine stands and one spruce stand under the conditions of declining SO2 and base cation emissions.

2 Material and methods

2.1 Site characteristics

The study was carried out in the forested subcatchment area (109 ha) of Lake Saare in eastern Estonia (58º39´ N, 26º45´ E), hereafter referred to as Saarejärve, and in Vilsandi, Estonia’s westernmost island (58º34´ N, 21º50´ E). At Saarejärve water and litter samples were collected from two permanent plots of the Scots pine (120 years old) and the Norway spruce (90 years old) stands. The permanent plots are situated on nearly flat surfaces, in areas representative of dominant forest site types in the catchment area. The Rhodococcum type Scots pine stand is located at an elevation. The Vaccinium type Norway spruce stand is




situated lower than the pine stand, near the bottom of a slope. At both stands the parent material is glaciofluvial sand, on which moderately eluviated Haplic Podzols have developed.

The permanent plot in Vilsandi is a 100-year-old Scots pine stand (Fragaria type) on Calcari-Gleyic Leptosol.

Mean annual precipitation during the study period (1995-2004) was 630 mm at Saarejärve and 535 mm in Vilsandi.

Table 1: Stand characteristics of sample plots.

Permanent plot Vilsandi pine stand

Saarejärve pine stand

Saarejärve spruce stand

Site type Fragario-Pinetum

Rhodococco-vitis-idaeo-Pinetum

Vaccinio-myrtilli-Piceetum

Soil type Calcari-Gleyic Leptosol

Haplic Podzol Haplic Podzol

Age of dominant trees 100 120 90 Number of trees in ha-1 440 551 672 Mean diameter (m) 0.29 0.37 0.26 Mean height (m) 16.9 28.7 22.5 Annual average throughfall amount (mm)

282 513 430

Annual average soilwater amount from depth of 10 cm (mm)

101 57 113

Annual average soilwater amount from depth of 40 cm (mm)

98 25 74

2.2 Sampling and chemical analyses

Bulk deposition was collected using two NILU-type collectors [5] at both monitoring areas. Throughfall deposition was collected by polyethylene funnel-type bulk collectors (20 cm in diameter, at a height of about 150 cm) in snow free time, and by buckets during winter. Spiral silicone collectors fitted to three trees per plot were used for collecting stemflow. Water volumes were measured on the field by graduated cylinder. Sampling frequencies were once a fortnight in summer and once a month in winter. Sampling areas and stand characteristics are given in table 1.

Soil water was sampled with zero-tension plate lysimeters of 0.1 m2 [6]. At Saarejärve pine and spruce stands the lysimeters were inserted into depths of 5 to 10 cm under organic horizon and about 40 cm under eluvial horizon with 6 replications per depth. At Vilsandi the lysimeters were installed under humus horizon (into depth of 5 to 10) and illuvial horizon (BC(g) into depth of about 35 cm). At both sites percolation water was collected approximately at 1-month intervals during the snow-free period along with deposition samples.

Water samples from Vilsandi and Saarejärve were analyzed in the Estonian Environmental Research Centre in Tallinn and in the Environmental Studies




Laboratory in Tartu, respectively. Both laboratories have continuous quality control programmes, and they participate regularly in international intercalibration exercises.

Major anions (Cl-, NO3-, SO4

2-) in precipitation and soilwater were analysed by ion chromatography. Base cations were determined using atomic absorption spectroscopy or ion chromatography depending on laboratory, pH was measured potentiometrically.

The nonparametric Mann-Kendall test was used to estimate significance (p<0.05) of the trends in annual values. The slope of a linear trend was estimated with the nonparametric Sen’s method [7].

3 Results

3.1 Decline of sulphur deposition

During 1995-2004 a significant decline was observed in concentration and deposition of sulphate in bulk precipitation and in pine and spruce stands’ throughfall and stemflow at the integrated monitoring sites in Estonia (Table 2). Mean annual concentration of SO4-S in bulk precipitation decreased from 2.2 mg l-1 to 0.37 mg l-1 at Saarejärve and from 1.4 to 0.51 mg l-1 at Vilsandi. Sulphur load dropped from 9.1 to 3.2 kg ha-1 yr-1 at Saarejärve. At Vilsandi the highest load was 5.1 kg ha-1 yr-1 in 1995 and the lowest 2.4 kg ha-1 yr-1 in 2002.

The prevailing anion in throughfall of the pine and spruce stand at Saarejärve was SO4

2-, which formed 58% and 66% of summed anions, respectively. Sea salt correction shows that marine fraction of sulphur is less than 8% in bulk- and throughfall precipitation at Saarejärve. At the remote island Vilsandi, however, the prevailing anion in throughfall was Cl-1 (53%), sulphate formed only 29% of summed anions. Marine fraction of sulphate was on average 10% in bulk deposition and 25% in throughfall flux.

The throughfall flux of S decreased significantly from 8.2 and 18.8 kg ha-1, in 1995 to 4.2 and 4.8 kg ha-1 in 2004 in Saarejärve pine and spruce stands, respectively. At Vilsandi pine stand the decrease in sulphur deposition was about the same as in Saarejärve pine stand (50%) (Fig. 1). The concentration of the other two main anions (NO3 and Cl) also seems to have decreased along with the decline of sulphate at Vilsandi pine stand (Fig. 1). Comparison of TF from the two monitoring sites shows that NO3-N deposition is higher at Vilsandi (2.3 kg ha-1) than Saarejärve (1.0-1.4 kg ha-1) pine stand. The decrease of nitrate deposition on Fig.1 could be explained by variation between the selected monitoring years. Furthermore, throughfall NO3 content is more influenced by internal cycling in canopy [8] than the SO4 content.

Actual annual S input flux into the soil comes from S of throughfall, stemflow and litterfall. Stemflow water was usually negligible in volume forming 2-5% of throughfall in the spruce and pine stand at Saarejärve and less than 1% in the pine stand at Vilsandi. However, due to the high concentration of SO4-S, significant transport of S took place by stemflow, at least at Saarejärve, in the first years of the monitoring period. Sulphur input by litterfall formed on average one third of total S input (TF+SF+LF) in the studied stands.




Figure 1: Decline of throughfall+stemflow deposition (keq ha-1 y-1),

comparison of first and last monitoring years in spruce stand and pine stand at Saarejärve and Vilsandi on the left side. Sulphate decline in deposition (throughfall+stemflow) and in soilwater under organic (org.) and eluvial (E) horizons of spruce and pine stands at Saarejärve and under organic and illuvial (illuv.) horizons in Vilsandi pine stand on the right side.




During the study period the actual S input decreased about 2 times in the pine stands and 4 times in the spruce stand. It should also be pointed out that, judging by annual litterfall amounts (needles only); the crown needle surface of spruce canopy is about three times bigger than that of pine canopy. The sulphur content in litter will be a more important part in the S-circle due to decreased atmospheric loads of S.

3.2 Decline of base cation deposition

The decreasing trend of SO42- concentration in throughfall+stemflow is in good

accordance with decreasing trends of the concentrations of most important base cations, Ca and Mg at Saarejärve coniferous stands. The share of Ca2+ and Mg2+ was on average 45% and 15% of summed cations, respectively. Decreasing flux of summed base cations (Ca+Mg+K+Na) during the study period was statistically significant at Saarejärve (Table 2).

Table 2: Trends in data series (“-“decreasing and “+” increasing) of annual mean concentration in throughfall (TF+SF) and soil water (SW) from 1995-2004. Estimated by Mann-Kendall nonparametric test (significance levels ***p< 0.001; **p<0.01; *p<0.05).

Stand Saarejärve spruce stand Saarejärve pine stand Vilsandi pine stand TF+

SF SW 10 cm

SW 40 cm

TF+SF

SW 10 cm

SW 40 cm

TF+SF

SW 17 cm

SW 35 cm

SO4 -***

-*** -*** -* - -* -* -* -*

Ca -** -* + -* - - Mg -** -* -*** -* - -* -* K + -* -** + - - -* Sum of cations

-** -** -* -* - -* -*

Sum of anions

-** - -* -* - -*

AlTot + +** +* + Al3+ - -* +* +*** H+ - +* + *+ + -* -* +*

There was a clear change in proportion of various cations in throughfall.

Proportions of Ca2+ and Mg2+ in summed cations decreased by 0.1% and 0.4% per year, accordingly, in pine throughfall, while the proportion of K increased by 1.6% per year (p<0.05) obviously due to leaching (K did not increase in open area precipitation). In the spruce stand the decrease of the share of Ca2+ and Mg2+ in summed cations was 1.2% and 0.38% per year, accordingly, and the increase of K+ due to leaching was 2.8% per year (p<0.05).

Due to its location the dominant cation in Vilsandi TF is Na (40% of summed cations on average), the remaining 60% is Ca+Mg+K in almost equal parts. There was no significant decreasing trend of base cation concentration or deposition at the Vilsandi pine stand.




At Vilsandi deposition of base cations neutralised the acidic deposition in bulk precipitation to a larger extent than in throughfall, while H+ flux also decreased in the throughfall of the pine stand. At the same time bulk precipitation was more acidic than total deposition in the pine and spruce stands at Saarejärve. The annual mean pH decreased in the bulk deposition (from 4.8 to 4.7) as well as in the throughfall (from about 5.4 to 5.2) during the study period, and the increase of H+ flux in the pine stand was significant (Table 2).

3.3 Responses of soil solution chemistry to decline of deposition

There were statistically significant declines in sulphate concentration, sum of cations and pH (from 5.5 to 4.9) in soilwater under organic horizon at Vilsandi. At Saarejärve statistically significant declining trends in sulphate concentration, sum of cations and sum of strong anions (SO4+Cl+NO3) were estimated in soilwater under both horizons of the spruce stand, and under eluvial horizon of the pine stand. Concentrations of SO4 and of most cations also decreased below the organic horizon of the pine stand but the trends were not significant (Table 2). SO4 concentrations were about 10% higher in soilwater under the eluvial horizon than under the organic horizon at Saarejärve. Likewise, a statistically significant decreasing timetrend of sulphate amount in soil water output from organic horizon was observed in the spruce and pine stands at Saarejärve. The output of sulphate flux from organic horizon decreased from 0.25 to 0.05 keq ha-

1 y-1 (about 5 times) in the spruce stand and from 0.14 to 0.025 keq ha-1 y-1 (6 times) in the pine stand at Saarejärve area. The decline of sulphate output fluxes from deeper horizons was not statistically significant, although leaching from eluvial horizon in the spruce stand decreased from 90 to 30 eq ha-1 y-1 and in the pine stand from 67 to 18 eq ha-1 y-1. At the remote island Vilsandi sulphate output fluxes decreased from 0.12 to 0.08 keq ha-1 y-1 (1.5 times) under organic horizon and from 0.16 to 0.08 keq ha-1 y-1 (about 2 times) under illuvial horizon. SO4 content under deeper (B) horizon was remarkable higher in Vilsandi pine stand than at Saarejärve (Fig.1 on the right).

The decline in soilwater sulphate was to a great extent matched by a decrease in most cations in organic horizons of the spruce and pine stands, and to a lesser extent also in the eluvial horizon of both stands at Saarejärve. In the soilwater of carbonate-rich soil at Vilsandi the cation content decreased in good accordance with SO4 decline under organic horizon but was not influenced by a decline of sulphate in illuvial horizon.

Total soluble Al contents increased in the soil water from both depths of both stands at Saarejärve (Table 1), although the trends were statistically significant only below organic horizon in the spruce stand and below eluvial horizon in the pine stand.

4 Discussion

Decrease in sulphur deposition is often accompanied by decline of base cations. At Vilsandi pine stand the total deposition of summed base cations did not




significantly decrease while the cation flux under humus horizon decreased significantly during the monitoring period. Thereby, it could be suggested that a substantial share of cations got deposited not along with sulphate ions at Vilsandi. On the other hand SO4 from deposition plays important role as an accompanying anion for cations, which readily moves down with percolation water through organic horizon.

Throughfall composition forming processes (dry deposition, interception, leaching) in canopy changed in both stands at Saarejärve during the study period. Deposition of anions via throughfall and stemflow decreased by about 1 keq ha-1 in the spruce stand and 0.8 keq ha-1 in the pine stand. At the same time deposition of cations at Saarejärve decreased by 1.5 and 1 keq ha-1 in the spruce and pine stands, respectively. Since the input of dust-associated base cations decreased more than the acid anion input, acidity of throughfall should increase, and cause increased leaching of K from canopies.

Good accordance between sulphate decline in deposition and topsoil water is partly due to the use of zero-tension lysimeters. Percolation water obtained by zero-tension lysimeters is the soil water fraction that is primarily involved in soil formation processes, e.g. transport of ions down the soil profile, from one horizon to another prior to e.g. buffering processes [9].

The decline of sulphate fluxes in deposition and under organic horizon of soil water at Saarejärve area resulted in a decrease of SO4 retention in organic horizon. The retention of sulphate via adsorption processes decreased from about 0.44 to 0.24 keq ha-1 in the spruce and pine stands suggesting a decrease in consumption of protons (potential increase of H+) by about 0.2 keq ha-1 in the spruce and pine stands’ organic horizon. The decline in the retention of sulphate in eluvial horizon occurred to a lesser extent, and resulted in the potential increase of H+ by about 0.15 and 0.07 eq ha-1 in the spruce and pine stand, respectively. Via proton consumption the sulphate retention had a decisive role in proton budgets in organic layers of the spruce and pine stands at Saarejärve in the first five years of the monitoring period [10]. The decline in SO4 retention in Vilsandi organic horizon was 0.16 keq ha-1, which is about the same as in Saarejärve. Higher output of SO4 from illuvial horizon than from organic horizon indicates suitable conditions for adsorption/desorption processes. The increase of H+ in percolation water of both horizons in the spruce stand and under the organic horizon of the pine stands could be due to changes in SO4 retention, as well as due to the decline of summed cations in soilwater, which reflects intensive accumulation of base cations into biomass or/and a shortage of exchangeable cations in soil. Although desorption of SO4 anion appeared to be much slower than adsorption [11], the previously adsorbed sulphate would have got released and thus delayed the decrease of SO4 concentration in deeper horizons.

Only additional H+ can increase both total Al and soluble free Al3+ concentration in soil water. The pH of solutions in the studied soil water was normally well below 5 at Saarejärve and, in fact, the pH of throughfall was commonly higher than that of the soil solution receiving it. The increase of total soluble Al in podzolized soil indicates an ongoing process of podzolization due




to dissociated organic acids derived from mineralisation of conifer litter, and is probably attributable to decreased retention of sulphate in soil horizons.

5 Conclusions

The decreasing trends of deposited sulphate and cations at Saarejärve integrated monitoring area are in good agreement with the decline of SO2 and fly ash emissions in Estonia, as well as with SO4 time series from local precipitation monitoring stations during the study period. At Vilsandi pine stand, which has the westernmost location and is under marine influence, the decrease of sulphur deposition was comparable with that of the eastern Saarejärve pine stand, but the total base cation load in Vilsandi pine stand remained unchanged over the entire monitoring period resulting in a decrease of throughfall acidity. Deposition of base cations at Saarejärve coniferous stands decreased more than that of acid anions causing an increase in K leaching from canopies.

A good accordance between sulphate decline in deposition and in topsoil water was accompanied by base cation decline in soilwater under organic horizon at both IM areas. In podzolized soil at Saarejärve the decline of SO4 and base cations resulted in increased levels of total soluble Al.

Acknowledgements

This study was supported by Estonian Science Foundation grant No. 6083 and Target Funding Project No. 0182534s03 of the Ministry of Education and Science of Estonia.

References

[1] Karofield, E., The effects of alkaline fly ash precipitation on the Sphagnum mosses in Niinesaare bog, NE Estonia, Suo 47, pp. 105-114, 1996.

[2] Treier, K., Pajuste, K. and Frey, J., ‘Recent trends in chemical composition of bulk precipitation at Estonian monitoring stations 1994-2001‘, Atmospheric Environment 38, pp. 7009-7019, 2004.

[3] TemaNord, Establishing of the Baltic Integrated Monitoring (BIM) sites 1992-1994, in, Nordic Council of Ministers, Copenhagen, pp. 1-44, 1996.

[4] Pajuste, K., Kimmel, V., Kohv, N. and Truuts, T., Assessment of the Estonian EMEP data, in, Assessment of the EMEP measurements and modelling work in Europe from 1977 until today, http://www.emep.int/assessment/estonia.pdf, 2003.

[5] EMEP: EMEP manual for sampling and chemical analysis, in, EMEP/CCC-Report 1/95, Chemical Co-ordinating Centre, Norwegian Institute for Air Research, Kjeller, pp. 1-199, 1996.

[6] Voll, M. and Roots, O., Soil water sample collector, Environmental Monitoring and Assessment 54, pp. 283-287, 1999.




[7] Salmi, T., Määttä, A., Anttila, P., Ruoho-Airola. T. and Amell, T., Detecting trends of annual values of atmospheric pollutants by Mann-Kendall test and Sen’s slope estimates – the Excel template application MAKESENS, in, Publication on Air Quality 31, Finnish meteorological Institute, Helsinki, pp.1-35, 2002.

[8] Pajuste, K., Frey, J. and Asi, E., Interactions of atmospheric deposition with coniferous canopies in Estonia. Environmental Monitoring and Assessment 112, pp. 177-196, 2006.

[9] Derome, J., Lindroos, A.J. and Niska, K., Comparison of soil water and percolation water quality, in, Forest Condition Monitoring in Finland. National report 1998, 748. The Finnish Forest Research Institute, Helsinki, pp. 132-137, 1999.

[10] Frey, J., Frey, T. and Pajuste, K., Input-output analysis of macroelements in ICP-IM catchment area, Estonia, Landscape and Urban Planning 67, pp. 217-223, 2004.

[11] McBride, M.B., Chemisorption and Precipitation of Inorganic Ions‘, in, Environmental Chemistry of Soils, Oxford University Press, New York, pp. 121-164, 1994.




SAKWeb© – Spatial Autocorrelation and Kriging Web Service

J. Negreiros & M. Painho

Portugal

first Web application to offer spatial autocorrelation and association measures, spatial exploratory tools, variography, Ordinary Kriging (OK) and Simple Kriging with global mean (SK). SAKWeb© version 2.0 exploits IE®, ASP®, PHP® and IIS® capabilities and was designed in an attractive and straightforward way for everyone’s use. Major features are presented in summary. As confirmed by the survey to GIS users, researchers suggest that this Web service and its enhanced developments are a promising alternative for the near future, including its use in an E-learning environment. Keywords: geostatistics, spatial autocorrelation, kriging, web services, E-learning.

1 Introduction

Spatial analysis technology is now beginning to reach the stage where advanced users are struggling with new methods of spatial exploratory data analysis relating to possible explanations for any spatial phenomena. From an explanation of the unitary to a description of the collective, all views of space are correct if the where question is answered, what spatial correlations exist and why spatial distribution is structured in that way. Hence, this article mainly focuses on methods that address the inherent stochastic nature of patterns, interpolation and relationships based on measured point properties such as spatial autocorrelation and Kriging rather than deterministic ones such as buffering areas.




doi:10.2495/GEO060091

autocorrelation and Kriging as well as in Web services. This paper presents the

Instituto Superior de Estatística e Gestão de Informação – UNL,

Abstract

SAKWeb is the result of research integrating developments in spatial

Many GIS users do not think in spatial terms and do not ask questions about spatial pattern and spatial autocorrelation. While conventional users prefer to use GIS modules because of the trademark that underlies the software or the direct need to topology access, experts tend to use statistical products and independent geo-software where interfaces are a non-concern issue. Also, conventional users continue to manage large volumes of spatial data with very deterministic spatial operations while experts are more concerned with stochastic issues such as fuzzier boundaries. If a high quality solution depends on geo-statistical and geographical knowledge then a W3 service with hypermedia help is an excellent way to fill this knowledge gap between both communities Negreiros and Painho [8].

With regard to the implementation of new spatial analysis tools, the future does not seem very bright on account of the undeclared divorce between the research community and the software developers. Therefore, the W3 environment emerges as an alternative solution when compared with the traditional statistical, independent and GIS modules approaches (Negreiros and Painho [8]). Johnston et al. [5]. From a technological standpoint, cooperative work, computation distribution and networking contribute to the widespread dissemination of geographical knowledge with more diverse embedding technologies. From the commercial perspective, the Internet can be viewed as an extension of the traditional competitive marketplace Laudon and Laudon [6]. From the science organization point of view, the synergy effect of the online educational material in which the overall information may be greater than the sum of its parts is a conviction. Under the GIS view, the greatest W3 impact is to close the data access gap among users with free and direct retrieval of spatial analysis tools in an E-learning context. Hence, in this paper we put together four components that will allow an enhancement and a wider use of spatial analysis, new and simpler spatial autocorrelation measures, alternatives to handle the Kriging nugget-effect implementation – everything in an universal access platform (the Internet).

Thus, this expose is organized as follows: The four sub-sections of the following section justifies SAKWeb project by highlighting major geo-statistical innovations, E-learning issues, the poor knowledge concerning spatial autocorrelation, Kriging and simulation topics by many GIS users and the challenge to build an alternative framework to the traditional geostatistical modules (Geostatistical Analyst of ESRI®), independent software (VarioWin®, GeoMS®, GSLib®, STAC®, Regard®, SAGE®, TRIWaco®) and statistical packages (SAS-GIS®, SPSS-X®, Glim®, Systat®

or Minitab®). The basic requirements facing technologies are presented in section three while section four contains a brief description of SAKWeb© functionalities. Section five analyses the user evaluation of SAKWeb©. The concluding section summarizes the main inferences of this paper and future prospects.




2 Why develop SAKWeb?

2.1 New features

It is significant to stress that all new features referenced here are being developed in other articles and documentation such as the inclusion of spatial autocorrelation measures as a valid exploratory spatial data analysis (ESDA) tool, the enhancements of the conventional Moran scatter plot, the inclusion of the four nugget-effect approaches and the implementation of the local error variance. Nevertheless, a brief overview of some of those will be mention next.

With regard to the weight neighbourhood matrix construction, inverse distance weighting (IDW) is indicated as a poor process because it is not sensitive to its direct neighbours Negreiros [7]. Hence, SAKWeb© presents the covariogram contiguity for which neighbours’ boundaries are setup by the variogram range. Any other weight can be applied provided it means something to the researcher and can be justified theoretically Ebdon [3]. In accordance with various lag distances, Moran I is presented as a good search factor of the neighbour radius. Based on the range of the highest Moran I, the weight matrix is defined while the Moran location scatter plot is created (the IDW deterministic interpolation is also based on the same procedure). This mapping computing framework allows the study of spatial patterns, outliers, changeover areas and trends. Furthermore, the Moran I correlogram may lead to a re-estimation process of the variogram sill, model, nugget-effect and range Negreiros [7].

The discontinuity of the variogram origin can be viewed as a non-sense situation if spatial reality is considered a continuous surface. Among all variogram factors, the nugget-effect is also the most unpredictable because of the lack of close samples Clark and Harper [2]. So, SAKWeb© offers four strategies to handle the nugget-effect: (A) γ(0) = 0 (model 1); (B) γ(0) = 0 but including C0 for superior lags (model 2); (C) γ(0) with micro-scale, γ1(h), and long-range, γ2(h), assessment (model 3); (D) If measurement error is given, this attribute will be incorporated within the Kriging system (model 4). It is assumed that the sill is a good global population variance estimate, a condition that most of the times is not true Soares [12]. Just as the arithmetic sample mean is a poor global mean estimate, preferential sampling with proportional effect leads to a false variability value. It is crucial to readjust the variogram sill to reflect the true global variance. Using the nearest neighborhood analysis for the samples weights, it is possible to find the relationship between the Estimated Global Variance (EGV) and the original variogram sill. It also avoids the difficulties of weighting samples at the edge. If this ratio multiplies the variogram model, a rescaling operation, then the Kriging classical variance and region confidence interval, both local and global, can be enhanced.

2.2 E-learning

Recently, a lot of interest has been put on E-learning methods. Quite often, common users request spatial analysis knowledge in a self-learning view because




of global cost reduction, both time and money. In addition, implementing the technological structure that supports E-learning platform is a scalable solution (Painho et al. [9]). Spatial analysis wizards, multimedia tools (including animation and hyperlinks), on-line help, software courses, videos, E-Learning and M-Learning with WML technology are ingredients of this demand. CRM of the Universidade Autónoma de Lisboa and GIS&Sc Master of ISEGI, certified by UNIGIS, are some examples of the Learning Space strategy (Semana Informática [11]). The International Center for Distance Learning (http://www-icdl.open.ac.uk), the AT&T Learning Network Virtual Academy (http://www.att.com/learningnetwork/virtualacademy), La Escuela de Negocios a Distancia de la Universidade Politecnica de Madrid (http://www.cepade.es), Le Centre National d’Enseignement à Distance (http://www.cned.fr) and the Universidade Aberta (http://www.univ-ab.pt) are others. As confirmed by Santos [10], five million US students were taking distance degrees in 1994 and two million in the EC in 1997.

In Europe, this learning trend has been successfully tapped into by EC programs such as Delta (1988-1990), Telematics Applications (1994-1998) and Multimedia Task-Force (1999). Strong future development is therefore expected, supported by current brain-ware know-how, and confirmed by the following examples: (A) The UNAVE project from the Universidade de Aveiro (http://www.unave.pt/fd) reveals greatly encouraging results with its multimedia Web interfaces program: 70% of the 350 students passed, 20% were advised to repeat and 10% failed. (B) With PROF2000 LAN/WAN management classes, the final grades of nineteen high-school professors ranged between 77% and 90% (Santos [10]). (C) During 1999, the estimate for Portugal Telecom’s cost savings with e-Learning (http://formare.ptinovacao.pt) reached €270,000 for ISDN, LAN/WAN and telecommunications courses. (D) Santos [10] presents a p-Fisher statistical study comparing conventional (276 participants) and distance learning (42) methodologies on the basis of age (26-35, 36-45 and 46-55) and school qualifications (grades 6, 7-9, 10-11 and 12), revealing a positive homogeneity of both approaches with a 95% confidence level. This conclusion was based on two representative populations of the same number, 30. A positive Pearson correlation was also found between final grades and school qualifications (+0.313) and a negative one with age (-0.413).

2.3 Modest expertise in sophisticated spatial analysis

Many GIS users do not possess the skills necessary to make use of sophisticated tools such as Kriging, spatial autoregressive models, autocorrelation measures and simulation uncertainty, thus creating a limitation to the understanding of further explanations to their problem. In a survey conducted to Portuguese users, ninety-two people were surveyed at six academic and research GIS institutions reveals a poor situation regarding the knowledge of spatial autocorrelation software: Statistical software – 5%; GIS software – 1%; Geo-software – 7%; No knowledge – 89% (Negreiros et al. [8]).

The First Law of Geography is only known by 5% of all respondents, with 20% possessing theoretical knowledge of some kind of spatial autocorrelation




and clustering measures: Variogram cloud (18% of the total population); Moran I (8%); Gi(d) and Geary C (5%). There are several explanations for this present situation: (A) The present teaching subjects do not involve these recent quantitative geography matters. (B) Constant training and updating are an absolute GIS requirement. (C) The hardware requirements to run these types of packages can be high. (D) There is a lack of difficulty of access to non-cost GIS software in high-schools, colleges and universities. (E) Geostatistical foundations, for instance, are statistical complex for major users.

According to this survey, all respondents believe that the implementation of a spatial analysis Web service can be a wonderful strategy although there is no software available. To fill this present spatial analysis software vacuum is a major aim of this work. After all, E-learning and expertise lack of sophisticated spatial analysis regarding GIS problems are two different issues of the same coin. Therefore, the ambition of SAKWeb© is to develop the first spatial autocorrelation and Ordinary Kriging approach in a Web environment for people with data observation problems that can be handled by both methodologies.

2.4 The W3 challenge

What the users appreciate and interact with is the GUI (graphical user interface), because they do not care about the technical structure, as long as the results are trustworthy, prompt and compatible with their operating system and hardware (Negreiros and Painho [8]). Users want intuitive and easy-to-use software in order to give immediate results without having to read pages of documentation. Luckily, the standard Web browser, irrespective of the background computer code adopted, fulfills this strategy quite well. It is cost free and already provided in any operating system. Ultimately, the user will work with all available software in the same way, regardless of the location of the data or its purpose: (A) In the beginning, the trend was to find the data with the browser; (B) Today, download the software and installed it; (C) In the near future, just run the programs from the Internet.

This represents a significant advance in the user interface because users will no longer have to worry about the software location and the technical knowledge necessary to connect to the data. Even today, the keyboards available on the market hold special keys for the browser such as home, search, back, mail and refresh. The standard Web browser is the future interface. This solution relies on an enhanced standard front-end acting as a protected wall against the computer code in an embeddable, extendable and reusable development. The capability to undercover technical implementation to the final user via WWW is essential. After struggling for years for digital information, spatial analysis needs to concentrate on what the information means, sharing it through the W3 and distributed architectures (Negreiros and Painho [8]). In effect, the wireless Web is developing furiously nowadays. Web-enabled wireless devices will enable millions of people to access the Internet while on the go. Therefore, the challenge of improving spatial data exploration by applying the available tools of spatial analysis and back-office technology integration with the Internet (mobile or not) is also critical.




3 SAKWeb© technology

When common users begin to handle and understand the background concepts of spatial autocorrelation and Kriging, the difficulty of getting specialized software becomes all too clear (particularly with spatial autocorrelation and spatial autoregressive methods). It is important to forget installation procedures and operating system requirements and to embrace a direct, friendly and known GUI with mobile access. If the inferences presented in the previous section are added to this process, the final major SAKWeb© specifications are, thus, setup.

To make this project come to life, several W3 technologies were used including ASP®, PHP® and Dreamweaver® (cf. figure 1). Data Access Components®, WebChart® from Component One®, FrontPage® Server Extensions Service Release, Flash®, HTML, Java Applets®, VBScript® and JavaScript® are some of the other components required. On the bottom of everything lies the client-side and server-side structure that supports this dynamic application.

This Web service is, thus, a collection of ASP® pages, server components and a Website where the distinction among the applications is made by the root directory within the site. All content within its directory structure is considered part of the scope of the same application. Further, each application has its own set of variables and attributes that define its current state and these are maintained throughout the application lifetime. Another primary concern with Web based applications is that HTTP has no memory and retains no information from one client request to the next. ASP® gets around this issue by using Application and Session objects to store information during a user’s session.

Internet Information

Server

DHTMLJavaScript

From CGI to ASP and PHP

ComponentOne WebChart And

ActiveX Controls

Java Applets

Flash

SAKWeb Technologies

Server and Client-side Technology

(SAKWeb Programming)Client-side Technology

(SAKWeb Help)

Figure 1: SAKWeb© technologies.




The 2D and 3D WebChart® gives also power to any ASP® environment by generating different graphic charts. Internally, the server-side component shares a common Application Program Interface (API) with the client-side that generates an .OC2 or .OC3 control to be passed to the client with the appropriate HTML tags. If the client does not have olch2x8.cab and olch3x8.cab, the server cab file generates a runtime-only copy of WebChart® on the client's machine (these CAB files are compressed versions of Component One® Chart). Control management of the site is achieved by means of three options: 1) DOS Commands - The capability to browse the contents of DOS® commands; 2) PHP Configuration - The ability to display the PHP® configuration of SAKWeb©; 3) Session Variables – It gives the CGI parameters and the contents of the session variables created in the current user session, a useful tool for the programmer to control the application flowchart. It also plots a horizontal 3D graphic with the distribution of the visitors to this site throughout the day by dividing it into six groups: 0h-4h, 4h-8h, 8h-12h, 12h-16h, 16h-20h and 20h-24h. This information should guide the Web users to select the best time to connect to SAKWeb©.

4 SAKWeb main features

SAKWeb© is not a full geostatistical and simulation software package. Its functionality is depicted on the basis of four critical procedures that implement together known spatial autocorrelation measures as well as newly developed ones:

• Data Input and Exploring View, which focuses on MS-Excel® input, control management of the user session and exploratory data analysis. New features include the estimation of the global mean and global variance whose weights are based on the nearest neighbourhood measures.

• ESDA, Spatial Autocorrelation and Variography, which concerns variogram setup, the Moran I correlogram, the Moran location and the variance scatter plot. Recent characteristics include the Moran location and variance scatter plots.

• SAKWeb© Ordinary Kriging, which concentrates on OK calculus and surface mapping in accordance with four nugget-effect strategies. Validation with an extra dataset, 3D-2D surface profiles, cross-validation and region plumes based on threshold values and confidence levels are also included.

• SAKWeb© Help, presents ten options regarding software and geostatistical help with the Web driven interface using hypermedia with sound and movement and an e-Learning structure that includes an IRC, Email and News service, to close the knowledge gap between experts and common GIS users.




5 SAKWeb© end-user evaluation

In 2003, a survey was carried out to twenty GIS users. SAKWeb© was setup in an IIS® server, for ten clients using the intranet of Rainha Dona Leonor college in Lisboa. The same GS+ Pb dataset was given so that they could test the SAKWeb© features (cf. Figure 2). Previously, a 45 minutes refresher overview of spatial autocorrelation, trend surfaces, nearest neighborhood and OK subjects was presented. However, no tips were given on how to use SAKWeb©.

3.74.1

3.6

4.5 4.8

1

2

3

4

5

Figure 2: The average classification of the impact of SAKWeb© major features.

On the basis of the survey presented and with post-statistical processing, several interpretations can be drawn:

a. The SAKWeb© Moran variance scatter plot was too slow to be computed. In fact, 80% of the users complained about this issue, stating that this routine leads to impaired performance, that is, the slowness of the server impacts on concurrent operations of other users. Thus, the ASP® script language that underlies this procedure was changed to C. However, the strength of SAKWeb© in concurrent processes did not reveal any other major flaw.

b. The capability to brush and link different maps was pointed out as seriously deficient by 60% of the interviewees, a major trend in the ESRI® Geostatistical Analyst module (particularly for assessing classical statistics in a moving average context). Accordingly, two new options were setup within SAKWeb©: local interactive statistics and 3D-2D OK profile.

c. In this survey, the limited number of OK models available, especially when compared with the ESRI® Geostatistical Analyst module, is another issue to be taken into consideration. This includes different lag spacing depending on different directions, smaller separation distances for the first lag and other variography tools such as the correlogram, the madogram, the general relative variogram and the pairwise relative variogram. Consequently, these last four features were included after this survey was carried out.

d. Surprisingly, 95% of the users demand cross-validation.




e. The questionnaire reveals that 45% of the respondents think that it is important to include the Box-Cox, Normal and logarithmic transformations in geostatistical software.

f. This questionnaire indicates that 30% of the users think that topographical and hill shade background maps enhance final Kriging maps. For this reason, image mapping in a new window was added afterwards. Curiously, no users mention the absence of the traditional semivariance/covariance cloud and the QQ-plot in SAKWeb©.

g. All answers reveal that these users support the SAKWeb© Moran location scatter plot for verifying trends and local pockets of stationary. The bivariate scatter plot is also a good option. However, the histogram tool in conjunction with the brushing and linkage location map represents a wonderful tool for this purpose.

Figure 3: The final ranking for the main advantages of the Web solution.

After the survey was completed, certain issues were discussed in a brief meeting. The respondents confirmed previous suggestions: they really appreciate the capability to generate several Kriging surfaces and compare them without any extra work. It is expected that future software developments will enhance this feature. Therefore, the capability to extend this on-the-fly potential based on different range, model and anisotropy factors could be a remarkable improvement since the present solution is only based on the four nugget-effect strategies. In terms of the interface, it was unanimously agreed that this Web service is extremely easy to use for two main reasons: 1) The hints given and the validation check for the input parameters; 2) The inability to proceed to the next step if the previous one is not correct (it works like a process wizard).

Two users reclaim the migration from ASP® 3.0 to ASP.Net® architecture. However, the introduction of the Moran location scatter plot (instead of the conventional one) and the Moran variance scatter plot represents a major step in spatial autocorrelation measures. According to the questionnaire, the Moran location scatter plot was ranked first in terms of usability and visual impact for

0

1

2

3

4

Low Marginal Costs Mobile And Easy Access Hardware and SoftwareManagement

Standard Web GUI




outliers, trends and patterns search. The Moran I and Geary C took second place while the variogram cloud ranked third. But, this latter measure requires the brushing and linking feature. Another interesting comment made by two users was: “Until now, the emphasis on spatial autocorrelation measures by geo-software has been limited and almost forgotten. It is hoped that SAKWeb© restores in our minds the need to implement them with Kriging software.”

Another point is that the SAKWeb© Help approach presents alternative help software to that provided by the conventional Contents-Search-Index pattern. According to most inquiries, laying out the overall help structure with hyperlinks among subjects is really an attractive model. As some users confirmed, “it works like a puzzle where each piece corresponds to each issue, though at the same time a complete view is given of the structure topics and how they are related to each other.” Future help applications must be encouraged to follow this concept too.

6 Conclusions

There is an increased demand for systems that do more than display and data organization Ebdon [3]. The set of potential applications for spatial analysis is enormous, for example, accident patterns, victim profiles within a residential population and spread rates for pollution levels. Thus, spatial statistics must hold a specific spatial framework to apply quantitative and statistical methods for a better understanding of spatial relationships. If the finding of spatial structures is fundamental then autocorrelation, interpolation and autoregressive models are three major spatial methods that fulfill this constraint.

However, when common users begin to handle and understand the background concepts of spatial autocorrelation and Kriging, the difficulty of getting specialized software becomes all too clear. It is important to forget installation procedures and operating system requirements, to embrace a direct, friendly and known GUI with mobile access, to incorporate standard exchange for data input, to validate input parameters including suggestions and hints on-the-fly, to add new features available and to provide the capability of discussing these issues on an on-line basis with good hypermedia help. With the advent of Web technology and modern wireless computing, it has become necessary to develop a W3 service for interpolation to understand the often complex spatial autocorrelation that exist among the samples collected in space. The E-Learning geostatistical solution and, particularly, the close linkage between spatial autocorrelation and Kriging were never considered. The inclusion of these two major components provided another inspiration for SAKWeb© version 2.0.

Although this project is still a work in progress, the future of SAKWeb© can be bright. Already beyond the scope of this paper, the present SAKWeb© infrastructure can be applied easily as a WWW interface with GSLib® routines to avoid the reinvention of the wheel for other geostatistical tools, a situation required by some respondents in the survey.




The Next Kriging Menu selection has this aim by leading the user to a new choice of options that include SK, UK, IK and CK. In effect, SK with global mean has already been implemented with the following options: SK Calculation, Default SK, SK With No C0, SK With Two Structures, Exact SK Differences, SK Versus OK, Validation With Extra Dataset and Cross-Validation (as with OK, this selection includes the location of positive-negative/true-false samples). A third menu screen has already been sketched for simulation and spatial auto-regression procedures. Adapting Anselin [1] contemplation, the respectability of the academic research concerning implementation should be a reality, where the results are reflected in appealing pictures while the arduous work is taking place beyond-stage in a W3 environment.

References

[1] Anselin, L., Exploratory Spatial Data Analysis in a Geocomputational Environment in Geocomputation A Primer. John Wiley and Sons, New York, 1998, pp. 77-94.

[2] Clark, I., Harper, W., Practical Geostatistics Answers to the Exercises, Ecosse North America, 2000, pp. 298.

[3] Ebdon, D., Statistics In Geography, Blackwell, 1998. [4] Goovaerts, P., Geostatistics For Natural Resources Evaluation, Oxford

University Press, 1997. [5] Johnston, K., Hoef, J., Krivoruchko, K., Lucas, N., Using ArcGIS

Geostatistical Analyst, ESRI, 2001, pp. 287. [6] Laudon, K, Laudon, J, Management Information Systems. Prentice-hall

International, 2002, pp. 523. [7] Negreiros, J., SAKWeb - Spatial Autocorrelation and Kriging Web, a W3

Computing Perspective, Unpublished Ph.D. Thesis, ISEGI-UNL, 2004, pp. 449.

[8] Negreiros, J., Painho, M., The Web Platform for Spatial Statistical Analysis, Proceedings of the 6th Conference of the Portuguese Association in Information Systems (CAPSI), Bragança, Portugal (26-28 Nov 05).

[9] Painho, M., Cabral, P., Peixoto, M., Pires, P., E-teaching and GIS: ISEGI-UNL learning experience, Third European GIS Education Seminar EUGISES, 2002.

[10] Santos, A., Ensino à Distancia e Tecnologias de Informação: E-Learning, FCA, 2000.

[11] Semana Informática, http://www.lotus.com/home.nsf/welcome/learnspace, 2002.

[12] Soares, A., Geoestatistica para as Ciências da Terra e do Ambiente, IST Press, 2000, pp. 206.




Section 3 Environmental pollution

and remediation

After treatment of landfill leachate in peat filters

M. Kõiv1, M. Kriipsalu2 & Ü. Mander1 1Institute of Geography, University of Tartu, Estonia 2Department of Technology, University of Kalmar, Sweden

Abstract

The main objective of this study was to determine the treatment capacity of well mineralised Sphagnum peat in order to reduce BOD and COD values and nutrient concentration in landfill leachate. The peat filters were suitable for the reduction (up to 93%) of ammonia nitrogen. Good results were obtained in the reduction of total phosphorus from both raw and pre-treated leachate (up to 81% and 70-99% respectively). The purification rate of the landfill leachate depended on the contamination rate – the outflow results were better with pre-treated leachate, and the results also improved due to the lowering of the flow rate (on average by 60 times). Therefore, it is recommended that peat filters be used in combination with conventional treatment methods, e.g. as soil filters of subsurface flow constructed wetlands for the secondary or tertiary treatment of the leachate. Keywords: ammonia-nitrogen, BOD, COD, peat filter, pre-treated landfill leachate, raw landfill leachate, total nitrogen, total phosphorus.

1 Introduction

Peat is partially fossilized decomposed plant matter that transforms in wet areas in the absence of oxygen. Compared to mineral soils, peat has a very high organic content (60% carbon). Peat has a surface area of >200 m2 g-1 and is highly porous (80-90%) [1]. In Estonia, peat lands cover 22% of total land area. Estonian peat resources are estimated at 2.4 billion tonnes, of which 0.2 billion tonnes are less decomposed, and 1.4 billion tonnes are well decomposed [2].




doi:10.2495/GEO060101

The potentially large availability of peat and its unique combination of biological, physical and chemical properties make it suitable for a wide variety of uses, including environmental protection [3]. For example, peat has been used in the treatment of wastewaters of various origin and quality. Several laboratory and field experiments have demonstrated that peat as a filter material for constructed wetlands, and also for bio-filters and other conventional treatment systems can effectively reduce nitrogen concentration and remove suspended solids [4, 5, 6], pathogenic bacteria [4, 5], mineralise organic material, retain phosphorus [7, 8] and other heavy metals [1, 9]. Most peat filters are designed for the treatment of domestic wastewater [10], and several systems have also shown very good performance in the treatment of landfill leachate [11, 12]. Likewise, results of experiments on floodplain fens in Estonia have shown the high potential of peatlands in the after-treatment of wastewater [13]. Although peat is an inexpensive and attractive material, such advantages must be balanced against the importance of peatland conservation and the maintenance of habitat diversity [14]. The composition of landfill leachate varies greatly, being dependent on waste quantity and quality (the age of the waste and also its decomposition rate and landfilling technologies). Leachate is considered difficult to treat due to its typically high concentration of P (up to 100 mg L-1 [15]), ammonia nitrogen (up to 300 mg L-1 [11]), high COD value (up to 60,000 mg L-1 [15]), and heavy metals. In Estonia there is a lack of long-term experience in the treatment of landfill leachate. There are only a few landfills where leachate is collected and purified using various methods. The main objective of this study was to determine the treatment capacity of well mineralised Sphagnum peat in order to reduce BOD and COD values, and nutrient concentration in landfill leachate from different stages of the leachate treatment system. We also investigated whether peat filters improve the efficiency of conventional leachate treatment systems. In addition, the effect of the duration of the experiments on treatment efficiency, as well as changes in the composition of the peat, was studied.

2 Materials and methods

Two experiments were conducted at Väätsa landfill in Estonia, the first experiment (E1) in summer 2003 and the second experiment (E2) in 2005.

2.1 Site description

The Väätsa landfill is the first sanitary landfill in Estonia that meets the requirements of the EU Council Directive [16] and Estonian landfill directives [17]. The first stage of the landfill (1.0 ha) was in service from 2000-2005, and the second stage (1.5 ha) is in operation since November 2005. The landfill serves approximately 40,000 inhabitants. By the present time (February 2006),




60,000 t of mixed waste has been deposited [18]. The landfill has a proper lining, leachate collection system and two-stage biological leachate treatment system consisting of an activated sludge treatment plant and aerobic-anoxic pond (Fig. 1).

Figure 1: Aeration of landfill leachate in aerobic-anoxic pond in Väätsa, Estonia.

2.2 Experimental design

In both experiments, custom-designed peat filters (F) (total four filter bodies) were used. In experiment 1 (E1) the two metal filter bodies (filters 1 and 2 – F1, F2) had a volume of 1 m3, were rectangular in shape, and had a permeable floor (Fig. 2). In the second experiment (E2), two filter bodies (filters 3 and 4 – F3, F4) with a volume of 0.2 m3, were made of PVC pipe (Ø 372 mm, h=1200 mm) (Fig. 2).

2.2.1 Peat type All filters were filled with well mineralised fluffy Sphagnum peat collected from Lokuta peat bog, which is located near Väätsa landfill. The well mineralised peat was obtained from the lower deposits of depleted industrial peatlands. The peat was sieved through a 26 mm sieve to remove stones and roots.




Figure 2: Design of the filter bodies and leachate distribution to the peat filters in

experiment 1 (F1; F2) and in experiment 2 (F3; F4).

2.2.2 Distribution of leachate and loading of filters For the loading of F1 raw leachate and for the F3 the leachate from the activated sludge treatment plant was used. For the F2 and F4 we used treated leachate from the pond (the outflow of the treatment system). The even distribution of leachate into the F1 and F2 (E1) filters was achieved using perforated elastic pipes placed above the filter bodies in a spiral pattern (Fig. 2). The even distribution of leachate into the F3 and F4 (E2) filters was achieved using perforated pipe inside crushed granite (Ø 6-8 mm) on top of the filter material (Fig. 2). All filters were loaded using timer-adjusted pumping. In the first experiment F1 and F2 were loaded for 36 days, with 2.9 m3 m-2 in a day (in total about 104 m3 per filter). In the second experiment, the filter 4 was loaded for a total of 5 months – from June to October 2005 – and the F3 since April 2005, for a total of 9 months. The loading rate for F3 was 0.083-0.033 m3 m-2 d-1 (total amount about 13.7 m3), and for F4 0.05 m3 m-2 d-1 (about 7 m3).

2.3 Sampling and statistical analyses

The leachate samples from the inflow and outflow of F1, F2 were taken on days 2, 4, 6, 8, 15 and 22, and once a month from the inflow and outflow of F3, F4: eight times from F3 and four times from F4. In both experiments, BOD7, COD, total N, NH4-N, NO3-N, NO2-N, total P and pH were determined in the certified laboratory using standard methods. In experiment two we also analysed: PO4

3-, conductivity, TSS, SO42-, Ca2+, Mg2+

and total hardness. In peat, the content of organic matter (%), N (%), pHKCl and P, K, Ca, Mg (mg kg-1) was determined.




The normality of variables was checked using the Lilliefors and Shapiro-Wilk tests; for normally distributed variables (Total P, BOD, COD, NH4-N, pH), the inflow and outflow values in different peat filters were compared via the pairwise t-test. When the distributions were skewed, the nonparametric Wilcoxon pairwise test was used. When the assumptions of ANOVA were fulfilled, a Fisher LSD test was used for multiple comparisons of mean removal efficiencies in different filters. For the remaining variables, Kruskal-Wallis ANOVA and the multiple comparison of mean ranks for all filters was used. The STATISTICA 7.0 software was used and the level of significance of α=0.05 was accepted in all cases.


3.1 Landfill leachate treatment at Väätsa

At Väätsa landfill, approximately 10-20 m3 of leachate is produced per day and treated in a biological treatment system. The treatment efficiency of conventional activated sludge treatment, followed by aerobic-anoxic treatment in a pond, is high. However, the average values of BOD7, COD, total N, and total P in the outflow from the aerobic-anoxic pond (Table 1) exceeded the limits required for treated wastewater in Estonia [19]. Table 1: Average treatment efficiency in Väätsa landfill leachate treatment

system in 2004 and 2005 [20, 19].

Two-stage treatment system inflow outflow

Overall treatment efficiency (%) Para-

meter Unit 2004 2005 2004 2005 2004 2005

BOD7 mg L-1 2193 993 52 93 90 91 COD mg L-1 4133 1880 451 847 90 55 Total N mg L-1 401 282 117 115 60 59 Total P mg L-1 1.7 6.5 3.6 4.6 30 pH 7.4 8.2 8.7 8.8

In Estonia there are special target values for leachate treatment [17]: contamination rate 25 mg L-1 (purification rate ≥90%) for BOD7; 125 mg L-1 (≥75%) for COD; 2.0 mgL-1 (≥80%) for total P and 75 mg L-1 (≥75%) for total N. Pre-treatment of leachate in the Väätsa biological treatment system is not sufficient to fulfil prescribed values for effluent.

3.2 Leachate treatment in experimental peat filters

In the first experiment, biologically pre-treated leachate from the second stage of treatment purified about 2-5% easier than the raw leachate. When we compare E1 and E2, however, the treatment efficiency is better in the second. One explanation of why E2 had better results in the removal of contaminants than E1 may be that the selected loading rate was too high for the filters in E1 [11].




Average values of contaminants in the inflow and outflow of the peat filters, and differences between inflow and outflow values in all filters (significant differences when p<0.05, according to pairwise t-tests) are presented in Table 2. Table 2: Average values and standard deviation of the contaminants in the

inflows and outflows of the peat filters. Significant differences between the inflow and outflow values of the contaminants in all of the peat filters studied (pairwise t-test results): * - p<0.05; ** - p<0.005.

F1 F2 F3 F4 Parameter

(mg L-1) in out in out in out in out

BOD7 1953 ±459

1728 ±505*

54 ±19

27 ±19*

79 ±32

24 ±15**

19 ±16

6 ±3

COD 3812 ±446

3462 ±555**

521 ±49

466 ±60**

1158 ±239

843 ±253*

598 ±37

592 ±150

Total P 1.3 ±0.2

0.7 ±0.3**

1.2 ±0.2

0.6 ±0.1**

5.6 ±1.9

2.0 ±1.7*

*

3.3 ±0.7

0.2 ±0.2*

Total N 447 ±24

401 ±63*

112 ±22

95 ±37

339 ±121

302 ±101

127 ±55

115 ±23

NH4-N 369 ±60

327 ±44*

52 ±10

37 ±7*

254 ±115

97 ±85**

9.5 ±9

7.2 ±2

NO3-N 5.8 ±2.2

5.8 ±1.3

18 ±6.7

17 ±2.7

38 ±53

143 ±93*

50 ±14

43 ±14

pH 7.5 ±0.1

7.8 ±0.2*

8.8 ±0.1

8.3 ±0.2**

8.5 ±0.2

8.2 ±0.2*

*

8.8 ±0.1

8.9 ±0.6

The average value of pH in leachate decreased in F2 and F3, and an increase took place in F1 and F4. The changes in pH value according to the pairwise t-test were significant in filters 1, 2 and 3 (Table 2). As stated by Patterson et al. [21], the slight change in pH could be related to the organic acid components flushed from the peat.

3.2.1 Reduction of BOD and COD values A decrease in BOD value was observed after treatment in all filters (Table 2). The treatment efficiency in filters was 12-70% on average, and maximum reduction was achieved in F3 (95%). The outflow value that fulfils the prescribed limit of 15.0 mg O2 L-1 [17] was achieved in F4, where the outflow value was 6.4 mg O2 L-1 on average. The most effective removal of COD was achieved in F3, where the values were on average almost 30%. The reasons why BOD and COD removal was some times not successful enough may be that factors were not favourable (pH range, the existence of inhibitors, the lack of substrate and phosphorus, temperature, contamination rate etc). For instance, the BOD and COD values in raw leachate were too high (Table 2) for it to be treated only with a peat filter.




3.2.2 Removal of nitrogen The best results for nitrification were achieved in the F3 (Table 2), where ammonia nitrogen was reduced by 62% on average (Fig. 2). Only in F4 did no significant nitrification take place. The reduction of total N concentration was noticeably higher in F2 (35% on average). However, significant differences between inflow and outflow were only determined in F1. The reduction was not sufficient to fulfil the Estonian discharge limit [17]. Maximum removal of ammonia nitrogen (62% on average) was achieved by F3 (Fig. 2). Significant differences between inflows and outflows were determined in F1, F2 and F3 (Table 2). An approximately 12 and 30% reduction is obtained in F1 and F2 respectively. Mæhlum [22] indicated that nitrification may be limited due to lack of oxygen. The amount of oxygen in peat was not directly measured, but the intermittent loading of leachate during the experiment was selected in order to increase the natural inflow of oxygen through the top and the bottom of the filter body. In experiment 2 the amount of oxygen was measured from the inflow and outflow leachate of the peat filters. The results show that the peat filters increase the amount of oxygen in leachate (e.g. 95% on average in F3). Nitrification occurs at an optimum temperature of about 30ºC [23]. At 5ºC, the nitrification rate is only 15% of the rate at 20ºC [24]. The temperatures of raw and pre-treated leachate during the summer period in Väätsa were 14.8±0.3ºC and 18.0± 1.6ºC respectively. Thus there should be favourable conditions for the creation of nitrifying bacteria in that period. The nitrification of ammonia also depends on the hydraulic loading [25]. We can confirm this fact, because E2 had better results in nitrification then E1. Also, when we lowered the hydraulic loading rate into the F3 filter several times during experiment 2, the ammonia nitrogen decreased, and NO3-N and NO2-N increased significantly.

3.2.3 Removal of phosphorus The removal of total phosphorus from Väätsa leachate (Fig. 2) was very successful, and average outflow values were below the limits [15]. According to pairwise t-tests, the inflow and outflow values demonstrated a significant difference (p≤0.005) in all filters. The total phosphorus reduced in F1 and F2 was on average 49% and the F3 and F4 filters removed 64 and 92% respectively. Mann [26], Kadlec and Knight [23] and Richardson et al. [27] demonstrated that the reduction of total P could be caused by the sorption, sedimentation and combination of complex compounds. A certain proportion of P may be bound to the biofilm [26, 27]. Phosphorus transforms easily from organic to inorganic forms and constitutes chemical complexes with organic and inorganic ligands, which can be adsorbed by or sedimented into the soil. In aerobic wetland conditions the P constitutes in dissolved complexes with oxidised Ca and Mg in alkaline conditions, and with Fe or Al in soil with acidic to neutral pH [22]. The content of the Mg and Ca in peat increased significantly, whereas the pH of the peat was neutral.




According to Mæhlum [22], it can be concluded that the fixation of P resulted in adsorption with Ca and Mg compounds, and these were settled in the peat filter. P content increased in peat, supporting this statement.

Figure 2: Average inflow and outflow concentrations of NH4-N and total P in all

filters. * - p<0.05, the inflow value is significantly different from outflow values, according to the pairwise t-test.

3.3 Composition of peat

The initial concentrations of Ca, Mg and K in the well mineralised peat are relatively high (Table 3). Before the experiment, the content of organic matter in peat was almost twice lower (37%) than on average, which could be due to the small stones that were found in the peat. During the experiment there was a slight increase in the content of organic matter. After the experiment, Ca, Mg and especially K concentrations increased significantly (96%). Table 3: Composition of peat in filters 1 and 2 before and after the first

experiment. Units are mg kg-1 (unless otherwise noted).

pHKCl N (%) P K Ca Mg Organic matter (%) Before 6.7 0.5 23 93 8294 1416 37.2 After 7.1 0.6 37 2177 9178 2149 36.9




The removal of contaminants from leachate can be linked to chemical changes in the peat. The results are affected by interactions between peat type, water quality, loading rates, duration of treatment etc. The removal of total P with the filters may be caused by the high content of Mg- and Ca-compounds in peat.

4 Conclusions

In Estonia, the use of peat as a filter material has good potential. It has shown sufficient purification efficiency and can be considered to be an ecologically sound and economically beneficial material in leachate treatment. The pre-treatment of landfill leachate considerably reduces the pollutant load on the peat filter, which increases its performance. A smaller flow rate and respectively longer retention time will be required for the optimal performance of peat filters. We can conclude that peat filters are well suited to the reduction of Total P from raw leachate (up to 81%), as well as from pre-treated leachate (up to 70-99%). The remarkable efficiency of well-mineralised peat in the reduction of BOD values (up to 95%) and NH4-N concentrations (up to 93%) encourages us to use peat filters in combination with conventional treatment methods, e.g. as soil filters of subsurface flow constructed wetlands for secondary or tertiary treatment of landfill leachate.

Acknowledgements

This study was supported by Estonian Science Foundation grant No. 6083 and Target Funding Project No. 0182534s03 of the Ministry of Education and Science of Estonia. Pille Kängsepp, Tõnu Salu, the staff from Väätsa landfill and colleagues from the Institute of Forestry and Rural Engineering at the Estonian University of Life Sciences are acknowledged.

References

[1] Brown, P.A., Gill, S.A., Allen, S.J., Metal removal from wastewater using peat. Water Res, 34 (16), pp. 3907-3916, 2000.

[2] Statistical Office of Estonia Web site, Extraction of mineral resources. http://www.stat.ee

[3] McLellan, J.K., Rock, C.A., Pretreating landfill leachate with peat to remove metals. Water Air Soil Poll, 37, pp. 203-215, 1987.

[4] Lens, P.N., Vochten, P.M., Speelers, L., Verstraete, W.H., Direct treatment of domestic waste-water by percolation over peat, bark and woodchips. Water Res, 28 (1), pp. 17-26, 1994.




[5] White, K.D., Byrd, L.A., Robertson, S.C., O’Driscoll, J.P., King, T., Evaluation of peat biofilters for onsite sewage management. J Environ Health, 58(4), pp. 11-17. 1995.

[6] Gunes, K., Ayaz, S.C., Wastewater treatment by peat filtration. Fresen Environ Bull, 7 (9A-10A). Sp.iss. SI: 777-782, 1998.

[7] James, B.R., Rabenhorst, M.C., Frigon, G.A., Phosphorus sorption by peat and sand amended with iron-oxides or steel wool. Water Environ Res; 64 (5), pp. 699-705, 1992.

[8] Roberge, G., Blais, J.F., Mercier, G., Phosphorus removal from wastewater treated with red mud-doped peat. Can J Chem Eng; 77, pp. 1185-1194, 1999.

[9] Davila, J.S., Matos, C.M., Calcavanti, M.R., Heavy-metals removal from wastewater using activated peat. Water Sci Technol; 26 (9-11), pp. 2309-2312, 1992.

[10] Geerts, S.M., McCarthy, B., Axler, R., Henneck, J., Heger Christopherson, S., Crosby, J., Guite, M., Performance of peat filters in the treatment of domestic wastewater in Minnesota. 9th National Symposium on Individual and Small Community Sewage Systems, Fort Worth, TX, March 11-14, 2001, Amer Society of Agricultural Engineers, Michigan, USA, 2001.

[11] Kadlec, R.H., Integrated natural systems for landfill leachate treatment. In: Vymazal, J. editor. Wetlands – Nutrients, Metals and Mass Cycling. Backhuys Publishers, Leiden, Netherlands, pp. 1-34, 2003.

[12] Kinsley, C.B., Crolla, A.M., Kuyucak, N., Zimmer, M., Laflèche, A., A pilot peat filter and constructed wetland system treating landfill leachate. 9th International Conference on Wetland Systems for Water Pollution Control, Avignon (France), 26-30th of Sept. 2004, Cemagref. Antony, pp. 635-642 2004.

[13] Öövel, M., Tarajev, R., Kull, A., Mander, Ü., Tertiary treatment of municipal wastewater in a floodplain peatland. In: de Conçeicao Cunha, M., Brebbia, C.A. editors. Water Resources Management III. WIT Transactions on Ecology and the Environment, Vol. 80 WIT Press, pp. 433-444, 2005.

[14] Ma, W., Tobin, J.M., Determination and modelling of effects of pH on peat biosorption of chromium, copper and cadmium. Biochem Eng J, 18, pp. 33-40, 2004.

[15] Tchobanoglous, G., Theisen, H., Vigil, S., Integrated solid waste management: Engineering principles and management issues. McGraw-Hill, Inc. 2nd Edition, New York, 1993.

[16] European Union Council Directive, 1999/31/EC of 26 April 1999 on the landfilling of waste. Official Journal of the European Communities, L182/1-19, 0001-0019, 16 July. 1999.

[17] RT I 2003.83.565, Requirements for wastewater discharged into water bodies or into soil. Regulation No. 327 of 19 December 2003 of the Government of the Republic of Estonia, Tallinn, Estonia, In: State Gazette, 2003.




[18] Lõhmus, A. Personal communication, 20 January 2006, Chairman of Väätsa landfill Ltd., Estonia.

[19] Sooäär, M., Activated sludge treatment of leachate from Väätsa landfill. Estonia, Eco-tech 03, Kalmar, Sweden, 47 pp, 2003.

[20] Maves Ltd., Väätsa landfill leachate, surface water and ground water monitoring data. Tallinn, 2005.

[21] Patterson, A.A., Davey, K., Farnan, N., Peat bed filters for on-site treatment of septic tank effluent. In: Patterson, R.A. and Jones, M.J. editors. Advancing On-site Wastewater Systems 25-27th September 2001, Armidale, Lanfax Labs Armidale, pp. 315-322, 2001.

[22] Mæhlum, T., Cold-climate constructed wetlands: Aerobic pre-treatment and horizontal subsurface flow systems for domestic sewage and landfill leachate purification. Agricultural University of Norway, PhD thesis, 1998.

[23] Kadlec, R.H., Knight, R.C., Treatment Wetlands. Lewis Publishers, New York, pp. 3-893, 1996.

[24] Ødegaard, H., Treatment of wastewater. (Rensing av avløpsvann. Tapir forlag. In Norwegian), Trondheim, pp. 133-284, 1992.

[25] Heavey, M., Low-cost treatment of landfill leachate using peat. Waste Manage, 23, pp. 447-454, 2003.

[26] Mann, R.A., Phosphorus removal by constructed wetlands: substratum adsorption. In: Cooper, P.F., Findlater, B.C. editors. Use of Constructed Wetlands in Water Pollution Control, Pergamon Press, Oxford, pp. 97-105, 1990.

[27] Richardson, C.J., Qian, S.S., Craft, C.B., Predictive models for phosphorus retention in wetlands. In: Vymazal, J. editor. Proc. Conf. Nutrient Cycling and Retention in Wetlands and Their Use for Wastewater Treatment, Prague, Inst. Of Botany, Trebon, Czech Republic, pp. 125-150, 1996.




Studies on the distribution of heavy metal Cd in contaminated soils of various particle sizes and removal efficiencies of heavy metal using acid washing

S. F. Cheng1, C. Y. Huang2 & L. S. Hsiao1

1Department of Environmental Engineering and Management, Chaoyang University of Technology, Taiwan 2Ordnance Readiness Development Center, Taiwan

Abstract

Improper treatment and disposal of industrial wastewater and solid wastes result in serious heavy metal contamination of soil. A commonly used method for remediation of the soil contaminated by heavy metals is acid washing. This method is simple in principle, easy to operate and efficient in achieving the removal of heavy metals. The experience gained on soil remediation in Taiwan reveals that the soil particle distribution influences the heavy metal removal efficiency to a large extent. Hence, in this research, soil samples were collected from the contaminated sites and used in the investigation on the distribution of heavy metal Cd in soils of various particle sizes and the efficiencies of acid washing to remove the Cd from the soil of various particle distributions. The results will be used for future engineering implementation of the acid washing technique. The research results indicate that soils containing particles with sizes below 0.150 mm will hold increasing quantities of Cd and organic matter at decreasing particles sizes. As far as the acid washing efficiency is concerned, smaller particle diameters and higher organic matter contents result in decreasing efficiencies. Keywords: soil, Cd, particle size, CEC, organic matter, acid washing.




doi:10.2495/GEO060111

1 Introduction

The rapid industrial development in Taiwan during last several decades have not gone together with adequate planning and management of industrial wastewater treatment and disposal. Irrigation channels have often been used to discharge industrial waste effluent thus causing serious pollution to large tracts of farmland. Among all problems, the contamination caused by heavy metals, which are accumulative and non-degradable, is the most serious. The acid-washing method is commonly used for remediation of heavy metal contaminated soils. This method is simple in principle and easy to implement to come up with good results in a short period of time [1]. It is considered as a preferred method for remediation metal contaminated farmland in Taiwan. However, the experience with several recent cases of soil remediation using the acid-washing method shows that the efficiency depends on the type of soil to a great extent. For sandy loam soil, this method can achieve metal removal efficiency of as high as 90% while for soils containing fine clay particles, a poor separation of the soil particle and the washing liquid causes a great deal of operational problems. Additionally, the soil that contains fine soil particles exhibits poor permeability thus making it impossible to have effective contacts between the washing liquid and soil particles. Hence, the on-site treatment has removal efficiency lower than 30%; it may sometimes not show any removal efficiency at all. The literature information [1] shows that most pollutants exist in the portion of soil particles having less than 63 µm with relatively larger specific surface area of the contaminated soil. Hence, sieving the soil into two portions prior to the acid treatment is recommended. The larger particles have a lower metal contamination and are considered as clean soil; only the smaller particles need to be washed in the subsequent acid-washing treatment [2]. Anderson et al. [3] pointed out that most of the contaminants concentrate in the portion consisting of silts and clays with less than 50 µm diameters. The volume of this portion is a small percentage of the overall soil volume. Thus, the prior sieving procedure may reduce 60~80% volume of the soil to be treated thus reducing the treatment cost. Results obtained by Sheets and Bergquist [4] on PCB and lead contents of soil samples collected at superfund sites indicate that soil particles with diameters less than 74 µm constitute about 8~15% of the total soil mass. Nevertheless, most PCBs exist in the soil particles with less than 74 µm diameters. Smaller particles, which have higher specific surface area and contain more organic substance to enhance the adsorption of contaminants, show higher PCB concentrations. The distribution of heavy metals generally follows the similar trend that smaller particles have higher metal concentrations. Particles with diameters between 0.425 to 2 mm have much higher metal concentrations. Thus, the presence of small lead particles is regarded as the cause of the observed high metal concentration in this soil particle diameter range. Di Palma and Medici [5] combined sulfuric acid and EDTA as the washing agent to wash copper contaminated soil and reported that soils with higher content of organic




substances exhibit lower copper removal efficiencies mainly because the complex reactions between copper and the organic substances. In this research, soil samples collected from the heavy metal contaminated field in Taiwan are used to study particle distribution as well as concentrations of heavy metal Cd and organic substances in soil particles of various sizes. The objective is to understand the correlation between distribution of heavy metals and soil particle size as well as organic substance contents. Additionally, the efficiency of extracting the metals from the contaminated soils using various acid solutions will be studied and the results will be used as references for future implementation of the acid-washing method.

2 Materials and methodology

2.1 Soil preparation

Samples used in this research were collected from the top 15 cm layer of a cadmium-contaminated farmland located at Huwei Township, Yunlin Hsieh in central Taiwan. The collected sample was thoroughly mixed; a portion of it was subject to basic characteristic analyses on pH, CEC, organic substances, water content and heavy metals. The remaining portion was dried, crushed and screened through a 10-mesh (2 mm) sieve to remove large pebbles and impurities. The screened samples were then separated into 4 groups of different particle ranges using the wet screening and the gravitational sedimentation methods. First, aqueous (NaPO3)6 solution was added to the sample to disperse soil particle. After thorough mixing, the soil solution was passed thorough 100 mesh (0.150 mm) and then 270 mesh (0.053 mm) wire screens for separating the particles into “sand 1” with particle diameters between 2 and 0.150 mm and “sand 2” with particle diameters range of 0.150 to 0.053 mm. The solution that had passed through the wire screens was poured into a 1000 ml sedimentation container. Based on Stock’s law, the particles settled at the bottom after 7 hours and 36 minutes are silt with diameters between 0.053~0.002 mm. The remaining particles suspended in the solution are clay with diameters below 0.002 mm; they were filtered out through 0.20 µm and then dried.

2.2 Soil characteristic analyses

Both the original and the processed soil samples were analyzed for organic substance and CEC. The organic substance analyses were done using the combustion method by heating 20 grams of the dried soil sample at 500~600oC for 4 hours. The weight loss after heating is taken as the soil organic substance content. For measuring the soil CEC, 4 grams of the dried sample was mixed with sodium acetate solution such allowing Na+ ion to replace all exchangeable cations contained in the soil sample. The mixture was then added with ammonium acetate to replace all Na+ ions by ammonium ions; the concentration of replaced Na+ ions is determined using the induced couple plasma (ICP) method as the soil CEC.




2.3 Soil heavy metal analyses

The dried soil samples were digested with aqua regia by adding 21 ml concentrate hydrochloric acid and 7 ml nitric acid with 3 g of soil slowly. After mixing, the sample was undisturbed for 16 hours. It was then diluted to 100 ml and filtered through 0.20 µm filter paper. The filtrate was analyzed for cadmium using the ICP method.

2.4 Acid-washing of soil

Soil samples containing various particle sizes were washed with single acids including two inorganic acids (hydrochloric acid and nitric acid) and one organic acid (citric acid) as well as the mixture of hydrochloric and citric acids to study the efficiencies of heavy metal removal from the acid-washed soil. The acid concentrations were 0.01 M, 0.05 M and 0.1 M; the mixed acid solution was prepared by mixing 1 volume of hydrochloric acid and 1 volume of citric acid to final concentrations of 0.01 M, 0.05 M and 0.1 M. An aliquot of 10 ml washing solution was added to 1 g soil sample. After mixed on a 150 rpms vibrating shaker for 10 minutes, the mixture was filtered through 0.20 µm filter paper. The washed soil sample was then analyzed for heavy metals to calculate the metal removal efficiency. All analyses were carried out with triplicate samples and the results were used to calculated average and standard deviations.


3.1 Soil characteristics

Table 1 lists the results of basic soil characteristics. According to the USDA soil classification standards, the contaminated soil is sandy loam containing 125.2 ~ 128.2 mg/kg cadmium. It has average 22.7 meq/100 g CEC and contains 5.2% organic matter.

Table 1: Properties of the Cd-contaminated soils.

pH 6.83 (0.09)

Organic content (by weight)% 5.15 (0.05)

Water content (by weight)% 2.81 (0.06)

CEC (meq per 100 g) 22.7 (1.30)

Sand (by weight)% 68.5 (1.63)

Silt (by weight)% 27.0 (0.21)

Clay (weight)% 5.5 (0.46)

Values in parentheses denote standard deviations (n=3).




3.2 Soil particle and Cd distributions

Figure 1 shows variations of the particle and cadmium distributions. The results indicate that “sand 2” particles; i.e. particle with diameters between 0.150 mm to 0.053 mm, constitute a major portion of the soil of nearly 48%. Additionally, the cumulative mass of all soil particles with diameter greater than 0.053 mm is 70% of the total soil mass. Fig. 1 also indicates that the cadmium content in these 70% soil particles is much lower than those in the clay and silt portions. The “sand 2” particles have the highest mass percentage but the lowest cadmium content of 48.74 mg/kg. On the contrary, the clay portion has the lowest mass percentage but the highest cadmium concentration of 557.37 kg/mg that is 11 time the cadmium concentration contained in “sand 2”. During the acid-washing process, clay and silt cannot be effectively separated from the liquid due to their small particle sizes thus lowering the washing efficiency and raising the treatment costs. Adopting the screening procedure to separate smaller particles that contain higher cadmium concentrations to be treated separated will greatly enhance the remediation efficiency for heavy metal removal.

Figure 1: Distribution of particle sizes and Cd content of various particle sizes for contaminated soil. Error bars are ± standard deviations for n=3.

3.3 Correlation between soil characteristics and heavy metal contents

The correlation between organic matter content and heavy metal concentration for soil samples containing particles of various sizes is shown in Figure 2. For sand 2 to clay, smaller particles contain higher concentrations of organic

20.46

48.04

26.98

5.52

77.2248.74

358.20

557.37

0

200

400

600

800

Sand1 Sand2 Silt Clay

0

10

20

30

40

50




Weight % Cd conc.

Cd conc.(mg/kg)

weight distribution %

substances. Since the average soil has higher humic substances associated with particles of smaller diameters, the organic content is expected to be higher in soils containing smaller particles. This is contrary to the observation made in this study that “sand 1” soil has the organic content more than “sand 2”. However, most of these organic substances are non-decomposed animal and vegetation impurities. Results shown in Figure 2 indicate similar trends of variations for heavy metal concentration and organic content in soils of various particles. For “sand 2” to clay, smaller particles contain more organic substances and heavy metals. While “sand 1” shows the increasing organic content is not seen by a proportional increase of cadmium although the cadmium is seen to increase. This observation may be caused by the fact that most organic substances contained in “sand 1” are large non-humic thus they do not have high affinity for heavy metals as those contained in “sand 2”.

Figure 2: The correlation between organic matter content and heavy metal concentration for soil samples containing particles of various sizes.

The correlation between CEC and heavy metal content for soils containing particles of various sizes as shown in Figure 3 indicates that “sand 1” soil has the lowest CEC. Further, variations of CEC and organic content show similar tendency. In “sand 2” and Clay, higher CECs are associated with small soil particles while “sand 1” has higher CEC than “sand 2”. The small increase of CEC in “sand 1” proves that portions of the organic substances contained in “sand 1” are non-decomposed large particles and they do not have much adsorption capacity for heavy metals.

0

100

200

300

400

500

600

700

800


Cd

conc

. ( m

g/kg

)

0.00

1.00

2.00

3.00

4.00

5.00

6.00

orga

nic

cont

ent %

Cd conc. organic matter




Figure 3: The correlation between CEC and heavy metal content for soils containing particles of various sizes.

69.57 70.95

52.51

25.51

86.73 84.85

51.76

30.05

91.42 96.57

54.03

33.78

0 10 20 30 40 50 60 70 80 90

100


Cd

rem

oval

eff

icie

ncy

%

0.01 M 0.05 M 0.1 M

Figure 4: The metal Cd removal efficiencies from soil containing particles of various sizes using hydrochloric acid.

3.4 Efficiencies for heavy metal removal by acid-washing for particles of various sizes

The metal Cd removal efficiencies from soil containing particles of various sizes using hydrochloric acid, nitric acid, citric acid and hydrochloric acid/citric acid mixture are shown in Figure 4 to 7, respectively. Results shown in all figures

0

100

200

300

400

500

600

700

800


Cd

conc

. (m

g/kg

)

0

20

40

60

80

100

120

140

160

CEC

(meq

/100

g )

Cd CEC




illustrate that the “sand 1” and “sand 2” particles have much higher removal efficiencies than clay and silt with as high as 65% difference. Using the single acid for washing, inorganic acids are more efficient than organic acids; hydrochloric acid is more efficient than nitric acid. For “sand 1” and “sand 2” particles, the acid-washing method (by 0.1 M HCl or HNO3) can remove more than 90% cadmium while the efficiency is around only 50% for silt and 32% for clay. Figure 7 shows the cadmium removal efficiency using the hydrochloric/citric mixture. The results show that the mixed acid solution is 2.5% more effective than single acids in removing cadmium from soils of large particles.

Figure 5: The metal Cd removal efficiencies from soil containing particles of

various sizes using nitric acid.

56.40 55.43

12.088.57

80.77 72.43

44.88

15.98

86.22 75.69

53.69

25.22

0 10 20 30 40 50 60 70 80 90

100


Cd

rem

oval

eff

icie

ncy

%

0.01 M 0.05 M 0.1 M

Figure 6: The metal Cd removal efficiencies from soil containing particles of various sizes using citric acid.

65.89 69.58

35.42

22.44

86.50 82.67

41.25

25.99

90.47 90.35

44.51

31.31

0 10 20 30 40 50 60 70 80 90

100


Cd

rem

oval

eff

icie

ncy

%

0.01 M 0.05 M 0.1 M




Figure 7: The metal Cd removal efficiencies from soil containing particles of

various sizes using the hydrochloric/citric (1:1) mixture.

4 Conclusions and recommendations

Results obtained in this research show that the organic content, CEC and heavy metal content in the soil are somewhat consistent. Particles with diameters greater than 0.053 mm constitute more than 70% of the total soil mass but contain the lowest metal contents. They have the high metal removal efficiency and are easily separated from liquid. On the contrary, sol particles with diameters less than 0.053 mm contribute to 30% of the total soil mass especially the clay portion that is only 5.5% of the total mass. This portion of soil has the highest metal content and the lowest removal efficiency in addition to being more difficult to be separated by sedimentation from the liquid. The results obtained in this research indicate the use of a prior screening procedure to removal particles of small diameters that have the lowest acid-washing efficiencies. The acid-washing carried out on large particles will be efficient in removing heavy metal while the separated smaller particles can be treated using other treatment methods such that the overall treatment efficiency can be optimized.

References

[1] Evanko, C.R. & Dzombak, D.A., Remediation of metals-contaminated soils and groundwater. Technology evaluation report, TE-97-01, Ground-water Remediation Technologies Analysis Center, 1997.

[2] Rosetti, P.K., Possible methods of washing fine soil particles contaminated with heavy metals and radionuclides. M.S. Thesis, Carnegie Mellon University, Pittsburgh, PA, 1993.

74.57 81.84

37.47

20.47

82.47 89.58

48.57

25.41

93.84 97.48

56.99

32.57

0 10 20 30 40 50 60 70 80 90

100


Cd

rem

oval

eff

icie

ncy

%

0.01 M 0.05 M 0.1 M




[3] Anderson, R., Rasor, E. & Ryn, F.V., Particle size separation via soil washing to obtain volume reduction. Journal of Hazardous Materials, 66, pp. 89-98, 1999.

[4] Sheets, R.G. & Bergquist, B.A., Laboratory treatability testing of soils contaminated with lead and PCBs using particle-size separation and soil washing. Journal of Hazardous Materials, 66, pp. 137-150, 1999.

[5] Di Palma, L. & Medici, F., Recovery of copper from contaminated soil by flusing. Waste Management, 22, pp. 883-886, 2002.




Mitigation of alkaline mine drainage in a natural wetland system

J. Kelly1, P. Champagne2 & F. Michel1 1Department of Earth Sciences, Carleton University, Canada 2Department of Civil Engineering, Queen’s University, Canada

Abstract

Many studies have focused on the generation and mitigation of acidic drainage generated when sulfide bearing material is exposed to the atmosphere and undergoes oxidation. Neutral or alkaline mine drainage can be produced from mining waste containing little or no sulfides, and such drainage can also contain elevated metal concentrations, potentially impacting receiving environments. The goal of this study was to characterize the biogeochemical interactions occurring throughout a natural wetland located in the Farr Creek drainage area in Cobalt Canada and to evaluate the ability of the system to effectively attenuate alkaline mine drainage. The biological characterization of the sediment samples demonstrated the presence of acid producing bacteria in consistent numbers with sulphate reducing bacteria and iron reducing bacteria. The data suggested that the acid produced by these active bacterial populations was immediately neutralized by the dissolution of carbonate minerals within the tailings, yielding a neutral to alkaline drainage. The distribution of metals including As, Co, Cu, and Zn throughout the sediments, pore water, and vegetation samples collected at various core locations the metal mass was primarily bound in the sediments or adsorbed onto organic matter or oxide fractions of the sediments. Phytoremediation processes involving Typha latifolia were shown to attenuate metals, particularly Cu and Zn. Adsorption onto organic matter and oxides was another attenuation pathway that significantly improved metal retention. Iron and sulfate reduction were also found to lead to the formation of metal sulfide precipitates, thereby immobilizing the metals. Keywords: wetland, alkaline mine drainage, heavy metals, sequential extraction, phytoremediation, sulphate reducing bacteria, iron reducing bacteria, acid producing bacteria, Typha latifolia.




doi:10.2495/GEO060121

1 Introduction

Anthropogenic sources of metals, including mining wastes, runoff waste streams, and air fall deposition from industrial operations, are becoming increasingly an issue for contamination of downstream environments and health impacts to aquatic species and humans. Acid mine drainage (AMD), which results from the oxidation of sulfide bearing material in waste rock or tailings, has received considerable attention. The oxidation of sulfides, which may be catalyzed by microbial populations, produces sulfuric acid, which can result in extremely low pH waters and enhance heavy metal dissolution. Metals are typically more soluble under low pH conditions due to the increased competition between protons and metal cations for organic ligands (Schnoor [1]). The dissolved metals have the potential to migrate offsite resulting in potential negative impacts to both humans and aquatic organisms. The acidity produced can potentially be neutralized if there is adequate carbonate minerals present in the tailings and mixing waters, resulting in a net neutral or alkaline mine drainage (NAMD). NAMD can also be produced from mining wastes containing little to no sulfides. Such drainage, although non-acidic, can also contain elevated dissolved metal concentrations, which can potentially impact receiving environments. Historically many mining companies have used active treatment options for mitigating impacts to the environment as a result of acid mine drainage. These options typically involved chemical additions of neutralizing agents such as lime, which was quite expensive in the long-term. This sparked the movement towards passive treatment options, one of which is the use of constructed or natural wetlands to attenuate the metals. As a result, it is important to characterize metal distributions and transformations in aquatic environments to better understand the geochemical and biological processes regulating transformations in these environments. The main objective of this research was to characterize the biogeochemical interactions occurring throughout a natural wetland receiving NAMD and to evaluate the ability of the wetland in the attenuation of metals. The study area is located in the Farr Creek drainage area, in Cobalt, Canada. The wetland area investigated is confined to a relatively narrow northeast oriented valley that is bounded to the south by Crosswise Lake and the remnants of a gravel dam, and bounded to the north by a water level control dam. Alkaline tailings underlie the entire study area as well as upgradient of the study area. Mill Creek, which transports metal loadings from several upstream tailings deposits and organic loadings from the municipal wastewater lagoon flows into Farr Creek as shown on Figure 1. The northern portion of the area is maintained under a water cover for much of the open water season, whereas the southern portion of the area is relatively dry throughout the summer and fall. This is further evidenced by the establishment of grasses and sedges in the drier areas, while waterlogged areas are primarily populated with cattails (Typha latifolia).




Figure 1: Site plan indicating core sample locations.

2 Experimental approach

2.1 Field sampling

Five 5 cm diameter x up to 1.4 m long sediment cores were collected from the study area in September 2004 (Figure 1). The cores were capped and sealed from the atmosphere throughout the extraction process. They were kept in a refrigerator and transported in coolers with icepacks to ensure the samples remained below 4°C until they were prepared for analysis.

2.2 Laboratory methods

Sediment samples were collected from the 5 cores at approximately 25 cm intervals and analyzed for acid producing bacteria (APB), iron reducing bacteria (IRB) and sulfate reducing bacteria (SRB), sequentially extracted metals (SEM), water content (WC), and organic matter (OM) content. Pore water was extracted from the remaining sediment samples by ultracentrifugation and analyzed for alkalinity, pH, dissolved oxygen (DO), Fe(II), sulfate, sulfide, and dissolved metals. Cattail shoot samples, collected near each of the monitoring wells, were analyzed for metals. The most probable number (MPN) technique was used to estimate the number of APB, IRB, and SRB in each sample. MPN values were calculated for each set of samples from statistical tables (Cochran [2]) and the results were expressed as colony forming units per gram of sediment dry weight (CFU/gdw). Ten different dilutions, at 1 mL in 10 mL were completed with 5 replicates of each dilution.




Sixteen sediment section samples taken at depths corresponding to the microbiology samples, were analyzed in duplicate for metals using a SEM procedure outlined by Tessier et al. [3] where the extracted metals are separated into five fractions: exchangeable, bound to carbonates, bound to Fe and Mn oxides, bound to OM, and residual fraction. Seven cattail (T. latifolia) shoot samples were washed with deionized water, broken into small pieces and crushed with a mortar and pestle. Thirty mL of 4:1 HNO3:HCl mixture was added to 200 mg of ashed cattail shoots. The mixture was digested at 100-130°C until the solution became clear. Prepared sequential extraction and vegetation samples were stored below 4°C until further analysis and the metals in the samples were analysed by ICPMS.

3 Results

3.1 Sequential Extraction of Metals (SEM)

Figures 2(a)-(d) present the results of the SEM for selected metals from the 5 cores. Arsenic was predominantly associated with the residual fraction, ranging from 35 to 75%, while the oxide fraction ranged from less than 5% to 31% and generally decreased with depth. The fraction of As associated with the OM fraction of the sediment ranged from 10 to 35%. Cobalt was strongly associated with the OM fraction of the sediment, ranging from 22 to 80%. Cobalt was also associated with the residual fraction, ranging from 5 to 40%. Cobalt associated with the oxide fraction ranged from less than 5% to 32% and generally decreased with depth. Cobalt associated with the carbonate fraction of the sediment ranged from 8 to 30%. Cobalt has been shown to have an affinity for carbonates, thus, it is likely that there was competition for sorption sites between the carbonate and other fractions of the sediment. Copper was primarily associated with the residual fraction of the sediment, ranging from 10 to 50%, as well as the OM fraction of the sediment, 40 to 85%. Zinc was predominantly associated with the residual fraction, ranging from 15 to 65%. Zinc associated with the oxides fraction ranged from less than 10 to 25% and generally decreased with depth. Zinc associated with the OM fraction of the sediment ranged from 35 to 45%.

3.2 Pore water chemistry

Figures 3(a)-(h) display porewater concentration profiles with depth. In core S2, both As and Co concentrations increased with depth, while in cores S3 and S5, As and Co concentrations increased to a depth of 50 cm and then decreased. In cores S1 and S4, no observable trends were noted. For cores S3, S4, and S5, Cu concentrations increased to a depth of 50 cm and then decreased. There were no observable Cu trends with depth in cores S1 and S2. Zinc concentrations increased with depth in cores S1 and S3. In core S5, Zn showed increasing concentrations to a depth of 50 cm, followed by decreases, while in core S4 Zn decreased with depth. In core S2, Zn showed no observable trend. Calcium values generally remained fairly constant with depth at all core locations.




01020304050607080

S1-

25

cm

S1-

50

cm

S1-

75

cm

S2-

25

cm

S2-

50

cm

S2-

75

cm

S3-

25

cm

S3-

50

cm

S3-

75

cm

S4-

25

cm

S4-

50

cm

S4-

75

cm

S4-

100

cm

S5-

25

cm

S5-

50

cm

S5-

75

cm

Location

Perc

ent D

istr

ibut

ion

(%)

Exchangeable Carbonates Fe, Mn-oxides Organic Residual (a) Arsenic

0102030405060708090

S1-

25

cm

S1-

50

cm

S1-

75

cm

S2-

25

cm

S2-

50

cm

S2-

75

cm

S3-

25

cm

S3-

50

cm

S3-

75

cm

S4-

25

cm

S4-

50

cm

S4-

75

cm

S4- 1

00 c

m

S5-

25

cm

S5-

50

cm

S5-

75

cm

Location

Perc

ent D

istr

ibut

ion

(%)

Exchangeable Carbonates Fe, Mn-oxides Organic Residual

(b) Cobalt

0102030405060708090

100

S1- 2

5 cm

S1- 5

0 cm

S1- 7

5 cm

S2- 2

5 cm

S2- 5

0 cm

S2- 7

5 cm

S3- 2

5 cm

S3- 5

0 cm

S3- 7

5 cm

S4- 2

5 cm

S4- 5

0 cm

S4- 7

5 cm

S4-

100

cm

S5- 2

5 cm

S5- 5

0 cm

S5- 7

5 cm

Location

Perc

ent D

istr

ibut

ion

(%)

Exchangeable Carbonates Fe, Mn-oxides Organic Residual (c) Copper

0

10

20

30

40

50

60

70

S1-

25

cm

S1-

50

cm

S1-

75

cm

S2-

25

cm

S2-

50

cm

S2-

75

cm

S3-

25

cm

S3-

50

cm

S3-

75

cm

S4-

25

cm

S4-

50

cm

S4-

75

cm

S4- 1

00 c

m

S5-

25

cm

S5-

50

cm

S5-

75

cm

Location

Perc

ent D

istr

ibut

ion

(%)

Exchangeable Carbonates Fe, Mn-oxides Organic Residual (d) Zinc

Figure 2: Sequential extraction results.

w

ww

.witpress.com

, ISSN 1743-3541 (on-line)

©2006 W

IT PressW

IT Transactions on Ecology and the Environment, V

ol 89,

Geo-Environm

ent and Landscape Evolution II 119

0

20

40

60

80

100

120

140

0 100 200 300 400 500 600 700

C onc e nt r a t i on ( mg/ L)

Core S1 Core S2 Core S3 Core S4 Core S5

0

20

40

60

80

100

120

140

0 0.2 0.4 0.6 0.8

C onc e nt r a t i on ( mg/ L)


(a) Sulphate (b) Sulphide

0

20

40

60

80

100

120

140

0 0.1 0.2 0.3 0.4 0.5 0.6

Conc e nt r a t i on ( mg/ L)


0

20

40

60

80

100

120

140

160

0 100 200 300 400

C oc ne nt r a t i on ( mg/ L)


(c) Ferrous iron (d) Calcium

0

20

40

60

80

100

120

140

160

0 2 4 6 8 10

Conc e nt r a t i ons ( mg/ L)


0

20

40

60

80

100

120

140

160

0 0.2 0.4 0.6 0.8 1



(e) Arsenic (f) Cobalt

0

20

40

60

80

100

120

140

160

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35



0

20

40

60

80

100

120

140

160

0 0.1 0.2 0.3 0.4 0.5 0.6



(g) Copper (h) Zinc

Figure 3: Porewater concentrations.




A reduction in sulfate concentrations with depth was observed in all of the cores. Increasing or measurable Fe(II) and HS- concentrations with depth were only noted in cores S3, S4, and S5, a further indication of the development of reducing conditions with depth. Porewater DO concentrations remained fairly consistent with depth, ranging from 1.50 to 2.15 mg/L. The pH values were also relatively consistent throughout the porewater profile, averaging 7.0.

3.3 Vegetation

Table 1 presents the results of selected metal extractions from the T. latifolia leaf samples collected. Significant concentrations of selected metals were taken up into the leaf samples. Higher metal concentrations were extracted from leaf samples collected near the wetland inlet, which was expected due to higher metal loading in this area, and is supported by other studies (O’Sullivan et al. [4]; Miller et al. [5]).

Table 1: Extracted metal concentrations from T. latifolia.

Table 2: Microbiological enumeration results.

3.4 Microbiology

Table 2 presents the results from the bacterial enumerations. There appear to be equally strong numbers of acid producing (oxidizing) and reducing bacteria, with populations ranging from 102 to 106 MPN/gram dry weight. The presence of all three types of microorganisms throughout the wetland and at relatively




Mr Smith

consistent concentrations with depth at each core location may suggest that both oxic and anoxic conditions exist throughout the wetland.

4 Discussion

This natural wetland appears to be an overall sink for metals, with the majority of the metals being tied up in the sediments or being taken up into the leaves of Typha latifolia. A detailed mass balance would be required to confirm this. In this study, results from the SEM indicated that most of the metals were associated with the residual and OM fractions of the sediment. The residual fraction accounts for silicate minerals and sulfides. This is important since one of the sinks for select metals is the formation of metal sulfide precipitates under reducing conditions. Metals associated with the exchangeable fraction are typically highly mobile and are often correlated with the cation exchange capacity of the sediment (Gambrell et al. [6]). The surfaces of oxides and OM are also prime adsorption sites for metals. Previous studies (Fuller [7]; Lagmuir et al. [8]; Soprovich [9]) have shown that As readily adsorbs to or coprecipitates with Fe and Mn oxides. The results from the SEM indicated that much of the Fe was associated with the residual fraction of the sediment and not the oxide fraction, while much of the As was associated with the residual and OM fractions of the sediment. Thus, it would appear that given the presence of adequate Fe and Mn oxides under oxidizing conditions and adequate sulfides or OM under reducing conditions, As is likely to be well attenuated in this wetland. Zinc and Cu were predominantly associated with the OM and residual fractions of the sediments. Cobalt was primarily associated with the residual fraction, however, it was also significantly associated with carbonates. Cobalt has known affinities for carbonates (Brookings [10]) and therefore competition for sorption sites between OM and carbonates are likely. The porewater sulfate and sulfide concentration profiles indicated that sulfate reduction was occurring throughout the wetland. The microbiology results identified the presence of APB populations in consistent numbers with the SRB and IRB populations. This may indicate that all these bacteria are active and the acid produced by these populations is immediately neutralized by the dissolution of carbonate minerals within the tailings. Above pH 6.3, which was the case throughout much of the study area, the dissolution of one mole of calcite consumes one mole of H+ (Blowes et al. [11]). This would also account for the elevated dissolved concentrations of Ca2+ in the pore water. It is also possible that APB populations are supported at depth throughout the wetland in the vicinity of vegetation root zones. Oxygen is transported by wetland plants to the root zones generating localized zones of oxidation. This could allow for sustainable populations of APBs throughout the wetland. In regions located away from the root zones, conditions are likely anoxic, which are favourable for biogenic iron and sulfate reduction. It is also likely from the microbiology results that there is considerable competition between bacterial species for organic substrates. This was supported by changes in population trends at different locations in the study area, where small changes in sulfate and Fe(II)




concentrations occurred. Prior studies have shown competitiveness between SRB and IRB, which appears to be dependent on electron donor availability and pore water chemistry (Fortin, pers. comm..). It should also be noted that Fe reduction can occur both biotically and abiotically, while sulfate reduction can only be mediated by SRB. The results from the extraction of metals from T. latifolia leaf samples indicated that up to 20% of the total mass of metals was taken up by the leaves. Previous studies have reported metal uptake in both the leaves and root zones of T. latifolia and found much higher percentages of metals associated with the roots than the leaves (Jackson et al. [12]). This suggests that phytoremediation may be an even more powerful source of metal immobilization in this wetland. There has been conflicting research on the ability of wetland plants to immobilize metals in the long term. This is in part due to the fact that plants transport oxygen to their root zones generating a zone of radial oxygen loss, which results in the generation of localized oxic conditions in the immediate vicinity of the roots, resulting in the oxidation of reduced sulfur and iron species. Under these conditions metals such as As and Co may adsorb or coprecipitate with the Fe oxides, however, metals such as Zn would likely remain mobilized in the dissolved state (Jacob and Otte [13, 14]). Background conditions were not known for the T. latifolia in the area, which would have been useful in determining the degree of metal uptake by T. latifolia as a direct result of the mine tailings deposits. However, the vegetation was flourishing in the area and thus it was not expected that the presence of metals at these concentrations were having a negative effect on T. latifolia. It is possible, however, that the growth of other wetland plants was inhibited due to elevated metal loading to the wetland system. There was also considerable variability between sample locations. This could most likely be attributed to variations in metal loadings at each location.

5 Conclusions

The data collected suggest that this is a relatively stable system. Should the current state of the system change, such as the input of a waste stream elevated in organic substrate, it is likely that a considerable amount of the metals retained within this system would become mobilized. Over time, the geochemical processes regulating metal mobilizations throughout the system would change, resulting in different biogeochemical controls on the metals throughout this system. Alkaline drainage systems are geochemically different than acidic drainage systems. Alkaline systems can immobilize metals under oxic conditions, such as adsorption onto OM or oxide precipitates or uptake by wetland vegetation, and anoxic conditions via microbial transformations, such as biogenic sulfate reduction, generating reduced metal sulfide precipitates. Acidic drainage systems typically require the formation of strongly reducing conditions in order to immobilize metals. These results have demonstrated the importance of considering the geochemical conditions of the wetland system treating the mine drainage, as well as having a detailed understanding of the metals of




concern within the mining waste, as these will geochemically interact based on redox conditions, presence of sulfides, Fe and Mn oxides, and organic matter.

References

[1] Schnoor, J., Modeling trace metals (Chapter 8). Environmental Modeling: Fate & Transport of Pollutants in Water, Air, Soil, eds. J. Schnoor & A. Zehnder, Wiley, New York, 381-451, 1996.

[2] Cochran, W.G., Estimation of bacterial densities by means of the most probable number. Biometrics, 6, pp. 105-116, 1950.

[3] Tessier, A., Fortin, D., Belzile, N., DeVitre, R, & Leppard, G.G., Metal sorption to diagenic iron & manganese oxyhydroxides & associated organic matter: narrowing the gap between field & laboratory measurements. Geochimica et Cosmochimica Acta, 60, pp. 387-404, 1996.

[4] O’Sullivan, A.D., Moran, B.M., & Otte, M.L., Accumulation & fate of contaminants in substrates of wetlands constructed for treating mine wastewater. Water, Air, & Soil Pollution, 157, pp. 345-364, 2004.

[5] Miller, W., McFee, W., & Kelly, J., Mobility and retention of heavy metals in sandy soils. J.Environmental Quality, 12(4), pp. 579-584, 1983.

[6] Gambrell, R.P., Wiesepape, J., Patrick Jr., W., & Duff, M., The effects of pH, redox, and salinity on metal release from a contaminated sediment. Water, Air, & Soil Pollution, 57-58, pp. 359-367, 1991.

[7] Fuller, J., Surface chemistry of ferrihydrite: Part 2. Kinetics of arsenate adsorption and coprecipitation. Geochimica et Cosmochimica Acta, 57, pp. 2271-2282, 1993.

[8] Lagmuir, D., Mahoney, J., MacDonald, A., & Rowson, J., Predicting arsenic concentrations in the porewaters of buried uranium mill tailings. Geochimica et Cosmochimica Acta, 63(19/20), pp. 3379-3394, 1999.

[9] Soprovich, E., Arsenic Release from Oxide Tailings Containing Scordite, Fe-Ca Arsenates, & As-Containing Geothites. Environmental Protection Branch-Yukon Division, Environment Canada, Whitehorse, Yukon, 1995.

[10] Brookings, D.G., Eh-pH Diagrams for Geochemistry. Springer-Verlag, USA, 1988.

[11] Blowes, D.W., Jambor, J.L., & Hanton-Fong, C.L., Geochemical, mineralogical and microbiological characterization of a sulfide-bearing carbonate-rich gold mine tailings impoundment, Joutel, Quebec. Applied Geochemistry, 13(6), pp. 687-705, 1998.

[12] Jackson, L.J., Kalff, J., Rasmussen, J.B., Sediment pH & redox potential affect the bioavailability of Al, Cu, Fe, Mn & Zn to rooted aquatic macrophytes. Can. J. Fish & Aquatic Science, 50, pp. 143-148, 1993.

[13] Jacob, D.L., Otte, M.L., Long-term effects of submergence and wetland vegetation on metals in a 90 year old abandoned Pb-Zn mine tailings pond. Environmental Pollution, 130, pp. 337-365, 2004.

[14] Jacob, D.L., Otte, M.L., Influence of Typha latifolia & fertilization on metal mobility in two different Pb-Zn mine tailings types. Science of the total environment, 333, pp. 9-24, 2004.




A sequential aerated peat biofilter system for the treatment of landfill leachate

P. Champagne1 & M. Khalekuzzaman2 1Department of Civil Engineering, Queen’s University, Canada 2Department of Civil and Environmental Engineering, Canada

Abstract

In recent years, researchers have identified peat as an alternative low-cost filter medium for on-site wastewater treatment, including landfill leachate. Peat possesses several physical, chemical and biological characteristics that make it a favorable filter medium for the mitigation of contaminants. The effectiveness and the impact of clogging of peat biofilter in terms of organic (COD, CBOD5), ammonia (NH3-N) and total suspended solid (TSS) loading are crucial in the operation of such systems. The main purpose of this research was to evaluate the performance of a bench-scale sequential aerated peat biofilter system treating landfill leachate at different hydraulic loading rates (HLRs) under continuous flow condition. The system consists of two major components: an aeration chamber with an attached growth media, followed by a peat biofilter. The leachate was aerated at a constant air flow rate of 3.40 m3/day for a hydraulic retention times (HRTs) of 2 or 5 days. The aerated leachate was then fed to two sets of triplicate peat columns, which were operated at average HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day. The result of the study showed that similar CBOD5, COD, NH3-N and TSS removal efficiencies and column life expectancies could be obtained from the two different hydraulic loading rates to the peat biofilter. However, the HRT in the aeration basin was found to significantly increase the life expectancy of the peat biofilter by reducing the overall contaminant loading to the biofilter. For a HRT of 5 days and constant air flow rate of 3.4 m3/day 99% NH3-N was removed in the aeration tank after 3 weeks. Removal efficiencies above 80%, 90% and 86 % were noted for COD, CBOD5 and NH3-N, respectively, in the peat columns after 6 weeks of operation. Keywords: peat, landfill leachate, aeration, biofilm, hydraulic loading, leachate treatment.




doi:10.2495/GEO060131

1 Introduction

The Trail Road landfill in the City of Ottawa, commissioned in 1980, generates an average rate of 190 m3 of leachate per day. Currently, leachate from Trail Road landfill is hauled by tanker truck for treatment and discharge at the Robert O. Pickard Environmental Center (ROPEC), the City’s wastewater treatment facility. However, the concentrations of several contaminants of the leachate exceed or closely approach the City’s Sewer Use By-law limit, particularly TKN, TSS, BOD5, H2S, boron, chloride, xylene, toluence, and barium. As such, the solid waste disposal facility must pay a surcharge for those contaminants that exceed the City Sewer Use By-Law limits. An on-site treatment system pre-treating the landfill leachate could reduce operational costs to the landfill, by bringing the landfill leachate to compliance with the Sewer Use By-law limits. In recent years, studies (Heavey, [1]; Kinsley et al., [2]; Kennedy and Van Geel, [3]; Lyons and Reidy, [4]; Talbot et al., [5]; Viraraghavan and Ayyaswami, [6]; Rock et al., [7]) have identified peat as an alternative low-cost filter medium for on-site wastewater treatment, including landfill leachate. Peat possesses several characteristics that make it a favorable filter medium for contaminants removal, such as high water holding capacity (Bergeron, [8]), low density (Buttler et al., [9]), large surface area (>200 m2/g) (McLellan and Rock, [10]), high porosity (Mclellan and Rock, [10]; Buttler et al., [9]; Mitsch and Gosselink, [11]), and excellent ion exchange properties (Sharma and Forster, [12]; Mckay, [13]). The properties of peat depend on several factors, including the ambient conditions during its formation, the extent of its decomposition and the method of harvesting (Couillard, [14]). To date, limited information regarding the behavior of peat filter systems under varying contaminant, as well as hydraulic loading rates when operated in a biofilter configuration is available. In addition, the treatment efficiency and the total operational life of the peat filter systems, are vulnerable to varying contaminant loads, particularly organic (COD, BOD5), NH3-N, and TSS concentrations, as well as hydraulic loading rates.

2 Methodology

Contaminant removal efficiency and the total operational life of the peat biofilter system are dependent on organic (COD, BOD5), NH3-N, and TSS constituent, as well as hydraulic loading rates (HLR). The removal efficiency and operational life of a peat biofilter preceded by an aeration chamber with a support media to promote the growth of an attached biofilm were investigated, under different hydraulic and contaminant loading rates and continuous flow condition. The attached growth medium provides a large active surface area and texture, which can promote the rapid growth of a biofilm, thereby, reducing contaminant loads, particularly NH3-N and BOD5, on the peat filter leading to an increase in the operational life of the peat biofilter system. Laboratory investigations were conducted using the bench-scale experimental set-up illustrated in Figure 1.




Figure 1: Sequential aerated peat biofilter experimental set-up.

Raw leachate was collected in 20 L (28 cm X 23 cm X 40 cm) containers from the City of Ottawa Trail Road landfill and stored in a refrigerators at 4ºC. The raw leachate was passed through a cylindrical aeration tank (64 cm X 44 cm ID) using a peristaltic pump at a flow rate 4.5 L/day, which was equal to the sum of the influent rates of the peat filter columns. An air pump, MAP2X Maxair 2XL, was utilized to inject air into the leachate at a constant air flow rate of 3.40 m3/day. To aerate effectively, a 28 cm long perforated hose with a 1 cm OD was placed at the base of the aeration tank. In addition, a spun plastic medium was used in the aeration basin to get a better performance of the aeration basin by providing a support medium for biofilm growth. The aerated leachate was fed from the aeration basin to two sets of triplicate peat columns by two sets of peristaltic pumps. One set of triplicate columns was fed at a rate of 8.28 cm3/cm2/day, while the other set was fed at a rate of 10.82 cm3/cm2/day. Masterflex® TYGON tubing was used to connect each of the stages: from the raw leachate container to the aeration basin, from the aeration basin to the peat column inlet, and from the distilled water container to the control peat column inlet. There were also two calibrated Masterflex® peristaltic pumps, maintaining constant flows from the raw leachate to the aeration basin, and from the distilled water container to the control peat column. Two sets of

Triplicate Peat Column Avg. 10.82 cm3/cm2/day

Distilled Water

Peristaltic Pump

Peristaltic Pump

Control Column Avg.10.82 cm3/cm2/day

Triplicate Peat Column Avg. 8.28 cm3/cm2/day

Raw Leachate

Aeration Basin

Air Pump

Diffuser

Attached Growth Media




pumps were engaged to feed the two sets of triplicate peat columns with aerated leachate from the aeration basin. Each pump set was assembled with three Masterflex® Easy-Load® pump heads attached to a Masterflex® peristaltic pump to ensure a constant flow rate. The pumps were attached to a GRASSLIN (model CP-924) timer which turned all pumps on intermittently five times a day, for a total of ten minutes per day. Samples of the raw leachate, aerated leachate and column effluents were collected and analyzed for COD, CBOD5, NH3-N, NO3-N, and TSS removal in order to assess the performance of the aeration basin with biofilm growth for contaminants removal at 5-day and 2-day HRTs, as well as the removal efficiencies and life expectancies of the peat biofilters. A blank column was operated with distilled water in the same manner as the higher HLR an average 10.82 cm3/cm2/day to observe the potential leaching of constituents from the peat and the behavior of the peat filter under control condition. The contaminant load in the Trail Road landfill leachate was much higher than is typically reported for untreated domestic wastewater especially in terms of the higher ammonia-N, TSS, COD, and CBOD5 concentration. The average influent COD, CBOD5, NH3-N, NO3-N, and TSS concentration were 899 mg/L, 340 mg/L, 511 mg/L, 2 mg/L, and 51 mg/L, respectively for the 5-day HRT. The average influent COD, CBOD5, NH3-N, NO3-N, and TSS concentration were 1052 mg/L, 534 mg/L, 392 mg/L, 2 mg/L, and 135 mg/L, respectively, for the 2-day HRT. Therefore, these high contaminant concentrations indicate that the leachate is a high-strength wastewater in comparison to municipal wastewater.


3.1 Aeration basin

The results of this study showed that the aeration basin did not significantly remove COD from the raw leachate for both the 5-day and 2-day HRTs, while CBOD5 concentrations in the aeration basin were observed to decrease from an average 340 mg/L and 534 mg/L to 98 mg/L and 139 mg/L for the 5-day and 2-day HRTs, respectively as shown in Figure 2(a). The TSS concentrations of aerated leachate were observed to decrease prior to days 70 and 78 for the 5-day and 2-day HRTs, respectively. Then the TSS concentration of aerated leachate was found to exceeded the raw leachate TSS concentration, which is likely due to the fact that sludge in the aeration basin was not collected and disposed of throughout the course of each experimental HRT (Fig. 2(b)). From Figure 3(a), it can be noted that steady-state removal of NH3-N was observed for the 5-day HRT after approximately 2 weeks of operation, while similar NH3-N removal was not observed for the 2-day HRT even after 3 weeks of operation. The 5-day HRT also exhibited better nitrification than the 2-day HRT. In addition, an average NO3-N generation of 108 mg/L was found for the 5-day HRT compared to 21 mg/L for the 2-day HRT (Fig. 3(b)). Denitrification was also noted in the aeration basin after 44 and 42 days of operation for the 5-day and 2-day HRTs,




0 20 40 60 80 100 120

0

100

200

300

400

500

600

Raw AB Control Column C-1: Avg. 8.28 cm3/cm2/day C-2: Avg. 8.28 cm3/cm2/day C-3: Avg. 8.28 cm3/cm2/day C-1: Avg. 10.82 cm3/cm2/day C-2: Avg. 10.82 cm3/cm2/day C-3: Avg. 10.82 cm3/cm2/day

2-day HRT:

CBO

D5 (m

g/L)

Day

0 20 40 60 80 100 120

0

100

200

300

400

500

6005-day HRT:

CBO

D5 (m

g/L)

Day

(a)

0 20 40 60 80 100 120

0

50

100

150

200

250

300


2-day HRT:

TSS

(mg/

L)

Day

0 20 40 60 80 100 120

0

50

100

150

200

250

300

350

400

5-day HRT:

TSS

(mg/

L)

Day

(b)

Figure 2: CBOD5 and TSS of raw leachate, aerated leachate and column effluents for the 5-day and 2-day HRTs.

w

ww

.witpress.com

, ISSN 1743-3541 (on-line)

©2006 W

IT PressW


ol 89,

Geo-Environm

ent and Landscape Evolution II 129

0 20 40 60 80 100 120

0

100

200

300

400

500

600

2-day HRT:

Amm

onia

-N (m

g/L)

Day

0 20 40 60 80 100 120

0

200

400

600

800

1000

1200


5-day HRT:

Amm

onia

-N (m

g/L)

Day

(a)

0 20 40 60 80 100 120

020406080

100120140160180200


2-day HRT:

Nitra

te-N

(mg/

L)

Day

0 20 40 60 80 100 120

0

50

100

150

200

250

300

350

400 5-day HRT:

Nitra

te-N

(mg/

L)

Day

(b)

Figure 3: NH3-N and NO3-N of raw leachate, aerated leachate and column effluents for the 5-day and 2-day HRTs.

w

ww

.witpress.com

, ISSN 1743-3541 (on-line)

©2006 W

IT PressW


ol 89,

130 Geo-Environm

ent and Landscape Evolution II

respectively. The aeration basin allowed for NO3-N removal as a result of the rapid formation of a biofilm onto the attached growth media in the aeration basin. As the microorganisms grow, the thickness of the biofilm layer increase, and the diffused oxygen is consumed before it can penetrate the full depth of the biofilm layer. Thus, an anaerobic environment is established near the surface of the media, which is likely the main mechanism for NO3-N removal in the aeration basin after an extended period of operation. The concentration of NO3-N was observed to decrease from 319 mg/L (day 44) and 96 mg/L (day 42) to 90 mg/L (end) and 1 mg/L (end) for the 5-day and 2-day HRTs, respectively.

3.2 Peat columns

In this study, two sets of triplicate columns were operated at HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively, for both the 5-day and 2-day HRTs. The averages of the triplicate column effluents are used in the discussion of the results. The average effluent concentrations from the peat columns were 356 mg/L and 383 mg/L at HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively, for the 5-day HRT; while average effluent COD concentrations of 413 mg/L and 415 mg/L were observed for the 2-day HRT at the same HLRs. From the control column effluent COD concentrations, it was found that the peat contributed COD to the effluents, with average COD concentrations of 39 mg/L for both the 5-day and 2-day HRTs. As a consequence, the overall COD removal was limited. Biodegradation of organic matter was observed in the peat columns through CBOD5 removal (Fig. 2(a)). Average effluent CBOD5 concentrations of 22 mg/L and 24 mg/L for 8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs, respectively, were noted for the 5-day HRT. Concentrations of 18 mg/L and 29 mg/L were obtained for the 2-day HRT, for the 8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs, respectively. Average effluent TSS concentration of 9 mg/L and 6mg/L in 5-day HRT, and 34 mg/L and 42 mg/L in 2-day HRT, were found for the HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively (Fig. 2(b)). The TSS removal was achieved through adsorption and physical filtration via its porous structure. Effluent NH3-N concentrations were less than 2.18 mg/L and 2.15 mg/L after 30 days of operation at 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively, for the 5-day HRT (Fig. 3(a)). Comparatively, NH3-N effluent concentrations were below 4.29 mg/L and 5.30 mg/L after 36 days of operation at 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively, for the 2-day HRT. The concentrations steadily increased at the end of the experimental run before clogging of the columns were observed. The likely mechanisms of NH3-N removal were adsorption of NH4

+ onto the peat up to the CEC saturation for NH4

+, followed by leaching of NH3-N, nitrification and denitrification. An average, NO3

--N concentration of 121 mg/L and 119 mg/L in 5-day HRT, and 38 mg/L and 48 mg/L in 2-day HRT were observed for the HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively (Fig. 3(b)). Denitrification in the peat columns was generally observed after 49 and 36 days of operation; this might be due to the formation of anoxic zones at the bottom of the columns.




One of the main objectives of this research was to investigate the lifetime of the peat biofilter system under different contaminant loadings, HRT in aeration basin, as well as hydraulic loading rate. The operational life of each of the peat filters was defined as the number of days of operation between when the peat columns were initially fed with leachate to the time clogging was observed as exhibited by surface ponding. The total cumulative COD, BOD5, and TSS removal of peat columns at the time of clogging for the two sets of triplicate columns were computed under the different operational conditions. The results are summarized in Figure 4 and Table 1. A single factor ANOVA was conducted with an alpha value of 0.05 to statistically compare the performances of peat columns operated under different conditions. The results of this study indicated that statistically similar total cumulative organic (COD, CBOD5) removals were observed in the peat columns under different HLRs and HRTs since the F values were always less then Fcritical values in the ANOVA test. However, the higher 5-day HRT of the aeration basin increased the operational life of the peat biofilters when compared to the 2-day HRT through the lowering of the contaminant loading onto peat biofilters.

0 20 40 60 80 100 120

4

5

6

7

8

9

10

11

12

Control Column Avg. 8.28 cm3/cm2/day: Col. Avg.

Avg. 10.82 cm3/cm2/day: Col. Avg.

Hyd

raul

ic L

oadi

ng R

ate

(ml/d

ay)

Day

0 20 40 60 80 100 120

3

4

5

6

7

8

9

10

11

12

2-day HRT:

5-day HRT:

Hyd

raul

ic L

oadi

ng R

ate

(ml/d

ay)

Day

Figure 4: Operational life and cumulative contaminant removal of peat columns.




Table 1: Operational life and cumulative contaminants removal of peat filters.

Cumulative Removal (mg/ g of Peat)

Phase Column ID

Total Operational

Life (day) COD BOD TSS

Controlled Column(DW) No Clogging ― ― ―

Column 1 104 34.68 6.42 10.92

Column 2 108 46.88 9.42 15.28 Avg. 8.28 cm3/cm2/day

Column 3 115 48.12 8.86 15.59

Column 1 108 41.31 7.54 14.96

Column 2 101 48.74 10.42 16.71

5-da

y H

RT

Avg. 10.82 cm3/cm2/day

Column 3 101 42.06 8.17 14.37

Controlled Column(DW) No Clogging ― ― ―

Column 1 82 30.04 7.65 2.91

Column 2 64 20.90 5.51 1.40 Avg. 8.28

cm3/cm2/day Column 3 93 37.79 9.57 4.23

Column 1 93 51.68 13.50 5.20

Column 2 64 31.10 5.80 1.32

2-da

y H

RT

Avg. 10.82 cm3/cm2/day

Column 3 82 46.77 10.60 3.26

4 Conclusions

One of the main objectives of this research was to investigate the total operational life of the peat biofilter under varied contaminant loading and hydraulic loading rates. The contaminant loadings to the peat columns were considered to be a function of the HRT in the aeration basin. The results of this research showed that the impact of the hydraulic loading rate was less significant than the effect of contaminant loading rate leading to a longer life of the peat filters. Statistically similar organic (COD, CBOD5) removal performances and life expectancies could be obtained at hydraulic loading rates of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day in both 5-day and 2-day HRTs. However, the higher HRT, 5 days, increased the life expectancy of the peat biofilter by approximately one month, due to the considerable decrease in the organic, NH3-N, and TSS loading through the aeration basin. The results also suggested that the contaminant removal efficiencies of the peat biofilter columns were similar for the 8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs. The results indicated that the peat columns were unstable during the first month of operation, since leaching of COD to effluents by peat itself and saturating of CEC for ammonia-N followed by leaching of ammonia-N was observed during the first month of operation. The aeration basin with support




media for biofilm growth was primarily effective for the removal of NH3-N and NO3-N through nitrification and denitrification. Steady-state nitrification was initially observed in the aeration basin after approximately 2 to 3 weeks of operation as this was likely the time required for the steady-state development of a biofilm on the attached growth media to which NH3-N removal was attributed. Therefore, an anaerobic environment was established near the surface of the media, which was mainly responsible for denitrification in aeration basin after approximately 1.5 months of operation at both the 5-day and 2-day HRTs. From this study, it can be noted that HRT was a limiting factor affecting the contaminants removal efficiencies of aeration basin. Therefore, an increase in HRT would increase the removal of contaminants.

References

[1] Heavey, M., Low-cost Treatment of Landfill Leachate Using Peat. Waste Management, 23, pp. 447-454, 2003.

[2] Kinsley, C., Crolla, A., & Fernandez, L., Treatment of Landfill Leachate using a Peat Filter: Final Report. City of Ottawa, Ottawa, ON, 2003.

[3] Kennedy, P. & Van Geel, P., Hydraulic of Peat Filters Treating Septic Tank Effluent. Transport in Porous Media, 41, pp. 47-60, 2000.

[4] Lyons, H.J. & Reidy, T.J., The Use of Peat in Treating Landfill Leachate. Humic Substances in Soils, Peats and Waters: Health and Environmental Aspects, eds. M.H.B. Haynes & W.S. Wilson, UK Royal Society of Chemistry, pp. 475-485, 1997.

[5] Talbot, P., Bélanger, G., Pelletier, M., Laliberté, G., & Arcand, Y., Development of a Biofilter Using an Organic Medium for On-site Wastewater Treatment. Water Science & Technol., 34, pp. 435-441, 1996.

[6] Viraraghavan, T. & Ayyaswami, A., Use of Peat in Water Pollution Control; A Review. Canadian J. of Civil Engineering, 14, pp. 230-233, 1987.

[7] Rock, C.A., Brooks, J.L., Bradeen, S.A., & Struchtemeyer, R. A., Use of Peat for On-Site Wastewater Treatment: I. Laboratory Evaluation. J. Environ. Qual., 13(4), pp. 518-523, 1984.

[8] Bergeron, M., Peat. Canadian Minerals Yearbook, pp. 37.1-37.8, 1987. [9] Buttler, A., Dinel, H., & Levesque, P.E.M., Effects of Physical, Chemical,

and Botanical Characteristics of Peat on Carbon Gas Fluxes. Soil Sci., 158(5), pp. 365-374, 1994.

[10] McLellan, J.K. & Rock, C.A., Pre treating Landfill Leachate with Peat to Remove Metals. Water, Air, & Soil Pollution. 37; pp. 203-215, 1988.

[11] Mitsch, W.J. & Gosselink, J.G., Wetlands, Van Nostrand Reinhold Co., New York, 1993.

[12] Sharma, D.C. & Forster, C.F., Removal of Hexavalent Chromium Using Sphagnum Moss Peat. Water Research, 27(7), pp. 1201-1208, 1993.

[13] Mckay, G., Use of Adsorbents for the Removal of Pollutants from Wastewater, CRC Press, 1996.

[14] Couillard, D., Review: The Use of Peat in Wastewater Treatment. Wat. Res., 28(6), pp. 1261-1274, 1994.




Are crop and range lands being contaminated with cadmium and lead in sediments transported by wind from an adjacent contaminated shallow lake? H. O. Rubio1, T. R. Saucedo1, M. R. Bautista2, K. Wood3, C. Holguin3 & J. Jiménez4 1Campo Experimental la Campana-Madera del Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias, México y Facultad de Zootecnia de la Universidad Autónoma de Chihuahua, México 2Universidad Juárez Autónoma de Tabasco 3New Mexico Water Resources Research Institute, Las Cruces, New Mexico, USA 4Facultad de Zootecnia de la Universidad Autónoma de Chihuahua, México

Abstract

Soil contamination by cadmium (Cd) and lead (Pb) in range and crop lands can occur from polluted sediments carried by the wind. Soils samples were obtained in three different areas on the leeward (east) side of a shallow lake called the Laguna de Bustillos: (1) sediment in the shallow lake (S), (2) soil in the range area adjacent to the shallow lake (RA), and (3) soil in agriculture land (AL) adjacent to the range area. In each area, three composite soil samples were randomly collected at depths of 0-15, 15-30 and 30-50 cm. Therefore, 45 soil samples were analyzed as a 3 (area) x 3 (depth) factorial arrangement. The samples were air dried, passed through a 2.0 mm sieve, ground and passed through a 0.355 mm sieve, and digested with concentrated nitric acid. The metals Cd and Pb were determined using ICP optical emission spectrometry. In addition pH and CE were evaluated. Cadmium concentration was statistically different among areas and showed a strong interaction between depth x area. Maximum Cd concentration was found in S with 0.94 mg kg-1 in the surface horizon (0-15 cm). Lead concentration was different among areas and the interaction was not as strong as the one noted for Cd. Highest Pb concentrations were observed in the surface horizons (0-15 cm) of both S and RA with 74.19 mg kg-1 and 55.09 mg kg-1. Nevertheless, for the AL system the maximum Pb concentration was found in the subsurface horizon (30-50 cm) with 40.23 mg kg-1. It is concluded that Cd and Pb are contaminating the RA and AL through wind movement with Pb contamination being the worst. Keywords: cadmium, lead, soil contamination, laguna Bustillos, Chihuahua, México.




doi:10.2495/GEO060141

1 Introduction

Cadmium (Cd) and lead (Pb) are considered to be at the top of the five most important elements in terms of food-chain contamination. These two metals, when present in highly contaminated soils may have a negative human reproductive outcome [1]. It is generally recognized that the presence of these metals in soils is usually thorough antrophogenic activities such as agriculture (fertilizers, liming materials, agrochemicals), industry (atmospheric deposition, wastes) and urban activities (sewage sludges, drainage deposition) whereas other elements in soils are often derived from the parent rock material which the pedo-genetic processes are carried out. Once in soil, the metals as contaminants, often attach to soil particles that can migrate from one place to another. Evaluation of heavy metals contamination is often based on comparison with the background concentration in a given environment [2], evaluating the effect on plant growth or soil utilization [3, 4] and determining the hazard to human health [5, 6]. It is generally accepted that soil sediment is a more complex medium than water. This is true because soils include water as well as solid and gas phases, and soils do not move from one place to another as fast as does water [7], and determining concentrations of heavy metals is a difficult task [8]. The water of the Laguna de Bustillos in Chihuahua, México is highly contaminated [9, 10] and it is suspected that its soil sediments are also contaminated. In the particular case of the Laguna de Bustillos environment, the sediment is exposed because the water of the shallow lake is absent during large periods of time. For example, during the year 2005 (January to July) most of the shallow lake was completely dry. In addition, strong wind events are present in the period from February to April every year. Therefore, the question that arises is whether the soil sediment from the shallow lake is contaminating the close rangeland and agriculture land. The objective was to evaluate the level of soil contamination with Cd and Pb metals in three different environments; lake sediments, rangeland, and agriculture land. To our knowledge, this is the first report on contaminant risk assessment of rangeland and cropland soils from a polluted soil in Mexico.


The study was conducted on the east side of the Laguna de Bustillos, located in the state of Chihuahua, México. Its location is in the polygon: latitude 28° 58´ 12”-28° 15´ 00” N; longitude 107° 09´ 36”-106° 15´ 00” W and at 2,300 meters elevation (Figure 1). Annual precipitation averages about 480 mm mostly as rain during the summer (July to September), but some snow events occur during the winter (December to February). The source is a shallow lake, heavily silted due to soil erosion coming from a denuded watershed which becomes dry during drought periods. Wind storms moving from the west to the east side usually occur every year from late winter (February) through the beginning of spring (April).




Figure 1: A panoramic view of the Laguna de Bustillos, Chihuahua, Mexico.

Soil samples where obtained from three different areas on the leeward (east) side of the Laguna de Bustillos. The first system was the sediment of the shallow lake (S) where the water used to be, but at the time of collecting the sample was dry. The second area was the rangeland (RL) adjacent to the shallow lake where the domestic livestock graze during the whole year. The third area was the crop land (AL) that is close to the RL. The domestic livestock belong to the commonly owned communities (Ejidos) of Cuitlahuac, La Selva, Centro Calles and Fabela. An “Ejido” is a community where the law established that any grazing land must be collectively shared [11] which results in the worldwide phenomena commonly called “The Tragedy of the Commons”; therefore, the RL is an extremely deteriorated grassland. The most important crops in the AL are beans (Phaseoluos vulgaris), corn (Zea mays) and oats (Avena sativa) growing in the precipitation season under dryland condition. Five points were randomly selected in the shallow lake (S). Then, at each point of S a transectal line to the east was selected with five randomly located points in RL as well as in AL. At each point, three composite soil samples were taken at 0-15, 15-30, and 30-50 cm profile depths. Hence, 15 soil samples were taken in each system, giving a total of 45 soil samples. The samples were air dried and passed through a 2.0 mm sieve. After this, the samples were ground and passed through a 0.355 mm sieve. They were evaluated for Cd and Pb concentration as well as pH and EC. The digestion of soil samples for Cd and Pb evaluation was realized with concentrated nitric acid in the laboratory of the Faculty of Zootechnic of the Autonomus University of Chihuahua, Mexico, following the sampling and analysis protocol of Canada (MAF). Cd and Pb




concentrations were determined using an ICP-OES (Inductively Coupled Plasma-Optical Emission Spectrometry) 2100, Perkin Elmer. Values of pH and EC were determined in a saturated paste using a standard glass electrode and a conductivity meter, respectively. The statistical analysis used a factorial treatment design 3 x 3; where Factor A was the system (S, RL, AL) and Factor B was the depth (0-15, 15-30, 30-50 cm). The differences were noted using a 0.01 significance level.


The Cd concentration was different among systems. Not surprisingly the higher Cd concentration was noted in S, with the high concentration noted in the upper horizon (0-15) where it was detected at 0.946 mg kg-1 in comparison with 0.672 mg kg-1 found at the 15-30 cm depth and 0.614 mg kg-1 noted in the 30-50 cm depth. In contrast, in RL and Al, the higher Cd concentration was not detected in the upper profile (0-15 cm); in fact, the lowest concentration in these two areas was found in the upper horizon (Figure 2). A strong interaction was noted in Cd concentrations (Figure 2). These results are similar to the levels found in Cd concentrations in a soil of West Bengal, India where it was detected at 0.37 mg kg-1 [12], soil samples form England and Wales had a Cd concentration of 0.8 mg kg-1 [13] and China soils had a Cd concentration of 0.07 mg kg-1 [14]. In other studies, Cd levels were 0.4-0.8 mg kg-1 in Spain [15], levels of 0.6-1.4 mg kg-1 in asphalt roads in Germany [16] and levels of 0.41 mg kg-1 in Japan [17].

Figure 2: Mean of Cadmium level under three systems and different soil depth.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 10 20 30 40 50 60

Agriculture

Sediment

Rangeland

X axis = Soil depth (cm) Y axis = Concentration (mg kg-1)




A study in the United States of America [18] found 3,045 surface soil samples contained Cd values from <0.01 to 2.0 mg kg-1 with a mean of 0.265 mg kg-1. In any case, the levels of Cd found in the present study do not represent a hazard to the soils in the three areas, because Dudka et al. [18] concluded that soils with Cd concentrations of up to 30 mg/kg are still safe for production of some crop plants. Moreover, the soil contamination due to wind-born Cd was insignificant, being our hypothesis not tested for this element. The Pb concentrations were different among areas and no statistical differences were observed for depth and for the interaction (Figure 3). Maximum levels of Pb concentration was noted in S with the higher concentration in the profile 15-30 cm depth with about 75.99 mg kg-1 while the lesser amount of Pb was observed in AL system in the upper profile with 33.49 mg kg-1 (Figure 3). A study evaluating surface soils in a greenhouse [15] found Pb concentrations in a range of 2.5 to 89.9 mg kg-1 which are similar to the results presented in this study; but the concentration range was shorter from 33.49 to 75.99 mg kg-1. Another study [19] reported a Pb concentration as high as 59 mg kg-1 in the 0-20 cm depth of the soil profile irrigated with wastewater. Moreover, in a study carried out in India [12] they reported a mean concentration of 10.4 mg kg-1 in a soil affected with arsenic. A study carried out in Japan [17] noted a 21 mg kg-1 of Pb in an uncultivated soil. The values reported here are of practical importance, because some of the soil fauna like the earthworm (Eisenia fetida Andrei) may be severely affected with Pb concentration higher than 30 mg kg-1.

Figure 3: Mean of lead concentration under three different systems and different soil depth.

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50 60

Agriculture

Sediment

Rangeland

X axis =Soil depth (cm) Y axis =Concentration (mg kg-1)




A study conducted on road dust in Sydney [20] suggested that deposition and removal of road dust is not a static process. We could speculate that our findings suggest that Cd and Pb had similar origin and this is a continuous process. This means that an equilibrium may be reached when the rate of deposition in a system is equal to the rate of removal.

4 Conclusion

This study showed that our hypothesis that sediments of the Laguna de Bustillos are contaminating the rangelands and crop lands with Cd and Pb is true. However, the Cd concentration does not represent a hazard at least in the area tested. On the other hand, the Pb concentration in the contaminated area may represent a potential hazard in the short term.

Acknowledgements

We are deeply grateful with the Produce Foundation of Chihuahua (Fundacion Produce Chihuahua, A.C.) for partial financial support as well as with the Faculty of Zootechnic of the Autonomous University of Chihuahua. The INIFAP (National Research Institute of Forestry, Agriculture and Rangeland) deserves special thanks for financial support and for facilities given to this research.

References

[1] Laudanski, T., Sipowicz, M., Modszelewski, P., Bolinski, J., Szamatowicz, J., Razniewska, G., and Akerlund, M. 1991. Influence of high lead and cadmium soil content on human reproductive outcome. Int. J. Gynaecol. Ostet. 36(4):309-315.

[2] Zhang, X.P., Deng, W., and Yang, X.M. 2002. The background concentrations of 13 soil trace elements and their relationship to parent material and vegetation in Xizang (Tibet), China. J. Asian Earth Sci. 21:167-174.

[3] Arao, T de N., Sugiyama, M., and Takahashi, M., 2003. Genotypic differences in cadmium uptake and distribution in soybeans. Plant Soil 251:247-253.

[4] Sheppard, S.C. 2005. Assessment of long-term fate of metals in soils: Inferences from analogues. A review. Canadian Jour. of Soil Sci. 85(1):1-18

[5] Storelli, M.M., Barone, G., and Marcotrigiano, G.O. 2005. Cadmium in cephalopod molluscs: implications for public health. Journal of Food Protection 68(3):577-588.

[6] Linde, A.R., Sanchez-Galan, S., and García-Vazquez, E. 2004. Heavy metal contamination of European Eel (Anguilla anguilla) and brown trout (Salmo trutta) caught in wild ecosystems in Spain. Journal of Food protection 67(10): 2332-2336.




[7] Peijnenburg, W.J.G.M. and Jager, T. 2003. Monitoring approaches to access bioaccessibility and bioavailability of metals: matrix issues. Ecotoxicol. Environ. Saf. 56:63-77.

[8] Chen, M., Ma, L.Q. and Harris, W.G. 1999. Baseline concentrations of 15 trace elements in Florida surface soils. J. Environ. Qual. 28:1173-1181.

[9] CNA, 2003. Comision Nacional del Agua. Estadísticas del agua en México. Reporte.

[10] Rubio, A.H., Saucedo, R.A., Lara, C.R., Wood, K., and Jimenez, J. 2005. Water quality in the Laguna de Bustillos of Chihuahua, México. Water Resources Management III. Editors M. De Conceicao Cunha and C.A. Brebbia. 155-160.

[11] Cornehls, J.V. 1964. Mexico´s rural road to progress: an analysis of agrarian reform and agricultural development. Univ. of Texas, Austin, USA.

[12] Roychowdhury, T., Uchino, T., Tokunaka, H., abd Ando, M. 2002. Arsenic and other heavy metals in soils from an arsenic-affected area of West Bengal, India. Chemosphere 49(6)605-618.

[13] McGrath, S.P. and Loveland, P.J. 1992. Heavy metals in soils. In the soil geochemical atlas of England. B.J. Alloway (ed.). Blackie Academic and Professional, Glasgow, Scotland.

[14] Wang, Y., and Wei, F.S. (eds.), 1995. Soil environmental element chemistry. Chinese Environmental Science Press, Peking, People´s Republic of China.

[15] Gil, C., Boluda, R., and Ramos, J. 2004. Determination and evaluation of cadmium, lead and nickel in greenhouse soils of Almeria (Spain). Chemosphere 55(7)1027-1034.

[16] Munch, D. 1992. Soil contamination beneath asphalt roads by polynuclear aromatic hydrocarbons zinc, lead and cadmium. Science of the Total Environments 126(1-2):49-60.

[17] Akira, T., Hirofumi, T., Masami, N., Yuichi, T., Toyokazu, U., Shun´ichi, H., Jiro, I. 2005. Effect of long-term fertilizer application on the concentration and solubility of major and trace elements in a cultivated andisol. Soil Sci. Plant Nutr. 51(2)251-260.

[18] Dudka, S., Piotrowska, M., and Terelak, H. 1996. Transfer of cadmium, lead, and zinc from industrially contaminated soil to crop plants: a field study. Empoissonnement Pollution 94(2)181-188

[19] Mapanda, F., Mangwayana, E.N., Nyamangara, J., and Giller, K.E. 2005. The effect of long-term irrigation using wastewater on heavy metal contents of soils under vegetables in Harare, Zimbabwe. Agriculture, Ecosystem&Environment 107(2-3)151-165

[20] Ball, J.E., Jenks, R., Aubourg, D. 1996. Dry weather build-up of constituents on road surfaces. In proceedings of the 7th International Conference on Urban Storm Drainage. Hanover, Germany 785-790.




The sorption characteristics of PAHs onto soils in the presence of synthetic and bio surfactant

J.-H. Chang1, Z. Qiang2 & C.-P. Huang3 1Department of Environmental Engineering and Management, Chaoyang University of Technology, Taiwan 2Environmental Research Center, Department of Civil, Architectural and Environmental Engineering, University of Missouri-Rolla, USA 3Department of Civil and Environmental Engineering, University of Delaware, USA

Abstract

Contaminants such as polycyclic aromatic hydrocarbons (PAHs) are of special concern due to their high toxicities and hydrophobicities. The high hydrophobic natures of PAHs enable their strong sorption onto soil or sediments, which makes it difficult for their removal from the environment. This study was to explore the sorption characteristics of PAHs with synthetic nonionic surfactant and biosurfactant in the soil. Two surfactants (one synthetic and one naturally occurring) Triton X-100 and rhamnolipid were used. Results show that the partition coefficient of PAHs in the Triton X-100 system decreases with increasing surfactant concentrations. Throughout the whole concentration range of studied biosurfactant (0 ~ 30 CMC), the partition coefficient of PAHs in rhamnolipid system is roughly constant, around 400.0 L/Kg. Keywords: sorption, soils, PAHs, surfactant, biosurfactant.

1 Introduction

The polycyclic aromatic hydrocarbons (PAHs) are of special concern for the natural environment due to their high toxicities and hydrophobicities. On the basis of the characteristic of high hydrophobicity, PAHs are strongly sorbed onto soil or sediments [1]. As such, it is difficult to remove PAHs from the soils. Among many remediation techniques, surfactants have been induced in soil-




doi:10.2495/GEO060151

washing, soil-flushing, and pump-and-treat technologies to increase the solubilization of sorbed hydrophobic contaminants [2]. The sorption-desorption process has been considered as the most critical mechanism with respect to the soil remediation. To study the sorption reaction of PAHs in the soil-water system in the presence of surfactants one is able to gain insight into the surfactant remediation process of contaminated soils. The purpose of this investigation is to explore partition coefficients of different PAHs in the presence of surfactants. In addition to the synthetic surfactant, the biosurfactant is also evaluated in this study.

1.1 Surfactant fundamentals

Surfactants can be categorized into four groups including cationic, anionic, nonionic and zwitterionic (both cationic and anionic groups). These classifications are typically based on the nature of their head group (hydrophilic head). In aqueous solution, surfactant molecules have the formation of aggregates known as micelles when the surfactant concentrations exceed a certain value. The certain concentration is called the critical micelle concentration, namely, CMC. The CMC value of surfactant is a function of surfactant type (nonionic type generally has lower CMC than ionic one), and system conditions (e.g., temperature, hardness) [3]. Compared to the cosolvent (e.g., alcohols), surfactants exhibit a much higher degree of surface activity and present the high potential for remediation techniques. In aqueous solution, micelles consist of two portions: one portion is the hydrophilic exterior (the hydrophilic heads are oriented to water phase) and the other is the hydrophobic interior (the hydrophobic tails are oriented to the center of micelles). Since hydrophobic organic compounds can be dissolved into the interior of micelle, the solubilities of organics are then enhanced in the solution. The increased “aqueous solubility” of organic compounds in micelles is referred to as solubilization. The degree of organic solubilization depends on the class and the concentration of surfactants.

1.2 Biosurfactant

Biosurfactants are produced by many different bacterial genera. There are a variety of biosurfactants manufactured and most of them can be classified as anionic or nonionic biosurfactants. In addition, biosurfactants can also be classified into several broad groups: glycolipids, lipopeptides, lipopolysaccharides, phospholipids, and fatty acids/neutral lipids [4]. The largest and best-studied group of the biosurfactant is glycolipids, which includes the sophorose-, rhamnose-, trehalose-, sucrose-, and fructose-lipids. In this study, the rhamnolipid biosurfactant has been chosen as the dissolution agent. The production and purification of biosurfactants are difficult, which are controlled by many factors including growth conditions, culturing medium nutrients, temperature, pH, and agitation method [5]. Molecular weight of biosurfactant ranges from approximately 500 to 1,500 mw and CMC value of biosurfactant typically ranges from 1 to 200 mg/L [6].




1.3 PAHs sorption-desorpiton phenomena

PAHs sorption onto the soil in the presence of surfactants can be distributed in three phases including water, micelles, and the soil. During the sorption-desorption process, there are mainly three reactions occurring such as: (1) the dissolution of PAHs in the micelle phase, (2) the sorption of surfactant in the soil phase, (3) the partition of PAHs between aqueous and soil phase [7]. As all of three reactions reach equilibrium, the partition coefficient is defined as the ratio between the concentration of the PAHs in the soil phase and in the liquid phase (water+micelle), respectively. Although the liquid phase also comprises the other sub-phases including the dissolved organic matter (DOM) and colloids, the effect of DOM and colloids in the liquid phase is relatively small and can be neglected compared to the micelle concentration. The measured partition coefficient is still an apparent rather than true coefficient between aqueous and soil phase.

2 Methodology

The soil samples collected from a specific waste site were air-dried and sieved through a No. 10 standard sieve (2 mm openings). Table 1 shows the physical-chemical properties of this soil sample with corresponding analytical methods adopted. The PAHs organic chemicals were purchased from Aldrich Co. The purity of fluorine, phenanthrene, fluoranthene, and pyrene is 98.0% and that of anthracene is 97.0%. Table 2 lists the major features of these PAHs organic chemicals including molecular formula, molecular weight, saturation water solubility, and logKow. The selected surfactants were Triton x-100 and biosurfactant (rhamnolipid).

Table 1: Basic physical-chemical characteristics of the soil.

Batch sorption experiments were conducted with fluorene, phenanthrene, anthracene, fluoranthene, and pyrene at various CMC values ranging from 1 to 5. The soil samples were contaminated by PAHs in advance, which concentrations in soils ranged from 25 µg/g to 200 µg/g. To a series of glass tubes, 2 g of PAHs contaminated soil samples and 10 ml of various concentration (0 CMC to 30 CMC) surfactant solutions were added. Two surfactants were used to conduct the sorption experiments including Triton x-100 and rhamnolipid. The tubes were placed in a shaker and shaken constantly for 72 hours to reach the

Physical-chemical Characteristics

Result Method

Sand (%) 14.0 Hydrometer Silt (%) 38.0 Hydrometer Clay (%) 48.0 Hydrometer pH 7.6 In 0.01M CaCl2 Organic Matter (%) 1.7 Heating at 105 oC for 2 hours, then

at 360 oC for 2 hours




equilibrium condition. The mixtures were then centrifuged at 2,500 rpm (1,000 g) for 10 minutes using a Precision Scientific Co. model K-9 centrifuge to separate the coarse particles from the supernatant. Then, a Sorvall superspeed refrigerated centrifuge model RC-5 was used to remove fine particles in the supernatant at the rotation speed of 12,000 g. The concentration of PAHs organic compounds in the centrate was analyzed with a HPLC/FLD (Hewlett-Packard, model 1100 series), respectively. Table 2: The physical properties of fluorene, phenanthrene, fluoranthene,

pyrene, and anthracene.


Figure 1 shows the sorption isotherms at various Triton X-100 CMC values for fluorene, phenanthrene, and anthracene, respectively. Results of fluoranthene and pyrene present the similar trend (not shown here). All of these isotherms can be fitted by linear equations. It implies that the partition process is the main factor for the PAHs distribution in the micelle-soil-water system. The driving forces of partition process are believed as the hydrophobicity of PAHs, surfactant, and soils. Because the partition coefficient is used to describe the competition of PAHs in the liquid and soil phases, the magnitude of various PAHs partition depends not only on their water solubilities but also on the sorption ability of the soils and surfactants. As such, no significant relationship presents between partition coefficient and solubility in all figures. Figure 1 also shows that partition coefficients of PAHs decreases with increasing the surfactant concentrations. This observation is in agreement with the results reported by Sun and Boyd [8]. It is noticed that the difference between isotherms of blank and 5 CMC is insignificant for these five PAHs compounds. It can be attributed to the sorption of surfactant on soil surface; therefore, the solubilization capacity of PAHs is not significant. In addition, results present that large amount of surfactants were sorbed in soil phase, the sorbed surfactants may increase the sorption capacity of the soil. Figure 2 shows sorption isotherms at various rhamnolipid CMC values for fluorene, phenanthrene, and anthracene, respectively. Results of fluoranthene and pyrene with a similar trend are not shown. In the figure, the differences among various surfactant concentrations are insignificant.

Physical Property

fluorene phenanthrene anthracene fluoranthene pyrene

Molecular Formula

C13H10 C14H10 C14H10 C16H10 C16H10

Molecular Weight

166.2 178.2 178.2 202.3 202.3

Solubility (mg/L)

1.84 1.09 0.06 0.23 0.13

logKow 4.18 4.57 4.54 5.22 5.13




Figure 1: The PAHs isotherms with various CMC values in Triton X-100 solution-soil system such as fluorine, phenanthrene, anthracene.

0

50

100

150

200

0 5 10 15 20 25

blank

5 cmc10 cmc

15 cmc20 cmc

30 cmc

Fluo

rene

in S

oil (µ

g/g)

, Qe

Fluorene in Solution (mg/L), Ce

0

50

100

150

200

0 1 2 3 4 5 6 7 8

blank

5 cmc

10 cmc

15 cmc

20 cmc

30 cmc

Phen

anth

rene

in S

oil (µ

g/g)

, Qe

Phenanthrene in Solution (mg/L), Ce

0

50

100

150

200

250

0 2 4 6 8 10

5 cmc10 cmc15 cmc20 cmc30 cmc

Ant

hrac

ene

in S

oil (µ

g/g)

, Qe

Anthracene in Solution (mg/L), Ce




Figure 2: The PAHs isotherms with various CMC values in rhamnolipid solution-soil system such as fluorine, phenanthrene, anthracene.

0

20

40

60

80

100

120

140

160

2 3 4 5 6 7 8 9 10

5 cmc10 cmc15 cmc20 cmc30 cmc

Fluo

rene

in S

oil (µ

g/g)

, Qe

Fluorene in Solution (mg/L), Ce

0

50

100

150

200

0.1 0.2 0.3 0.4 0 .5 0.6 0.7

5 cm c

10 cm c

15 cm c

20 cm c

30 cm c

Phen

anth

rene

in S

oil (µ

g/g)

, Qe

Phenanthrene in Solu tion (m g/L), Ce

0

50

100

150

200

0 0.2 0.4 0.6 0.8 1

10 cmc15 cmc20 cmc

30 cmc

Ant

hrac

ene

in S

oil (µ

g/g)

, Qe

Anthracene in Solution (mg/L), Ce




Most equilibrium concentrations of PAHs in these data are below 1 mg/L, which are far lower than their solubilities (10 ~ 20 mg/L) under the same rhamnolipid concentration in water. This implies that most biosurfactant molecules are sorbed in soils and just few in the solution to aid dissolving PAHs compounds from soil to aqueous phase. However, these isotherms still can roughly be fitted by linear equations. Throughout whole concentration range studied of biosurfactant concentration (0 ~ 30 CMC), the partition coefficient of PAHs in rhamnolipid system is constant, around 400.0 L/Kg.

4 Conclusion

The partition coefficient of PAHs between soil and micelle aqueous phase can be described by the linear equations regardless of synthetic or bio surfactants. In the Triton X-100 system, the partition coefficient of PAHs decreases with the surfactant concentrations. In addition, Triton X-100 is sorbed in soil; the sorbed surfactants may influence the sorption capacity of the soil. Throughout whole biosurfactant concentration (0 ~ 30 CMC), the partition coefficient of PAHs in rhamnolipid system can be regarded as a constant, around 400.0 L/Kg. The biosurfactant also exhibits strong attraction onto the soil.

References

[1] Karickhoff, S.W., Organic Pollutant Sorption in Aquatic System. Journal of Hydraulic Engineering, 10(6), pp. 707-735, 1984.

[2] Mackay, D.M. & Cherry, J.A., Groundwater Contamination: Pump-and-Treat Remediation. Environmental Science and Technology, 23(6), pp. 630-636, 1989.

[3] Rosen, M.J., (eds). Surfactants and Interfacial Phenomena, John Wiley & Sons Inc.: New York, 1989.

[4] Fiechter A., Biosurfactants: Moving towards Industrial Application. Trends Biotech, 10, pp. 208-217, 1992.

[5] Miller R.M., Biosurfactant-Facilitated Remediation of Metal-Contaminated Soils. Environmental Health Perspective, 103(Suppl 1), pp. 59-62, 1995.

[6] Lang, S. & Wagner, F., (eds). Structure and Properties of Biosurfactants. In: Biosurfactants and Biotechnology, New York, 1994.

[7] Shaoo D., Smith, J.A., Imbrigiotta, T.E., & Mclellan, H.M., Surfactant-Enhanced Remediation of a Trichloroethene -Contaminated Aquifer. 2. Transport of TCE. Environmental Science and Technology, 32(11), pp. 1686-1693, 1998.

[8] Sun, S. & Boyd, S.A., Sorption of Nonionic Organic Compounds in Soil-Water Systems Containing Petroleum Sulfonate-Oil Surfactants. Environmental Science and Technology, 27(9), pp. 1340-1346, 1993.




Carbon dioxide sequestration in coal: implications for CO2 disposal and CH4 displacement from coal seams

M. Mirzaeian & P. J. Hall Department of Chemical and Process Engineering, University of Strathclyde, Glasgow, UK

Abstract The sequestration of CO2 within unmineable coal seams is one of the most attractive options for reducing atmospheric CO2 levels. Thus there is currently considerable interest in the interactions of coal with CO2 for its long-term disposal. This paper reports the analysis of coal / CO2 interactions at pressures of up to 30 bar. The results obtained from differential scanning calorimetry (DSC) show that the interactions of CO2 with coal leads to strongly bound carbon dioxide on coal. It was also found that the temperature of the second order phase transition of coal decreases with increase in CO2 pressure significantly, indicating that high pressure CO2 diffuses through coal matrix, causes significant plasticization effects, and changes the macromolecular structure of the coal. Desorption characteristics of CO2 from coal were studied by temperature programmed desorption mass spectrometry (TPD-MS). It was found that CO2 binds more strongly to coal and demands more energy to desorb from coal at higher pressures. Keywords: CO2 sequestration, coal, irreversible adsorption, high pressure interactions, macromolecular structure.

1 Introduction

Increased atmospheric CO2 concentrations due to fossil fuel combustion cause entrapment of solar radiation in the atmosphere and induce a gradual warming of the Earth’s surface (Greenhouse effect). This problem is now recognized as one of the most important environmental issues facing society. Therefore there is




doi:10.2495/GEO060161

currently considerable interest in the permanent disposal of CO2 and one option is to sequester it into uneconomic coalfields.

Coal is a chemically heterogeneous solid containing mainly carbonaceous material with very lower amount of mineral mater. It is a microporous macromolecular material containing a wide range of highly reactive chemical functional groups [1]. Its porosity results in the entrance of fluids into its structure and its polymeric nature accompanied with the presence of various functional groups, leads to the chemical interactions of fluids with its matrix.

It is well known that coal swells if it is in contact with solvents, such as pyridine, which break hydrogen bonds [2]. There is some evidence that coal can swell in high pressure CO2 [3], presumably due to the quadrupolar nature of the molecule disrupting weak electrostatic bonds within the coal structure. Dilatometric studies on coals in contact with CO2 showed significant increase in sample size [4]. It was suggested that CO2 swells coal because the solubility parameter of CO2 is close to the solubility parameter of coal. In the most significant study of coal behaviour in high pressure CO2 atmospheres Reucroft and Sethuraman [3] have shown that coals swell after exposure to CO2 and the amount of swelling increases with increasing pressure. It is supposed that the increased swelling effect with increased pressure might be due to the solubility parameter of CO2 approaching a value closer to that of the coals.

In the present study we have applied differential scanning calorimetry (DSC) for the first time to the investigation of coal/ CO2 interactions and the effect of high pressure CO2 on the coal structure. Temperature programmed desorption mass spectrometry (TPD-MS) has been also applied to the study of the desorption characteristics of CO2 from coal at pressures of up to 30 bar.

2 Experimental

Wyodak coal obtained from the Argonne sample bank was used in this study.

2.1 The DSC procedure

The DSC measurements were performed with a Mettler DSC 30 instrument in conjunction with Mettler software TA72PS for data acquisition and processing. Standard aluminium pans were used, with two pinholes in order to minimise mass transfer limitations in evaporation of water or contact of gas with sample during DSC scans. Nitrogen flowing at 10 ml/min was used as a carrier gas to keep the cell free of oxygen during measurements. Typically 10 mg of sample was used in an experiment. The DSC measurements were performed at a heating rate 10ºC/min. Cooling of the furnace between consecutive heating scans was carried out using a liquid nitrogen cooling accessory directly beneath the furnace.

2.1.1 Adsorption of N2 and CO2 on Wyodak coal

To study the irreversible adsorption of CO2 on the coal two different series of scans were conducted on the Wyodak coal sample.




1- A sample of fresh coal was purged with N2 flowing at 10 ml/min for 10 min in DSC chamber then heated to 110ºC at 10ºC/min and held for 30 min, cooled at nominal rate of 100ºC/min to -60ºC. The sample was then heated from -60 to 200°C at 10ºC/min three times in succession with a cooling rate of 10ºC/min between heating runs. In all scans the sample always remained in the DSC chamber under nitrogen flowing at 10 ml/min.

2- A sample of fresh coal was purged with N2 flowing at 10 ml/min for 10 min in DSC chamber. then heated to 110ºC at 10ºC/min and held for 30 min, cooled at nominal rate of 100ºC/min to -60ºC under N2 flowing at 10ml/min. At this point gas was switched to CO2 and the sample was then heated from -60 to 200ºC under CO2 flowing at 10 ml/min at 10ºC/min three times in succession with a cooling rate of 10ºC/min between heating runs.

2.1.2 Effect of high pressure CO2 on the coal structure Differential scanning calorimetry was also employed to study the effect of high pressure CO2 on the macromolecular structure of Wyodak coal:

1- A sample of Wyodak coal was purged with N2 flowing at 10 ml/min for 10 min in DSC chamber. Then heated from 30ºC to 110ºC at a heating rate of 10ºC/min, held at 110ºC for 30 min and cooled to 30ºC in N2 atmosphere at a flow rate of 10 ml/min. This was performed to remove water from coal sample before determining its glass transition temperature. The dried sample was heated from 30ºC to 200ºC at 10°C/min in N2 atmosphere flowing at 10 ml/min three times in succession. This experiment was carried out to determine the phase transitions of Wyodak coal in N2 atmosphere.

2- A sample of Wyodak coal was purged with 20 bar Ar three times in succession, to flush adsorbed gases present in the pores in coal and also oxygen form the high pressure cell. Then heated to 110ºC, held for 30 min at 110ºC, and cooled to 30ºC at 10°C/min under Ar atmosphere in the high-pressure cell. The sample was loaded with CO2 to the desired pressure at room temperature. It was exposed to this CO2 environment for 24 h. After that, the CO2 pressure was rapidly released, the sample was purged with 20 bar Ar and transferred to the DSC chamber. The sample was purged with N2 with a flow rate of 10 ml/min for 10 min in the DSC chamber and DSC was carried out from 30ºC to 200ºC at 10°C/min in N2 atmosphere three times in succession. The purpose of this sequence was to determine the effect of high-pressure CO2 atmospheres on the phase transition of the coal sample.

2.2 The TPD-MS procedure

Approximately 150 mg of sample was placed in a sample tube and loaded with CO2 to the desired pressure at room temperature in a high-pressure cell. The sample was exposed to this high-pressure CO2 atmosphere for a certain period of time. Then CO2 pressure was rapidly released and the sample was transferred to the desorption chamber and purged with high purity helium for 10 min at 298 K, before commencing the TPD run. The gas flow rate was 100 ml/min. To perform a TPD –MS scan, the sample was heated by linearly increasing the temperature,




20 K/min, and the evolution of CO2 from the sample was monitored by a Hiden Analytical HAL/HPR20 Quadrupole Mass Spectrometer (QMS).


3.1 Irreversible adsorption of CO2 on Wyodak coal

The results of the calorimetric measurements of the adsorption process for nitrogen and carbon dioxide on Wyodak coal are shown in figures1-2.

Figure 1: DSC scans for adsorption of N2 on Wyodak coal sample.

Figure 2: DSC scans for adsorption of CO2 on Wyodak coal sample.

-4

-2

0

2

4

6

8

-4

-2

0

2

4

6

8

-50 0 50 100 150 200

Heat Flow (W/g) Heat Flow (W/g)

Temperature ( oC)

first scan

second and third scans

exo

endo

-4

-2

0

2

4

6

8

-4

-2

0

2

4

6

8

-50 0 50 100 150 200

Heat flow (W/g) Heat flow (W/g)

exo

endo

first scan

second and third scans

Temperature ( oC)




DSC results for the adsorption of N2 on coal sample in figure 1 show that coal/N2 interactions are very weak and the adsorption of N2 on coal occurs physically and reversibly.

Exothermic peaks for the adsorption of CO2 on coal in figure 2 are associated with the uptake of CO2. This is an activated process and presumably at the temperature of the exotherms there is enough thermal energy to overcome the activation energy for diffusion. A comparison between figure1 and figure 2 shows that exotherms evident at low temperatures are absent when the experiments were conducted under N2. This suggests that interactions between coal and CO2 are much stronger than those between coal and N2.

The integrated values for the exotherms associated with the adsorption of CO2 on Wyodak coal are given in table 1. These values are indicative of the amount of CO2 sorbed during the experiment. The reduction in the value of the exotherm between the first and second runs suggests that some CO2 is irreversibly bound to the structure even after heating to 200ºC.

Table 1: Differential enthalpies of the adsorption of CO2 on Wyodak coal.

First scan

Second scan

Third scan

Irreversible Sorption capacity

∆ H (J/g) 27.40 20.60 18.72 6.8

3.2 Structural change in coal caused by high pressure CO2

Figure 3 shows the DSC thermograms for dried Wyodak coal from 30ºC to 200°C. The first scan show an irreversible process, which might be attributed to the structural rearrangement and relaxation in the coal when heated above its glass transition temperature. Subsequent two scans after the first scan show a reversible second order process. This process has the characteristics of glass transition [5]. Before the transition coal is a glassy solid with severely restricted macromolecular motions and diffusion of gases and liquids in its structure is slow. When heated to a certain temperature which is called glass transition temperature, a significant increase in coal’s macromolecular motions happens. Above the transition coal becomes rubbery and diffusion into its structure becomes much faster.

Figure 4 shows the DSC thermograms for dried Wyodak coal held under 30 bar CO2 atmosphere for 24 hours prior to DSC measurements. In this case, the first scan illustrates two endothermic effects. The first one might be related to the evaporation of moisture adsorbed by sample during transferring from the high pressure cell to the DSC chamber and also continuous release of sorbed CO2 since desorption process is endothermic. The second effect might be attributed to the fast release of sorbed CO2 from the coal sample at the vicinity of its glass transition temperature. As the coal in the DSC chamber is heated it will continuously release CO2. At the vicinity of glass transition temperature the




desorption rate may suddenly be accelerated since the chain mobility of the coal suddenly increases. This process is irreversible and has disappeared on the second and third scans.

Figure 3: DSC for dried Wyodak coal from 30ºC to 200ºC in N2 atmosphere.

Figure 4: DSC for dried Wyodak coal (held under 30 bar CO2 atmosphere for 24 hours) from 30ºC to 200ºC in N2 atmosphere.

Figure 5 shows the change in the glass transition temperature of the coal with CO2 pressure. Depression in glass transition temperature of coal at CO2 atmosphere might be due to the solubility of CO2 into the coal matrix and plasticization of coal by CO2.

-2

-1.5

-1

-0.5

0

0.5

1

1.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

40 60 80 100 120 140 160 180 200

Temperature ( oC )


exo

endo

third scan

second scan

first scan

-5

-4

-3

-2

-1

0

1

2

3

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

40 60 80 100 120 140 160 180 200

Temperature ( oC )

exo

endo


first scan

third scan

second scanDelta H = 13.2 J/g




Figure 5: The dependence of glass transition temperature of Wyodak on the pressure of CO2.

Characteristics of the glass transition of Wyodak coal at various pressures of CO2 are given in table 2. The further decrease in the glass transition temperature with CO2 pressure might be due to the further solubility of gas into the coal which is partly due to the solubility parameter of CO2 approaching to a value closer to that of the coal at higher pressures.

It is supposed that the acidic and basic properties of CO2 which allow it to form hydrogen bonds or other acid-base bonds play a major role in its solubility in coal [3]. It dissolves in the organic coal matrix, thus modifying the physical and possibly the chemical structure of the coal matrix by disrupting hydrogen bonds in the coal structure during its interaction with coal [6].

Table 2: Glass transition temperature of Wyodak coal at different pressures of CO2.

Second scan Third scan P CO2 (bar) Tg (°C) Tg (°C)

0 121.1 121.6 5 86.6 81.7 10 83.2 78.2 20 82.4 77.8 30 81.1 75.7

3.3 Desorption characteristics of CO2 from Wyodak coal

Figure 6 shows the TPD-MS spectrum of CO2 desorption from Wyodak coal. The coal sample was held at 5 bar CO2 atmosphere for 24 hours prior to the

-3

-2

-1

0

1

2

3

40 60 80 100 120 140 160 180 200


-3

-2

-1

0

1

2

3

Temperature ( oC )

30 bar

5 bar

10 bar

20 bar

exo

endo




TPD-MS measurements. The spectrum has two regions of interest. The first one at low temperature region (in the range of 300-350 K) and second one after the peak temperature where desorption rates decrease with increasing temperature.

0

5 10-11

1 10-10

1.5 10-10

2 10-10

300 350 400 450

Des

orpt

ion

rate

(arb

.un.

)

Temperature (K)

Redhead Equ.

Exp. Data

0

5 10 -11

1 10 -10

1.5 10-10

2 10 -10

300 310 320 330 340 350

Des

orpt

ion

rate

(arb

.un.

)

Temperature (K)

E xp. Da ta

Redhead Equ. Res idual

Figure 6: TPD-MS spectrum of CO2 adsorbed on Wyodak coal at 5 bar.

To analyze the desorption data, the low temperature part of the TPD-MS spectrum has been modelled using the assumption of a first order desorption process with a single activation energy for desorption (Redhead Equation) [7].

111exp11ln2

−

−−

+

−=

P

des

PP

des

P TTRE

TT

TTRE

NN

NP is the maximum desorption rate at peak temperature, TP, and N is the

desorption rate at any temperature. The embedded graph in figure 6 shows that the first order model provides a

good fit to the experimental data in the range of 300-350 K. The value of the activation energy for desorption, Edes, is calculated from the fitting of the low temperature part of the desorption spectrum to the model. The value of Edes for desorption of CO2 from Wyodak coal is estimated as Edes ≈ 15 kcal/mole. This value is much smaller than the ~ 30 kcal/mol value for activation energy for desorption of pyridine from Illinois #6 coal measured by Hall and Larsen [7]. Pyridine is an excellent hydrogen bound acceptor and a strong organic Lewis base [8], it interacts with coal through strong noncovalent interactions and hydrogen bonding with coal hydroxyls [9]. A comparison between these values shows that the estimated value for activation energy of CO2 desorption from Wyodak coal is sensible.

The trail of the spectrum after the peak shows that desorption rate does not follow the first order desorption model. The process of CO2 desorption from coal is a mass transfer process involving several elementary processes. These processes may include diffusion through macropores, diffusion through micropores and diffusion through coal matrix. In small micropores, diffusional




resistance limits mass transfer process. The mass exchange can also be limited due to the large energy barrier for desorption. At higher surface coverages, the diffusivity is high and the rate of desorption reaches a maximum. As temperature increases and the desorption process proceeds, high energy sites in the microporous structure are released. Due to the large energy barriers the rare of desorption form these sites is very slow [10]. Therefore the deviation of desorption spectrum from the fist order kinetic model might be due to the contributions of activated diffusion effects, micropore diffusional resistance, barrier resistance of high energetic sites and experimental error in the instrument.

TPD-MS spectra of CO2 adsorbed on Wyodak coal at various pressures are compared in figure 7. The total area under a TPD-MS spectrum is proportional to the amount of adsorbed CO2

[11]. The values of Edes and the integrated areas under the spectra of CO2 for various pressures are shown in Table 3. It can be noted that the desorption intensities increases with pressure indicating the amount of CO2 sorbed in the coal is greater at higher pressures. Consequently at higher pressures the relatively larger amount of CO2 would be desorbed.

0

2 10-11

4 10-11

6 10-11

8 10-11

1 10-10

300 35 0 40 0 45 0

No CO25 bar CO215 b ar CO 220 b ar CO 2

Tem perature (K)

Des

orpt

ion

rate

(arb

.un.

)

Figure 7: TPD-MS spectra of CO2 adsorbed on Wyodak at various pressures.

Table 3: Desorption parameters for CO2 adsorbed on Wyodak coal at various pressures for 24 hours.

P (bar) E des (cal/mol) Integrated area 5 14521 1.5744 × 10 -9

15 15210 3.341 × 10 -9 20 15399 4.673 × 10 -9

4 Conclusions

It has been found that CO2 binds to the structure of Wyodak coal strongly and irreversibly even after heating to 200°C. Glass transition of coal decreases with




CO2 pressure significantly, so there is a strong suggestion that high pressure CO2 diffuses through coal matrix, causes significant plasticization effects and changes the macromolecular structure of coal. Desorption characteristics of coal loaded with high pressure CO2 show that the amount of CO2 uptake increases with increasing CO2 pressure. Increase in activation energy for CO2 desorption from coal with pre-adsorbed CO2 pressure suggests that high pressure interactions will demand more energy to desorb from coal probably due to the further access of CO2 to the coal microporous structure.

The results of this study prove that Wyodak coal have a great affinity and high irreversible sorption capacity for CO2 and could be excellent for CO2 sequestration.

References

[1] Green, T. L.; Kovac, J.; Brenner, D.; Larsen, J. W. In Coal Structure; Meyers, R. A. , Ed., Academic Press, New York, pp. 199-282, 1982.

[2] Larsen, J. W.; Mohammadi, M. Energy & Fuels, 4, pp. 100-106, 1990. [3] Reucroft, P. J; Sethuraman , A. R. Energy and Fuels, 1, pp.72-75, 1987. [4] Reucroft, P. J.; Patel, H., Fuel, 65, pp. 816-820, 1986. [5] Mackinnon, A. J.; Hall, P. J. Energy & Fuels, 9, pp. 25-32, 1995. [6] Glass, A. S.; Larson, J. W. Energy & Fuels, 8, pp. 629-636, 1994. [7] Hall, P.J.; Larsen, J.W. Energy & Fuels, 7, pp. 47-51, 1993. [8] Arnett, E. M.; Joris, E.; Murty, T. S. S. R.; Gorrie, T. M.; Schleyer, P. v.

R. J. Am. Chem. Soc. 92, pp. 2365-2377, 1970. [9] Larsen, J. W.; Baskar, A. J. Energy & Fuels, 1, pp. 230-232, 1987. [10] Reid, C. R.; Thomas, K. M. Langmuir, 15, pp. 3206-3218, 1999. [11] Habenschaden, E.; Kuppers, J. Surf. Sci, 138, pp. L147, 1984.




Section 4 Climatological processes

Response of dissolved organic carbon in a shallow groundwater ecosystem to a simulated global warming experiment

K. P. Wilson & D. D. Williams Department of Life Sciences, University of Toronto at Scarborough, Canada

Abstract

Dissolved organic carbon (DOC) in marine and freshwater ecosystems represents an immense reservoir of organic matter with varied and significant ecological value. Global warming poses a significant threat in that it has the capacity to alter the concentration and distribution of DOC. Since groundwater constitutes approximately two-thirds of the available freshwater on earth, it is crucial to determine how global warming may affect its DOC balance. However, in higher latitudes carbon cycling is poorly understood, and ecosystem-scale studies are urgently required. We conducted an in situ temperature manipulation of a shallow groundwater system in Ontario, Canada that simulated temperature increases predicted by general circulation models for this region. Specifically, treatment block temperatures in spring, summer, and fall were elevated 3.9 ± 0.6 SD °C, whereas winter temperatures were elevated 5.0 ± 0.6 °C compared with a control block. We found no significant difference in DOC between control and treatment blocks during the pre-manipulation study period. However, there was a significant increase in DOC with groundwater depth in both blocks: 4.54 ± 0.25 mg/l at –20 cm to 5.79 ± 0.24 mg/l at –100 cm. During this period there was also a difference in DOC among seasons: fall and winter concentrations were lower than spring and summer. During the manipulation period there was also no difference in DOC between the control and treatment blocks, however, a positive trend in the treatment block was observed for all collections. Also, seasonal and depth differences between blocks were still apparent. Although during the manipulation period nitrate and total phosphorus showed no difference between control and treatment blocks, ammonia showed a significant decrease in the treatment block. We discuss the implications of these findings to the biogeochemistry and ecology of shallow aquifers. Keywords: global warming, dissolved organic carbon, DOC, shallow groundwater, Canada, temperature manipulation, ecosystem experiment.




doi:10.2495/GEO060171

1 Introduction

The importance of dissolved organic carbon (DOC) as a potential source of carbon and energy for subsurface and surface metabolism has been widely studied. Fisher and Likens [1] first examined how subsurface waters can effectively transport DOC from terrestrial ecosystems to stream ecosystems. Hynes [2] showed that subsurface DOC contributes significantly to stream ecosystem metabolism. These influences of DOC are due to its ability to affect a variety of physical, chemical, and biological processes. DOC entering shallow groundwater can be from terrestrial environments via soils (Chappelle [3]), through streams via the hyporheic zone (Williams and Hynes [4]), or laterally from riparian zones (Meyer and Tate [5], Storey et al. [6]). Once DOC enters shallow groundwater ecosystems, it can be oxidized to CO2 (Chappelle [3]), or if consisting of labile substances it can be rapidly utilized by microorganisms (Wetzel [7], Storey et al. [6]). DOC interacts with dissolved nutrients influencing nutrient concentrations and can also act as a buffer by affecting pH (Pace and Cole [8]). There is growing concern of the fate of DOC in higher latitudes because these temperate regions are expected to respond to global warming more so than lower latitudes (Schindler [9]).

General circulation models (GCM) predict increases in temperature to vary between 1.5 and 5.0 °C by the year 2040 for southern Ontario, Canada (Hengeveld [10]). These models also project a differentially higher temperature increase in winter than in summer in latitudes >30° (Hengeveld [10]). Temperature is one of the most important factors that affect life history characteristics and distributions of bacteria (Chapelle [3]) and aquatic insects (Sweeney [11], Vannote and Sweeney [12]). General circulation models have also shown that increased CO2 will change precipitation and temperature patterns, and therefore likely impact the abundance and distribution of species (McCarty [13]). Changes in the community structures of shallow groundwater ecosystems may affect DOC concentrations through changes in metabolic activity and hence aspects of the global carbon-cycle.

Global warming studies examining DOC concentrations have been conducted mainly in terrestrial ecosystems, such as soils (MacDonald et al. [14]), bogs and fens (Pastor et al. [15]), and arctic tundra and sedge ecosystems (Neff and Hooper [16]). Some aquatic ecosystem studies have also examined the response of DOC to global warming and climate change. For example, Schindler et al. [17] examined the physicochemical properties of boreal lakes, with emphasis on DOC and Clair et al. [18] modelled DOC loss from a small temperate wetland under a doubling of CO2. To our knowledge there has not been any direct whole-ecosystem manipulation of a shallow groundwater ecosystem that has examined DOC concentrations both spatially and temporally. To address the possible effects of global warming on DOC concentrations in a shallow groundwater ecosystem, we conducted an in situ temperature manipulation of a shallow groundwater ecosystem in accordance with global warming projections for southern Ontario, Canada. Specifically, we examined DOC concentrations: 1) spatially along a natural vertical gradient from surface water to –100 cm




sediment depths; 2) temporally by comparing seasons; and 3) by perturbing the groundwater temperature over a period of 13 months. We predict two possible responses of DOC concentrations to our ecosystem manipulation: 1) DOC concentrations will not change between control and treatment blocks (ambient DOC concentrations maintained by soil C storage); or 2) DOC concentrations will decrease in the treatment block compared with the control block (consumption of DOC through increased metabolism).

2 Methods and materials

2.1 Study site

This study was conducted on a small first order spring-brook (Valley Spring) located in southern Ontario, Canada (43°45’ N, 79º15’ W). Valley Spring is located at an elevation of 152 m and is approximately 60 m in length and 0.5-1.5 m wide with a discharge ranging between 1800-2300 l/hr. The surrounding vegetation consists of coniferous and deciduous trees and mixed grasses and shrubs. For further descriptions see Williams and Hogg [19] and Hogg and Williams [20].

2.2 Temperature manipulation

The temperature manipulation study area was located approximately 5 m from the spring source in an 8 m x 4 m grid. The study area was separated into a control and experimental treatment block, each 4 m x 4 m. Heating pipes were installed on 8 October 2003. Eight pipes (110 cm long, 4 cm outer diameter) pinched at the bottom (last 5 cm), were placed at 0.5 m intervals perpendicular to the stream channel along four transects at 1 m intervals longitudinally with the stream channel (n = 32 for each of the control and treatment blocks, fig.1). A treatment divide (5 m in length and 1 m deep) was installed along a natural groundwater divide in October 2003 using sixteen gauge galvanized steel sheet metal.

To increase the temperature in the treatment block, a 61 m long EASYHEAT® roof and gutter de-icing cable (Model ADKS, 120 V, 1000 W) was installed along each transect (n = 4). Starting at the end of each cable, 6 m of cable was wrapped length wise and tied together with cable-ties and placed into each pipe with the lead end running to the next pipe (in series). In order to control seasonal temperatures, each cable was attached to a variable transformer Powerestat® model # 3PN117C, 120 V, 12 A. Heat was applied on 5 March 2004 and continued until 5 May 2005. From general circulation models our targeted temperature differences between the control and treatment blocks for summer, spring, and fall range from 3.5-4.5 °C and in winter from 5-6 °C.

2.3 Collecting regime

Bimonthly samples of DOC concentrations, nutrients, and water chemistry were collected in June, August, October, December 2002, and February, April, June,




August, October, and December 2003 and 2004, and February, and April, 2005 (n = 18 collection dates). Collections between June 2002 and February 2004 (n = 11) were used to generate pre-manipulation data to determine if there were differences between the control and treatment blocks. Collections between April 2004 and April 2005 (n = 7) were used as manipulation response samples to examine possible temperature effects on DOC concentrations.

Figure 1: Schematic of experimental design showing the control block (no heat)

and treatment block (heating). Bold circles (o) represent heating pipes, regular circles (o) represent pipes with no heating. Dashed line is location of groundwater divide. Large circles with an (X) are the location of nested piezometers. Arrows are direction of surface water flow.

2.4 DOC sampling protocol

For each sampling date, duplicate samples of DOC concentrations were collected from five depths, –20, –40, –60, –90, and –100 cm and from the surface. Water samples were collected from nested mini-piezometers (Freeze and Cherry [21]) located within the control and treatment blocks. DOC determinations were performed on a Tekmar Dohrmann™ Phoenix 8000 UV-persulfate oxidation TOC analyzer.

2.5 Statistical analyses

All data were analyzed using JMP-start statistics software, (SAS Institute Inc.). A Shapiro-Wilks W-test of normality was used to test for homogeneity of variances (Zar [22]). DOC and temperature data were analyzed using two-way analysis of variance (ANOVA). The response variables (DOC and temperature) were compared with the predictor variables of depth, season, control and treatment blocks, and pre-manipulation and manipulation collections. To determine differences among depths and seasons Tukey-Kramer HSD comparison tests were conducted. Winter seasons included December and

o o o o o o o o

o o o o o o o o

o o o o o o o o

o o o o o o o o

o o o o o o o o

o o o o o o o o

o o o o o o o o

o o o o o o o o

TreatmentControl

Groundwater Divide




February collections, spring consisted of April collections, summer included June and August collections, with fall including October collections. Stepwise regression analyses were conducted on the response variable (DOC) to 9 predictor variables (total phosphorus O-PO4

3-, mg/l), nitrate (NO3—N, mg/l),

ammonia (NH3-N, mg/l), sulfide (S2-, mg/l), dissolved oxygen (DO, mg/l), conductivity (µS), pH, total dissolved solids (TDS), and temperature (°C).

3.1 Temperature data: pre-manipulation

Temperature showed no statistical difference for all depths between control and treatment sites during the pre-manipulation. Fall and winter temperatures were most uniform among depths compared with spring and summer. Fall and winter temperatures increased with depth with fall values ranging from 12.6 ± 0.3 SD to 13.2 ± 0.3 °C and winter values ranging from 7.4 ± 0.3 to 8.0 ± 0.3 °C at –20 and –100 cm, respectively. In contrast, spring and summer temperatures decreased with depth. Spring temperatures ranged from 8.2 ± 0.3 at –20 cm to 6.3 ± 0.3 °C at –100 cm and summer temperatures at –20 cm ranged from 14.3 ± 0.3 to 12.9 ± 0.3 °C at –100 cm.

3.2 Temperature data: manipulation

Temperature showed statistical differences between control and treatment blocks during the experimental manipulation. Temperatures between control and treatment blocks were significantly different for each depth in each season (p < 0.0001, n = 468, F = 2314.5). In winter, mean temperature for the control at –20 cm was 7.9 ± 0.1 °C, compared with 11.9 ± 0.1 °C for the treatment, an average difference of 4 °C, fig. 2. In winter, temperature differences between control and treatment plots were slightly higher with depth. At –40 cm the control was 9.5 ± 0.1 °C and the treatment was 14.6 ± 0.1 °C, a difference of 5.1 °C. Differences between control and treatment at –60, –80 cm, and –100 cm were 5.5, 5.5, and 4.9 °C, respectively, fig 2.

Spring, summer, and fall temperature differences between control and treatment blocks also varied with depth. Differences at –20 cm for spring, summer, and fall were 3.2, 3.3, and 3.3 °C, respectively. At –40 cm, differences were 4.1, 4.1, and 4.0 for spring, summer, and fall, respectively. At –60, –80, and –100 cm, differences were also close among seasons; –60 cm differences were 4.3, 4.3, and 4.0 °C, at –80 all were 4.2, and at –100 cm differences were 3.7, 3.5, and 3.7 °C. In fall, highest temperatures for the treatment block were 18.6 and 18.7 °C at –80 and –100 cm, respectively.

3.3 DOC concentration patterns: pre-manipulation

There was no significant difference in DOC concentrations between control and treatment blocks for surface and subsurface pre-manipulation collections.




3 Results

However, seasonal patterns were detected for most collection sites. Pre-manipulation concentrations of DOC at the surface were significantly higher in spring (4.17 ± 0.63 SD mg/l) compared with summer (1.21± 0.27 mg/l), fall (1.35 ± 0.40 mg/l), and winter (1.51 ± 0.30 mg/l; p = 0.002, n = 3, F = 6.43; fig. 3A).

Figure 2: Temperature (°C) at –20 and –100 cm sediment depths, for control and treatment blocks during the heat-manipulation phase. Heat was turned off on 5 May 2005.

Furthermore, seasonal patterns during the pre-manipulation collections were not consistent with depth. At –20 cm, winter DOC was significantly lower (3.48 mg/l ± 0.31 SE) than spring (5.28 ± 0.68), summer (4.75 ± 0.30 mg/l), and fall (4.60 ± 0.37 mg/l) concentrations (p = 0.016, n = 3, F = 3.91; fig. 3A). At –40 cm, spring and summer DOC concentrations were significantly higher than both fall and winter (p < 0.0001, n = 3, F = 10.60). Spring and summer DOC were 5.18 ± 0.58 mg/l and 5.24 ± 0.29 mg/l, respectively, compared with 3.77 ± 0.37 mg/l for fall and 3.06 ± 0.30 mg/l for winter. In contrast, seasonal trends were not significant for the –60 and –80 cm depths. Mean DOC concentrations at –60 cm ranged from a low of 4.22 ± 0.57 mg/l in winter to a high of 6.46 ± 0.91 mg/l in spring, and at –80 cm, DOC concentrations ranged from a low of 4.36 ± 0.34 mg/l in winter to a high of 5.52 ± 0.30 mg/l in spring. However, at –100 cm seasonal trends were similar to –20 and –40 cm. Spring, summer, and fall DOC concentrations were all higher and significantly different from winter concentrations (p = 0.0003, n = 3, F = 7.73).

02468

101214161820

Tem

pera

ture

( C

)

-20 cm Treatment

Control

02468

101214161820

-100 cm

Control

Treatment

0

Spring 2004 Summer 2004 Fall 2004 Winter 2005

Heat turnedoff

Heat turnedoff




Figure 3: for (A) pre-manipulation and (B) manipulation collections.

3.4 DOC concentration patterns: manipulation

Increased experimental temperatures did not have a significant statistical effect overall between control and treatment blocks. However, manipulation collections showed similar patterns to pre-manipulation collections. Similar to pre-manipulation collections, surface DOC concentrations were highest in the spring (3.56 ± 0.08 mg/l) and significantly different from winter (2.95 ± 0.08 mg/l), summer (1.03 ± 0.10 mg/l), and fall (1.00 ± 0.11 mg/l; p < 0.0001, n = 3, F = 213.84; fig. 3B). However, different patterns were detected at several subsurface depths. At –20 cm, spring showed the highest DOC levels (6.09 ± 0.28 mg/l) which were significantly different from winter (4.25 ± 0.28 mg/l), summer (4.18 ± 0.28 mg/l), and fall (3.05 ± 0.40 mg/l; p < 0.0001, n = 3, F = 15.27). In contrast with pre-manipulation results, manipulation patterns from –40 to –100 cm showed statistical differences. At –40 cm, DOC concentrations for spring, summer, and winter seasons were highest and significantly different from the fall (p = 0.27, n = 3, F = 3.64). Seasonal patterns during the manipulation were also detected at –60 and –80 cm unlike pre-manipulation collections. At –60 cm, fall was significantly different from the other seasons (p = 0.010, n = 3, F = 4.69; Table 1). In contrast, at –80 and –100 cm, summer DOC concentrations were significantly higher (6.21 ± 0.35 mg/l, and 6.61 ± 0.35 mg/l, respectively) than fall (3.22 ± 0.5 mg/l, 4.41 ± 0.51 mg/l,

0

2

4

6

8D

isso

lved

Org

anic

Car

bon

(mg/

l)Pre-manipulation

0

2

4

6

8

Winter Spring Summer Fall

Season

Manipulation

-20 cm

-100 cm

-40 cm

Surface

-60 cm-80 cm

(A)

(B)




(DOC, mg/l) Seasonal concentrations of dissolved organic carbon

respectively; p = 0.005, n = 3, F = 8.44), whereas winter and spring concentrations were similar at these depths.

Although, there were no statistically significant differences between control and treatment blocks detected during the manipulation, a positive trend was detected. At –20, –40, and –60 cm, DOC concentrations were slightly higher in the treatment block compared with the control block. Differences between control and treatment block concentrations were 4.3 to 4.7 mg/l at –20 cm, 3.7 to 4.2 mg/l at –40 cm, and 4.5 to 5.1 mg/l at –60 cm. Positive increases in DOC concentrations were 0.4 mg/l, 0.5 mg/l, and 0.6 mg/l for –20, -40, and –60 cm, respectively, for all manipulation collections combined. However, this trend was not detected at –80 and –100 cm.

3.5 Stepwise regression: pre-manipulation and manipulation

The response variable DOC was analyzed with several predictor variables (nutrients and water chemistry, n = 9) to examine which predictor(s) best explain DOC concentrations. Examining pre-manipulation data for all collections and depths combined, the highest r2 was 0.36 for all predictors (n = 9). After examining all possible models, NH3-N mg/l, total-phosphorus (TP, mg/l), and temperature (°C) were the best predictors, r2 = 0.33. The same analysis using manipulation data (all collections and depths) had an r2= 0.31 for all predictors. However, the best predictors for the manipulation period were NO3 mg/l, NH4 mg/l, and total dissolved solids (TDS), with an r2 = 0.27, n = 3.

Interestingly, predictor variables explain more variation when examining individual depths. For example, at –100 cm for the manipulation period, all predictors produce r2 = 0.73, n = 9. With the best model, including NO3-N mg/l, NH3-N mg/l, pH, and TDS, r2 = 0.71, n = 4.

4 Discussion

Surface concentrations of DOC were temporally variable during the pre-manipulation and manipulation study periods. Highest seasonal DOC concentrations at the surface occurred in spring compared with other seasons, fig 3A, 3B. A similar pattern was detected in an alpine catchment in Colorado, U.S.A. Boyer et al. [23] determined that during spring snowmelt, stream DOC concentrations increased and were highest (approximately 4.5 mg/l) during this time compared with the rest of the year. The primary source of increasing DOC concentrations in spring was attributed to subsurface flow of water through soils in the Deer Creek catchment (Boyer et al. [23]). In contrast, Bernal et al. [24], determined that surface DOC concentrations were highest in a transition period, dry to wet (September to November, our fall season), and lowest in a wet and dormant period (December to February, our winter) and a vegetated period (March to May, our spring). In the Bernal et al. study [24], seasonal differences may be attributed to the study having been conducted in an intermittent Mediterranean stream with autochthonous and allochthonous carbon sources being available at different times of the year.




Shallow groundwater (–20 to –100 cm) concentrations of DOC below the Valley Spring streambed were more seasonally variable than surface concentrations. Seasonal pre-manipulation and manipulation concentrations of DOC in the subsurface were always higher (3-7 mg/l) than the surface (< 4.2 mg/l), fig. 3. Highest subsurface DOC concentrations were also during spring and summer, fig 3A, 3B. Rutherford and Hynes [25] sampled DOC from three agriculturally impacted Ontario streams from the surface to a depth of –140 cm. Variations in DOC concentrations were highly variable with depth and over time. Concentrations were typically higher at the surface and at –20 cm, but some stations had higher concentrations at the deepest depths (–140 cm) than intermediate depths (–40 to –60 cm; Rutherford and Hynes [25]). Kaplan and Newbold [26] determined from the literature that concentrations of DOC in shallow groundwater can either decrease or increase with depth. These patterns are suggested to be due to extensive abiotic and biotic processing of terrestrial DOC sources in the vadose zone (zone of water limited above by the land surface and below by the water table) creating low phreatic zone DOC levels, whereas high phreatic zone DOC implies the opposite, low processing rates (Kaplan and Newbold [26]).

Temperature differences between the control and treatment blocks during the temperature manipulation were successfully maintained. The spring through fall difference, combining all depths, was 3.9 ± 0.5 SD °C, and combining all depths in winter the difference was, 5.0 ± 0.6 °C, fig. 1. Although targeted temperatures were reached, no significant differences in DOC concentrations were found between control and treatment blocks. Thus, our first of two predictions that no change in DOC concentrations would occur was supported. There could be a couple of reasons for this. Firstly, DOC concentrations in aquatic environments, especially in hyporheic and shallow groundwater ecosystems can be very heterogeneous. For example, Pabich et al. [27] examined DOC concentrations below the water table of a shallow estuary on Cape Cod, U.S.A. DOC concentrations of shallow groundwater were spatially variable, but temporally stable suggesting that local heterogeneity plays an important role in DOC delivery to shallow groundwaters (Pabich et al. [27]). Rutherford and Hynes [25] also suggest that the heterogeneity of DOC concentrations may be due to the complex flow patterns of hyporheic zones, and the typically mixed nature of bed sediments. Our data also suggest that subsurface DOC concentrations are heterogeneous due to the spatial and temporal variability in our system, thus possibly masking temperature-induced differences, fig. 3B. Secondly, the molecular structures that make up DOC are varied and highly complex. DOC is usually categorized into two main groups, non-humic or labile substances and humic substances (Wetzel [7]). Our study site is located in a temperate forest in southern Ontario, Canada. Such headwater streams are typically characterized by plant matter that accumulates after leaf-drop in the fall. Humic substances form most (70-80%) of the organic matter in soils and water, are structures with high-molecular-weight (HMW) and are typically the result of microbial activity on plant material (Wetzel [7]). Humic substances released from leaves may decrease turnover rates of DOC and provide a continuous supply of DOC to the shallow




groundwater ecosystem of Valley Spring. Further, since humic substances are recalcitrant to biological activity and tend to have low turnover rates, the effects of increased metabolic activity on DOC concentrations under the increased thermal regime induced in Valley Spring may not have been detectable over the short time period of 13 mo. In contrast, labile DOC typically consists of carbohydrates, proteins, amino acids, and other low molecular weight (LMW) substances. These LMW substances are readily used by microorganisms, creating conditions for rapid flux in aquatic ecosystems.

Manipulation concentrations of DOC were correlated most with NO3-N mg/l, NH3-N mg/l, and total dissolved solids (TDS), r2 = 0.27, p < 0.0001, n = 3. DOC was positively correlated with ammonia, both concentrations increased with depth, and DOC was negatively correlated with nitrate; DOC concentrations increased with depth while nitrate concentrations decreased. This shows a tight coupling between DOC and nitrate and ammonia. For example, if there are small changes in DOC concentrations there is the possibility on non-linear effects on watershed N retention (Goodale et al. [28]). These latter authors examined the spatial patterns of nitrate and DOC concentrations of 100 northeastern U.S. streams, and showed that as DOC concentrations increase, nitrate concentrations decrease.

Our study did not see a change in DOC concentrations under a simulated global warming experiment, although a trend of slightly higher DOC concentrations was detected in the treatment block. We advocate that further large-scale ecosystem manipulations should be conducted to more fully understand the role and transformations of DOC in streams and shallow groundwaters. In particular, these studies should examine the molecular fractions or species of DOC, in order to detect changes in allochthonous and autochthonous sources (e.g. Sachse et al. [29], Sobczak and Findlay [30]). For example, it has been shown that allochthonous DOC is the most important determinant of thermocline depth in small boreal lakes (Perez-Fuentetaja et al. [31]). Further, if the type of DOC entering such aquatic systems is not known (allochthonous verse autochthonous) correct management policy decisions may be very difficult to make.

References

[1] Fisher S.G. & Likens G.E., Energy Flow in Bear Brook, New Hampshire: an integrative approach to stream metabolism. Ecological Monographs 43(4), pp. 421-439, 1973.

[2] Hynes, H.B.N., The stream and its valley. Internationale Vereinigung fur Theoretische und Angewandt Limnologie, 9(1), pp. 1-15, 1983.

[3] Chapelle, F.H., Ground-water microbiology and geochemistry, John Wiley & Sons, Inc.: New York, pp. 263, 2001.

[4] Williams D. D. & Hynes H.B.N., The occurrence of benthos deep in the substratum of a stream. Freshwater Biology, 4(1), pp. 233-256, 1974.




[5] Meyer J.L. & Tate C.M., The effects of watershed disturbance on dissolved organic carbon dynamics of a stream. Ecology, 64(1), pp. 33-44, 1983.

[6] Storey, R.G., Howard, K.W.F. & Williams, D.D., Factors controlling riffle-scale hyporheic exchange flows and their seasonal changes in a gaining stream: A three-dimensional groundwater flow model. Water Resources Research, 39(2), pp. 2003.

[7] Wetzel, R.G., Limnology: Lake and River Ecosystems, Academic Press: San Diego, pp. 735, 2001.

[8] Pace, M.L. & Cole, J.J., Synchronous variation of dissolved organic carbon and color in lakes. Limnology and Oceanography, 47(2), pp. 333-342, 2002.

[9] Schindler, D.W., The cumulative effects of climate warming and other human stresses on Canadian freshwaters in the new millennium. Canadian Journal of Fisheries and Aquatic Sciences, 58(1), pp. 18-29, 2001.

[10] Hengeveld, H.G., Projections for Canada’s climate future. Environment Canada Document CCD 00-01, pp. 1-27, 2000.

[11] Sweeney, B.W., Bioenergetic and developmental response of a mayfly to thermal variation. Limnology and Oceanography, 23(2), pp. 461-477, 1978.

[12] Vannote R.L & Sweeney B.W., Geographic analysis of thermal equilibria: a conceptual model fro evaluating the effects of natural and modified thermal regimes on aquatic insect communities. The American Naturalist, 115(5), pp. 667-695, 1980.

[13] McCarty J.P., Ecological consequences of recent climate change. Conservation Biology, 15(2), pp. 320-331, 2001.

[14] MacDonald N.W., Randlett, D.L. & Zak, D.R., Soil warming and carbon loss from a lake states spodosol. Soil Science Society of America Journal, 63(1), pp. 218-221, 1999.

[15] Pastor, J., Solin, J., Bridgham, S.D., Updergraff, K., Harth, C., Weishampel, P. & Dewey, B., Global warming and the export of dissolved organic carbon from boreal peatlands. Oikos, 100(2), pp. 380-386, 2003.

[16] Neff, J.C. & Hooper, D.U., Vegetation and climate controls on potential CO2, DOC and DON production in northern latitude soils. Global Change Biology, 8(9), pp. 872-884, 2002.

[17] Schindler, D.W., Bayley, S.E., Parker, B.R., Beaty, K.G., Cruikshank, D.R., Fee, E.J., Schindler, E.U. & Stainton, M.P., The effects of climatic warming on the properties of boreal lakes and streams at the Experimental Lakes Area, northwestern Ontario. Limnology and Oceanography, 41(5), pp. 1004-1017, 1996.

[18] Clair, T.A., Arp, P., Moore, T.R., Dalva, M. & Meng, F.R., Gaseous carbon dioxide and methane, as well as dissolved organic carbon losses from a small temperate wetland under a changing climate. Environmental Pollution, 116(1), pp. S143-S148, 2002.




[19] Williams D.D. & Hogg I.D., The ecology and production of invertebrates in a Canadian coldwater spring. Holarctic Ecology, 11(1), pp. 41-54, 1988.

[20] Hogg, I.D. & Williams, D.D., Response of stream invertebrates to a global-warming thermal regime: An ecosystem-level manipulation. Ecology, 77(2), pp. 395-407, 1996.

[21] Freeze, R.A. & Cherry J.A., Groundwater, Prentice-Hall, Inc.: New Jersey, pp. 23-24, 1979.

[22] Zar J.H., Biostatistical Analysis, Prentice Hall: New Jersey, pp. 88-89, 1999.

[23] Boyer, E.W., Hornberger, G.M., Bencala, K.E. & McKnight, D.M., Response characteristics of DOC flushing in an alpine catchment. Hydrological Processes, 11(12), pp. 1635-1647, 1997.

[24] Bernal, S., Butturini, A. & Sabater, F., Seasonal variations of dissolved nitrogen and DOC: DON ratios in an intermittent Mediterranean stream. Biogeochemistry, 75(2), pp. 351-372, 2005.

[25] Rutherford, J.E. & Hynes, H.B.N., Dissolved organic carbon in streams and groundwater. Hydrobiologia, 154(1), pp. 33-48, 1987.

[26] Kaplan, L.A. & Newbold, J.D., Surface and subsurface dissolved organic carbon (Chapter 10). Streams and Ground Waters, ed. J.B. Jones & Mulholland P.J., Academic Press: San Diego, pp. 237-253, 2000.

[27] Pabich, W.J., Valeila, I. & Hemond, H.F., Relationship between DOC concentration and vadose zone thickness and depth below water table in groundwater of Cape Cod, U.S.A. Biogeochemistry, 55(3), pp. 247-268, 2001.

[28] Goodale, C.L., Aber, J.D., Vitousek, P.M & McDowell, W.H., Long-term decreases in stream nitrate: Successional causes unlikely; Possible Links to DOC? Ecosystems, 8(3), pp. 334-337, 2005.

[29] Sachse, A., Henrion, R., Gelbrecht, J. & Steinberg C.E.W., Classification of dissolved organic carbon (DOC) in river systems: Influence of catchment characteristics and autochthonous processes. Organic Geochemistry, 36(6), pp. 923-935, 2005.

[30] Sobczak, W.V. & Findlay, S., Variation in bioavailability of dissolved organic carbon among stream hyporheic flowpaths. Ecology, 83(11), pp. 3194-3209, 2002.

[31] Perez-Fuentetaja A., McQueen D.J. & Ramcharan C.W., Predator-induced bottom-up effects in oligotrophic systems. Hydrobiologia, 317(2), 163-176, 1999.




Regional analysis of climate and bioclimate change in South Italy

A. Capra, P. Porto & B. Scicolone Department of Agro-forestry and Environmental Sciences and Technologies, Mediterranean University of Reggio Calabria, Italy

Abstract

In this paper, monthly values of rainfall (P) and temperature (T) recorded in Southern Italy (Calabria and Sicily) during the period 1921-2000 are investigated. In particular, a series of 211 raingauge and 53 temperature stations are analysed for evidence of trend by using the linear regression and the Kendall non-parametric test. The tests are applied at a seasonal and annual scale; a spatial analysis is also carried out at both a regional and sub-regional scale in order to check the effect of different sub-areas on these trends. An additional investigation useful for checking the climate change effects on vegetation is also included analysing bioclimatic parameters such as evapotranspiration and aridity index. The results obtained confirmed, for the two investigated regions, the importance of the climatic analysis carried out at a regional scale. In fact, the tests showed for both the P and T records a strong difference between the two investigated regions. In particular, the total annual P showed a decreasing and increasing trend with -344 and 197 mm/100 years respectively in the Tyrrhenian and Ionian sub-regions of Calabria and a decreasing trend with -179 mm/100 years in Sicily. A decreasing trend is evident in the Tyrrhenian sub-region for the mean annual values of Tmax (-3.2 °C/100 years) and Tmin (-2.9 °C /100 years); a different trend is shown in the Ionian sub-region where the mean annual values of T showed an increase of 2.2 and 0.9 °C /100 years for Tmax and Tmin respectively. In contrast, a clear increasing trend (1.8 and 2.2 °C /100 years) is shown for both Tmax and Tmin in Sicily. A detailed analysis involving the calculation of 10-year moving averages, showed a significant change of trending after the 1950s for P and after 1970s for T. Keywords: climate change, precipitation, temperatures, bioclimatic parameters.




doi:10.2495/GEO060181

1 Introduction

It has been observed that during the last 100-150 years the Italian climate has changed and it has resulted in a rise of temperature and aridity. According to Brunetti et al. [1], during the period 1865-2000, the mean annual temperature (Tya) indicates a 0.4°C/100 years rise within the northern areas (N) of the country (continental zone) and a 0.7°C/100 years rise in the central (C) and southern (S) Italy (peninsular zones). More particularly, at seasonal scale, the slopes of the regression line are greater during the winter season, ranging from 0.7°C/100 years (N) to 0.9°C/100 years (S), while for the summer season they are lower and in some cases not significant. A negative trend of the annual rainfall (Py) is evident within both N and S areas with a slope equal to -47 mm/100 years and -104 mm/100 years respectively which mean, for the investigated period, 7% of the mean rainfall for the northern area and 18% for the South. This decreasing trend occurred particularly after 1950 since then the number of wet days (NP) has also decreased. The maximum annual temperature (Tymax) series show a slope of the regression line ranging between 0.4°C/100 years for N and 0.6°C/100 years for S while the minimum annual temperature (Tymin) series show slope values between 0.3°C/100 years for N and 0.5°C/100 years for S. According to Brunetti et al. [1], the most important contribution to this positive trend is due to roughly the last 20 years (before 1996) for N and to roughly the last 50 years for S. In respect to extreme events, it has resulted an increase of rainfall intensity within both N and S areas [1], [2] and a tendency toward an increase in drought [3], [4], [5], [6]. Other investigations carried out in Italy [4], [7] showed a different behaviour of the same climatic series if a subdivision of the N and S areas into smaller sub-regions is taken into account. For this reason, it seems necessary to investigate the climate change on a regional scale where different geographical factors (e.g. distance from the sea, elevation, aspect) are likely to influence the magnitude of these trends. Such investigation can be useful not only in terms of improving scientific knowledge but also to provide the Italian government with useful information in order to make the right choices in planning future activities. Calabria and Sicily (Fig. 1) are two regions located in South Italy and are particularly prone to be investigated at local scale because of their geographic characteristics and local orographic features. In particular, Calabria (Ca) is a narrow (with a width ranging from 30 to 95 km) and long peninsula extending from North to South for about 250 km; a mountain range (Apennine) runs in latitude and divides the region into two opposite areas: the Tyrrhenian and the Ionian zones. This mountain range is located almost perpendicularly to the direction of the dominant moisture-bearing winds and for this reason it causes a very strong variability in terms of rainfall and temperature patterns considering also the different altitudes and aspects [8]; that is why it is easy to find flat and semi-arid coastal areas (Tya and Pya equal to 17.4°C and 683 mm respectively) as well as mountain zones with Tya equal to 9.1°C and Py equal to 1242 mm (the




highest in the South). Locally, the annual rainfall Py ranges from 1107 mm over the Tyrrhenian area (CaT) to 945 mm over the Ionian zone (CaI).

Figure 1: Location of investigated regions, sub-regions and weather stations. The two Calabrian sub-regions named Tyrrhenian and Ionian are marked with CaT and CaI respectively; SiN, SiS, and SiE represent North, South, and East zones of Sicily.

Sicily (Si) is the largest island of the Mediterranean Sea. According to traditional geographical distinctions for making sub-regional analysis, Sicily was divided into three homogeneous sub-regions (North, East, and South in Fig. 1). Each sub-region is separated from the others by mountain ranges and for this reason the rainfall pattern is different from area to area. Locally, the annual rainfall Py is 591 mm in the South area (SiS), 735 mm over the East area (SiE) and 717 within the North sub-region (SiN). Because of its particular geographic location in the centre of the Mediterranean Sea, Sicily was often investigated as a key region in order to explain the climate evolution within the Mediterranean basin [8].




To date, the studies carried out in this area are limited on a few stations (5 for temperatures and 11 for rainfalls – see Brunetti et al., [1], [2]) and for the reasons explained above it could be of interest to extend this investigation on a greater number of stations in order to give useful information on climate change in Mediterranean areas. The study proposed here aims at analysing temperatures and precipitations over the two regions (Calabria and Sicily) for the period 1921-2000, using weather stations with high spatial resolution. The analysis also includes the use of bioclimatic indexes because in agro-forestry environments the consequences of climatic change depend on the interaction between temperatures (maximum, Tmax, minimum, Tmin, and mean, Ta) and rainfalls (Pm), that can be summarised by appropriate indexes which account for aridity and plant water demand.

2 Data and methods

This paper investigates monthly values of temperature (Tmax, Tmin, and Ta) and rainfall (Pm) collected by the Italian Hydrographic Service (IHS) during the periods extending respectively from 1926 to 2000 and 1921 to 2000. The IHS database concerning the two investigated regions includes about 120 temperature and 600 raingauge stations. Even if a considerable number of stations was considered, many of these sites were neglected because long periods of malfunctioning. The analysis was firstly carried out on a dataset including 53 temperature and raingauge stations (25 of which in Ca and 28 in Si). These sites were selected considering the dataset continuity as well as geographic and altitude representativeness; datasets with more than 20% of lacks (with some exceptions) were also neglected. An additional database including only rainfall series (monthly, seasonal and annual) and composed of 158 Calabrian stations was also used in the analysis in order to get more details about the spatial variability over the investigated regions. For this database an additional analysis involving the number of wet days, Np, was carried out for the period 1951-2000. The analysed weather stations are located as in figure 1. Although no specific tests were applied, we suppose that the series used here are homogeneous because they were collected using the same criterion that did not change during the analysed period; also, specific tests involving Sicilian data only, did not show particular anomalies due to malfunctioning [6]. The monthly values of Tmax, Tmin, Ta and Pm were also used to calculate the following indexes: a) the monthly reference evapotranspiration, ET0m, estimated by Hargreaves formula [9] that, for sites where no direct radiation measurements are available, assumes the following form:

( )0 max min0.0023 17.8m aET Ra T T T d= + − (1)




where Ra (mm d-1) is the extraterrestrial radiation, calculated on the basis of latitude and calendar day [10]; d is the number of days of the month. The Hargreaves equation was chosen because it represents a good compromise between simplicity (because based on temperature data only) and goodness of results in many environments [10] including Sicily [11]; b) the aridity index, AI, calculated using the following equation:

12

0 ,1

y

m jj

PAI

ET=

=

∑ (2)

The monthly series of Tmax, Tmin, Ta and Pm were also analysed at sub-regional scale within the two Calabrian (CaT and CaI) and the three Sicilian (SiN, SiE and SiS) homogeneous sub-regions. Seasonal and annual mean values of temperature and rainfall for each sub-region were checked with the Mann-Kendall non-parametric test, as described in Hirsch et al. [12] to look for a trend. The slope of the trends was calculated by least-square linear fitting. An additional analysis involving the 10-years running averages was carried out with the aims to show the long period tendencies. The indices ET0m and AI were analysed by splitting the study period into two sub-periods: from 1926 to 1962 and from 1963 to 2000. For each period, we calculated the slope of the regression lines and the percentage of stations for which the t-test [12] at the 0.05 level was significant.

3 Results

3.1 Yearly and seasonal temperature and precipitation analysis

The mean values b of the calculated slopes of the regression lines for the investigated stations are listed in Table 1 together with the proportion of weather stations for which the Mann-Kendall test showed significant values (at the 0.05 level of significance). The Tya values showed increasing trends in all zones and sub-zones, with b values ranging from +0.1 to +2.2 °C/100 years, with the only exception of CaT (where b = - 1.4 °C/100 years). In general, the b values calculated for Si were greater than those resulted in Ca. The higher increments occurred during the winter season in Si and during the spring in Ca where Tya also showed a decreasing trend in autumn and, only for CaT, in summer and winter. The analysis carried out for Tmin showed in most cases b values greater than those related to Tmax. The Tmin increase was similar in all seasons; conversely, the Tmax increase was greater in autumn and in winter. At seasonal scale, the analysis showed the highest b values in Si and in particular for SiS. The percent of weather stations for which the Mann-Kendall test (K, %) showed significant values was always greater than 50% for the annual values of Ta, Tmax and Tmin. At seasonal scale, the highest percentages occurred in Si, with values greater than 80% particularly for Tmin.




Table 1: Mean values of linear regression coefficients (b), and percent of stations with significant trend (Kendall’s test, K, %) for temperatures (T) and rainfalls (P) of Sicily (Si) and Calabria (Ca).

T Tmax Ta Tmin

P

zone b K b K b K b K

Spring Si 1.1 39 1.7 46 2.3 71 -7 4 SiN 1.2 29 1.4 43 1.6 57 2 0 SiE 1.1 22 1.8 56 2.5 89 -30 11 SiS 1.0 58 1.8 42 2.5 67 5 0 Ca 1.4 44 2.1 40 0.8 56 -43 15 CaT -1 43 2.3 43 -1.2 43 -62 13 CaI 2.7 44 0.5 22 1.1 39 7.9 0 Summer Si 1.3 29 1.9 57 2.4 75 8 7 SiN 1.7 29 1.5 43 1.3 57 3 0 SiE 1.2 22 2.0 56 2.7 89 2 0 SiS 1.2 33 2.0 67 2.9 75 15 17 Ca 0.7 48 1.4 36 0.3 56 37 22 CaT -3.2 43 -0.5 43 -0.3 43 30 42 CaI 2.1 50 0.1 33 1.1 33 106 21 Autumn Si 2.1 54 1.9 68 1.6 64 -82 36 SiN 1.5 29 1.4 43 1.4 57 -80 29 SiE 2.1 67 2.1 78 2.0 87 -97 22 SiS 2.4 58 2.0 75 1.5 67 -72 50 Ca 1.1 44 -0.4 56 -0.1 32 -128 39 CaT -0.1 29 -3.9 71 -2.6 14 -135 40 CaI 1.9 44 -0.4 50 0.4 11 43 1 Winter Si 2.8 71 2.6 75 2.5 68 -98 57 SiN 1.8 43 1.8 43 1.7 43 -105 57 SiE 2.6 89 2.6 89 2.6 78 -91 56 SiS 3.4 75 3.1 83 2.8 75 -100 58 Ca 1.1 28 0.8 28 0.6 48 -139 37 CaT -1.5 29 -1.0 29 -1.7 43 -176 47 CaI 2.2 22 0.6 22 1.1 28 41 0 Year Si 1.8 61 2.0 57 2.2 79 -179 43 SiN 1.5 57 1.5 57 1.5 71 -181 43 SiE 1.7 56 2.0 89 2.4 89 -212 44 SiS 2.0 67 2.2 75 2.4 75 -153 42 Ca 1.1 56 0.1 56 1.1 68 -272 47 CaT -3.2 43 -1.4 57 -2.9 43 -344 49 CaI 2.2 51 0.2 44 0.9 33 197 1

The annual values of rainfall showed decreasing trends in all zones and sub-zones, with b values ranging from -153 to -344 mm/100 years; the decrease was




E = East; N = North; b = °C/100 years for T and mm / 100 years for P ;I = Ionian. S = Sud; T = Tyrrhenian;

greater for Ca while it covered about 21% of the mean annual rainfall of the investigated period for both the regions. Where b was positive (CaI) the percent of stations with significant trends was very low (1%) and can be neglected. Negative trends occurred in autumn and winter, with b values ranging from -72 to -176 mm/100 years, with the only exception of CaI. During these seasons, although almost all the stations showed decreasing trends, these were significant, for more than 50% of the sites, only for winter rainfalls of Si. A greater variability occurred for spring rainfalls, while the summer ones showed positive values of b (between 2 and 106 mm/100 years). The rainfall decreasing trend seems to depend principally on the decrease of Np (number of wet days); in particular, for Ca, during the period 1951-1998, Np decreased of 32 days (b = 67 days/100 years); this decrease occurred especially during winter and spring. Referring to long-period tendencies, the 10-years running averages showed complex and different behaviours from zone to zone and between the investigated climatic variables. Tya showed a strong increase since the 70s (Fig. 2) for Si; the trends are very similar from zone to zone. Conversely, the corresponding trends in Ca, since the same period, seem to be decreasing (Fig. 2). In this case, if the Tay series are divided into two sub-periods of equal duration, it can be seen that in CaI, b is increasing during the first sub-period (1926-1962) and almost equal to zero during the second one (1963-2000); in CaT, b was almost equal to zero during the first period and negative during the second one.

14

14.5

15

15.5

16

16.5

17

17.5

18

1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

Year

Mea

n te

mpe

ratu

re, °

C

CalabriaSicilyCaICaT

Figure 2: Ten-years running averages of annual mean temperature.

Py showed a decreasing trend, more clearly since the 50s, in both Si and Ca. Even in this case, the 10-years running averages showed similar behaviours within the Sicilian sub-regions; conversely, in Ca the regional results seem to hide the increasing trend of CaI (Fig. 3).




0

200

400

600

800

1000

1200

1400

1600

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

Year

Prec

ipita

tion,

mm

Calabria SicilyCaICaT

Figure 3: Ten-years running averages of annual precipitation.

Table 2: Mean linear regression coefficients (b), and percent of stations with significant trend (t-test, %) for Evapotranspiration (ET0) and Aridity index (AI) for Sicily in the sub-periods 1926-62 and 1963-2000.

ET0 Spring Summer Autumn Winter Year AI

1926-62 b 29 82 21 22 154 -0.2 t, % 25 25 25 25 25 18 1963-2000 b 58 35 7 40 150 -0.3 t, % 50 25 39 46 46 14

0. -1

3.2 Bioclimatic indices analysis

The annual values of ET0 in Si showed b values always positive and almost the same for the two investigated sub-periods during which its increase was about 5% of the ET0 values calculated for the same sub-periods. At seasonal scale, the greatest rise occurred in summer, during the first sub-period (1926-1962) and in spring during the second one (1963-2000). The proportion of weather stations that showed significant trends (t-test at 0.05 level of significance) was greater during the second sub-period and this behaviour can be explained by the greater temperature increase occurred during the same period. The aridity index AI showed a decrease in both the sub-periods, particularly during the second one (1963-2000), where it assumes a mean value equal to 0.62 (typical for sub-humid dry climate) that is less than the first period where its calculated mean value resulted equal to 0.70 (typical for sub-humid




b = mm/100 year for ET ; mm mm /100 years for AI.

climate). The trends are significant (at the 0.05 level) for less than 20% of the considered weather stations.

4 Conclusions

The analysis carried out in this study which involved temperature and precipitation data covering a period of about 80 years, showed an increase of temperature (ranging from 0.1 and 2.2°C/100 years for the mean annual temperature) and a decrease of precipitation (ranging from 153 to 344 mm/100 years for the annual rainfall) over the two investigated regions (Calabria and Sicily) located in South Italy. The precipitation decrease can be explained by the decrease of the number of wet days NP according to other studies performed in different Italian regions. The calculated values of evapotranspiration ET0, which accounted for mean (Ta), maximum (Tmax) and minimum (Tmin) values of temperature, showed an increasing trend which was lower than that related to the temperature because Tmax and Tmin did not increase with the same magnitude. The aridity conditions are not encouraging during the analysed period: the calculated Aridity Index (AI) showed values which tend to decrease from a first sub-period (1926-1962) to a second one (1963-2000). This resulted in a climate change in Sicily from a sub-humid climate to a sub-humid dry climate. Because of the geographic location and the particular orographic features of the two investigated regions which extend over 3° of latitude, a strong difference in terms of climate change occurred. These main differences can be summarised as follows: a strong increase in temperature has occurred in Sicily since the 1970s, while in Calabria the same variable showed steady values or a light decrease; Tmax and Tmin values have clearly decreased during all the seasons over the Calabrian Jonian sub-region (-3.2 e -3.9 °C/100 years respectively for Tmax and Tmin); rainfalls have increased over the Calabrian Ionian sub-region (197 mm/100 years for Py). The overall results showed the importance of the climate change analyses at a regional scale and provide the Italian authorities with useful information to begin to assess the impacts of such climate changes on human activities.

References

[1] Brunetti, M., Buffoni, L. & Mangianti, F., Temperature, precipitation and extreme events during the last century in Italy, Global and Planetary Change, 40, pp. 141-149, 2004.

[2] Brunetti, M., Buffoni, L., Maugeri & M, Nanni, T., Trend of minimum and maximum daily temperatures in Italy from 1865 to 1996, Theor. Appl. Climatol., 20, pp. 1017–1031, 2000.

[3] Capra, A., Indelicato, S., Li Destri Nicosia, O. & Scicolone, B., Evaluation de la sécheresse d’aprés les données de précipitation. Une application au Sud d’Italie, Proc. of the 16th European Regional Conf., ICID, Budapest, Hungary, June, pp. 41-51, 1992.




[4] Capra, A., Li Destri Nicosia, O. & Scicolone B., Application of fuzzy sets to drought classification, Proc. of the 2nd Int. Conf. On Advances in Water Resources Technology and Management, Eds. G. Tsakiris & M.A. Santos, Balkema, Rotterdam, pp. 479-483, 1994.

[5] Piccarreta, M., Capolongo, D. & Boenzi, F., Trend analysis of precipitation and drought in Basilicata from 1923 to 2000 within a southern Italy context, Int. Journal of Climatology, 24, pp. 907-922, 2002.

[6] Rossi, G. & Cancelliere, A., Problemi e prospettive del monitoraggio e delle mitigazione della siccità, Editoriale Bios, Cosenza, Quaderni di Idrotecnica, pp. 9-44, 2002 (in Italian).

[7] Capra A., Malara, S.L. & Scicolone, B., (2004), Analisi delle temperature e delle piogge mensili in Calabria nell’ultimo cinquantennio, Economia Montana- Linea Ecologica, 3, pp. 31-36, 2004 (in Italian).

[8] ENEA, Cambiamenti climatici e rischi di siccità e desertificazione in area mediterranea ed in Sicilia, Progetto speciale clima globale, Roma, Italy, giugno, 2002 (in Italian).

[9] Hargreaves, G.H. & Samani, Z.A., Reference crop evapotranspiration from temperature. Transactions of the ASAE, 1(2), 96–99, 1985.

[10] FAO, Crop evapotranspiration. Guidelines for computing crop water requirements, FAO Irrigation and drainage Paper, 56, 1998.

[11] Di Stefano, C. & Ferro, V., Estimation of Evapotranspiration by Hargreaves Formula and Remotely Sensed Data in Semi-arid Mediterranean Areas, J. Agric. Engng Res., 68, 189–199, 1997.

[12] Hirsch, R.M., Helsel, D.R., Cohm, T.A. & Gilroy, E.J., Statistical analysis of hydrological data, Handbook of Hydrology, Ed. Maidment, D.R., McGraw-Hill, pp. 17.1-17.55, 1993.




Section 5 Geo-environment in

urban settings

The urban geo-science model: an essential tool to support planning and sustainable development

British Geological Survey, Nottingham, UK

Abstract

New technology for viewing and sharing digital information means that national geological surveys can present geo-science information in innovative ways that take better account of the third dimension. The British Geological Survey is piloting the development of 3-D geo-science models in four conurbations across the UK. These models are being developed with a view to supporting sustainable development, particularly in the context of urban groundwater management and development planning. The work is research based but is being guided by the very specific needs of regulatory bodies and local authorities. The success of this initiative will depend ultimately on creating greater awareness amongst the user community of the value of 3-D models, and ensuring that relevant information is made available in advance of major infrastructure projects. Keywords: urban geoscience, 3-D modelling, groundwater, development planning.

1 Introduction

Most of the 60 million people that make up the population of Britain live in urban and suburban areas that comprise only 10 per cent of the countries’ surface area. The principal conurbations (Greater London, West Midlands, Greater Manchester, West Yorkshire, and the Central Lowlands of Scotland) expanded dramatically during the industrial revolution, in part, due to the ready availability of local raw materials (coal, water, ironstone). Some of these conurbations directly overlie productive aquifers where groundwater is a resource; others are underlain by less permeable rocks where urban drainage and the disposal of water from the urban infrastructure to the subsurface is a problem. Both require




doi:10.2495/GEO060191

D. McC. Bridge, B. L. Morris & J. R. A. Giles

an understanding of subsurface geology and hydrogeology to guide urban planners, inform urban infrastructure investment and enable sustainable development.

The Government’s promotion of urban regeneration and redevelopment of brownfield sites places an additional premium on understanding the conditions of the ground before development, and the likely impacts both at the construction stage and subsequently, when the land is restored to use.

2 The challenge

The provision of information on ground conditions is not a new idea. William Smith’s first geological map, published in 1820, was compiled to meet a practical need – to document the principal mineral resources of the country. Subsequent refinements by the British Geological Survey (BGS) over the past 170 years mean that the UK is now covered by the most detailed set of geological maps of any country in the world. One of the problems facing any national survey, however, is how best to make information available at a level of detail that meets the needs of decision makers, not all of whom are necessarily geologists.

The applied mapping programme of urban conurbations, commissioned by Central Government during the 1980s and 1990s, was an ambitious attempt to address this problem. It aimed to provide accessible advice on the relevance of geological conditions to strategic land use planning e.g. Forster et al [1]. Although the initiative was successful at the technical level, the hard copy outputs were expensive to produce and there was no provision for updating the underlying databases or the map outputs.

Recent advances in modelling software and web delivery options make the task of integrating and delivering geo-science information a practical reality. However, raising awareness amongst policy makers and planners of the role of geo-science information remains a significant challenge.

3 The 3-D approach

As part of its remit to make geological information available to a wider audience, the British Geological Survey (BGS) has embarked on a national programme of 3-D modelling. Within this programme, work has begun in four major development areas (Thames Gateway, Manchester, Clyde Basin and Belfast) to build high-resolution geo-science models capable of supporting decisions on a range of geological, engineering and hydrogeological topics. These urban models concentrate on the shallow subsurface (zone of human interaction) and are constructed primarily from downhole information collected in the course of site investigation studies. A typical urban model may incorporate information from several thousand boreholes, fig 1. Modelling of Quaternary sediments is carried out using proprietary software developed at the University of Cologne, Kessler and Mathers [2]. In more complex, faulted bedrock




sequences, Gocad® is the preferred package. There is complete interoperability between the two systems and outputs can be combined and exported in standard format.

Figure 1: 3-D model of part of Trafford Park, Manchester showing selected Quaternary elements.

4 The role of the 3-D model

The outputs from the 3D geological model can be tailored to a range of end uses, and have direct application in several key areas, including urban groundwater management, planning for development and preliminary site appraisal. Three case studies illustrate these different applications.

4.1 Groundwater management, Trafford Park, Manchester

One consequence of the European Water Framework Directive is that, as surface and groundwater safeguards become more stringent, there will be a greater need to manage urban water resources in a more sustainable manner. The use of geological models to inform and better represent the complexities of the subsurface in groundwater flow models and contaminant studies is now beginning to interest regulatory bodies and hydrogeological consultants. In the Trafford Park industrial zone of Manchester, the sensitivity of the Triassic




sandstone bedrock aquifer to pollution and the extent to which recharge may occur have been analysed through detailed characterisation of the underlying superficial deposits. Potential hydrogeological pathways from ground surface to the sandstone have been identified, and thematic outputs show the importance of the Manchester Ship Canal and related waterways as potential sources of recharge and pollution of the bedrock aquifer, figs. 2 & 3.

Figure 2: Synthetic section derived from 3-D model showing inferred recharge pathways.

Figure 3: High vulnerability /recharge zones (shown cross-hatched) in Trafford Park, Manchester.




4.2 Development planning

‘Unforeseen ground conditions’ are frequently cited as the main cause of project overruns and escalating development costs. In Scotland, Glasgow City Council, working in collaboration with BGS, has taken the innovative step of making available their detailed borehole databases for incorporation in a 3-D superficial and bedrock model now being built for the city centre. This model, when completed, will provide a representation of ground conditions at local and regional scale and inform planning of subsequent investigation phases. Emphasis is being placed on the acquisition of mining information with a view to delineating areas within the city where development may be compromised by unrecorded, shallow coal workings.

Figure 4: Property damage caused by shallow undermining for coal. Belmont Street, Glasgow.

4.3 Preliminary site assessment

At project concept stage, modern visualisation techniques and virtual models are used increasingly by design consultants to illustrate the environmental impact of major new infrastructure. This technology can be equally applied to the subsurface, where the 3-D model can provide a picture of probable ground conditions to facilitate preliminary design and costing. The example, fig. 5, shows how, by model interrogation, a range of geological and geotechnical information can be quickly assembled to plan a focused and cost-effective site investigation programme.




Figure 5: Rapid site assessment based on interrogation of 3-D model, attributed with geotechnical data.

5 Information management: the real cost of 3-D modelling

3-D geological models are only as good as the data that underpins them. The cost of acquisition of new data is prohibitive, therefore effort is focused on prior information, Wood and Curtis [3]. Borehole information is acquired by a wide range of exploration industries (e.g. hydrocarbon, coal, water, site investigation). Such information has to be collected from the companies, either under statute (Water Resources Act), or by donation from the company after careful negotiation. Once the original records have been acquired, the information management process begins. This is a non-trivial task, which requires considerable resources and a wide range of skills from a number of non-




geological disciplines. Because of the volumes of data involved (over one million records held in the National Geoscience Records Centre) data management has to follow clear procedures as outlined by Feinemann [4]. From the 3D modelling viewpoint, the key requirements are to:

• populate the databases • manage the domains associated with the databases to ensure

interoperability • make the data available via appropriate applications

6 Emerging issues

The primary output of the modelling process is a 3-D geological model with a resolution designed to support the urban map scale of 1:10 000. For many applications, the geological model is sufficient in itself to provide answers to many development problems. The extent to which other information (geotechnical, hydrogeological, geochemical) is incorporated in the model is likely to be driven by legislative requirements. In the context of groundwater, the emerging topic include:

6.1 Infiltration from rainfall, surface drainage and the urban pipe infrastructure

One of the major uncertainties of urbanisation is the urban water budget, which is much more complicated to calculate than its rural equivalent. Get it wrong and the resulting uncertainty can be costly as groundwater levels rise beyond their historical levels, causing major problems of flooding (tunnels, basements, underground car parks) and foundation stability.

6.2 Point-source pollution from contaminated land

As brownfield sites are targeted for redevelopment, planners are increasingly interested in the ability of the subsurface to detain, attenuate or dilute the products of former polluting activities, especially if the contaminants are toxic.

6.3 Sustainable urban drainage systems

SUDS technologies are the subject of much discussion in urban engineering fields, being seen by protagonists as a cost-effective way to limit the complexity and consequences of runoff control in built areas and by critics as having the potentially harmful side-effect of dispersing pollutant loads. The answer to many questions over their use lies not merely in the mechanics of the diverse engineering structures being developed as SUDS systems but in the ability of the subsurface (soils, regolith/unsaturated zone and saturated zone) to remove, attenuate or mediate substances of concern.




7 Conclusions

3-D urban geo-science models complement and enhance the information portrayed on traditional 2-D geological maps. Although access to, and use of 3-D urban geo-science models is currently limited, there is increasing recognition among users of the added value offered by acquiring information in 3-D rather than 2-D. The challenge is to make the models available at an early stage in the urban regeneration process so that they may better inform strategic planning options, ground investigation and reclamation strategies. This is reliant on the models being appropriate in terms of content and scale, and that inherent uncertainties associated with the model are appropriately quantified.

Acknowledgement

This paper is published with the permission of the Executive Director of the British Geological Survey (NERC).

References

[1] Forster, A., Arrick, A., Culshaw, M. G. & Johnston, M., (eds) A geological background for planning and development in Wigan, British Geological Survey Technical Report No. WN/95/3: Keyworth Nottingham, 1993.

[2] Kessler, H. & Mathers, S. Maps to models. Geoscientist, 14(10) pp. 1- 6. [3] Wood, R. & Curtis, A. Geological prior information and its application to

geoscience problems. Geological prior information: informing science and engineering , Wood, R. and Curtis, A. (eds) Geological Society, London, Special Publication, 239, pp.1-14, 2004.

[4] Feineman, D. R., Data management : Yesterday, today and tomorrow. Petex ‘92 Conference. London, 1992.




Developing design-oriented strategies to combat regional scale climate change

B. Stone, Jr Georgia Institute of Technology, City and Regional Planning Program, USA

Abstract

This paper presents a study of residential parcel design and surface heat island formation in a major metropolitan region of the southeastern United States. Through the integration of high-resolution multispectral data (10 m) recorded by the National Aeronautical and Space Administration (NASA) with property tax records for over 100,000 single-family residential parcels in the Atlanta, Georgia metropolitan region, the influence of the size and material composition of residential land use on an indicator of surface heat island formation is reported. In contrast to previous work on the urban heat island effect, this study derives a parcel-based indicator of surface warming to permit the impact of land use planning regulations governing the density and design of development on the excess flux of heat energy to be measured. The results of this study suggest that the contribution of individual land parcels to regional surface heat island formation could be reduced by approximately 40% for average sized parcels through the adoption of specific land use planning policies, such as zoning and subdivision regulations, and with no modifications to the size or albedo of the residential structure. Keywords: climate change, urban heat island effect, land use, urban planning.

1 Introduction

The significance of land use to climate change has only recently begun to receive serious attention within the technical literature. As reported by Kalnay and Cai [1], analyses of both surface and atmospheric temperature trends over the last half century suggest that land use change may be responsible for as much as 50% of the observed reduction in the mean diurnal temperature range – a phenomenon



doi:10.2495/GEO060201


largely attributed to increasing minimum temperatures throughout the continental United States. Two principal causes of this increase in minimum temperatures, agriculture and the urban heat island effect, are identified as significant drivers of ongoing changes in climate. In short, an increase in surface temperatures resulting from land cover changes, independent of ongoing increases in greenhouse gas concentrations, appears to be a significant driver of climate change. In a similar vein, Foley et al [2] conclude from a global analysis of land use that while “[l]and use has generally been considered a local environmental issue … it is becoming a force of global importance” (p. 570). Despite a growing recognition of the importance of land use to climate change, the suite of mitigation strategies embodied in the still-evolving Kyoto framework remain almost entirely technological in nature, with little attention to strategies rooted in the design of human settlements. Because the process of climate change is driven by two distinct mechanisms – an increase in the heat absorbing potential of the atmosphere and an increase in surface heat emissions – there is a critical need to focus not only on greenhouse gas reductions but also on strategies to reduce surface heat emissions at the local level. To address this issue, this article presents the results of a study of residential parcel design and surface warming patterns in the Atlanta, Georgia (USA) metropolitan region. With the aid of high resolution thermal infrared (IR) imagery and a parcel level geographic information system, the thermal properties of over 100,000 single family parcels are associated with the spatial structure and material composition of residential development throughout a 48 km2 study region. In measuring the thermal properties of individual land parcels, an “emissivity adjusted” indicator of the parcel black body flux is developed to account for the influence of variable surface emissivity and sky radiation captured in the thermal IR imagery. Furthermore, the predevelopment radiant flux is estimated and subtracted from the observed, post-development radiant flux to differentiate the warming impacts of urban development from natural, background sources of surface heat.

2 Background

In order to associate remotely sensed thermal IR data with land use planning policies it is essential that a “policy-relevant” indicator of surface warming be developed. While remote sensing technologies can provide a valuable tool for assessing the thermal performance of urban environments, these must be used with great caution to yield reliable information for land use planning. Most importantly, there is a critical need to isolate the influence of specific classes of land use on surface warming patterns. While previous work has demonstrated a clear distinction between the thermal properties of “urban” and “rural” land cover types, each of these broad categories of development is composed of a number of specific land use classes that are subject to an array of development regulations. If we are to mitigate heat island formation through the modification of specific zoning and subdivision regulations, we must be able to isolate the influence of these individual policies on surface warming. To be policy-relevant,




an indicator of surface warming must be capable of quantifying the thermal characteristics of distinct and legally defined classes of land use. In this section of the paper, I discuss three criteria that must be satisfied to derive a policy-relevant measure of heat island formation. These include: 1) the measurement of surface warming at the level of the individual land parcel; 2) the adoption of a flux-based rather than a temperature-based indicator to capture the effects of land area on warming; and 3) the adjustment of the radiant flux to more closely correspond to the sensible flux of heat.

2.1 Measuring thermal performance at the parcel level

In assessing the surface thermal properties of urban and rural land covers, many studies have adopted a unit of analysis consistent with the resolving dimension of a satellite radiometer, such as 1.1 km or 120 m (e.g., Roth et al [3]; Nichol [4]). An important limitation of adopting such a spatially uniform unit of analysis, however, is that it fails to conform to the irregularly shaped boundaries of individual land parcels, the unit of area at which land use is controlled in the United States. For example, in dense and well-mixed urban districts, a spatially uniform unit of analysis is likely to capture both commercial and residential land uses, confounding any attempt to isolate the thermal properties of a single land use class. Such an outcome may be categorized as a special instance of mixed pixel error – mixed, that is, in terms of land use class rather than land cover type. The limitation of this type of mixed pixel error for planning analysis is that it obscures the underlying policy drivers of environmental performance. If we are interested in evaluating the environmental impacts of a particular planning policy, such as maximum impervious area regulations, we must first identify those zones subject to the policy. For example, within a square kilometer of urban land there is likely to be a number of unique land use classes, each with its own zoning regulations governing the permissible area of impervious cover per parcel. In measuring the environmental characteristics of an area of land subject to multiple zoning regulations, it is difficult to assess the relative influence of any single policy on environmental performance. For the purposes of planning research, a more methodologically sound approach entails the measurement of surface warming at a dimension compatible with land use regulation, such as a zoning district or individual land parcel. As adopted herein, a parcel-based approach to surface warming analysis permits parcel-specific land use policies, such as zoning regulations, subdivision regulations, and building codes, to be associated with thermal performance at the same spatial dimension.

2.2 Accounting for the effects of land area on surface warming

While parcel-based measures of surface warming are needed to associate heat island formation with land use policies, the adoption of a spatially irregular unit of analysis complicates the use of surface temperature as an indicator of warming. As noted by Price [5] over 25 years ago, temperature-based measurements, such as the urban—rural temperature differential, often fail to reflect the magnitude of heat islands created by urban areas of different sizes.




Price states, “[f]or investigation of urban heating, the peak temperature is less significant than a summation (area integral) of the excess power radiated as a result of the surface temperature elevation” (p. 1557). In other words, a measure of the excess flux of thermal energy from a city provides a more accurate indicator of the total magnitude of warming within an urbanized area than does the urban—rural temperature differential. In this sense, the magnitude of surface warming between urban and rural zones (defined herein as the excess flux of sensible heat energy) is more relevant to urban planning policy than heat island intensity (the urban – rural temperature differential). The reason for this is that the area-integrated flux of heat energy emitted from a city or parcel provides a more accurate indicator of the volume of the near surface atmosphere influenced by elevated surface temperatures than does the surface temperature differential. In this sense, two cities of similar climate and development type but dissimilar land area are likely to exhibit similar average surface temperatures but different fluxes of energy. As the area of a city expands, the volume of the heat island “dome” – a volume subject to elevated air temperatures and enhanced pollution formation – expands as well. Measures of heat island intensity across cities may not reflect a differential in the volume of heat island domes. While smaller in scale, the same is true when comparing two parcels of different area.

2.3 Approximating the sensible heat flux

While aerial and satellite radiometers provide the most viable means of obtaining continuous surface thermal data across a large study region, a principal limitation of remote sensing studies of the urban heat island effect is a failure to account for the variable distribution of surface emissivity. Given the wide array of surface material types found in urban areas, ranging from heavily canopied areas to the impervious materials of driveways and rooftops, radiant flux densities recorded by thermal sensors are prone to systematically underestimate thermodynamic temperatures from the heterogeneous surfaces of urban environments. As a result, the surface radiant flux density must be adjusted to more closely approximate the black body flux density, which is a better indicator of sensible heat. In summary, to be useful in planning research, an indicator of the urban heat island effect must be spatially compatible with the legal dimensions of land use control, and must accurately approximate the quantity of energy contributed by a single land parcel to elevated temperatures within cities. What follows in the remainder of the paper is an application of this conceptual approach to the Atlanta, Georgia metropolitan region and a presentation of policy insights that may be gleaned from this analysis.

3 Methods

In May of 1997, the National Aeronautic and Space Administration (NASA) obtained high resolution multispectral imagery over the Atlanta, Georgia




metropolitan region for the purpose of investigating the influence of land use on urban heat island formation. Recorded from an aerial platform with the Advanced Thermal and Land Applications Sensor (ATLAS) at a ground resolution of 10 meters, 15 channels of multispectral data were obtained over a 48 km2 region by NASA during the course of two flights and made available to a range of research efforts known collectively as “Project Atlanta.” The exceptionally high spatial resolution of this data permits the thermal characteristics of high-density urban parcels to be reliably measured and integrated into a parcel level geographic information system. With the aid of the Project Atlanta data, a measure of the parcel “net black body flux” was derived for every single family residential lot in the Atlanta study region. What follows is a brief overview of the methods employed to derive a measure of the net black body flux based on remotely sensed data; for a full discussion of this methodology, please see Stone and Norman [6].

3.1 Deriving the net black body flux

The first step in estimating the parcel net black body flux requires that the thermal characteristics of individual land parcels be measured. To do so, parcel boundary information for the approximately 104,000 single-family parcels located in the study region was obtained from the City of Atlanta and Fulton County offices of tax assessment. By registering the multispectral imagery and parcel geography data layers to the same coordinate system, it was possible to spatially overlay the two datasets in a geographic information system. Once integrated, the average apparent radiant flux density of each parcel was derived through a zonal summary function. In addition to the apparent radiant flux density, information on the composite parcel emissivity and emitted sky radiation was needed to derive the black body flux density. Employing the method of Carlson et al [7], the fraction of vegetative cover for each land parcel was estimated through the derivation of a normalized difference vegetation index (NDVI) with ATLAS bands in the red and near infra-red spectral regions. Based on published data for vegetative and impervious materials characteristic of residential zones, an emissivity value of 0.96 was assigned to the vegetative component of each parcel and a value of 0.92 was assigned to the impervious component (Jensen [8]; Oke [9]). The composite parcel emissivity was then calculated based on the area of pervious and impervious materials per parcel and used, in concert with the average value of emitted sky radiation for the day of ATLAS data collection, to compute the estimated parcel black body flux density. As a final step, the quantity of radiant energy attributable to predevelopment land characteristics – the rural “background” radiation – was subtracted from the measured average parcel black body flux density. In deriving an indicator of the urban heat island effect, we are interested in the thermal differential between developed and undeveloped land features. When measuring heat island intensity, this variable is quantified as the temperature differential between developed and undeveloped zones. When measuring heat island magnitude, this variable becomes the flux differential between developed and undeveloped zones. To




calculate this flux differential at the parcel level, we first calculate the average black body flux density for a developed parcel and then subtract from this quantity an estimate of the predevelopment black body flux density that would result were the same parcel occupied by a natural land cover, such as forest canopy cover. This flux density differential is then multiplied by the parcel area to estimate the parcel net black body flux.

3.2 Measuring parcel design

Following the development of the parcel net black body flux measure, four parcel design variables were derived for this study based on city and county property tax records and the multispectral ATLAS imagery. These included the area of impervious materials, such as driveways and building footprints, the area of lawn and landscaping, the proportion of the parcel overlaid by tree canopy, and the number of bedrooms in the residential structure. This final variable was used to control for residential capacity. In short, I am interested in comparing the thermal characteristics of parcels varying in size and material composition but similar in the number of residents the parcel was designed to accommodate. The inclusion of this variable in our analysis ensures that differences in thermal performance by region of the city are not attributable to differences in residential capacity. Information on the area, footprint size of residential buildings (houses and detached buildings), and number of bedrooms in the residential structure was obtained from the City of Atlanta and Fulton County offices of tax assessment. As driveway areas are not recorded in the property assessment process, a sample of parcels stratified by age and size was selected to estimate the area of these paved surfaces. The driveways of selected parcels were then measured directly to determine how paved areas scale with parcel size throughout the study region. In combination, the estimated area of driveway paving and the area of the building footprint constitute the impervious component of the parcel. The area of lawn and landscaping was derived by subtracting the estimated impervious area from the total parcel area. Finally, the proportion of each parcel overlaid by tree canopy was estimated from the NDVI.

4 Analysis and results

With the aid of the Statistical Package for the Social Sciences (SPSS), an analysis was performed to assess the influence of parcel design on the parcel net black body flux. The average single family residential parcel in the database is characterized by a three bedroom house and is approximately 1,360 m2 in area, of which 208 m2 or 15% is occupied by the impervious components of the house and driveway, with the remaining 85% occupied by lawn and landscaping. Approximately 45% of the average single-family residential parcel is overlaid by tree canopy cover. To evaluate the graphic correlation between parcel area, material composition, and surface warming, values of increasing lot size are plotted




against impervious area, lawn area, and the net black body flux in Figure 1. As illustrated in this figure, the magnitude of surface warming scales closely with lot size, with the mean net black body flux increasing by a factor of about six between the highest and lowest density classes. While this descriptive evidence supports my hypothesis of a negative relationship between the density of single-family residential development and surface warming, it is important to note that this simple covariation does not account for the distribution of residential capacity or tree canopy cover throughout the region. If larger lot sizes are also associated with greater residential capacities (e.g., four and five bedroom houses) and a less mature tree canopy cover, then the relationship between the density of development and thermal performance may be attributable to incompatible design objectives or to the age of development in different areas of the city. In order to control for these important influences, a multivariate statistical model is developed and evaluated in the following section.

Figure 1: Parcel attributes by lot size.

4.1 Explanatory analysis

In order to quantify the potential thermal benefits of specific design based strategies, ordinary least squares was used to develop a predictive model linking parcel design to the net black body flux. Four independent variables were incorporated into the model to explain the net black body flux. These included the area of impervious cover, the area of pervious cover or lawns and landscaping, the percentage of the parcel overlaid by tree canopy, and the number of bedrooms in the residential structure. With the aid of SPSS, a set of parameter estimates and model summary statistics were generated for the approximately 104,000 single family residential parcels located in the Atlanta study region. The results of this modeling process are presented in Table 1.

0

500

1000

1500

2000

2500

3000

3500

4000

0 - 500 500 - 1,000 1,000 - 2,000 2,000 - 4,000 > 4,000

Lot Size (sq. meters)

Are

a of

Lan

d C

over

(s

q. m

eter

s)

0

1000

2000

3000

4000

5000

6000

7000

8000

Net

Bla

ck B

ody

Flux

(W)

Imperv Area Lawn Area Net Black Body Flux




The results of the regression analysis indicate that the various parcel design attributes are significantly related to the parcel net black body flux. As expected, the area of impervious and pervious materials were found to exhibit a significant positive relationship with parcel warming when controlling for other parcel design attributes believed to be associated with the dependent variable. Also as expected, the presence of tree canopy cover was found to significantly decrease the magnitude of warming per parcel. These relationships were found to hold when controlling for the number of bedrooms in the residential structure, suggesting that the elevated surface warming is not solely attributable to variation in residential capacity.

Table 1: Regression analysis results for parcel net black body flux (W)*.

Variable B-Coefficient Standardized Coefficient

Significance

Impervious Area (m2)

0.079 0.274 < .001

Lawn Area (m2)

0.013 0.553 < .001

Tree Canopy Cover (%)

-23.93 -0.419 < .001

Number of Bedrooms

0.438 0.019 < .001

Summary Statistics Adj. R-Square 0.57

F-Statistic 35,217

F Significance < .001

* The square root of the net black body flux was used in this model as a variance stabilization transformation.

5 Discussion and conclusions

The results of this analysis provide compelling evidence that the size and material composition of single family residential parcels are significantly related to the magnitude of surface warming in the Atlanta study region. Specifically, smaller, higher density parcels were found to be associated with a lower net black body flux than larger, lower density parcels when controlling for the class of land use and the number of bedrooms in the residential structure. While both the area of impervious materials and lawn and landscaping were found to be positively related to the parcel net black body flux, the area of lawn and landscaping – a strong correlate of parcel size – was found to be the strongest predictor of excess surface warming. In contrast to previous work on the urban heat island (e.g., Hoyano [10]), the results of this study support the hypothesis that lower density, dispersed patterns of urban residential development contribute more surface energy to regional heat island formation than do higher density, compact forms. While specific to the Atlanta metropolitan region, due to the uniformity of building materials and development patterns found throughout North American cities, these findings are




generalizable to a number of large, mid-latitude cities characterized by a similar climatological and physiographic characteristics. For such cities, this work illuminates at least three specific planning strategies that may be employed to offset the excess surface heat energy emitted from both existing and future residential development, including reductions in the area of residential lawn space, reductions in the area of impervious cover, and an increase in tree canopy cover. Each of these strategies is discussed briefly in turn. The results of this analysis suggest that the most effective design-oriented strategy for mitigating the thermal impacts of new housing development is a reduction in the area of the residential lawn. For the average parcel, a reduction in the lawn area of 25%, holding other design attributes constant, would be associated with a 14% reduction in the parcel net black body flux. By serving to increase the density of land use, a reduction in the lawn area of new residential development would offset enhanced surface warming through at least two mechanisms. First, by reducing the average lot size of development at the urban periphery, the area of rural land converted from forest, Atlanta’s natural land cover, to urban land uses is reduced, constraining the zone of enhanced warming. Second, by limiting the total zone of urban expansion, higher density development limits the extension of roadways and other thermally intensive infrastructure, also serving to preserve rural landscapes. The finding of a positive relationship between the area of impervious cover and the parcel net black body flux confirms the well established role of mineral-based building and paving materials in urban heat island formation. The most widely advocated approach to mitigating the thermal impacts of impervious materials is through the application of high albedo surface treatments to increase surface reflectivity – an approach that is cost effective and applicable to both new and existing development. For new construction, a reduction in the total area of impervious cover provides an additional tool for offsetting an increase in surface warming. These findings suggest that a 25% reduction in the impervious cover of an average single-family parcel would be associated with a 15% reduction in the net black body flux. When combined with an equivalent reduction in lawn area, parcel warming was found to be 27% lower. While the potential to reduce the thermal impacts of new growth are significant, it is important to note that changes to a city’s land development regulations will have only a limited influence on existing development. As the vast majority of the Atlanta region’s 2020 built area is already in place, changes in future peripheral development will serve to offset continued growth rather than to abate present day heat island formation. In light of this observation, the most effective design-oriented strategies for reducing total regional warming must address both new and existing development. As noted above, one strategy for doing so would entail the replacement of traditional driveways with driveway runners at the time of routine resurfacing. A second critical strategy for existing development entails the planting of trees. As indicated by our analysis, for the average single family parcel, an increase in tree canopy cover from 45 to 60% reduces the parcel net black body flux by 13%. For trees strategically planted along roadways and in proximity to houses, the thermal benefits are likely to be




greater. In combination, a 25% reduction in the areas of pervious and impervious materials, and an increase in parcel tree canopy cover from 45 to 60%, were found to reduce the net black body flux from a parcel of average size by approximately 40%.

References

[1] Kalnay, E. & Cai, M., Impact of urbanization and land use change on climate. Nature, 423, 528-531, 2003.

[2] Foley, J., Defries, R., et al, Global consequences of land use. Science, 309, 570-574, 2005.

[3] Roth, M., Oke, T., Emery, W., Satellite-derived urban heat islands from three coastal cities and the utilization of such data in urban climatology. International Journal of Remote Sensing 10, 1699-1720, 1989.

[4] Nichol, J., High-resolution surface temperature patterns related to urban morphology in a tropical city: A satellite-based study. Journal of Applied Meteorology 135, 135-46, 1996.

[5] Price, J., Notes and correspondence: Assessment of the urban heat island effect through the use of satellite data. Monthly Weather Review 107, 1554 -1557, 1979.

[6] Stone Jr., B., Norman, J., Land use planning and surface heat island formation: A parcel based radiation flux based approach. Atmospheric Environment, forthcoming. Study methodology reprinted with permission from Elsevier.

[7] Carlson, T., Capehart, W., Gillies, R., A new look at the simplified method for remote sensing of daily evapotranspiration, Remote Sensing of Environment 54, 161-167, 1995.

[8] Jensen, J., Remote sensing of the environment: An earth resource perspective. Prentice Hall, Upper Saddle River, NJ, 2000.

[9] Oke, T., Boundary layer climates, Routledge, New York, NY, 1987. [10] Hoyano, A., Relationships between the type of residential area and the

aspects of surface temperature and solar reflectance (based on digital image analysis using airborne multispectral scanner data). Energy and Buildings 7, 159-73, 1984.




Land subsidence evolution and controlling

M. A. Ortega-Guerrero Centro de Geociencias, Universidad Nacional Autónoma de México, México

Abstract

Mexico City is situated in the Basin of Mexico on a highly compressible lacustrine aquitard that overly highly productive aquifers of both volcanic and sedimentary origin. This aquitard contributes with regional recharge through leakage to the underlying alluvial-pyroclastic regional aquifer, from which about 50 m3/sec are pumped to supply water for the Metropolitan Area of Mexico City, to nearly 25 million inhabitants. The extensive lacustrine Chalco Plain in the southeastern part of Mexico City is underlined by an aquitard up to 300 m thick. This area was a shallow lake until the 1940s when it was drained for agricultural use and human habitation. In the middle 1980s, when major groundwater extraction began with the onset of the Santa Catarina Well Field in the middle of the Chalco Plain, subsidence of 2 m occurred between 1984 and 1989, causing a shallow lake to form and gradually expand. Land subsidence in the central part of the Chalco Basin has increased to 0.4 m/yr since 1984 and by 1991 total subsidence had reached more than 10 m at present. The rapid land subsidence in this area is causing the accumulation of meteoric waters during the rainy season resulting in extensive flooding of farmland and potentially urban areas. Three different stages of lake growth were identified: (i) ponding (1984-1989), (ii) lake formation (1990-1993), and (iii) bifurcation (1994-present). In this last stage, the local presence of a basalt layer in the middle of the Santa Catarina Well Field is causing differential consolidation and forming a local topographic elevation that permits the development of two separated lakes. Since the middle of the 1990s, the new lake conditions are permitting the growth of few aquatic native plants and arrival of different species of migratory birds: therefore, environmental planning and management for the southernmost end of the Basin of Mexico is proposed through the development of a series of connected artificial lakes to control urban growth, to improve environmental conditions of the MAMC, to develop ancient agricultural production systems such as the floating gardens or “Chinampas” and recreational facilities with proportional benefits to the social and economical sectors. Keywords: land subsidence, consolidation, artifitial lake development, environmental planning.




doi:10.2495/GEO060211

environmental planning and management mechanisms near Mexico City:

1 Introduction

Most of Mexico City is constructed over this highly compressible aquitard within the Basin of Mexico and is surrounded by volcanic mountains (Figure 1). The Basin is hydrologically closed and as such, a quiescent series of interconnected lakes occupied the basin floor prior to the construction of an artificial drainage system in 1789 (Bribiesca [1]). Groundwater is extracted from a regional alluvial-pyroclastic aquifer and from a fractured volcanic aquifer for supplying almost 25 million inhabitants and some 30% of the nation’s industry in the Metropolitan Area of Mexico City (MAMC) INEGI and INE [1]. Both aquifers underlie a thick sequence (between 50-300 m) of clayey lacustrine sediments,

[2]. Approximately 52 m3/s of groundwater has been extracted from these aquifers, out of a total of 65 m3/s consumed in the MAMC [1] and NRC-AIC-ANI [3].

Figure 1: Location of the Chalco Basin within the Basin of Mexico and

Groundwater extraction from the regional alluvial-pyroclastic aquifer underneath the lacustrine aquitard started in the XIX century and became

was documented since the early 1940s in downtown Mexico City over several decades due to depressurization and consolidation of the aquitard caused by heavy pumping, particularly in the middle of the city Carrillo [5]. Indeed, the magnitude of the subsidence became so extreme in some locations that severe

transportation system.




problems arose related to building foundations, sewer drainage, and subway

considered in hydrogeological terms as a regional aquitard Ortega and Farvolden

extensive during the 1930s Hiriart and Marsal [4]. Extensive land subsidence

former lakes.

To minimize further damage from additional subsidence within the MAMC many of the downtown wells were taken out of production and new well fields were constructed on the outskirts of the city and groundwater from the basin was also imported to satisfy water demands. In the southern part of the Basin of Mexico, the most recent well field to be added to the urban supply system is located within the Chalco Basin (Figure 3), where one of the ancient lakes once existed (Figure 1). This new well development has resulted in the partial transfer of the land subsidence problem from within the MAMC to the nearly agricultural and urban areas of the Chalco Plain. The main goal of this work is to update the land elevation survey up to 2005 and to present hydrological evolution of the development of the New Chalco Lake and environmental conditions recorded since 1988 to 2005. A plan for environmental improvement is suggested throughout the development of a series of artificial lakes, in the southern end of the Basin, by controlled groundwater extraction from the regional aquifer, with different environmental, social and economic impacts.

2 Hydrogeology of the Chalco Basin

The Chalco Plain was a shallow lake until the 1940s when it was drained for agricultural use and human habitation. Historic information indicates that the Chalco Plain was an area of fresh ground-water discharge prior to the onset of heavy groundwater extraction from the semiconfined aquifer, in contrast with the former Texcoco Lake that was recognized since the Aztecs by its high salinity content. The thickness of the lacustrine aquitard increases from the edges of the former lake to a maximum thickness of about 300 m in the centre of the plain (Figures 2 and 3) Ortega-Guerrero et al. [6]. The lacustrine sediments are interbedded with Quaternary basalts and pyroclastic deposits near Sierra Chichinautzin and in the vicinity of the Santa Catarina Well Field (SCWF). Of particular importance, for this work, is the basalt interbedded with the lacustrine sequence in the middle of the SCWF, which has influenced differential land subsidence and that will be discussed below. A regional granular aquifer underlies the lacustrine aquitard; consisting primarily of alluvial and pyroclastic material and ranging in thickness between 200 and 400 m. Tertiary volcanic bedrock is below the granular aquifer (Figures 2 and 3). Detailed investigations of transient groundwater flow and land subsidence in the Chalco Basin reveal sever exploitation of the regional aquifer system, resulting in progressive lowering of piezometric levels in the aquifer and a fast decline in land surface. Aquifer depressurization and land subsidence began throughout the Chalco Basin prior to the construction of the SCWF in the middle of the plain in 1984. Since pumping started from the SCWF the regional piezometric surface in the aquifer has been drooping at a rate of about 1.5 m/year. The present rate of land subsidence in the middle of the clay plain is near 0.40 m/year, the highest within the Basin of Mexico. Consequently,




accumulation of rain and surface water is occurring in this topographic depression Ortega-Guerro et al. [7].

Figure 2: Location of the Santa Catarina Well Field and hydrogeologic cross-section (Modified from Ortega-Guerrero et al. [6]).

Figure 3: Hydrogeologic cross-section showing the distribution of the lacustrine clayey aquitard and the interbedded basalts (Modified from Ortega-Guerrero et al. [6]).

Land subsidence has caused elevation gradients to change, such a degree, that part of the surface water drainage network is not longer functional and a series of




pumps are used to drain surface water to the nearby Texcoco area (Figure 1). Consequently, more of the land surface is progressively inundated, particularly during the rainy season and flooding has not only taken agricultural land out of production but has also begun to threatened urban areas that are expanding from the mountain slopes and edges of the former lakes onto the plain. In 2000, extraordinary rain events caused the northeastern urban area of Valle de Chalco, to become flooded with a mixture of rainwater and industrial-urban wastewater, affecting the health of people and their property. Numerical simulations of land subsidence based on numerous observation piezometer sites carried out by Ortega-Guerrero et al. [6], suggest that under current pumping rates, total land subsidence in the area of thickest lacustrine sediment will reach 15 m by the year 2010 (Figure 4). If pumping were reduced to the extent that further decline in the potentiometric surface is prevented, total maximum subsidence would be significantly less, about 10 m, and the rate would nearly cease by 2010 (Figure 4). This analysis demonstrates that the new shallow lake conditions will prevail for several decades with enormous environmental effects. How can this hydrologic, landscape, ecologic and environmental situation can be controlled for environmental benefits is the topic of this work.

Figure 4: Predicted transient evolution of the total subsidence in the middle of the Chalco Plain if the rate of drawdown is reduced by, 0% (case 1), 25% (case 2), 50% (case 3), 75% (case 4) and 100% (case 5). (Modified from Ortega-Guerrero et al. [6]).

Of concern of this region therefore is the total magnitude of the land subsidence and the future subsidence rate that is likely to result from the current groundwater extraction system. There are also multiple geo-environment concerns at the Chalco Basin particularly on: the new hydrological and landscape




ecology, environmental planning and management based on sustainable groundwater resources exploitation, flooding hazards, between others. The information published in Ortega-Guerro et al. [7] and Ortega-Guerrero et al. [6] was augmented with new data such as land elevation surveys, evolution of hydraulic response of the aquitard and evolution of lake growing, identification of environmental modifications derived from the process of pumping-land subsidence, to provide elements for environmental planning and management.


3.1 Evolution of land subsidence and lake growth

Based on physical observations of the extension of the new Chalco Lake with time, three different periods of lake evolution are identified: (1) Ponding stage, (2) New lake stage, and (3) Bifurcation of the lake. (1) Ponding stage. In 1987 a series of nests of drive point piezometers where installed in the southern edge of the Chalco Plain to measure hydraulic heads in the upper 20 meters of the lacustrine sequence. In 1988 this site and other isolated sites became flooded as shown in Figure 5. Those sites were located at particular area that had a total drop in land surface elevation of 2 m in the period 1984 to 1989. The network of surface water canals to drain surface water became no functional, and in order to avoid flooding to the nearby urban areas on the Chalco Plain, a system of electric pumps were installed. The Xico ponds are due to the load of the Xico Volcano and existed before operation of the SCWF. (2) New Lake stage. Progressive consolidation of the aquitard due to aquifer pumping in the SCWF caused a total land subsidence of about 7 m in the central part of the plain. In August 1991 the lake extended to an area of about 4 km2 (400 ha) as shown in Figure 5. The roads parallel to the SCWF were progressively elevated to provide access to the wells. (3) Bifurcation of the lake. In 1994 a second lake began to form in the northern part of the Chalco Plain. The bifurcation is associated with the presence of a basalt unit interbedded with the lacustrine sequence in that part of the plain (Figure 3). This figure also shows the position of the basalt unit and the present position of the two lakes. This basaltic layer is causing differential consolidation of the aquitard and potentially will maintain the lakes separated for some period of time until a lateral connection may be reached. The extension of the lakes in 2005 is also shown in Figure 5, where development of two individual lakes is observed, covering a total area of about 6 km2 (600 ha).




Figure 5: Evolution of the shallow lakes. The bifurcation of the lake in 2005 is controlled by a basalt unit interbedded with the lacustrine sequence in the middle of the Santa Catarina Well Field Modified from Ortega-Guerrero et al. [7].

3.2 Environmental situation

In the first stage of the New Chalco Lake formation, land subsidence in the middle of the Chalco Plain caused elevation gradients to change and part of the surface water drainage network was not longer functional and the wastewater from urban areas and industry, without previous treatment, accumulated towards this shallow lake, causing sever problems in water quality with total absence of wild life. However, as land subsidence increased and extraordinary rainwater events flooded the topographic depression, in the 1990s, more favorable water quality conditions developed, permitting the growth of few aquatic native plants and arrival of different species of migratory birds. This unexpected situation may open new possibilities for environmental planning and management in the Basin of Mexico. Water balance at the New Chalco Lake is not well understood, it captures between 7E+06 to 10E+06 m3/yr of rainwater; real evaporation and contributions from sewage are not well known. Groundwater-surface water interaction and the new climatological conditions in addtion to chemical and isotopic characterization of the new lake are the focus of present research.




3.3 Developing a series of connected lakes in the southern end of the Basin of Mexico

There are different well fields in the southern lacustrine area of Mexico City. A well-planned and controlled extraction of groundwater from the underlying alluvial-pyroclastic and basaltic aquifers can, progressively, generate adequate amounts of consolidation of the overlying compressible aquitard, to develop a series of topographic depressions (Figure 6) that can be filled with rainwater, runoff and perhaps some treated water. The wells would continue to provide drinking water to the City while causing progressive consolidation. Groundwater extraction must be designed to progressively reduce hydraulic conductivity at the bottom of the aquitard as demonstrated by Rudolph and Frind [8] and avoid an excess of stress that may cause fracturing of the aquitard as demonstrated by Aguilar-Perez et al. [9]. Fracturing of the aquitard may cause important leaks. The area proposed for lake development is shown in Figure 6. The basic idea is to connect the Xico ponds, with the New Chalco Lake, Mixquic, Tlahuac and Xochimilco. The Mixquic and the Tlahuac areas require further consolidation development to connect them with the other lakes. In the Texcoco former lake, the Nabor Carrillo Lake was developed, in the 1960s, by pumping a 2-3 m thick sandy layer located about 20 m below a clayey aquitard to form a topographic depression, which was filled with water. This method required a large number of shallow wells that produced high salinity content in the water, which localy was used to produce caustic soda Marsal [10]. However cost benefit analysis may be needed to choose the method for consolidating the aquitard.

Figure 6: Proposed development of a series of connected lakes towards the southern end of the Basin of Mexico to improve environmental conditions.




This development of a series of connected lakes by controlled consolidation would progressively improve environmental conditions and ultimately improve social and economic situation of the inhabitants of this region of the Basin through payment of environmental services and ecotourism. Food production using the system of Floating Gardens or “Chinampas” from the Aztec tradition and still use in Xochimilco and Mixquic may represent an additional attraction of this project.

4 Conclusions

The extensive lacustrine Chalco Plain in the southeastern part of Mexico City is underlined by an aquitard up to 300 m thick. Land subsidence in the central part of the Chalco Basin has increased to 0.4 m/yr since 1984 and by 1991 total subsidence had reached 8 m and more than 10 m at present. The rapid land subsidence in this area is causing the accumulation of meteoric waters during the rainy season resulting in extensive flooding of farmland and potentially urban areas. Three different stages of lake growth were identified: (i) ponding (1984-1989), (ii) lake formation (1990-1993), and (iii) bifurcation (1994-present). In this last stage, the local presence of a basalt layer in the middle of the Santa Catarina Well Field is causing differential consolidation and forming a local topographic elevation that permits the development of two separated lakes. This area is still susceptible to the highest potential land subsidence effects as a result of groundwater extraction of anywhere in the basin. Environmental planning and management for the southernmost end of the Basin of Mexico is proposed through the development of a series of connected artificial lakes to control urban growth, to improve environmental conditions of the MAMC, to develop ancient agricultural production systems such as the floating gardens or “Chinampas” and recreational facilities with benefits to social and economical sectors.

References

[1] Bribiesca, C.J.L. Hidrología Histórica del Valle de México [Historic hydrology of the Valley of Mexico]. Ingeniería Hidráulica, Vol. XIV, No 1, México . pp. 107-125, 1960.

[2] Ortega, G.M.A. and Farvolden, R.N. Computer analysis of regional groundwater flow and boundary conditions in the Basin of Mexico. Journal of Hydrology 110, pp. 271-294, 1987

[3] NRC-AIC-ANI. National Research Council, Academia de la Investigación Científica and Academia Nacional de Ingeniería. México City´s water supply. National Academic Press, Washington DC, 1995.

[4] Hiriart, F. and Marsal, R.J. The subsidence of México City. In volumen Nabor Carrillo, Secretaría de Hacienda y Crédito Público. México, Vol II, pp. 109-147, 1969.

[5] Carrillo, N. Influence of Artesian Wells in the Sinking of Mexico City. Comisión Impulsora y Coordinadora de la Investigación Científica,




Anuario 47. In: Volumen Nabor Carrillo, pp. 7-14, Secretaría de Hacienda y Crédito Público, México, 1969.

[6] Ortega-Guerrero, M.A., Rudolph, D.L. and Cherry, J.A. Análisis of long-term land subsidence near México City: Field investigations and predictive modeling. Water Resources Research, 25(11), pp. 3327-3341, 1999.

[7] Ortega-Guerrero, M.A., Cherry, J.A. and Rudolph, D.L. Large-scale aquitard consolidation near Mexico City. Ground Water, 31(5), pp. 707-718, 1993.

[8] Rudolph, D.L. and Frind, E.O. Hydraulic response of hyghly compressible aquitards during consolidation. Water Resources Research, 27(1), pp. 17-30, 1991.

[9] Aguilar-Perez, L.A., Ortega-Guerrero, M.A., Hubp, J.L. and Ortiz, D.C. Análisis nemérico acoplado de los desplazamientos verticales y generación de fracturas por extracción de agua subterránea en las proximidades de la Ciudad de México. Revista Mexicana de Ciencias Geológicas. In Press, 2006.

[10] Marsal, R.J. Development of a lake by pumping-induced consolidation of soft clays. In volumen Nabor Carrillo, Secretaría de Hacienda y Crédito Público. México, Vol II, pp. 229-266, 1969.




Stability of slopes of municipal solid waste landfills with co-disposal of biosolids

M. Chopra, D. Reinhart, M. Vajirkar & B. Koodhathinkal University of Central Florida, Orlando, Florida, USA

Abstract

This paper deals with the impact of the addition of biosolids on the geotechnical properties of class I landfills with Municipal Solid Waste (MSW). The properties are estimated from a field exploration program using the Cone Penetration Test (CPT). The geotechnical shear strength parameters (angle of internal friction and cohesion) of MSW and biosolids mixture are estimated by correlation with the CPT results. These parameters are then used to study the impact of the addition of biosolids on the stability of the slope of the landfill. The slope stability analysis is conducted on various landfill models using the software SLOPE/W. The waste is assumed to act similar to a cohesionless soil. Based on the field explorations, the angle of internal friction was found to be about 29°. Previous research indicated that the most suitable approach to introducing biosolids into the landfill was in the form of trenches. From the slope stability analysis, it was found that the factor of safety reduces significantly with the introduction of biosolids due to a reduction in the shear strength and with increase in the overall moisture content. Keywords: landfills, biosolids, slope stability analysis, cone penetrometer.

1 Introduction

Landfills are well engineered facilities that are used for disposal of Municipal Solid Waste (MSW) and are located, designed, monitored, operated and financed to ensure compliance with federal regulations. Disposal of water treatment facility sludge and wastewater treatment facility biosolids presents significant challenges to facility operators, as this practice often violates loading of metals and exceeds the allowable levels of pathogens for land application. Thus,




doi:10.2495/GEO060221

landfilling of the biosolids is promising as it results in minimal contact with humans and can provide moisture in the landfill for it to act as a bioreactor. With the addition of sludge and biosolids, significant changes in the composition and characteristics of landfill may take place. The stability of waste slopes is particularly important and depends on the strength properties of the MSW with biosolids. The complexity of the problem is further increased by the heterogeneous nature of waste, placement conditions and level of decay of various constituents of the landfills. Proper analysis and design of the landfills have pushed the boundaries of geotechnical engineering practice, in terms of proper identification and assessment of strength and deformation characteristics of waste materials (Sharma et al. [1]; Sadek et al. [2]). Typical laboratory approaches to measure the geotechnical parameters of the waste such as density, moisture content, cohesion, and shear strength have limitations due the heterogonous nature of the MSW. The scarcity of data is due to the difficulties in sampling and testing the refuse. This difficulty is further compounded by the fact that refuse composition and properties are likely to change within a landfill and the waste is likely to decompose with time. An alternate method for estimating landfill geotechnical properties is using in-situ testing devices that provide accuracy with minimal efforts and costs. In the field of geotechnical engineering, a device that meets these criteria for soil testing is the Cone Penetrometer. The unit weight, moisture content, friction angle and cohesion influence the stability of landfill side slopes and interfaces among landfill components. True cohesion between particles is unlikely in landfills but there may be significant cohesion that results from interlocking and overlapping of the landfilled constituents (Singh and Murphy [3]). This paper aims to evaluate the performance of slopes in landfills that may allow the co-disposal of sludges and biosolids, using standard slope stability analyses. Field tests using a Cone Penetrometer are carried out, followed by a slope stability analysis on landfill models using the computer software SLOPE/W (Geo-Slope [4]).

2 Field site description

The field study was conducted at the Highlands County Landfill, located in Highlands County, Sebring, Florida. Two test MSW pads, namely Pilot Area (PA) and Control Area (CA) were constructed at this location. Each test pad was about 120 by 100 feet in plan dimensions at it base and 40 by 60 feet at the top. The height of each test pad was about 10 feet. The compacted waste below the test pads had an average thickness of about 13.2 feet. Liquid biosolids from the City of Sebring Wastewater Treatment Plant were added to the waste in the PA. Two tanker loads of unstabilized biosolids in the form of sludge were transported from the City of Sebring Wastewater Treatment Plant to the site. The liquid sludge had a solids content of 23 g/l or about 3%. The sludge volume was about 1,202 cubic feet which occupied about six inches of depth on an average. The measured decrease in the surface elevation of the impounded sludge due to infiltration of sludge moisture into the underlying and surrounding MSW over 48




hours was about 0.8 inches on an average. Based on the rate of evaporation for the site area for the month of May, approximately 0.45 inches of the measured decrease in sludge level over 48 hours could be attributed to evaporation losses from the surface. The resulting average thickness of the remaining sludge in the entire impoundment was about 4.75 inches. The bulldozer operator had no problem operating the machine over the top of the sludge. The bearing capacity for the wide track, low ground pressure bulldozer (36-inch wide tracks) was excellent with no machine tilting or soft spots were encountered. The three-foot thick layer of MSW was compacted into a layer about 1.5 feet thick, without any observed extrusion of sludge to the surface. No soft spots were encountered and no pumping action was observed as the compactor moved back and forth. Additional MSW was then placed, spread, and compacted on the pad. All four side slopes were constructed at a final grade of about 1:3. The entire surface of the test pad was covered with 24 inches of loosely spread intermediate cover soil. The Control Area (CA) was prepared in a similar manner without the addition of sludge. A total of 28 cone penetration tests were then conducted at these areas, five at CA and twenty three at PA. CPT locations were taken as close to edge of the slope as the truck could be placed.

3 Methodology

3.1 Cone penetration tests

The CPT utilizes electrical transducers rather than analog gauges to obtain a nearly continuous profile of point (tip) resistance, sleeve friction and pore pressure with depth. Specialized data acquisition hardware and software was used to record readings from the transducers at a frequency of approximately five readings per second. Electrical signals reading were then converted to engineering units of stress using device-specific calibration factors. Immediately after the sounding was completed, the borehole was filled with bentonite clay to the top of the intermediate soil cover to prevent water and gas transport from the landfill. The desired depth of soundings in pilot area and control area was about 20 feet from the intermediate soil cover. Ten soundings had to be terminated at depths shallower than the intended depth due to buried obstructions encountered in the path of the cone in the test areas. The data was stored and used in a subsequent statistical data reduction analysis to determine the geotechnical properties for use in subsequent slope stability modeling. A frequency distribution curve of the values of tip resistance was created for each sounding. In order to eliminate outliers in the tip resistance values, the cut off frequency values was selected as 20. Unrealistic tip resistance and pore pressure values are possibly due to the heterogeneous material and obstructions. Based on the field results for tip resistance and sleeve friction, an attempt was made to estimate the location and potential effect of the biosolids layers within the MSW mass. However, this layer could not be identified explicitly, as the




biosolids added had only about 3% solids content; it is likely that the wet material got very well mixed with the MSW. The CPT data provides a repeatable index of the aggregate behavior of the in situ soil in the immediate area of the probe (Robertson [5]). Based on the average values of the MSW parameters, the landfill material is seen to behave similar to “Coarse Grained Sandy-Silty Soil”. Since the material is classified as sand, charts proposed by Robertson and Campanella [6] may be used to estimate the friction angle for the MSW. The resulting friction angle values were adjusted proportionally for variation in density. This is a conservative approach and the actual friction angle (φ) may be somewhat higher. From the generalized equation for friction angle, the average value of friction angle was found to be 29º.

3.2 Slope stability analysis

The behavior of MSW is assumed to be frictional in nature and is governed by the Mohr-Coulomb criteria. The slope stability analysis was conducted using the software SLOPE/W. Daily or intermediate cover was not taken into account in the slope stability analysis. The biosolids or sludges were placed in trenches. The geotechnical properties needed for modeling the slope stability were determined in the field using CPT testing as discussed in the previous section. The landfills were analyzed under circular failure and block failure conditions, and the factors of safety against these modes of failure were evaluated. This paper only presents the results of the study with circular failure. A factor of safety is an index indicating the relative stability of a slope. It does not imply the actual risk level of the slope due the variability of input parameters. With probabilistic analysis, two useful indices are available to quantify the stability or the risk level of a slope. These two indices are known as the probability of failure and the reliability index. The input parameters required for the Mohr-Coulomb Failure Model are unit weight, angle of internal friction and cohesion with associated standard deviation (SD) values for each parameter. Table 1 shows the properties of each material used as input in the model. In Table 1, MSW+BS1, MSW+BS2 and MSW+BS3 refer to layers of different moisture content, which indicate the gradual migration of the biosolids’ moisture into the MSW. This is reflected in the gradation of shear strength parameters with addition of biosolids to the MSW. The shear strength parameters were estimated from the field data obtained using the CPT. For landfills modeled using the circular failure model, the radius of failure plane and the center of the circular failure plane are required to be defined. The radius of the failure plane was specified such that it covers a large range of radii, while the grid for the center of the circular failure plane was chosen so that the minimum factor of safety lies within the grid. Once the landfill model is developed and the grid and radii for the failure plane are specified, slope stability analysis is carried out. A typical result from the program is shown in Figure 1 indicating the failure plane and the minimum factor of safety for the chosen grid.




Table 1: Landfill material properties used in modelling.

Material Description

Unit Weight (pcf)

Cohesion (psf)

Friction Angle (degrees) Source

MSW 70 (SD=5) 0 29 (SD=5) Field CPT

BS 35 (SD=5) 0 0 Ref [7]

MSW + BS1 70 (SD=5) 0 7 (SD=5) Extrapolated Results



Sand Liner 30 (SD=3) 0 40 (SD=5 Ref [8]

Figure 1: Typical slope stability analysis - landfill of slope 1:3 with MSW only – circular failure.

Next, a landfill model was generated incorporating the placement of biosolids in trenches as shown in Figure 2. The width of the trench was about 2-2.5 feet and spacing between the trenches was about fifteen feet. The depth of the trenches was about six feet. Biosolids were added in the trenches. The trenches were then filled with regular MSW and compacted with regular effort. The compaction effort needed to compact the waste layer was reduced based on




discussions with the landfill operators. Slope stability analysis was then conducted on landfill models with side slope varying from 1:2 to 1:4. Although side slope 1:2 is not permitted in Florida state regulation, it is studied as an extreme condition.

Figure 2: Typical landfill profile with biosolids placed in trenches.

4 Results

The results are reported in two parts - results from the field tests and those from the subsequent slope stability study. Field tests conducted consisted of Cone Penetration Test on areas receiving biosolids, areas without biosolids. The average tip resistances range from 60 to 100 tsf, the average sleeve resistance is in the range of 0.7 to 2.4 tsf. These are similar to the range of tip resistance and sleeve resistance values from previous literature on CPT used in landfills. The region of solid waste is characterized by a highly variable set of test results. The CPT results indicated that the cone frequently encountered stiff objects, which produced sharp peaks in the tip resistance measurements resulting in highly variable readings. However, a trend of increasing lower bound tip resistance with depth, was apparent in most the tests. It is important, therefore, to reduce these data by eliminating unrealistic outliers based on cone tip resistance and pore pressure values. From the CPT probes, daily or interim cover soil could not be distinguished from the refuse. Using the method proposed by Robertson and Campanella [6], it is possible to derive the profile of friction angle as a function of depth from the piezocone penetration data for the landfill. The average value of friction angle is 29º. This value of friction angle (φ) is used for slope stability analyses.




Slope stability analysis was conducted using commercially available software SLOPE/W on models of landfills with and without biosolids. Biosolids were placed in trenches and different values of side slope such as 1:2, 1:3 and 1:4 were modeled. Three scenarios are considered.

1. Landfill with MSW only. 2. Landfill with MSW and biosolids which were placed in trenches. 3. Landfill with MSW and biosolids in trenches taken close to the edge of

the slopes.

Table 2 presents the results of the factor of safety (FOS) of landfill slopes for the different conditions that are modeled for the circular failure mode.

Table 2: Factor of safety for landfills with different side slopes.

Side Slope

MSW only MSW and Biosolids MSW and Biosolids with

trenches close to side slope

FOS Reliability Index



1:2 1.94 16.3 1.72 19.0 1.26 7.0 1:3 2.51 19.3 2.17 17.4 1.28 6.4 1:4 2.96 21.8 2.31 3.4 1.75 11.1

5 Discussion

Flattening the slope not only reduces the sum of the driving forces, but also tends to force the failure surface deeper into the ground. The change in length of the failure surface increases in the resisting forces because the shear strength is distributed over a wider area, thereby enhancing stability. Results for FOS suggest that the stability of the landfill slope has been reduced with the addition of biosolids. The factor of safety is clearly reduced if the biosolids trenches are placed close to the edges of the slopes. This scenario allows a weak plane to develop close to the side slope and encourages the failure plane to pass through this weak layer. This situation is considered unstable and needs to be avoided in the field. These results are similar to the previous work by Koodhathinkal [7] who tested MSW with biosolids in a laboratory study. The lower values may be attributed to a lower friction angle for biosolids that are placed as discrete layers and may not be completely mixed in with the MSW. Disposing of biosolids in trenches is a feasible solution from both slope stability point of consideration and ease of field application practice. However, trenches should not be close to the edge side of the landfill as the factor of safety reduces significantly. Berms at the toe of slopes contribute to resisting forces and as a result there is an increase in the factor of safety when considering local failures.




6 Conclusions

Based on the field explorations, the angle of internal friction was found to be about 29°. The most suitable approach to introducing biosolids into the landfill was in the form of trenches. From the slope stability analysis, it was found that the factor of safety reduces significantly with the introduction of biosolids due to a reduction in the shear strength and with increase in the overall moisture content.

References

[1] Sharma, H. D., Dukes, M.T. and Olsen, D.M. (1990). “Field Measurement of Dynamic Moduli and Poisson's Ratios of Refuse and Underlying Soils at a Landfill Site.” Geotechnics of Waste Fills - Theory and Practice ASTM STP 1070.

[2] Sadek, S., Abou-Ibrahim, A., Manasseh, C. and El-Fadel, M. (2001). “Stability of Solid Waste Sea Fills: Shear Strength Determination and Sensitivity Analysis.” International Journal of Environmental Studies 58: 217-234.

[3] Singh, S., and Murphy Bruce (1990). “Evaluation of the Stability of Sanitary Landfills.” Geotechnics of Waste Fills - Theory and Practice.

[4] Geo-Slope (2001). SLOPE/W for Slope Stability Analysis Version 5, Geo-Slope.

[5] Robertson, P. K. (1990). “Soil classification using the Cone Penetration Test.” Canadian Geotechnical Journal (27): 151-158.

[6] Robertson, P.K. and Campanella, R.G. (1983). “Interpretation of Cone Penetration Tests, Part I: Sands,” Canadian Geotechnical Journal, Ottawa, Vol. 20, No. 4, pp. 718-733.

[7] Koodhathinkal, B. R. (2003). Stability of Slopes in a Class I Landfills with Co-Disposal of Sludges and Biosolids. MS Thesis, University Of Central Florida.

[8] Das, B. M. (2000). Principles of Geotechnical Engineering, Brooks/Cole Publishers.




Section 6 Geoecology

Net photosynthetic O2 evolution and calcium precipitation in Chlamydomonas reinhardtii

Y. Y. Wu1, P. P. Li1, B. L. Wang2 & C. Q. Liu2 1Institute of the Agricultural Engineering, Jiangsu University, Zhenjiang, Jiangsu, People’s Republic of China 2The State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, The Chinese Academy of Sciences, Guiyang, People’s Republic of China

Abstract

The relationship between carbonic anhydrase activity, net photosynthetic O2 evolution rate and calcium precipitation in Chlamydomonas reinhardtii was studied. By adding bovine carbonic anhydrase and inhibitor of carbonic anhydrase-acetazolamide to SE medium to alter the activity of external carbonic anhydrase, the variation of net photosynthetic O2 evolution rate and calcium precipitation in Chlamydomonas reinhardtii was determined. Calcium precipitation of Chlamydomonas reinhardtii was determined by the difference of algal calcium content during the culture period. The result shows that net photosynthetic O2 evolution rate and calcium precipitation in Chlamydomonas reinhardtii were enhanced when bovine carbonic anhydrase was added to medium. Acetazolamide can inhibit calcium precipitation and photosynthetic O2 evolution. Carbonic anhydrase may be an important factor to accelerate algal calcium precipitation. Moreover, the geological significance of carbonic anhydrase in algal calcium precipitations and the relationship between algal calcium precipitation and biomineralization of carbonate precipitation are discussed. Keywords: calcium precipitation, carbonic anhydrase, Chlamydomonas reinhardtii, net photosynthetic O2 evolution rate.

1 Introduction

The biogeochemical cycle, which regulates all the biochemical reactions, is the important “driving force” of the geo-environment. The carbon cycle is one of the




doi:10.2495/GEO060231

most important element cycles. CO2 is the important link and ligament of global carbon cycle [1]. The photosynthetic fixation of CO2 and precipitation of CaCO3 are closely connected both spatially and temporally. Photosynthetic carbon assimilation is commonly thought as the cause of calcium carbonate precipitation in algae [2]. External HCO3

- is inorganic carbon resource of calcification in the algal cells [3, 4]. Carbonic anhydrase (Carbonic anhydrase, CA; carbonate hydrolyase, EC 4.2.1.1) is a ubiquitous enzyme catalyzing the reversible conversion of CO2 to bicarbonate. The ability to use HCO3

- for photosynthesis has been associated with the activity of external CA [5, 6]. Hence, there are the putative relationship between CA, the photosynthetic fixation of CO2 and calcification in algae. This work is intended to investigate the effect on calcification and photosynthesis by CA, and the significance of CA in biomineralization.


2.1 Culture growth

Chlamydomonas reinhardti, obtained from the Institute of Hydrobiology, Chinese Academy of Sciences, was grown axenically in continuously stirred artificial freshwater medium SE, which was provided by the Freshwater Algae Culture Collection (http://www.ctcccas.ac.cn). The flasks were closed with bacteriological cotton plugs. Two-day preculture was conducted in medium SE free of calcium (NaCl 20 mgl-1 instead of 20 mgl-1 CaCl2⋅2H2O) before treatment. It was harvested by means of centrifugation(1,700×g for 10 min); The pellet was suspended in medium SE. Growth of culture was initiated by introducing inoculum containing about 108 alga cells. Treatment 1: Cells were cultured in the medium SE containing an inhibitor of extracellular carbonic anhydrase--acetazolamide (AZ) (30 mM), at 25.0°C ±1.0°C, 150 µmol m-2 s-1

light on a 16-h/8-h day/night cycle. The control culture was in the same condition except that AZ was omitted and distilled water was used. Treatment 2: Cells were cultured in the medium SE containing bovine carbonic anhydrase (BCA) (10µg ml-1) (Sigma C2522), at 25.0°C±1.0°C, 150 µmol m-2 s-1 light on a 16-h/8-h day/night cycle. The control culture was the same above. After 6, 12, and 24 hours, the packed cell volume, the algal content of calcium, and the net photosynthetic O2 evolution rate were determined, respectively.

2.2 Determination of packed cell volume

Packed cell volume was determined by centrifugation of 5 mL of cell suspension in hematocrit tubes for 15 min at 2,200×g.

2.3 Assay for the algal content of calcium

100 ml algal solution is filtered through analytical filter paper (Whatman), the filter paper with algal cells is dried at 80oC. The dried sample is ashed at 500oC.




The ash is dissolved in a nitric acid solution (50 ml final volume), and analyzed on an ICP spectrometer for calcium.

2.4 Measurements of photosynthetic O2 evolution

The net photosynthetic O2 evolution rate was measured with a Clark-type oxygen electrode (YSI-5300, USA) under a photon flux density of 150µmol m-2 s-1 and a temperature of 25°C. The effect of acetazolamide on the cells O2 evolution was estimated after 6, 12, and 24 hours’ incubation in medium SE containing the acetazolamide (30 mM). The effect of bovine carbonic anhydrase (BCA) on the cells O2 evolution was estimated after 6, 12,and 24 hours’ incubation in medium SE containing the BCA (10µg ml-1). Cells separated from the medium SE (by centrifugation at 5000 × g for 5 min), were re-suspended in 2 ml of 25 mM HEPES-KOH (pH 8.2) and transferred into an electrode chamber. Before determining the dissolved inorganic carbon-dependent O2 evolution, cells were allowed to photosynthesize to deplete possible intracellular pool of “CO2” until no net O2 evolution was observed. Following the addition of the concentration of 2 mM bicarbonate (final concentration), or the concentration of 2 mM bicarbonate with 30 mM AZ (final concentration), or the concentration of 2 mM bicarbonate containing 10 µg ml-1 the BCA (final concentration), the rate of O2 evolution was measured.

2.5 Statistics

Each treatment consisted of five replicates. The mean and standard deviation are calculated for each treatment. One-way ANOVA and LSD tests are conducted for each group.

3 Results

3.1 The effect on the algal content of calcium

After 6, 12, and 24 hours’ incubation in medium SE containing BCA, the algal contents in calcium of Chlamydomonas reinhardtii are significantly higher than those of the control at the same period. After 24 hours’ incubation in medium SE containing AZ, the algal contents in calcium of Chlamydomonas reinhardtii are significantly lower than those of the control at the same period (Table 1).

Table 1: The algal contents of calcium in Chlamydomonas reinhardtii with time. Values are means ± SD of five replicates.

Time Hours

Control mg ml -1 PCV

+BCA mg ml -1 PCV

+AZ mg ml -1 PCV

6 7.44 ±0.67 10.36±1.12* 6.54±0.63 12 8.72 ±0.76 12.85±1.03* 7.25±0.75 24 10.36 ±0.83 14.78±1.34* 7.72±0.81*

* The mean difference is significant between the culture containing BCA/AZ and the control culture at the same period (P＜0.05).




3.2 The effect on the net photosynthetic O2 evolution rate

After 6, 12 and 24 hours’ incubation in medium SE containing BCA, the net photosynthetic O2 evolution rate is significantly higher than that of the control at the same period (Table 2). After 12 and 24 hours’ incubation in medium SE containing AZ, the net photosynthetic O2 evolution rates are significantly lower than those of the control at the same period (Table 2).

Table 2: The net photosynthetic O2 evolution rate of Chlamydomonas reinhardtii with time. Values are means ± SD of five replicates.

Time Hours

Control nmol O2 s-1 ml -1 PCV

+BCA nmol O2 s-1 ml -1 PCV

+AZ nmol O2 s-1 ml -1 PCV

6 15.45±1.74 26.35±3.25* 12.82±1.73 12 18.56±1.69 24.70±2.33* 10.87±1.52* 24 16.48±1.95 25.26±2.20* 10.25±1.29*

* The mean difference is significant between the culture containing BCA/AZ and the control culture at the same period (P＜0.05).

3.3 The effect on the net photosynthetic O2 evolution rate

From tables 1 and 2, the ratio of the Ca precipitation quantity (CaPQ) to the net photosynthetic O2 evolution rate (Pn) in Chlamydomonas reinhardtii can be calculated (Table 3). During 6-12 hours, the ratio of Pn/CaPQ is smaller than that during 12-24 hours. It shows that the Ca precipitation efficiency during 6-12 hours is greater than that during 12-24 hours.

Table 3: The ratio of the Ca precipitation quantity (CaPQ) to the net photosynthetic O2 evolution rate (Pn) in Chlamydomonas reinhartii.

Time Spaces

Control Pn/ CaPQ

+BCA Pn/ CaPQ

+AZ Pn/ CaPQ

6-12 h 12.5:1 8.6:1 13.2:1 12-24 h 17.4;1 22.7:1 36.9:1

4 Discussion

Calcification is one of the most important biological processes in the living world. Although, it has been proved that the calcification of corals provides CO2 source for its photosynthesis, the role of CA isn’t confirmed yet [7]. This experiment confirmed the relation between the activity of CA and calcium precipitation.

The difference of the algal content of calcium in Chlamydomonas reinhardtii with time reflects the calcium precipitation and calcification. The inhibitor of




extracellular carbonic anhydrase—AZ decreases the calcium precipitation and the net photosynthetic O2 evolution. Extraneous bovine carbonic anhydrase can accelerate the calcium precipitation and the net photosynthetic O2 evolution. i.e. external CA facilitates the calcium precipitation and the net photosynthetic O2 evolution.

CA

Calcification

CaCO3

H2OCO2

HCO3-

PhotosynthesisRespiration

Ca

Water bodies

Plant

Ca

Figure 1: The sketch map regulation of photosynthesis and calcification by CA.

The role of CA on the calcium ion precipitation in cell membrane and carbonate transportation as well as biomineralization of calcium carbonate has a vital significance. The activity of CA expresses the ability to regulate the carbon cycle in the ecosystem. The carbon cycle in the ecosystem is related to the transportation of calcium ion and carbonate. The variation of the ratio of the calcium precipitation to the net photosynthetic O2 evolution is great. i.e. the variation of rate of calcification is very great. It is similar to some other results [8, 9]. From this experiment, we can conclude that the calcium carbonate precipitation reacts rapidly at first, gradually slows down afterwards. It is because no calcium carbonate crystals such as coccoliths can be produced not to be capable of holding more calcium carbonate on the cell surface of Chlamydomonas reinhardtii. The rates of absorption and the precipitation on calcium were gradually decreased, even to zero. The calcium carbonate precipitation reaction is known to produce CO2 according to the following equation: Ca 2++2HCO3

- →CaCO3 +CO2. Therefore, algal calcification exerts a vital effect on calcium transportation and carbon cycle. CA influenced carbon cycle by regulating the reaction: H2O+CO2↔HCO3

-+H+[10]. Algal calcification is also related to the concentration of HCO3

-. Therefore, the relationship between algal calcification, the net photosynthetic O2 evolution and the CA activity can be found in aquatic ecosystem (Figure 1). From Figure 1, Calcium transportation was led by carbon cycle; CO2 is the core of carbon cycle. Some




biomineralization reaction was led by calcium transportation. CA regulates photosynthetic O2 evolution and calcification through the influence on the concentration of inorganic carbon. The regulation of algal photosynthesis and calcification by CA is, therefore, an important factor in the global carbon cycle and biomineralization.

Acknowledgements

This work was supported by grant no. 40273038 of National Natural Science Foundation of China, no. KZCX3-SW-140 of the Knowledge-Innovation Program of Chinese Academy of Sciences, and no.2001-6-3 of Karst Dynamic Laboratory, Ministry of Land and Resources, P.R.C.

References

[1] Schindler, D.W., Carbon cycling: The mysterious missing sink. Nature, 398, pp. 105–107, 1999.

[2] Arp, G., Reimer, A., Reitner, J., Photosynthesis-induced biofilm calcification and calcium concentrations in Phanerozoic Oceans. Science, 292:pp1701-1704, 2001.

[3] Dong, L.F., Nimer, N.A., Okus, E. et al, Dissolved inorganic carbon utilization in relation to calcite production in Emiliania huxleyi. New Phytologist, 123:pp679-684, 1993.

[4] Skies, C.S., Roer, R.D., Wilbur, K.M. Photosynthesis and coccolith formation: inorganic carbon sources and net inorganic reaction of deposition. Limnology and Oceanography, 25:pp248-261, 1980.

[5] Haglund, K., Bjök, M., Ramazanov, Z. et al, Role of carbonic anhydrase in photosynthesis and inorganic-carbon assimilation in the red alga Gracilaria tenuistipitata. Planta, 187: pp275-281, 1992.

[6] Mercado, J.M., Niell, F.X., Figueroa, F.L., Regulation of the mechanism for HCO3

- use by the inorganic carbon level in Porphyra leucosticta Thur. in Le Jolis (Rhodophyta). Planta, 201: pp319-325, 1996.

[7] Marshall, A.T. & Clode, P.L., Effect of increased calcium concentration in sea water on calcification and photosynthesis in the scleractinian coral Galaxea fascicularis. The Journal of Experimental Biology, 205:pp 2107-2113, 2002.

[8] Balch, W.M., Drapeau, D.T., Fritz, J.J., Monsoonal forcing of calcification in the Arabian Sea. Deep-Sea Research II, 47:pp1301-1337, 2000.

[9] Sekino, K., Shiraiwa, Y., Accumulation and utilization of dissolved inorganic carbon by a marine unicellular coccolithophorid Emiliania huxleyi. Plant and Cell Physiology, 35:pp353–361, 1994.

[10] Wu, Y.Y., Li, P.P., Wang, B.L. et al, Significance of carbonic anhydrase and its distribution in the Karst ecosystem. In: Tiezzi, E., Brebbia, C.A., Jørgensen, S.E. and Gomar, D.A. eds. Ecosystems and Sustainable Development V. WIT Press, Southampton, pp.135-141, 2005.




Dynamics of suspended sediment concentration and the impact on a lake-inhabiting bivalve (Corbicula japonica) in the Abashiri River basin, Hokkaido, northern Japan

S. Yanai1, Y. Nishihama2 & R. Tamura3 1Hokkaido Institute of Technology, Japan 2Tokyo University of Agriculture, Japan 3Hokkaido Fish Hatchery, Japan

Abstract

The dynamics of suspended sediment and the impact on a lake-inhabiting bivalve were studied in the Abashiri River basin, in eastern Hokkaido, northern Japan. The Japanese bivalve: Corbicula japonica is the most important fishery resource in Abashiri Lake. Torrential rain over 80 mm per day resulted in considerable sediment concentration during the summer of 2003. Suspended sediment (SS) concentrations >1600 mg/l were measured in the tributaries, where fragile rock formations and faults had developed. The factors influencing SS concentration were analyzed by multiple regression analysis. This led to the identification of soft bedrock and fault densities as significant factors influencing sediment concentration. Fine-sediment particles less than 0.063 mm in diameter accounted for more than 80% of the samples at all stations. Since the lake bivalve prefers coarse substrates, a negative relationship was observed between the percent of fine sediment and the C. japonica distribution and fatness. The substrate around the mouth of the river was also composed of fine sediment, which indicated that the suspended sediment produced from the mountain tributaries has deteriorated the lake environment, especially for C. japonica survival and growth. Keywords: suspended sediment, Abashiri River, substrate, Corbicula japonica and watershed management.




doi:10.2495/GEO060241

1 Introduction

Anthropogenic activities such as urbanization and agriculture alter water infiltration rate and storage capacity and, hence, susceptibility to soil erosion. These effects result in changes in run-off and sediment routing within watersheds [16]. Hokkaido, the most northerly island in Japan, has been subject to increasing human exploitation since the early 20th century. Land use has expanded dramatically, with concomitantly increasing concern about environmental degradation. The fine sediments produced in headwater areas flow downstream, eventually reaching river mouths, which may damage numerous fishery-based resources [11]. Despite increasing public awareness of the linkage between forest and ocean, there have been few studies on the ecological impacts of agriculture and deforestation on aquatic biota [10].

Figure 1: Location of the study watershed.

Here, we focus on a lake watershed (a relatively closed system in comparison with coastal watersheds, in which the terrestrial impact is felt primarily by fish populations). We first measured suspended sediment concentrations in rivers and lakes over the whole watershed, and we discuss determinants of sediment production. Subsequently, we observed the distribution and fatness of a lake-inhabiting bivalve. We consider bivalves to be good indicators of environmental impact, because they are largely sedentary and unable to escape the effects of environmental degradation.

2 Study site

The Abashirti River basin is located in western Hokkaido, has an area of 1380 km2, a channel length of 115 km, and drains into Lake Abashiri (Fig.1). This is a lagoon lake with a surface area of 32.3 km2, a maximum depth of 16.1 m and mean depth of 6.1 m. The fishery resource species in the lake are Corbicula japonica, Hypomesus nipponensis and Salangichthys microdon. The

!

!

!

Tokyo

SapporoAbashiri

0 150 300 450 60075Km

¯




salinity in surface water ranges from 1.0 to 3.5 PSU between May and December. A large portion of the lake is anoxic below a depth of 5 m. The Abashiri River flows into the lake on the south shore, and out of the lake northwards into the Sea of Okhotsk. The lower and middle reaches of the basin are cultivated for potato, wheat and beet, while coniferous forests cover the upper reaches.

The western portion of the watershed has fragile sedimentary rock, whereas volcanic pyroclastic flow deposits are found in eastern regions, forming a plateau along the middle reach of the river. Numerous landslide scars have been observed in eastern parts of the watershed.

The annual mean precipitation is around 800 mm, with snow cover from November to April. Snow melt usually occurs from early April to mid May. Heavy rainfall sometimes occurs during summers, and daily precipitation may exceed 100 mm.

Figure 2: Water sample collection points.

3 Method

3.1 Sample collection

Seven water collection stations were established along the main river channel (St.m1~m7), and 11 stations were established at the mouths of tributaries (St. t1~t11) (Fig.2). Samples were collected by the Abashiri Watershed Environment Protection Board during downpours in spring, summer and autumn from 2003 to 2005. Bivalves were collected with scooping net and bottom sediments were

0 4 8 12 162Km

¯ m1

m2

Abashiri L

m4

m6

m7

m5

t1

m3

t2

t3

t4

t5t6

t7t8 t9

t10

t11

TributaryMain channel

0 4 8 12 162Km

¯ m1

m2

Abashiri L

m4

m6

m7

m5

t1

m3

t2

t3

t4

t5t6

t7t8 t9

t10

t11

TributaryMain channelTributaryMain channel




collected with a Smith-McIntyre mud sampler (0.05 m2) at depths of 3.0 at 50 stations along the lake margins.

3.2 Sample analysis

Suspended sediment concentration in water samples was determined by percolation through glass fiber filters (0.5 µm mesh). The percolated samples were dried at 50oC for 48 h and weighed, then combusted at 550oC in a muffle furnace for determination of organic dry weight (percent of ash free dry mass). Sizes of particles included in water samples were analyzed using two methods:

1. particles larger than 0.063 mm were separated out in sieves with different mesh sizes, then dried at 60oC for 24h;

2. smaller particles than 0.063 mm were analyzed with a particle analyzer (CAPA-300 Horiba).

Bottom sediments were treated in the same way as suspended sediments. The C. japonica specimens collected in the sampler were counted, and the density per square meter was calculated. The width, height and length of each shell were measured with vernier calipers (Mitutoya, CD-20GM). Each soft tissue was dried and weighed. The condition of the animals was calculated as a dry weight per volume index (x 1000).

3.3 Spatial analysis by Geographical Information Systems

Land use on each tributary was measured using digital land use maps supplied by the national Ministry of Land Infrastructure and Transportation. Environmental factors, such as forest coverage, agricultural land, lithological hardness, slope gradient, and fault densities were also measured from these maps using GIS software (TNTmips ver.6.8, ESRI Arc View 3.2a and Spatial analyst).

4 Results

4.1 Suspended sediment concentrations in spring and summer

During the study period, we collected water samples in August 2003, April 2004 and April 2005. There were no large discharges in the summer of 2004. The highest concentrations of suspended sediment were observed in August 2003. Figure 2 shows the discharge during downpours in that month. This discharge was induced by torrential rainfall exceeding 79 mm on August 9-10. Sampling was carried out during peak, regression and normal discharges (Fig.3). Suspended sediment concentrations at all stations along the main channel, except at St. m1 and m7, were higher than 1600 mg/l, and the tributaries draining from the left bank transported sediment at even higher concentrations. These values declined remarkably with time to <50 mg/l at all sampling stations after ten days (Fig.4). Most of the sediment was inorganic [>80% except at St.m1 (uppermost headwater) and St.m7 (downstream from the lake)].




Figure 3: Discharge at August 2003.

Figure 4: SS concentrations observed in Aug. 2003.

4.2 Particle analysis

Particle distribution in the suspended sediments did not differ significantly between the main channel and tributaries. Finer particles, less than 10 µm, comprised 40% of all samples, followed in abundance by medium sized particles between10 µm and 0.063 mm. Coarser particles usually comprised <10% of all samples.

4.3 Bottom sediment and C. japonica in the lake

Bottom sediments were collected at three depths, at 50 stations around the lake. Fig. 6 shows the percentage of fine particles <0.063 mm at each sampling station. Fine particles were more abundant on the southern shore where the river flowed in, and on the northern shore where the river flowed out. The percent of finer sediments tended to increase with water depth.

10th Aug. SS 11th Aug. SS 19th Aug. SS10th Aug. SS10th Aug. SS 11th Aug. SS11th Aug. SS 19th Aug. SS19th Aug. SS

L. Abashiri L. Abashiri L. Abashiri

0

0.5

1

1.5

2

2.5

3

3.5

4

8/1

8/3

8/5

8/7

8/9

8/11

8/13

8/15

8/17

8/1

9

8/2

1

8/2

3

8/2

5

8/2

7

8/29

8/31

Wat

rtab

le (m

)

Date (2003)

Peak dishcarge

One day after9 days after

0

0.5

1

1.5

2

2.5

3

3.5

4

8/1

8/3

8/5

8/7

8/9

8/11

8/13

8/15

8/17

8/1

9

8/2

1

8/2

3

8/2

5

8/2

7

8/29

8/31

Wat

rtab

le (m

)

Date (2003)

Peak dishcarge

One day after9 days after




Figure 5: Particle distribution in suspended sediment at the all stations.

Figure 6: Percent of silty substrate (<0.063 mm) (left), density (N/m2)(center) and fatness index of C. japonica (right) in the Abashiri Lake.

C. japonica specimens were collected at the same sampling stations. Densities and fatness were higher on the western and eastern shores than on the northern and southern shores (Fig.6). There was a negative relationship between bivalve abundance and the percentage of finer particles. We also found a negative relationship between the fatness of C. japonica and the percent of fine particles.

4.4 Land-use and other factors affecting sediments discharge revealed by GIS analysis

Factors affecting sediment yield were calculated for the eleven tributaries (Table 1). The watershed area was 22.3 to 202.3 km2. The agricultural area

0%

20%

40%

60%

80%

100%St

.m2

St.m

3

St.m

4

St.m

5

St.m

6

St.t1

St.t2

St.t3

St.t4

St.t5

St.t6

St.t7

St.t8

St.t9

St.t1

0

St.t1

1

<10μm 10μm～0.063mm 0.063mm>

Main channel Tributary

0%

20%

40%

60%

80%

100%St

.m2

St.m

3

St.m

4

St.m

5

St.m

6

St.t1

St.t2

St.t3

St.t4

St.t5

St.t6

St.t7

St.t8

St.t9

St.t1

0

St.t1

1

<10μm 10μm～0.063mm 0.063mm>

Main channel Tributary

-0 2 41Kilometers

Fatness

-15.0

-16.5

- 18.0

- 19.5

- 21.0

Density(/m2)

- 400

- 1000

- 1800

- 2600

- 4000

Silt(%)

- 4

- 14

- 35

- 68

- 92

Km

Fatness indexDensity(N/m2)Silt(%)

River flow in

River flow out




ranged from 0 to 56.5% of each watershed, while the forested land occupied between 36.5 and 100%. The percentage of slopes steeper than 20 degrees was smaller for downstream tributaries than for upstream tributaries. Soft rocks dominated western parts of the watershed. Higher densities of faults developed in these tributaries.

The relationships between suspended solid concentration and percentage cover of agricultural land were unclear; the larger agricultural tributaries had lower sediment concentrations in August 2003, whereas forested tributaries, such as t9~t11, had the highest suspended solid concentrations. Slope steepness was not correlated with the particle concentration. The t5 tributary had the steeper slope, but not the highest suspended solid concentrations. Presence of soft rock strongly influenced the suspended solid concentration (tributary t10 showed the highest concentration). Multiple regression analysis was performed to determine proportional effects of environmental factors on variations in suspended solid concentration. Lithological hardness and the densities of faults (medium) were positively correlated with suspended solid concentrations (Table 2).

Table 1: Environmental variables in each tributary.

5 Discussion

Land use practice usually strongly influences sediment yield (suspended sediment concentration correlates positively with the proportion of agricultural land, and negatively with the proportion of forested land [15]). Our previous study showed that sediment concentrations exceeding 2000 mg/l were produced in highly agricultural watersheds; the sediment sources were landslide scars resulting from soil compaction by agricultural machinery [11, 13]. However, in the present study, we observed lower suspended yield in larger agricultural tributaries. Instead, suspended sediment concentration was strongly correlated with lithological conditions (Table 2). The western tributaries (where a large portion of watershed is on soft rock) are assumed to be the major source of suspended sediment for the main channels (Fig.4). Furthermore, higher suspended solid discharge may affect sedimentation in Lake Abashiri (Fig.6).

Area

(km 2) Agricultural land (%)

Forestcoverage (%) Soft rock (%)Higher density

of faults (%)Steep slope(>20°)(%)

t1 126.3 56.53 36.59 10.26 0 0.08 265.15 t2 39.7 35.57 59.46 12.39 0 3.7 377.2 t3 202.3 19.75 73.21 0 0 12.04 258.5 t4 54.6 9.7 89.49 11.59 11.63 48.63 335.05 t5 87.9 6.57 92.6 53.04 19.73 82.08 666.25 t6 152.2 8.25 86.81 14.87 0 59.49 208.05 t7 75.7 13.49 85.49 58.83 0 56.39 741.4 t8 52.5 5.79 91.02 55.07 35.37 82.35 405.25 t9 108 9.77 89.32 40.57 12.11 59.3 1499.4 t10 30 0.14 99.68 62.95 15.16 76 1872 t11 22.3 0 100 53.67 0 54.85 704.6

Tributary Environmental variable(%) Highest SS

concentration (mg/l)




The lithological characteristics of the watershed are such that land use must be carefully managed, even for forested regions.

In these mountains, numerous landslide scars were observed along channels (Sato unpublished data). These scars resulted from lateral scouring of stream banks during large water discharges; soil wash out can be a serious consequence of this. Other sources of stream sediments may be timber harvesting and logging road construction. The higher densities of logging roads were near St.t9 and t10 tributaries. Roads can affect streams directly by accelerating erosion and sediment loading, by altering channel morphology, and by changing the runoff characteristic of the watershed [5]. In addition, cut branches and trunks are piled along streams, and this debris is easily swept away during downpours. In order to protect land cover, precautionary forest practices should be followed. For instance, although there is no regulation on the cutting of riparian forest, a buffer strip of more than 20 m wide should be protected. Logging roads should avoid streams and inner valley gorges. The number of stream crossings should be minimized [17]. The exposed soil of cuts and fills should be planted with grass or rapidly growing shrubs. One effective method for decreasing surface water runoff is the installation of ditches and cross-drain culverts along logging roads [9]. Sediment ponds at the end of small catchments may be effective in trapping and reducing suspended matter downstream.

Table 2: Results of multiple regression analysis.

Suspended sediments have strong effects on aquatic biota [10, 17]. Inorganic

turbidity presents a wide range of problems, particularly for filter feeding organisms. The harmful effects on suspension-feeding bivalves include gill tissue fouling, increased pseudofaecal production [18], decline of ingestion rate resulted from reduction of clearance rate [3], depressed growth rates through overloading the gut and gills with inorganic solids [4, 12], reduced metabolic rates and feeding efficiency [1]. Respiration rates are significantly affected by acclimation temperature and turbidity [2, 14]). Based on these studies, it was likely that fine sediment derived from upstream tributaries had a significant influence on the lake inhabiting bivalve C. japonica. The higher suspended solid concentration (>1600 mg/l) occurred at St. m2, which was the nearest station upstream from the lake. Concentration declined drastically at St.m1, just downstream from the lake (Fig.4). Hence, most of sediment transported from upstream probably accumulates around the river mouth on the southern shore,

Un standerdrized coefficient

BConstant 26763.337 0.0573Agricultural land (%) -279.891 0.0621Forest coverage (%) -266.837 0.0522Steeper slope <20°(%) -19.366 0.422Soft rock (%) 33.933 0.0198Fault density (medium) (%) 95.412 0.0475Fault density (higher) (%) -103.864 0.0315

R2ProbabilityVariable

0.801(P=0.132)




and the lower density and lower growth of C. japonica in that area may be attributable to material produced upstream.

Salinity is a factor controlling distribution of the C. japonica [7]. In this study, high salinity considered to be one of negative factor to influence the density and fatness of C. japonica on the north shore (Fig.6) where seawater flows in from the Sea of Okhotsk during high tides. Because the frequency of open shells in C. japonica is generally lower in higher salinity conditions

Corbicula sp. contributes significantly to total benthic community respiration [6]. In particular, C. japonica uses organic material derived from terrestrial sources [8]. However, the proportion of organic matter was <15% in the sediments deposited around the river mouth, and it is likely that the negative effects of fine inorganic sediments outweigh the positive effects of any organic food source in the sediment. Management of upstream forests to decrease fine inorganic sediments and increase organic content may be important for conservation of fishery resources in the lake represented by C. japonica.

Although concerns for river management have increased recently in this region, a watershed network (a partnership of local residence involving various stakeholders such as fisherman, farmers, and foresters aimed at preserving water quality and regulating land use practice) has not been established in the province of Hokkaido. More detailed data on the influence of land use practice on fish resources will be required for better watershed management.

Acknowledgements

This study was performed with the cooperation of Marine science center of Abashiri, and Abashiri fisheries cooperation. We would also like to be grateful to Mr Kamada T., Kikuchi T.and Moriwaka H., who were students of Hokkaido Institute of Technology and assisted a field investigation and laboratory analysis.

References

[1] Aldridge D. W., Payne B. S. & Miller A. C., The effects of intermittent exposure to suspended solids and turbulence on three species of freshwater mussels, Environ. Pollut. 45, pp. 17-28, 1987.

[2] Alxander J E., Thorp J. H. Jr., and Fell R. D., Turbidity and temperature effects on oxygen consumption in the zebra mussel (Dreissena polymorpha), Can. J. Fish. Aquat. Sci. 51, pp. 179-184, 1994.

[3] Bricelj, V. M. & Malouf R. E., Influence of algal and suspended sediment concentrations on the feeding physiology of hard clam Mercenaria mercenaria. Mar. Biol. 84, pp. 155-165, 1984.

[4] Bricelj, V. M., Malouf R. E., & de Quillfeldt C., Growth of juvenile Mercenaria mercenairia and the effect of resuspeded bottom sediments, Marl. Biol., 84, pp. 167-173, 1984.

[5] Furniss M. J., Roelofs T. D., & Yee C. S., Road construction and maintenance. Influences of forest and rangeland management on salmonid fishes and their habitats, ed. in W. R. Meehan, pp. 292-323, 1991.




[6] Hakenkamp C. C. & Palmer M. A., Introduced bivalves in firewater ecosystems: the impact of Corbicula on organic matter dynamics in a sandy stream, Oecologia, 119, pp. 445-451, 1999.

[7] Ishida O., & Ishii T., Salinity tolerance and regional morphological differences for Corbicula japonica. Fishery culture 19(4), pp.167-182, 1971.

[8] Kasai, A. & Nakata A., Utilization of terrestrial organic matter by the bivalve Corbicula japonica estimated from stable isotope analysis. Fisheries Science, 71, 151-158, 2005.

[9] Moore G. D., Resource road rehabilitation handbook: Planning and implementation guidelines. Watershed Restoration Technical Circular no. 3. Ministry of environment, lands and Parks and Ministry of Forest. Province of British Columbia, Canada, pp.99, 1994.

[10] Nagasaka, A., Nakajima, M., Yanai, S. & Nagasaka, Y., Influences of substrate composition on stream habitat and macroinvertebrate communities: a comparative experiment in a forested and an agricultural catchment. Ecology and Civil Engineering, 3, pp. 243-254, 2000 (in Japanese with English abstract).

[11] Nagasaka A., Yanai S., Sato H. & Hasegawa S., Soil erosion and gully growth associated with cultivation in south-western Hokkaido, Japan. Ecological Engineering 24, pp. 503-508, 2005.

[12] Payne B. S., Lei J., Miller A. C. & Hubertz E. D., Adaptive variation in palp and gill sizen of zebra mussel (Dreissena polymorpha) and Asian clam (Corbicula fluminea). Can. J. Fish. Aquat. Sci. 52, pp. 1130-1134, 1995.

[13] Sato Hi., Yanai S., Nagasaka Y., Nagasaka A. & Sato Ha., Influences of land-use on suspended sediment discharge from watersheds emptying into Funka-Bay, south-western Hokkaido, Northern Japan, J. Japan Soc. Hydrol. & Water Resour., 15, pp. 117-127, 2002 (in Japanese with English abstract).

[14] Summers R. B., Thorp J H., Alexander J. E. & Fell R. D., Respiratory adjustment of dreissenid mussels (Dreinnena polymorpha and Dereissena bugensis) in response to chronic turbidity, Can. J. Fish. Aquat. Sci., 53, pp. 1626-1631, 1996.

[15] Walling D. E. & Webb B. W., Water quality 1. Physical characteristics. River Flow and Channel Forms, ed. G. Petts & P. Calow, Blackwell Science Publishers, pp. 77-101, 1995.

[16] Walling, D. E., Suspended sediment yields in a changing environment. Changing River Channels, ed. A. Gurnell & G., Petts, John Wiley & Sons Ltd., 1995.

[17] Waters T. F., Sediment in streams, sources, biological effects and control, American Fisheries Society Monoghraph 7, pp.251, American Fisheries Society, Bethesda, Maryland, 1995.

[18] Way C. M., Hombach D. J., Miller-way C. A., Payne B. S. & Miller A. C., Dynamics of filter feeding in Corbicula fluminea (Bivalvia: Corbiculidae). Can. J. Zool. 68, pp. 115-120, 1990.




Potential of poplar plantation for enhancing Polish farm sustainability

A. Czarnecki & A. Lewandowska-Czarnecka Institute of Ecology and Environmental Protection, Nicholas Copernicus University, Torun, Poland

Abstract

Farm sustainability depends on policy that is able to protect natural resources while being open to the market. Traditional crop pressing on natural processes creates an open input-output unstable system. Farming situated in a postglacial landscape cannot counteract erosion and does not engage in all natural resources. The poplar stand has an ability to renew the soil system and assimilate the surplus of mobile substances. This paper presents a proposal to combine poplar with a traditional crop to make the biosystem able to engage in all resources to develop efficient production while making the system more stable. The concept is related to the results of the investigation carried out on farms dealing with typical problems and poplar stands in a similar soil condition. Keywords: farming, poplar, crop, sustainability and policy.

1 Introduction

Until the 1980’s Polish agricultural production was controlled by the state, whose aim was to produce maximum crops, the results of which strongly impact on the environment, compromised natural resources and the functions of the ecosystem on arable land. Poland’s past agriculture intensity reduced soil structure [1] reaped the landscape of vital nutritional elements. Natural habitats, hedgerows, the majority of wetlands and small surface water basins have all but vanished [2], those which remain are (predominantly) highly polluted due to neighboring agro-ecosystems. The 1990s saw the external conditions for farming totally change. Production met the limited market, resulting in decreased prices for primary agricultural products; forcing farms to begin careful management of natural resources, and an initially expensive investment into farming methods, to




doi:10.2495/GEO060251

increase productivity. Land, which is more productive, is pushed to excessively productive, whilst less productive land has been abandoned. The government is proposing to re-forest, the so far, 2 million ha of abandoned arable land. For national and international interests the reforestoration is a positive action, however, the tree type requires careful consideration to achieve maximum economic benefits for farmers of this unproductive arable land. Poplar has this potential and timber as an initial crop. It grows fast, and reaches maturity within 25 years. Poplar requirements would combine not available for traditional crop resources [3]. Timber and biomass can revive local economy on the basis of soil productivity. Considering continuity of farm business the surplus of land and other environmental and social resources ought to be engaged in agricultural activity giving not only direct income but also combining into a process of attaining sustainability [4]. Thus, poplar has potential as an extensive crop. The aim of this work is to establish the total expected value from poplar cultivation on a traditional farm in view of a policy driver for long-term sustainability.

2 Sustainability and agriculture

The farm manages ecosystem products and services. Basic human needs that are expected to be provided by the farm are foodstuffs. However, economic viability requires farmers to attentively manage ecosystems and social services provided by the farm. Economic viability means reduction in manpower and implementation of agricultural machinery, but it also means diversification and looking for new markets. Biomass and timber are becoming important crops. Presence of poplars can improve landscape for adaptation and inclusion of land for leisure pursuits. Farmers should engage all resources proportional to the markets, and create new jobs, though economic gain must be balanced with natural processes of a cyclic renewal of resources, which dominate the location. The position of farmers within local community means that they are responsible for the state of resources [5] and the environment over a larger area than just the farm, as farming processes interact with ecosystems [6]. Therefore, sustainable management methods must respect natural resources, while retaining land productivity. By using renewable resources produced under proposed poplar plantation systems, such as biomass, farmers reduce productivity costs and enhance efficiency. Long term changes in agricultural methods require governmental and also farm policies to transform them into operational plans.

3 Methodology

Knowledge based decisions on farm tends to be sustainable and flexible, fully combining resources and farm productivity with respect for ecosystems. When considering the farm as a socio-economic unit, its state is important. An analysis requires a collection of several data sets i.e. economic, socio-economic, and ecological. This will produce a basis for establishing gaps between the actual and sustainable farm status, highlighting obstacles preventing actualization of sustainable methods. Problems encountered by farms in Polish lowlands were




used for this paper. The data were collected qualitatively over a 10-year period by the authors. Each problem the study raised is shown in fig. 1.

The diagram (fig.1) enables farm needs to be quantified, and the potential contribution to answer specific farm needs to before filled. Hydrological processes and interaction between soil and trees productivity in time, the retention of biodiversity and active ecosystems need to be considered when establishing the cost both economically and in a sustainable way. The case studies analyzed have enabled quantification of these problems, and the cost of traditional farming methods vs. poplar plantations. The results of the farming combining tradition and poplar crops will be discussed. The data are transformed into economic–efficiency of input, socio-economic-employment, and ecological-N circulation categories.

4 Site locations

All the farms are situated in postglacial terrain with differing relief within a moraine complex and a river valley bottom on the moraine plateau, in a depressing part of the ground within moraine, on the edge of a moraine and a valley as well as in the river valley. All farms need reduced overheads but increased efficient production by utilizing farmed derived biomass, to increase employment opportunities for families and the local community. Farm 5 is located in fragile areas; where surface erosion and loss of organic matter impair




Conceptual diagram presenting the approach to farm analysis.

Figure 1:

production predominantly. Farms 1 and 3 situated in the river valley are affected during periods of flood. Intense cattle farming in farm 2 and 4, within the river catchments caused nitrogen contamination. The characteristics of each farm are presented in table 2.

Table 1: Functional characteristics of farms.

Specifications Farm 1 Farm 2 Farm 3 Farm 4 Farm 5

farm areas (ha) 49,75 18,24 70,12 24,88 42,00

in that: forest 0,00 0,00 0,00 0,00 5,00

meadows and pasture

28,06 0,00 32,12 10,28 4,00

set aside 2,94 0,61 0,00 1,10 0,00 arable fields 18,75 17,63 38,00 13,00 32,50

Main crops Fb, M, B, O, W, T, P

T, P, B, M, Z, B

Sb, P, J, Po

T, B, P, Z, M R, Po, Mz, Z, Bp

Livestock Dc 25 0 6 14 4

C

14

4

0

0

0

Bc 4 4 0 14 10 P 5 80 0 100 150

Al IV-VI, Al IV-VI Al II- IV Al III-IV, soil bonitation class P IV-VI, M IV M V-VI M V, VI

Al IV-VI P

N - mineral a fertilizers b

16,28 43,20

49,12 50,82

130,49 240,79

73,83 138,46

95,34 107,08

N total input b yield contents

14,46 80,16 52,92 48,76 87,49

Energy input: (GJ/ha/year)

18,89 35,78 33,94 30,3 28,46

in that: renewal 8,38 13,3 8,99 9,39 10,75

Efficiency: Money (PZL/PZL)

1,41 3,54 2,56 5,79 1,95

cropping livestock Al – arable lands W – wheat P – pigs M – meadows T – triticale Dc – dairy cow P – pasture B – barley Bc – beef cattle bonitation class I –VI P – potatoes Sc – slaughter cattle a estimate on farm areas Fb – fodder beet C – calves b estimate on crop O – oats P – pigs R – rye M – maize

Poplar plantations situated on land exhibiting similar characteristics to that of the five farms of the study were analyzed. These comparative sites consisted of 21 poplar plantations on medium to poor quality of soil. The plantations were




established by state forest services on abandoned arable land and meadow (tab. 2). Research into poplar sites was conducted under project PAMUCEAF (FAIR6-CT98 4193).

Table 2: Poplar stands characteristics.

Characteristics

Description Comments Comparing with arable land

Suitable site, mineral soil, above 7% average clay content coarse-textured

Underlying coarse-textured subsoils

Possible to form poplar plantations

Low agriculture value

Very suitable Medium-texture

Underlying medium-textured subsoils

Deep subsoil Ground water supply

Also good for traditional crop Not access to ground water

Changes in soil N C

22,14 g/kg 1,82 g/kg

Litter layer is forming, Two times higher comparing with arable land

Bulk density soil subsoil

1,6 mg/m3 1,4 mg/m3

Soil is loosened by roots and macrofauna

Lower then arable soils lower compaction

AWC (available water capacity)

O,06 m3/m3 Lower bulk density and larger organic matter retention

Slightly higher then arable soils approximately 0,01-0,02 m3/m4

Contents in soil P K

0,03 g/kg 0,1 g/kg

Soil is clean up Soil matrix contents

In poplar stand three times less The same

Production per year/ha 10-20 m3 Depend on soil, available water

Similar productivity

N immobile per year To 200-300 kg Soil development, biomass enlarging

Accumulation and storing

5 Farmer’s needs vs. poplar contributions

Ecosystems confine the threshold of farm productivity within a sustainable system. This needs to be considered when discussing a farming decision for an economic output. This is represented by a group of variables: the state of the farm, farmer’s requirements, socio-economic and environmental factors. Poplar plantations are characterized by two main factors: natural species composition and long duration without human intervention allowing ecosystem development. From tables 2 and 3 it can be seen that poplar planted farms have more positive net gains [7] when compared to similar arable lands. The ability to store nutritional substances and water was higher due to soil developing an accumulation of organic matter and soil loosening volume. The poplars enabled soil to renew its profiles by accumulation and assimilation properties, and also yielding timber and biomass, all positive outcomes of the poplar system (tab.3).





6.1 Quantifying problems on the investigated farm

On each farm one of the key problems was erosion and surplus of resources. They were estimated quantitatively as losses and resources not fully engaged in production. Other categories assessed were: reduction of annual income, decrease in soil productivity (as natural capital), and an ability to cover discounted rates of machinery and equipment. Some factors were considered external factors; they included the effects on the local community and the environment. Surplus nitrogen combined with eroded material deposited in the valley bottom causes pollution and trophication. It can limit farm environment and indirectly local development. The loss of nitrogen, due to geographic location of the farm can be up to 200 kg N per ha/year (farm 5). All farms within the study do not create adequate conditions for crops, hence they reduce yield by approximately 40%. Farm 1 has very extensive production. Farm 3 with meadow situated in a high-risk flood zone causes the farmer to limit productivity (see table 1).

Table 3: Potential contribute of poplar to farm sustainability.

Problem/farmer needs Criteria Expectations from new crops

Decreasing loss Application for free resources Forming new resources/ Prospect for future Requirements Poplar stands

Better efficiency of input Fertility conservation Decreasing leaching of biogens Loss of productivity caused by erosion Biomass Fragile land application

Assimilation of N overdose

Application for surplus of human and man-made resources

Better efficiency of applied resources

Cheaper input Enhancing income Social acceptation Capital Labor

Fixed plant cover on fragile land Extensive crop Mixed crop of different requirements Biomass from extensive long term crop Assimilation of organic fertilizers New products/resources carpentry, tourism, leisure

Input decreasing

Differentiation of activity Improving landscape

Crop for abandoned land Replace present crop Silvopasture system




Resources, which are not applied, reduce total farm productivity by 40-60%. One of these is a sub-terrestrial hydrological system, which runs water from the moraine to the valley floor, below the level of accessibility for traditional crop roots. For all farms mechanical and human resources need to be applied more effectively. The two farms (2 and 4), which conduct a secondary productivity results in excess nitrogen being introduced into the valley catchments, externalize effect of farm management.

Table 4: Losses in natural capital and not applied resources in investigated farms.

factors criteria characteristics farm

1 2 3 4 5

Soil 3 2 3 1 3

Losses erosion Input 2 2 2

input Water 3 3 1 1

Yield 3 2 1 1

insensitivity Landscape 1 3 2 1

Biogens Surplus of N 3 3 3

Resources Productivity Meadows 3 3 1

not Land Field 3 3 2 1

applicable Labor Human resources 2 3 2 2

Man-made Machinery 3 1 3 1 2

capital 1 little 2 medium 3 substantial.

Table 5: Capacity of poplar to solve farm problems.

Process Farm 1 Farm 2 Farm 3 Farm 4 Farm 5 Apply all forms land Low input Flexibility Development Complement production Biomass Secondary production Raw materials/workshops Leaching and erosion Assimilation Extensive permanent crop Landscape diversifying Naturalness

3 3 3 3 3 2 3 1 2 3 1 1

1 2 1 1 3 2 1 3 1 1 3 2

3 2 3 3 3 2 2 3 1 2 3 2 2

1 2 1 1 2 1 1 3 1 1 3 1

2 3 2 3 2 2 3 2 2 3 2 2 1

Effect vs. problem 1 -3 little, medium, substantial.




6.2 Expected poplar contribution to farms

Farmers cannot afford to allow for further decrease in ecological processes and ecological efficiency. However, the farmer requires mutual gain from protecting these ecological systems, by means of free natural resources to improve production. The solution for many of these environmental and ecological problems is deep-rooted plant cover, which would stop and assimilate nutritional substances. Poplar plantations with solve main problems so far raised but in doing it with different efficiencies. Poplar crops yield raw material, which has a possibility to be processed on the farm; trees as an extensive crop improve soil development by creating better storage capacities for nitrogen and water. Poplars causing soil to become looser, allowing it to be used as strips among fields or part of crop rotation. It has been suggested compaction by heavy machinery causes a dead pan that vanished under poplars. In soil under poplar, phosphorus (P) quantities fall, likely to be caused by P’s immobility and assimilation into biomass. Farms proposing such a dynamic system need to consider markets, the environment, and ecosystem processes [8]. Management of these systems and socio-economic systems require specific knowledge and skills to maximize potential outputs. The movement and processing of lumber requires more equipment, machinery, and man-power, a consideration within cost effectiveness of poplar plantations, though the reduction in farm overheads by the use of biomass, may go someway to meeting these costs. Poplar plantations are planned for each farm not in view of above requirements but only due to the farm size, the problem experienced by the farm, both economically and soil degradation, and the proportion of resources free to be engaged within the project. In some cases a part of traditional cropping would be converted into poplars, giving benefits of increased income or meeting local community, or national legislation. Knowing what poplar stand would change in it were estimated proposal in term of stand characteristics considering free resources on farm. In relation to the main problem on particular farm and free resources they dispose, poplar stands differ substantially. Thus contribution to solve problems ranges between 20-80%. The biggest potential contribution of poplar is on farm 2 and 5 no least in farm 2 and 4. That’s mean that consequences to farmer needs differ between farms.

7 Conclusion

Comparison between the factors that arise from transformed external conditions for food production and the state of farms enabled to establish threats and gaps in the state that prevents adaptation process. Results of the investigations proved that poplar plantation would positively impact the state and money flow on farms if only the whole farm is designed according to natural processes as surface, interflow and base flow as well nutrients of water flow as well by fitting poplar stands to overall money flow in farm. Farms situated in postglacial landscape in Poland would use poplar as a crop to facilitate achieve the balance their performance with nature and external conditioning for farming. In dependence of




main problem and quantity of free resources on farm, poplar stand would contribute less or more to development of farm business as well as internalization of matter turnover to satisfactory degree.

References

[1] Dziadowiec H., A. Czarnecki, J. Jonczak. Roczna i wieloletnia dynamika wybranych właściwości uprawnych gleb płowych Glebowej Powierzchni Testowej ZMŚP w Koniczynce. X Ogólnopolskie Sympozjum Zintegrowanego Monitoringu Środowiska Przyrodniczego. Kampinoski Park.

[2] Jaworowski P., Cz. Sobków, A. Czarnecki, T. Celmer, J. Szablowski. Melioracje wodne, ich wpływ na środowisko przyrodnicze i gospodarkę rolną. wyd. UMK, p. 210,. 1996.

[3] Cannell, M.G.R., Van Noordwijk, M., Ong, C.K., The central agroforestry hypothesis: the trees must acquire resources that the crop would not otherwise acquire. Agrofor. Syst. 34, pp, 27–31. 1996.

[4] Schoorl J.M., A. Veldkamp. Linking land use and landscape process modeling: a case study for the Alora region (south Spain). Agriculture, Ecosystems and Environment 85, pp, 281-292. 2001.

[5] Riley J. Multidisciplinary indicators of impact and change. Agriculture, Ecosystems and Environment 87, pp, 245-259. 2001.

[6] Erickson J.D., J.M. Gowdy. Resource use, institutions and sustainability: a tale of two Pacific island cultures. Land Economics, 76(3) pp, 345-354. 2000.

[7] Cacho O. An analysis of externalities in agroforestry systems in the presence of land degradation. Ecological Economics 39, pp, 131–143. 2001.

[8] Torquebiau E. F. A renewed perspective on agroforestry concepts and classification. C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 323 pp, 1009–1017. 2000.




Section 7 Hydrological studies

Modified quantitative estimation model of erosion and degradation in three mountainous watersheds

D. Emmanouloudis1 & M. Kaikis2 1Department of Forestry and Natural Environment Management Technological Educational Institute of Kavala, Annex Drama, Drama, Greece 2Rhodes Municipal Environmental Organisation

Abstract

This paper presents a new version of a well-known stochastic model to quantitatively assess drainage basin degradation in order to determine erosion in three torrent basins. This new version, 3D-structured and using G.I.S. for data processing, can provide a quantitative assessment of the degradation of any torrent basin in a very short time. Furthermore, a system to prevent erosion and degradation by means of agricultural and technical forest works is presented; the system was originally designed for these three basins, but it is flexible and easy to adapt to any torrent basin. Keywords: stochastic model, GIS, degradation, 3D Digital Elevation Model.

1 Introduction

Drainage basin erosion and degradation is one of the most complex environmental problems worldwide. Thus, it has naturally been the subject of research for many decades in countries representative of all the main climatic types on the planet. The problem is fairly aggravated in the Mediterranean countries, due to their temperature range and other factors (irregularly distributed rainfall, insufficient plant coverage, etc.). In our country, plagued by forest fires in the past decades, it tends to acquire the dimensions of a real disaster for mountainous areas, due to the degradation of the soil it implies.




doi:10.2495/GEO060261

The assessment of mountainous soil degradation due to the erosive action of rainwater runoff is of two types:

a) Qualitative assessment (intensity, degree, speed, etc.) b) Quantitative assessment, usually calculated in total m /year/basin, or in m /year/Km, or finally in mm/year/basin.

Quantitative assessment is more difficult than the qualitative one, since it implies an attempt at calculating magnitudes and volumes which are hardly measurable in practice and are related to a diachronic effect. In the past few years several methods have been developed for the quantitative assessment of degradation. The more known methods are referred to be the USLE, WEPP, AGNPS1 [1], CREAMS, ANSWERS [2], EPIC, the method of Rydgren [3] and Terry [4], and the method of Mitas et al. [5], as well as the contemplative methods of Fournier, Corbel and Gavrilovic. The most of them have been used mostly in United States and less in Europe. However, according to Kotoulas [6] the main feature of these methods is that "they require the existence of basic data which however can not be often designated precisely so that the precision and the potential of their application are reduced". Therefore, according to the same author "the methods being developed to date are far from the point to provide precise and reliable forecasts". Moreover, in the chapter of erosion-degradation this is not the only difficulty. There is questioning even on how to confront the phenomenon particularly in situations during which we have intensive procedures (e.g. after a fire). In the present paperwork one from the above mentioned methods was selected, the stochastic Gavrilovic [7] model, by the help of which effort was made to estimate the degradation in the selected under research area. This estimation was conducted by the help of specific G.I.S. software (GRASS), which has helped out to simplify the model application on one hand and to increase the precision of the results to an impressive way on the other. After the estimation of degradation was implemented, the quotation of a protection and how to handle a problem was attempted, by applying a system easy-to-use, flexible and particularly effective, as proved by the experience gained to date.

2 Research area

As research area was selected the broader area of Kastoria Prefecture and especially four small torrential watersheds from which two were directly adjacent. As a reason of the above selection is quoted the fact that the aforesaid area -because of its geomorphologic condition- has many, in direct or less direct proximity, small runoff basins which particularly show a special vegetative and geological diversity. Of course, it is easily comprehensible why a model, in order to be elaborated, requires watersheds of small size. Regarding the highest possible proximity, an identification of the attack climatic factors in all parts of the sample is essential, so that the results obtained by the research are directly comparative. Following these thoughts we selected the above stated sample of four watersheds.




3 Research method

The morphometric characteristics of the selected watersheds have led us necessarily (because the so called method provides particularly good results when estimating the degradation in small basins of mountainous areas Gavrilovic [7], Pintar [9], Kotoulas [6]), to the selection of the Gavrilovic method. This method is uniquely appropriate to estimate the degradation in small mountainous watersheds, as the ones included in our sample. In Table 1 the morphometric characteristics of the sample watersheds are quoted, which define the mountainous character of the specific torrents.

Table 1: Morphometric characteristics of the sample watersheds.

Fwat. (Km2) Hmax (m) JmW (%) No Torrents Projective 3D" (max

elevation) (mean watershed

slope) 1 Riganorema 9.43 10.03 1420 30.90 2 Fotini 4.23 4.46 1190 29.41 3 Triantafillia 8.42 8.98 1580 33.32

F3D stands for the watershed area, not as it results from the calculation on the map (FPR0J) but for the actual area of the basin in three-dimensional format, as it results from the calculation after a 3D construction of the watershed by the help of GRASS. The F3D is slightly bigger than FPE0J and the deviation from each other depends on the average surface slope of the watershed, Emmanouloudis and Filippidis [10]. The above data was obtained from maps 1:50.000 of the G.A.S. (Sheets KASTORIA, MESOPOTAMIA). According to the Gavrilovic method eqn (1), the type providing the average annual degradation in a watershed is:

FzhTW ⋅⋅⋅⋅= 3π (1) where T, eqn (2), is coefficient of temperature given by the type:

1,0100 +=

tT (2)

with t0: average annual temperature in the watershed h: average annual height of rain (mm) π: 3.14 F: area of watershed (Km2) and z: coefficient of erosion given by the ratio )( Jφyxz +⋅⋅=

with χ, y, φ being the partial coefficients that depend on the vegetation, the geological base and the erosion degree of the watershed respectively whereas j is the average slope of the watershed area expressed as angle tangent. It is implied that the values of χ, y, and φ are given analytically in tables by Gavrilovic. The selections of the suitable values assigned to each case are at the




aptitude of the researcher who applies the formula when needed. By applying the method, each watershed under research was divided by the authors in spots depending on:

- the vegetation species prevailing in each part - the geological base it was consisted of - the seats of producing debris material.

Therefore, for example, there could be 6 vegetation spots distributed as follows: 2 spots of forest, 1 spot with shrubby plants, 1 spot with pastures, 2 spots with arid lands. There could also be 4 spots of geological base, 2 of them having geological base of granite, one gneiss and the other schist and so on. It is implied that the spots of vegetation, geology, seats of producing debris material are not identical as far as their boundaries are concerned. Given that often the spots of vegetation, geology, etc., were numerous, it becomes clear that this method was knotty, time-consuming and not particularly accurate, because the Gavrilovic method functions by the same slope for the whole watershed. To overcome this insufficiency, we stretched further the Gavrilovic model by improving the ways of spot separation. The G.I.S. GRASS [11] served greatly this effort. The following methodology was applied: In sheets of the G.A.S., the research area was scanned on GIS layers and converted to image with the use of image processing software. Then, converted into an appropriate image file format readable by GRASS, the scanned area was imported in GRASS in raster image format. Image rectification or registration of the research area was the next processing step. The rectification was originally carried out by digitization of 4 points, that is 4 angles, of the area on the HATT grid, and then by conversion of the raster co-ordinates imported in GRASS (which were in accordance to a local reference system) into projectional co-ordinates, with the use of linear affine transformation. Then, using the scanned rectified map, the research area was digitized [12] more specifically, digitization involved basin frontiers, peaks and contours (for peaks and contours the relevant elevation data were provided). Following this, the spots of vegetation, geology and erodibility were digitized, as suggested by the Gavrilovic formula on the basis of its x,y and (p. These spots were numbered and their respective values were provided according to the Gavrilovic tables. Following the above processing, all the data were exported in vector format and, after topology building, were converted in raster format, in view of producing all necessary maps and proceed to the necessary calculations. More specifically, from the 3 raster of the contours, the Digital Elevation Models (DEM) were produced, 3D-views of which can be seen on figures 1,2 and 3. Then, on the basis of these Digital Elevation Models, the slope maps were created. In order for the DEMs to be accurate, the very same slopes were used, which of course differed for every cell of the slope raster, i.e. an area of approx. 67 m (8.17m X 8.18m) for the torrents of Fotini and Riganorema and 36 m2 (6m X 6m) for the torrent of Triantafillia.




Figure 1: 3D View of watersheds Riganorema and Fotini.

Figure 2: 3D View of Triantafilia.

Figure 3: Grey-toned map of potential degradation with legendary classification (Triantafillia t.).




Then the GRASS software tool r.mapcalc [13] was used; this is a program that allows users to manipulate, analyze, and create map data by performing mathematical calculations on raster map layers. The under process and adjustment model by this way shows an improvement in the following points:

a. Precision of calculations: By the help of segmentation into cells we have a complete splitting of the basin in very small areas of 36-67 m2 which are worked out one by one as to the erosion coefficient, which is impossible without the G.I.S. process'. Besides, the most important American models such as ANSWERS, AGNPS, etc. run with such sort of splitting. In this adjustment model the watershed area slope is given "pointly" thus it is very important for the estimation of erodibility.

b. Ease at entering unlimited parameters: By entering the parameters in the form of raster maps in various layers, the participation process of many variables into the same model becomes much easier.

c. Convenience in processing the variable values: By applying the GRASS tool mapealc, any mathematical elaboration on the above mentioned raster maps is feasible resulting that the most complex combinations of parameter values become simple as to their calculation.

d. Appearance: 3D construction provided by the new model improves the appearance of the "real" picture of watersheds, depicting them as they are in nature and not as on the map.

4 Research results

The elaboration of the coefficients x, y, <p and the other factors of the method come up the erosion z coefficients for each watershed. The final coefficient z of each watershed was the weighed result of partial coefficients z, from which each one represents the erosion coefficient of a group" of cells. It is evident that all cells of the same group have the same z,. The groups for each watershed, depending on the case, were ranging from tens to several hundreds. Meanwhile, the size of the used analysis through the GRASS is extremely remarkable, considering that each group may include some tens of cells. In Table 2 the partial coefficients zx of the cells groups of Riganorema are quoted. In addition, in Table 3 the values of the final z estimation for all the watersheds of the sample but also the total respective annual sediment and debris yield W, are quoted. The analysis of the results gave the following findings: The bigger annual degradation is displayed by the Riganorema torrent and the smallest one by the Triantafillia. The first one has a geological base of silt-mixed marls and sandstones whereas the second one has pure granites. Also, the first one has relatively sparse vegetation of degraded coniferous forests and grasslands while the second one has extensive forest stands and shrubberies in good condition. Thus, the bigger average annual production of debris material of the first one in relation to the second is clearly explained. Finally, the other two watersheds, displaying




intermediate values of degradation, have respective intermediate values % and y of the vegetation and geology coefficients as well.

Table 2: Partial coefficients z, of the cells groups of Riganorema.

Table 3: Final z values and W values.

Torrents

No Name Coefficient

z

Average annual sediment and debris yield, W (m3/year)

Average annual

degradation (mm)

1 Riganorrema 0.78 15.82 1.57 2 Fotini 0.66 5.484 1.22 3 Triantafillia 0.75 6.823 0.75

Map and legend study indicates that the areas with light chromatic tones are the ones with higher risk from erosion and leaching and which need immediate protection. In other words, we notice that degradation from watershed to watershed differs quite, and surely in all cases is bigger than the average degradation given by Kotoulas for the Greek inland and which equals to 0.67 mm. Therefore, we realise that they are watersheds of mountainous character with a significant discharge of debris material. Finally, in order to have an even more integrated picture of the z distribution upon all watersheds of the sample, we drew some maps as follows: For the construction of maps we classified the values of Z; into 5 categories, arising from the tables such table 2. This classification has applied a different colour of grey tone for each category. Thus, surfaces with the same colour tone




corresponding to each particular category zi have arisen. In fact, by this way we had one more (but much more synoptically) grouping of z, so that to have a final result which is a coloured map of potential degradations with constant categorised annotated chromatology (figure 4).

5 Protection system

The protection system recommended in the present paperwork for the erodible areas, as mentioned earlier, is a mixed system of agro-plant-artificial works. The agro-plant-artificial works is a relatively old protection technique against erosion however specialised and with best results usually in the area of hillsides in respect to area of gullies. The selection of species and number of works depend on the slope of hillsides or gullies as well as on the soil type and the aspect. So, the standardisation of these works is possible (Emmanouloudis [14]) according to the case, when we are aware of the above factors in an interference area. However, the elaboration of the Gavrilovic model through the GRASS, in addition to its other mentioned earlier advantages, provides higher possibility of having in each area the surface slopes along with the aspect map. Thus, in conjunction to the aforesaid standardisation, according to Emmanouloudis [14], we may suggest per z category the following agro-plant-artificial works:

Area of hillsides

Z 0,6-0.8+ Construction of mosaic clusters with very tolerant species

Z0,4-0,6 Construction of envelopments

Area of micro gullies

Z 0.8+ Construction of picket fence belt Z0,6-0,8 Construction of envelopments belt Z0,4-0,6 Construction of clusters belt

6 Conclusions

Summarising all the above mentioned, we drew up the following conclusions: For the quantitative estimation of mountainous watersheds degradation a new model was used. The contemplative model of Gavrilovic served as a starting-point, which has been further stretched out by initiating appropriate adjustments by the use of the G.I.S. GRASS In the case of the four runoff basins, the average annual degradation has been calculated with a remarkably high precision due to this improved model. Actually, it is a combination of the Gavrilovic and the most




known American definition model. Also, an attempt was made so that the "Balkan" parameters of the first one (which resembles the Greek torrential environment) and the American G.I.S. detailed techniques are blended. Four maps of potentional erosion came out after calculating degradation; their study, along with a standardised system of protection works, can provide an integrated protection system planning. What's more important, this improved model of estimation and the protection works system may be applied on any watershed.

References

[1] Line, D., Foster, M., "User's Manual for the watersheds GRASS - AGNPS modelling tool", N. Carolina State University, U.S.A., 1996.

[2] Beasly, D., Huggins, L., "ANSWERS User's Manual", Purdue University, U.S.A., 1991.

[3] Rydgren, B. ,"Soil erosion: its measurement, effects and prediction. Case study from the southern Lesotho lowlands", z.f.Geom. N.F. 40, 429-445, Dec. 1996, Uppsala., 1996.

[4] Terry, J.P., "Erosion pavement formation and slope process interactions in commercial forest plantations", n. Portugal, z.f.Geom. N.F. 40, 107 - 115, Marz 1996, Bunbury, 1996.

[5] Mitas, L., Mitasova, H., Brown, W., and Astley, M., "Interacting fields approach for evolving spatial phenomena : application to erosion simulation for optimized land use", National Center for Geog. Inf. and Analysis, C.A., U.S.A., 1996.

[6] Kotoulas, D., Diefthetisis Himarikon Revmaton , A.U.T., Thessaloniki., 1997.

[7] Gavrilovic, SI., "Inzenjcring o bujich im tokovima i eroziji", Beograd, 1972.

[8] Emmanouloudis, D., "Natural depositional landforms of the Greek torrents", Ph.D. Thesis, Dpt. of Forestry and Nat. Environment, A.U.T., Thessaloniki., 1990.

[9] Pintar. J., "Grenzen und Moglichkeiten der vorbeugung vor Unwetterkatastrophen im alpinen Raum", INTERPRAEVENT, Villach, AUSTRIA., 1972.

[10] Emmanouloudis. D., Filippidis, E., "Torrents basins Morphometric features calculation through 3-D models", Institute for education and Technology of Drama. Special Edition, 2001.

[11] Shapiro, M., Westerveld, J, "GRASS user's manual", U.S. Army Construction Engineering Research Laboratory, Champain, Illinois, U.S.A. ., 1993.

[12] Neidig, C.A., Gerdes, D., Kos, Ch., "GRASS 4.0 Map Digitizing Manual : v.digit", U.S. Army Construction Engineering Research Laboratory, Champain, Illinois, U.S.A., 1991.




[13] Shapiro, M., Westerveld, J., "r.mapcalc. An algebra for G.I.S. and image processing", U.S. Army Construction Engineering Research Laboratory, Champain, Illinois, U.S.A. 1992.

[14] Emmanouloudis, D., "The contribution of watershed management to the integrated flood protection of the Athens basins", 2nd International Conference of C.N.W.R., Athens., 1994.




Bathymetric curve (75 years old) validation, using the soil erosion transportation at Cuitzeo Lake Watershed

J. Lafragua1, A. Gutiérrez1, A. Bahena1, G. Moriel2 & S. Férnández2 1Mexican Institute of Water Technology, Mexico 2Secretary of Communications and Transports, Michoacán Center, Mexico

Abstract

A new main road called Copandaro-La Cinta was built in the state of Michoacán, México; with a length of seventeen kilometers this road crosses Lake Cuitzeo (the section under study). In order to preserve the hydrological regime of this water area, the road structure was built with twenty sewers and four boat passages. With a 4000 km2 basin area, Lake Cuitzeo’s watershed transports 1,296,461 tons of sediment per year. Due to this great amount of material, a hydrological study was carried out. A hydraulic study was prepared including elevation-storage and elevation-area relationship curves since 1930. Using a longitudinal profile of the study section from April 2003, several bank levels are used to make an adjustment between the bathymetric curve and longitudinal profile, fixing some main structures points. In order to select the best alternative (to verify hydraulic works dimensions), soil erosion in the watershed is calculated using the Universal Soil Loss Equation (USLE); a sediment yield ratio is also obtainable. The results show that only 7 percent of the soil loss computed by the USLE appears as sediment yield in the watershed outlet. In the alternative selected, this quantity represents approximately 19 cm of sediment near the study section. With this alternative, the zero level in the longitudinal profile and in the bathymetric curve is at an elevation of 1823.34 and 1818.00 masl, respectively. Thus, all points of the bathymetric curve were adjusted at 5.34 m. Flow routing through Lake Cuitzeo shows that the structures hydraulic capacity was enough to allow free water flow. Keywords: soil erosion, hydrologic routing, inflow hydrographs, bathymetric curve, Lake Cuitzeo.




doi:10.2495/GEO060271

1 Introduction

The new Copandaro-La Cinta main road on the Morelia-Salamanca freeway crosses the western area of Lake Cuitzeo. This structure has twenty sewers and four boat passages, and a bridge called Dren La Cinta. In previous years, this bridge allowed the passage of flows from Lake Cuitzeo to Lake Yuriria. To check the dimensions of the existent hydraulic works and verify their proper hydraulic operation it is necessary to consider the volumes of inflow for different return periods. Lake Cuitzeo’s basin is divided into 25 sub-regions Lafragua et al., [6]. To carry out the hydraulic flow analysis, a bathymetric curve fundamental; however, the only one available was for the year 1930 (provided by the National Water Commission, Conagua). The objective of this study is to correct the bathymetric curve, taking into account the longitudinal profile of the road (2003), provided by the Secretariat of Communications and Transport (SCT). The amount of sediment deposited in the Lake is also required.

Figure 1: Location map of the study site.

2 Study area

Cuitzeo’s watershed is located in the central part of the state of Michoacan, Mexico, between the coordinates 19°24’ to 20°05’ north latitude and 100°41’ to 101°33’ west longitude, and is part of hydrologic region number 12 called Lerma-Santiago, figure 1. Total watershed surface is 4,000 km2, of which 409.82 km2 of the Lake include hydrophyte vegetation and open water. The main river is Rio Grande of Morelia, with a drainage area of 2,043 square kilometers. The climate is temperate and subhumid with summer rainfall. The annual mean




temperature ranges from 13°C to 20°C, and mean annual precipitation is 804.0 millimeters. The Copandaro-La Cinta new main road crosses Lake Cuitzeo along seventeen kilometers (section in study) and has twenty sewers and four boat passages, in addition to La Cinta drain, figure 2.

Figure 2: Hydraulic works location.

3 Procedure description

A bathymetric adjustment was conducted because there is no common level between the bathymetric curve (Conagua curve) and the longitudinal profile (SCT profile) of the study section. The procedure was as follows:

a) Drawing a longitudinal profile of the study section with the bathymetric curve information (Conagua Profile).

b) Analyzing possible alternatives of elevation similarities between the Conagua profile and the SCT profile.

c) Selecting the best alternative, by using the erosion value of the watershed outlet.

In order to calculate the annual erosion of Lake Cuitzeo’s watershed, the Universal Soil Loss Erosion (USLE) equation was used, Ponce [7], given by

PCSLKRA = (1) where A is the annual mean soil loss in tons ha-1, R (MJ mm ha-1 h-1) is the rainfall erosivity index, calculated according to Figueroa et al. [4], K (tons ha h ha-1 MJ-1 mm-1) is the soil erodibility factor, calculated according to the FAO [3] methodology, C is the cover and management factor, and P is the support practice factor, assumed as 1.0, Izurieta et al. [5]. L is the slope length factor




and S is the slope gradient, both are known as topographic factors and were calculated according to Izurieta et al. [5].

m

L

=

1.22λ (2)

where λ (m) is the slope length. For this study area, according to Cortes [1], value λ can be considered as 100 m and value m is the slope-dependent dimension factor, considered as 0.4.

3.0)(8.10 += sgsenS with %9<s (3a)

5.0)(8.16 −= sgsenS with %9>s (3b) where sg and s are the slope of the land expressed in degrees and in percent, respectively.

With eqn (1), the amount of solid material eroded due to rainfall in one year is estimated; however, the total volume of material is not necessarily arriving at the watershed outlet. Escalante [2] declares that there are differences between soil loss in the watershed and the amount of sediment that the watershed outlet is receiving, and suggests using the following equation:

ADRAS = (4) where AS (tons ha-1) is sediment in the watershed outlet, A is the value calculated by eqn (1), and DR is the sediment yield ratio, calculated by,

127097.0417662.0 134958.0 −= −AcDR (5) where Ac (mi2) is the watershed area.

Table 1: Fixed points.

Alternative Point number

Characteristic Elevation difference (m) (SCT minus Conagua)

A-1 5 High point and located outside the Lake.

5.83

A-2 41 High point. 4.43 A-3 47 Without major inflow. 5.70 A-4 27 High point and without

major inflow. 5.34

A-5 49 High point and without major inflow.

5.27


Using the Conagua elevation curve, 55 points were selected, including the 25 hydraulic works, to draw the longitudinal profile of the new main road. Many alternatives were analyzed by fixing points where erosion was not observed through time. Table 1 shows five alternatives indicating the fixed point




considered, and table 2 shows the differences between the Conagua profile and the SCT profile, the rest of the alternatives showed a similar behavior.

Table 2: Selected points and elevation differences.

No. Hydraulic work

Elevation (SCT)

Elevation (Conagua)

A-1

A-2

A-3

A-4

A-5

1 1827.18 1820.50 -0.85 -2.25 -0.98 -1.34 -1.41 2 1826.01 1820.00 -0.18 -1.58 -0.31 -0.67 -0.74 3 1825.63 1819.50 -0.30 -1.70 -0.43 -0.79 -0.86 4 1824.80 1819.00 0.03 -1.37 -0.10 -0.46 -0.53 5 1824.33 1818.50 0.00 -1.40 -0.13 -0.49 -0.56 6 1823.87 1818.00 -0.04 -1.44 -0.17 -0.53 -0.60 7 1823.86 1818.00 -0.03 -1.43 -0.16 -0.52 -0.59 8 1823.81 1818.00 0.02 -1.38 -0.11 -0.47 -0.54 9 1823.83 1818.00 0.00 -1.40 -0.13 -0.49 -0.56 10 1823.88 1818.50 0.45 -0.95 0.32 -0.04 -0.11 20 1824.38 1820.50 1.95 0.55 1.82 1.46 1.39 21 1824.48 1820.50 1.85 0.45 1.72 1.36 1.29 22 1824.76 1820.00 1.07 -0.33 0.94 0.58 0.51 23 1824.93 1819.50 0.40 -1.00 0.27 -0.09 -0.16 24 1824.56 1819.00 0.27 -1.13 0.14 -0.22 -0.29 25 1824.59 1819.00 0.24 -1.16 0.11 -0.25 -0.32 26 1824.66 1819.00 0.17 -1.23 0.04 -0.32 -0.39 27 1824.85 1819.51 0.49 -0.91 0.36 0.00 -0.07 28 1825.15 1819.86 0.54 -0.86 0.41 0.05 -0.02 29 1825.85 1820.00 -0.02 -1.42 -0.15 -0.51 -0.58 30 1825.59 1820.50 0.74 -0.66 0.61 0.25 0.18 40 VC10 1824.40 1819.62 1.05 -0.35 0.92 0.56 0.49 41 VC11 1824.55 1820.12 1.40 0.00 1.27 0.91 0.84 42 VC13 1824.85 1819.31 0.29 -1.11 0.16 -0.20 -0.27 43 VC14 1825.05 1819.17 -0.05 -1.45 -0.18 -0.54 -0.61 44 VC15 1824.45 1818.99 0.37 -1.03 0.24 -0.12 -0.19 45 VC16 1824.35 1819.00 0.48 -0.92 0.35 -0.01 -0.08 46 VC17 1824.60 1819.03 0.26 -1.14 0.13 -0.23 -0.30 47 VC18 1825.30 1819.60 0.13 -1.27 0.00 -0.36 -0.43 48 VC19 1824.80 1819.70 0.73 -0.67 0.60 0.24 0.17 49 VC21 1825.24 1819.97 0.56 -0.84 0.43 0.07 0.00 50 VC22 1825.24 1819.98 0.57 -0.83 0.44 0.08 0.01 51 PL1 1823.85 1818.02 0.00 -1.40 -0.13 -0.49 -0.56 52 PL2 1823.85 1818.02 0.00 -1.40 -0.13 -0.49 -0.56 53 PL3 1824.60 1819.00 0.23 -1.17 0.10 -0.26 -0.33 54 PL4 1824.85 1819.38 0.36 -1.04 0.23 -0.13 -0.20 55 PEMEX 1824.45 1819.10 0.48 -0.92 0.35 -0.01 -0.08

For each alternative, total sediment area was obtained by considering the differences shown in table 2. Figure 3 shows the sediment areas of alternative 4, of 34,800 to 45,504 km. Assuming a uniform distribution along the longitudinal profile, an average sediment height is estimated, table 3. In table 3, the negative values indicate that the fixed point was eroded; therefore, these alternatives were eliminated. The rest of the alternatives have an average (elevation) height of between 0.18 and 1.5 m. Consequently, we compared water levels measured by Conagua and SCT during the construction of the road, obtaining an average difference of 5.18 m. Otherwise, we considered the incrustation value designated by the SCT (0.60 m), then the points with




differences between 0 and -0.60 m were counted. In alternatives 4 and 5, more points, 30 and 32 points, respectively, were found. Therefore, alternatives 4 and 5 are good options. In order to select only one alternative, watershed outlet soil erosion was considered. The variable values for each watershed from eqn (1) are shown in table 4.

Figure 3: Sediment area along the new main road.

Table 3: Area and average sediment height.

Alternative Total sediment area, (m2)

Average sediment height, (m)

A-1 -5442.41 -0.31 A-2 18784.97 1.08 A-3 -3192.73 -0.18 A-4 3037.17 0.18 A-5 4248.54 0.24

The results of applying eqn (1) and eqn (2) are shown in the table 5. Soil loss in all watersheds in Lake Cuitzeo is 17,672,146 tons year-1, but only 7 percent appears as sediment yield at the watershed outlet. Assuming that 1,296,461 tons of sediment arrived at the Lake per year in average, then in 75 years, the Lake has received 97 hm3 of sediment. Due to unavailable information, we assume that the sediment is first deposited into the deeper areas, resulting in 70 cm of sediment in these areas and 19 cm approximately in the study area, figure 4. Finally, alternative 4 was selected because it exhibits the value closest to 19 cm, obtained by calculating the erosion. In the selected alternative, the zero level in the SCT and Conagua profiles is at an elevation of 1823.34 and 1818.00 masl, respectively, therefore, every point of the bathymetric curve was adjusted at 5.34 m. Figure 5 shows the adjusted bathymetric curve.




Table 4: USLE variables for each watershed.

Watershed number

Area (ha)

Rainfall (mm) S (%) R K LS C P

1 1194.11 705.79 17.0 2445.68 0.02 3.91 0.04 1.00 2 2549.69 833.20 16.9 2883.03 0.02 4.09 0.13 1.00 3 1293.56 845.52 16.6 2925.26 0.02 4.22 0.16 1.00 4 2081.71 843.01 15.7 2916.70 0.02 3.81 0.20 1.00 5 17517.45 802.72 13.0 2778.56 0.02 3.03 0.17 1.00

14 4052.99 670.76 7.2 2325.29 0.02 1.41 0.34 1.00 15 4862.76 625.53 6.1 2168.99 0.03 1.11 0.29 1.00 16 5777.69 640.41 4.9 2220.35 0.03 0.83 0.23 1.00 17 7207.63 715.52 10.2 2478.98 0.02 2.01 0.15 1.00 18 2578.24 720.54 18.2 2496.25 0.02 5.09 0.24 1.00 19 2898.54 720.54 17.6 2496.25 0.02 5.02 0.28 1.00 20 6807.55 734.05 12.9 2542.42 0.03 3.16 0.29 1.00 21 10859.16 784.39 13.7 2714.75 0.03 3.25 0.16 1.00 22 707.42 742.85 16.4 2572.35 0.02 3.95 0.10 1.00 23 55985.36 767.67 16.9 2657.86 0.03 4.47 0.21 1.00 24 204294.24 860.61 12.2 2977.92 0.02 2.93 0.23 1.00 25 2513.50 714.65 7.3 2476.29 0.02 1.68 0.42 1.00

Table 5: Erosion results for each watershed.

Watershed number

Eqn. (1) (tons ha-1 yr-1)

DR Eqn. (4) (tons yr-1)

Watershed number

Eqn. (1) (tons ha-1 yr-1)

DR Eqn. (4) (tons yr-1)

1 7.39 0.21 1877.97 14 27.47 0.16 17932.76 2 28.19 0.18 12912.61 15 17.77 0.15 13312.72 3 36.75 0.21 9940.18 16 10.67 0.15 9100.33 4 39.67 0.19 15538.25 17 17.36 0.14 17454.16 5 26.89 0.11 51531.12 18 67.23 0.18 31059.16 6 18.06 0.17 9160.92 19 84.92 0.17 42924.66 7 24.06 0.20 7189.23 20 67.41 0.14 64968.17 8 33.96 0.13 39552.37 21 42.58 0.13 57874.83 9 22.46 0.16 16011.22 22 25.87 0.24 4347.56

10 128.75 0.23 25097.48 23 80.84 0.08 339784.28 11 51.41 0.19 17831.81 24 46.85 0.04 408418.37 12 101.81 0.20 33020.48 25 37.71 0.18 17084.97 13 52.54 0.16 32535.29 Total 1098.61 0.16 1296460.88

The adjusted bathymetric curve was used by routing floods with the storm design of 50 years. We consider tree scenarios: a) inflow from the east and west during the same time, figure 6, b) only inflow from the east, and c) only inflow from west. The hydraulic results including the 25 hydraulic works located in the study site and the sewer located in the freeway are shown in table 6. Maximum elevation at the end inflow was 1824.53 masl from hydraulic work VC1 to VC5 and 1824.67 masl from VC10 to La Cinta drain. The maximum discharge obtained in the hydraulic works was 4.94 m3 s-1 and maximum velocity was 1.33 m s-1. In the freeway sewer, the maximum discharge was 11.89 m3 s-1 and maximum velocity was 2.8 m s-1 Lafragua et al. [6].




Figure 4: Scheme showing sediment deposit into Lake Cuitzeo.

50 100 150 200 250 300 350 400

1823.00

1823.50

1824.00

1824.50

1825.00

1825.50

100 200 300 400 500 600 700 800 900

Volume

Area

Volume, Mm3

Area, km2

Ele

vatio

n, m

asl

Figure 5: Adjusted bathymetric curve.

5 Conclusions

A bathymetric curve from 1930 was adjusted by using a longitudinal profile of the study section from 2003, and one sediment yield ratio was used in order to select the best alternative. All points of the bathymetric curve were adjusted at 5.34 m. Also considered were water levels measured by SCT and Conagua during the building of the new main road, furthermore, 60 cm from incrustation designed by SCT was considered.




Figure 6: Input floods for rainfall (50-years).

Table 6: Hydraulic results.

Hydraulic Footing Grade Initial End Initial End Area Discharge Velocity Work

Elevation

(masl) Elevation

(masl) Elevation

(m) Elevation

(m) Depth

(m) Depth

(m) (m2) (m3 s-1) (m s-1)

VC1 1823.95 1826.60 1824.50 1824.53 0.55 0.58 1.10 0.42 0.38

PL1 1823.85 1829.75 1824.50 1824.53 0.65 0.68 20.38 4.71 0.23

VC2 1823.80 1826.60 1824.50 1824.53 0.70 0.73 1.40 0.58 0.42

PL2 1823.85 1829.65 1824.50 1824.53 0.65 0.68 20.38 4.94 0.24

VC3 1823.85 1826.60 1824.50 1824.53 0.65 0.68 1.31 0.41 0.32

VC4 1823.95 1826.60 1824.50 1824.53 0.55 0.58 1.11 0.44 0.40

VC9 1824.55 1826.30 1824.50 1824.71 -0.05 0.16 0.30 0.30 1.03

VC10 1824.40 1826.30 1824.50 1824.67 0.10 0.27 0.35 0.41 1.15

PEMEX 1824.45 1826.30 1824.50 1824.67 0.05 0.22 3.43 1.21 0.35

VC11 1824.55 1826.30 1824.50 1824.67 -0.05 0.12 0.07 0.03 0.41

VC13 1824.85 1826.30 - - - - - - -

VC14 1825.05 1826.30 - - - - - - -

PL3 1824.60 1830.90 1824.50 1824.67 -0.10 0.07 1.90 0.07 0.04

VC15 1824.45 1826.30 1824.50 1824.67 0.05 0.22 0.26 0.24 0.95

VC16 1824.35 1826.30 1824.50 1824.67 0.15 0.32 0.45 0.60 1.33

VC17 1824.60 1826.30 1824.50 1824.67 -0.10 0.07 0.13 0.01 0.06

PL4 1824.85 1830.90 - - - - - - -

LA CINTA 1823.70 1829.40 1824.50 1824.67 0.80 0.987 24.35 0.28 0.01

Sewer

freeway (D=5 m) 1823.34 1824.50 1824.64 1.16 1.30 4.13 11.89 2.88




The adjusted bathymetric curve was used by routing floods with the storm design of 50 years. Flow routing through Lake Cuitzeo shows that hydraulic structures were enough to permit a free flow through them.

References

[1] Cortés T.H., (2005). Personal communication, 10 October 2005. Resarcher. Mexican Institute of Water Technology (IMTA).

[2] Escalante S.C., (2005). Capítulo 8. Efecto en la estimación del factor erosivo de la lluvia en el aporte de sedimentos. En: Rivera-Trejo F., Gutiérrez-López A., Val-Segura R., Mejía-Zermeño R., Sánchez-Ruiz P., Aparicio-Mijares J, Díaz-Flores L., (Editores). “LA MEDICIÓN DE SEDIMENTOS EN MÉXICO”. Ediciones IMTA-UJAT, México. 325 p. ISBN-968-5536-53-8.

[3] FAO, (1980). Metodología provisional para la evaluación de la degradación de los suelos. Roma. 86 p.

[4] Figueroa S.B, Amante O.A, Cortés T.H, Pimentel L.J, Osuna Ceja E.S, Rodríguez O.J.M and Morales F.F.J, (1991). Manual de predicción de pérdidas de suelo por erosión. Subdirección de Conservación del Suelo y Agua, SARH.

[5] Izurieta J., Huerto D.R., Medina M.R., Cortés T.H., Spillecke W.K.W., Brena Z.J. y Castillo R.C., (2002). Estimación del impacto de las cargas de contaminantes del Dren Zurumútaro en el lago de Pátzcuaro y propuestas de tratamiento. SGC-UAPS-MICH-02-006-RF-CC, CNA-IMTA.

[6] Lafragua C.J., Gutiérrez L.A., Báhena H.A., Leal B.G., and Peña P.T., (2005). Dimensionamiento de alcantarillas y pasos de lancha, en el tramo carretero Copándaro-La Cinta, Morelia, Michoacán. IMTA-SCT. Proyecto TH-0550.

[7] Ponce V.M., (1989). Engineering Hydrology: Principles and Practices. Prentice Hall.




Efficient watershed modeling using a multi-site weather generator for meteorological data

M. Khalili, R. Leconte & F. Brissette Department of Construction Engineering, École de Technologie Supérieure, Montréal, Québec, Canada

Abstract

The multi-site generation of precipitation data is developed using a Richardson (1981) WGEN-type weather generator. This approach is based on spatial autocorrelation to analyze patterns in space and investigate the dependence of weather data at multiple locations. Reproducing the dependence between meteorological data at several stations should make the hydrological model results more realistic. The Chute du diable watershed and surrounding area located in the province of Quebec, Canada was used to test the proposed approach. Daily spatial autocorrelations between precipitation occurrences and amounts were successfully reproduced as well as total monthly precipitation and monthly numbers of rainy days. A hydrological model has been used to quantify the natural inflow process. As envisaged, the multi-site generation of weather data produced more practical natural inflow hydrographs, compared to those obtained using a uni-site weather generator. Keywords: weather generator, precipitation, Markov chain, spatial

1 Introduction

Weather generators can be used to generate climatic data (precipitation, temperature, solar radiation...) with the same statistical properties as the observed ones. Most weather generators operate for a single site, e.g. [15], [2] and [7]. Therefore, they ignore the regional coherence and the spatial dependence between the stations, which entail many problems in the hydrological modeling results, obtained using the simulated time series of meteorological data.




doi:10.2495/GEO060281

autocorrelation, hydrological modeling.

A few models were developed for multi-site simulation of weather variables, in particular for daily precipitations such as space-time model of Bardossy and Plate [1], the non-homogeneous hidden Markov model of Bellone et al. [3] and Hughes et al. [11], the nearest-neighbor resampling approach developed by Buishand and Brandsma [5], and a method based on serially independent but spatially correlated random numbers developed by Wilks [16]. A regionalization approach based on spatial autocorrelation is proposed to improve the watershed modeling. It is applied to the Chute du diable watershed and surrounding area located in the province of Quebec, Canada. This technique performed successfully in simulating meteorological data and runoff process. The proposed methodology is presented in section 2 of this article. Section 3 describes the results obtained in the studied basin.

2 Methodology

2.1 Uni-site weather generator

Following the Richardson approach [13], a uni-site weather generator uses a first-order two states Markov chain to simulate daily precipitation occurrence

( )kX t at site k on day t. A uniform [0, 1] random number ( )kut is compared with a critical probability, which is equal to one of the transition probabilities depending on the state of the previous day:

( ) ( ) ( )( ) ( ) .

1,0,

111

101

==

=−

−

kXifkpkXifkp

kpt

tc (1)

A wet day is simulated if the random number is smaller than this critical probability:

( ) ( ) ( ).

,0,1

≤

=otherwise

kpkuifkX ct

t (2)

Another uniform [0, 1] random number ( )kvt is used to simulate the synthetic precipitation amounts by the inversion of the distribution function of amounts. In the case of an exponential distribution [14], the precipitation amounts ( )krt can be computed as: ( ) ( )( ) ).(/1ln kkvkr tt λ−−= (3)

where λ is the parameter describing the exponential distribution function: ( )[ ] ( )( ).exp1 krkrf λ−−= (4)




2.2 Multi-site weather generator

The Multi-site weather generator used here is based on the concept of spatial autocorrelation. This theory has been used in a wide array of applications in which the spatial dependence has to be accounted for, such as social, economic and physics sciences. Spatial autocorrelation is the correlation between values of a single variable in geographic space. The analogous statistic to spatial autocorrelation is serial autocorrelation which is the correlation between values of a single variable at different time. Spatial autocorrelation can be measured by statistical indicators such as Moran’s I [10, 12].

( ) ( )

( ) ( ).

//

/

1

2

1

2

1 1 1 1

∑∑

∑ ∑ ∑∑

==

= = = =

−−

−−

=n

ii

n

ii

n

i

n

j

n

i

n

jijjiji

nxxnxx

wxxwxx

I (5)

where ix denotes the observed value at location i, x is the average of the ix over the n locations and ijw is the spatial weight between two locations i and j. The matrix form of Moran’s I contains a spatial weight matrix whose elements are the weights ijw . Generally, these weights are in a row-standardized form, which means that all weights in a row sum to 1 and by convention 0=iiw . Moran’s I takes values greater than zero if the geographically nearby observations are similar, lower than zero if they are dissimilar and equal to zero if these observations are independent. The weather generator is modified to simulate precipitation data with daily spatial autocorrelations, measured by Moran’s I, identical to those observed. Indeed, the random numbers used in the weather generator, eqns. (2) and (3), are transformed to spatially autocorrelated ones whose spatial autocorrelations will reproduce the spatial autocorrelations computed between observed precipitation series. To generate spatially autocorrelated random numbers, a spatial moving average process [6, 8] is used: .uuWV +××= γ (6)

where ( )1,nV is a vector of n spatially autocorrelated random numbers to be used for n locations. ( )nnW , is a weights matrix. ( )1,nu is a vector of n independent and uniformly distributed random variables.

γ is the moving average coefficient. Differentγ values provide random numbers with different spatial autocorrelations and accordingly simulated precipitation processes exhibit




different daily spatial autocorrelations. Therefore, a relationship between the coefficients γ and the spatial autocorrelations of precipitation occurrences and amounts can be used to identify the particulars γ , which yield the set of random numbers reproducing the observed daily spatial autocorrelations of precipitation occurrences and amounts. These meteorological data will be then used as input into the hydrological model.

2.3 Hydrological modeling

The HSAMI model [4] is used for hydrological modeling. HSAMI is a lumped conceptual model consisting of three linear reservoirs in cascades currently used by Hydro-Québec to forecast and to simulate natural inflows to reservoirs or runoff at watershed outlets. A set of meteorological data consisting of rainfall, snow, minimal and maximal temperature and insulation [9] is required to simulate the hydrological processes. Fig. 1 gives a simplified diagram of HSAMI [9].

Figure 1: Simplified diagram of HSAMI [9].


Seven weather stations were used in Chute du diable watershed and surrounding area in the province Quebec of Canada (fig. 2): 1: Péribonca, 2: Normandin CDA, 3: Hémon, 4: Bonnard, 5: Chute du diable, 6: Chute des passes, 7: St-

Horizontal flow

evapotranspiration

Reservoir

Inflow

Interception

Vertical flow

According to ground saturation

and frost

Rainfall and snow




Leon-de-Labrecque. The uni-site and multi-site approaches described above were used to simulate precipitation data at these stations to be incorporated in the hydrological model. The uni-site approach simulates the data at each weather station independently from the others. Thus, the spatial dependence in precipitation data is ignored. The multi-site approach produces daily spatial autocorrelations that are identical to those observed at the stations.

Figure 2: Chute du Diable watershed with locations of weather stations.

Fig. 3 gives an example of observed versus simulated monthly numbers of rainy days at all stations for the month of March and fig. 4 gives the total monthly precipitation at all stations for the month of September. Similar results were obtained for all the other months. These curves show a good agreement between the observed and simulated precipitation occurrences and amounts. Furthermore, daily spatial autocorrelations of precipitation occurrences and amounts were well reproduced, as shown by fig. 5 and 6, which illustrate these results for March and September respectively. Again, similar results were obtained for the remaining months.




4

6

8

10

12

14

1 2 3 4 5 6 7Stations used

Mon

thly

num

bers

of r

ainy

day

s

Observed monthly numbers of rainy daysSimulated monthly numbers of rainy days

Figure 3: Observed versus simulated monthly numbers of rainy days at the

seven stations for March.

60708090

100110120130

1 2 3 4 5 6 7Stations used

Tota

l mon

thly

pre

cipi

tatio

ns

Observed total monthly precipitations (mm)

Simulated total monthly precipitations (mm)

Figure 4: Observed and simulated Total monthly precipitations at the seven

stations for September.

HSAMI hydrological model requires the mean precipitation over the studied area. Five weather stations were then selected within and around Chute du diable watershed according to their proximity to the watershed (stations 1, 4, 5, 6 and 7). An interpolation of precipitation values over this area is required to compute the mean precipitation using the Thiessen polygons. This operation was done using the precipitation data simulated by the uni-site and multi-site approaches. However, the temperature data were simulated only by the uni-site approach, as a multi-site model of temperature data is currently under development. The meteorological data were then incorporated in the HSAMI hydrological model. Five years of meteorological and natural inflow data were used for calibrating the HSAMI model. Natural inflows were then simulated using the two types of precipitation data. The simulation results from the input precipitation data generated using the multi-site approach were found to be more realistic than




using the input precipitation data generated from the uni-site approach. Typically, the simulated inflow hydrograph from the multi-site approach displays the spring as well as the late summer and early fall floods (August-October), while the simulated hydrograph from the uni-site approach indicates only the spring floods, see fig. 7.

00.10.20.30.40.50.60.7

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31March days

Dai

ly s

patia

l aut

ocor

rela

tions

Observed daily spatial autocorrelations

Simulated daily spatial autocorrelations

Figure 5: Observed and simulated daily spatial autocorrelations of

precipitation occurrences for March.

-0.2

-0.10

0.1

0.2

0.30.4

0.5

0.6

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

September days

Dai

ly s

patia

l aut

ocor

rela

tions

Observed daily spatial autocorrelations

Simulated daily spatial autocorrelations

Figure 6: Observed and simulated daily spatial autocorrelations of

precipitation amounts for September.

The hydrological model fed with multi-site generated precipitation data was able to better mimic the observed process because of spatial dependence of rainfall and snowfall, which was reproduced between the multiple sites in spite of the lumped nature of the HSAMI. In reality, the climate responsible of the precipitation process extends at the regional scale rather than at the station location and constrains the observations in a given station to be correlated to those in nearby area. Table 1 shows how the multi-site approach improves the




simulation of extreme events. Indeed, the extreme natural inflows from the multi-site approach (Multi) were more consistent with the observed ones (Obs) while those from the uni-site approach (Uni) were systematically underestimated.

Figure 7: Example for Multisite and unisite simulated natural inflows.

Table 1: Observed and simulated extreme events by the two approaches.

Natural inflows (m3/s)

Return periods Obs Multi Uni

2 1243.5 1321.45 1022.45 5 1419 1315.1 1211.2

10 1729 1813.55 1339.3

4 Conclusion

The Chute du diable watershed and surrounding area located in the province of Quebec, Canada is used to investigate the behaviour of hydrological modeling using uni-site and multi-site generation approaches for precipitation data. The multi-site approach performed successfully in simulating both precipitation occurrences and amounts. This result impacts on the hydrological modeling, as demonstrated using the HSAMI model, which better displayed the late summer-early fall flood when multi-site approach was used. Other performance criteria are under investigation to further demonstrate the potential and capabilities of coupling a multi-site precipitation model to a hydrological model.

References

[1] Bardossy, A. & Plate, E.J., Space-time model for daily rainfall using atmospheric circulation patterns. Water resources research, 28, pp. 1247-1259, 1992.

0

200400

600

8001000

1200

14001600

1800

1 25 49 73 97 121 145 169 193 217 241 265 289 313 337 361

Days

Nat

ural

inflo

ws

sim (multi) sim (uni)




[2] Bardossy, A. & Plate, E.J., Modeling daily rainfall using a semi-Markov representation of circulation pattern occurrence. Journal of Hydrology, 122, pp. 33-47, 1991.

[3] Bellone, E., Hughes, J.P. & Guttorp, P., A hidden Markov model for downscaling synoptic atmospheric patterns to precipitation amounts. Climate Research, 15, pp. 1-12, 2000.

[4] Bisson, J.L. & Roberge, F., Prévision des apports naturels: Expérience d’Hydro-Québec. Compte-rendu de l’Atelier sur la prévision du débit, Toronto, novembre 1983.

[5] Buishand, T.A. & Brandsma, T., Multisite simulation of daily precipitation and temperature in the Rhine basin by nearest-neighbour resampling. Water resources research, 37(11), pp. 2761-2776, 2001.

[6] Cliff, A.D. & Ord, J.K., Spatial processes: Models and applications. London: Pion, 1981.

[7] Corte-Real, J., Xu, H. & Qian, B. A weather generator for obtaining daily precipitation scenarios based on circulation patterns. Climate Research, 13, pp. 61-75, 1999.

[8] Cressie, N.A.C., Statistics for spatial data. Wiley series in probability and mathematical statistics, John Wiley & Sons, 900 p, 1993.

[9] Fortin, V., Le modèle météo-apport HSAMI: historique, théorie et application. Rapport de recherche, révision 1,5. Institut de recherche d’Hydro-Québec (IREQ), Varennes, Qué. 68 p, 2000.

[10] Griffith, D.A., Spatial autocorrelation and spatial filtering: Gaining understanding through theory and scientific visualization. Advances in spatial science, Springer, 247 p, 2003.

[11] Hughes, J.P., Guttorp, P. & Charles, S., A nonhomogeneous hidden Markov model for precipitation occurrence. Applied statistics, 48, pp. 15-30, 1999.

[12] Odland. J., Spatial autocorrelation. Sage Publications: Newbury Park, 87 p, 1988.

[13] Richardson, C.W., Stochastic simulation of daily precipitation, temperature, and solar radiation. Water resources research, 17(1), pp. 182-190, 1981.

[14] Todorovic, P. & Woolhiser, D.A., Stochastic model of daily rainfall. Proc. the symposium on statistical hydrology. Misc. Publ. 1275, U. S. D. A. Washington, Dc., pp. 232-246, 1974.

[15] Wilks, D.S. Conditioning stochastic daily precipitation models on total monthly precipitation. Water resources research, 25, pp. 1429-1439, 1989.

[16] Wilks, D.S., Multisite generalization of a daily stochastic precipitation generation model. Journal of Hydrology, 210, pp. 178-191, 1998.




Hydrological modelling for river basin management in a highly hydro-geological conditioned environment

D. Guida, A. Longobardi & P. Villani Department of Civil Engineering, University of Salerno, Italy

Abstract

Water resources management represents a present key issue in hydrology, and hydrological models generating streamflow time series are useful tools in this field. It is possible to refer, in the extreme, to lumped or fully distributed approaches, but when river basins with particular features have to be modeled it is possible to take advantage of a semi-distributed formulation. In this study we propose a semi-distributed conceptually based modeling approach, supported by field measurements collected within several seasonal campaigns, that has been set up for the Bussento river basin, located in southern Italy, characterized by soils and rocks with highly different hydraulic permeability and above all a highly hydro-geological conditioning. The proposed approach, which joins together all hydraulic, hydrological and geological data, is able to reproduce the river discharge mean characteristic. Keywords: rainfall–runoff model, water resources management, hydro-geological conditioning.

1 Introduction

Water resources management, more and more limited and poor in quality, represents a present key issue in hydrology. The development of a community is highly related to the management of the water resources available for the community itself and there is a need, for this reason, to rationalize the existing resources, to plan water resources use, to preserve water quality and, on the other had, to prevent flood risk. From this point of view, hydrological models, generating streamflow time series which are statistically equivalent to the observed streamflow time series,




doi:10.2495/GEO060291

become useful tools. When river basins with particular features have to be modeled, both traditionally conceptually based models and more recent sophisticated distributed models appear to give not very reliable results. In those cases it is possible to take advantage of a semi-distributed formulation, where every sub-catchment is modeled to account for its features and information coming from all the sub-catchments are related to each other in order to improve the system description. In this study we propose a semi-distributed conceptually based modeling approach, supported by field measurements collected within several seasonal campaigns, spanning over two years, that has been set up for the Bussento river basin, located in southern Italy, well know to hydrogeology and geomorphology scientists for its karst features, characterized by soils and rocks with highly different hydraulic permeability and above all an highly hydrogeological conditioning. The groundwater circulation is very complex, as it will be later discussed and frequently groundwater inflows from the outside of the hydrological watershed and groundwater outflows toward surrounding drainage systems occur. Even though the proposed approach has some similarity with a few well known conceptually schemes, based on the existence of linear reservoirs and liner channel to describe the different components the streamflow can be decomposed in, it is valuable because of the possibility, which is in this case the necessity, to join all together hydraulic, hydrological and geological data to achieve reliable results.

2 The Bussento river basin geomorphological and hydro-geological features

The Bussento River drainage basin, located in southern Italy, Campania Region, within the Cilento and Vallo di Diano National Park, is well know to geomorphologist and hydro-geologist for its widely and deeply karst features, as summit highland with dolines and poljes, lowland with blind valleys, disappearing streams into sinkholes, cave systems, and karst-induced groundwater aquifers. The main stream originates from Mount Cervati springs (1888 m), one of the highest mountain ridge in the Southern Apennine, then it flows downstream carving steep gorges and rapids, where further springs, along the streambed, increase the river discharge. The upper right area is characterize by marly-arenaceous rocks outcrop (M.nt Marchese hilly ridge), while the left upper area is characterized by limestone sequences (M.nt Rotondo highland and Serra Forcella). Further down, the Bussento river flows into “La Rupe “sinkhole channelling the surface flow into karst cave system and emerging four kilometers downstream, close to Morigerati town, at “Grotta Inferiore del Bussento”. A few hundred meters downstream, the Bussento river merges with the Bussentino creek, originating from eastern sectors of the drainage basin, flowing along canyons and deep gorges carved into Meso-cenozoic limestone sequences. In the western and southern sectors of the basin (Sciarapotamo creek sub-basin), marly-argillaceous successions of the Liguride and “affinità sicilide” Complex (Bonardi G. [5]) dominate the hilly landscape, whereas they underlie




the arenaceous-conglomerate sequences (Guida et al. [6]) at M.nt Centaurino (1511 m).

Figure 1: Hydro-geological scheme of the region (Celico [2]).

Based on the previous comments and given the presence of a complex hydroelectric system, the Bussento river basin cannot be considered a simple drainage basin, but a very complex Hydro-geological System (BHS). A number of homogeneous hydro-geological sectors can be recognized and outlined within it (figure 1).

3 Conceptual hydro-geological modelling

Because of the hydrogeological complexity of the BHS previously presented, a water deep circulation conceptual model has to be outlined before any hydrological computer-aided model can be built. The conceptual model is a preliminary physical-based model, accounting for an interconnected sequence of geologic substrates, permeability distribution, recharge areas and discharge points, that collectively provide a physical scheme of the recharge system, the storage system and the routing system. Karst aquifers modelling is not an easy task. Anderson and Woessner [3] indicate “karst” as one of the advanced topics in the groundwater researches, and summarize a few attempted models, none of which produces reliable results. One




of the earliest attempt of karst aquifers conceptualization (White [9]) focused on the variety of geologic settings and their controlling influence on groundwater flow patterns. This scheme was later expanded (White [11]) to take into account the overall area of the groundwater system. Based only on the type of permeability, Shuster and White [8] divided aquifers into “conduit flow” aquifers, which contain well developed conduit systems, and “diffuse flow” aquifers which do not. Following White [10], the basic components of the generic karst aquifer flow system can be sketched as in figure 3. Clearly, not all of these components are present in all aquifers, and their presence and relative importance is a fundamental point of distinguishing one aquifer from another. With reference to Iaccarino et al. [7] and White [10], this general conceptual model has been applied to the Bussento Hydro-geological System (BHS), recognizing the following recharge-discharge components (figure 2).

Allogenic Recharge

Surficial sinking stream infiltration

Deep sinking stream infiltration

Bedrock stream infiltration

Aquiclude

Deep fracture system

Aquiclude

Deep groundwater flow losses toward the sea

Deep groundwater flow

Diffuse infiltration

Internal runoff infiltration

Vadose Zones

Depression

Epikarst

Conduit System Delayed returns underflow springs

Quick returns overflow springs

Figure 2: Specific conceptual model of the karst aquifers in the Bussento Hydrological System (BHS).

Four sources of recharge for karst aquifers can be recognized: i) allogenic recharge: as surface water collected on outside aquifer basin and injected into the aquifer via sinking streams developed on surrounding aquiclude; ii) internal runoff recharge: as overland flow comes into closed depressions (dolines and poljes) where it enters the aquifer through sinkhole drainage; iii) diffuse infiltration recharge: as precipitation on the land surface, where it infiltrates through the soil and rock, remaining for days or weeks in the vadose epikarst zone, before it migrates downward, through the rock fractures, into the percolation zone, finally reaching the water table into the saturation zone; iv) recharge from fractured bedrock streams: as perched groundwater systems above carbonate aquifers, in which water reaches the main aquifer by means of vadose shafts and open fracture systems along the margins of the perched aquifers and carbonate fractured bedrock streams. A distinguishing feature of Bussento karst




aquifers is that most of the groundwater is discharged through a small number of large springs. Figure 3 illustrate a simplified scheme of the Bussento river network indicating, on a planimetric point of view, the complex interaction between the diffuse springs system, which generally determine an increase in river discharge, and the hydroelectric system, i.e. the anthropic impact, which generally determine a decrease in river discharge, retained and diverted within dams, artificial lake and weirs for human water uses.

-10

+10

+50+100+200+30

(-20) 3

1(-30)(-60)2

+30 +200

+2500+20

9(-20) (-3000)4

[+600]5

-3000 -10-50

+2000

+100

+50+10

+100

Policastro Gulf

+1000[+9000]-100 +5

+10

10

+200

+100 +50

(-50)7

(-30)8

(-600)5

+100

(-500)6

(-600)

+200

Figure 3: Bussento river network and interactions with the springs system and the hydroelectric system.

4 Semi-distributed coupled hydro-geological and hydrological modelling

4.1 The monitoring campaigns

Currently no working river flow discharge measurement stations are present over the catchment. Two stations were actually working, for a short period, over the decade 1960-1970, thus very short streamflow time series are indeed available. For this reason, on January 2003 the Regional Water Basin Authority, Sinistra Sele, started a monitoring campaign with the aim of measure in many different sections and on a monthly time scale the Bussento river discharge. Based on the Bussento catchment (313 km2) geomorphological and hydro-geological features described in the previous paragraphs, 13 stations were indicated as significant to individuate the river regime (figure 4). The monitoring campaigns is currently in progress managed by CUGRI, but so far only a two years periods has been




analyzed. In the following paragraphs some results relative to collected data analysis, particularly referring to modelling applications, are given.

4.2 Data analysis and modelling results

Given the Bussento catchment geomorphological and hydro-geological features described in the previous paragraphs, a lumped model cannot guarantee reliable results. For this reason and taking advantage of the dense monitoring campaign, a semi-distributed formulation, accounting for each sub-basin particular characteristics, seems to be more appropriate.

Figure 4: The Bussento river basin monitoring network.

When dealing with the monthly time scale each sub-basin can be described (figure 5) as two linear reservoirs in parallel, representing the groundwater flow and the deep subsurface flow, whereas the rainfall contributes which are characterized by delay times smaller then a month are supposed to reach the outlet through a linear channel (Claps et al. [1]). The scheme is also supported by the conceptual hydro-geological model described in the previous paragraph. In this case coupling the linear reservoirs balance equations with the whole system balance equation, total streamflow D at each time step is related to the net input by means of an ARMA (2,2) model, which stochastic formulations corresponds to:

)2t()1t()t()2t(D)1t(D)t(D 2121 −+−−=−−−− εΘεΘεΦΦ (1)

#

#

##

#

#

#

##

##

#

#

#

#

##

#

#

#

##

##

#

#

#

#

##

#

#

#

##

##

#

#

#

#

##

#

#

#

##

##

#

#

#

#

##

#

#

#

##

##

#

#

#

#

##

#

#

#

##

##

#

#

B01

B02B03

B04

B05

B07

B_AS02

B_AS04

B_AS05

B01

B02B03

B04

B05

B_AS06

B_AS02

B_AS04

B_AS05

B01

B02B03

B04

B05

B_AS02

B_AS04

B_AS05

B01

B02B03

B04

B05

B07

B_AS02

B_AS04

B_AS05

B01

B02B03

B04

B05

B_AS02

B_AS04

B_AS05

B01

B02B03

B04

B05

B06

B_AS01

B_AS03

B_AS04

B_AS05




where ε is the model residual, related to the net input I, that is then a periodic independent random process, and Φ1, Φ2, Θ1, and Θ2 are the model stochastic parameters, related to the model conceptual parameters K1, K2 (reservoirs response times), a and b (recharge coefficients). In its original formulation the model algorithm, starting from an observed streamflow time series, estimates the model parameters and, because of the univariate approach, with an inverse procedure, the net rainfall input. Collected streamflow data consist of mean monthly values, measured at each section, over two years. These are not enough to run the ARMA(2,2) model with its inverse procedure.

b It (1-b) It

DEEP SUBSURFACE

FLOW

Wt

b It fq

GROUNDWATER FLOW

Vt

a It

a It fk Ck Vt-1

Dt

It

Cq Wt-1

b It (1-b) It

DEEP SUBSURFACE

FLOW

Wt

b It fq

GROUNDWATER FLOW

Vt

a It

a It fk Ck Vt-1

DtDt

It

Cq Wt-1

Figure 5: Linear system of monthly streamflow time series.

To set up a modelling approach able to reproduce observed discharge values, we estimated a priori model parameters and net rainfall input. Equation (1) as been then used to generate 1000 years monthly streamflow time series at each section, comparing thus the discharge probability distributions, at each section and for each month, with the occurred values. The response times have been evaluated from streamflow collected data, applying the base flow recession equation:

K/t0eQ)t(Q −= (2)

whereas the recharge coefficients corresponding to each section, i.e. the outlet of a sub-basin, have been initially assigned on the basis of the relative sub-basin




hydrological and hydro-geological features and successively modified, with an iterative procedure, to achieve the better model results (table 1).

Table 1: Model parameters, response times and recharge coefficients.

Section K1 (days) K2 (days) a 1-a

B_AS 05 - 11.07 0.00 1.00

B 04** 149.71 7.37 0.70 0.30

B 04 153.66 119.75 0.70 0.30

B_AS 03 211.12 10.16 0.70 0.30

B_AS 02 81.21 25.42 0.30 0.70

B 01 280.94 31.12 0.60 0.40

B 02 280.00 32.61 0.70 0.30

B 03 293.69 46.62 0.80 0.20

B_AS 04 108.36 30.51 0.70 0.30

B 06 86.12 35.77 0.70 0.30

B 07 75.11 18.69 0.20 0.80

B_AS 01 - 30.07 0.00 1.00

B_AS 06 301.36 - 1.00 0.00 With regard to the model rainfall net input, the procedure we pursued was to generate it from its probability density distribution, with given parameters. The I(t) probabilistic representation is the Bessel distribution, which is the sum of a Poissonian number of events with exponentially distributed intensity:

0I])I(2[)I/(e)I(f

0Ie]0I[P

2/11

II >ℑ=

===

−−

−

υλυλυλ

υ

(3)

where λ=1/β is the exponential parameter, υ is the Poisson parameter and )x(1ℑ is the modified Bessel function of order 1. The rationale for such

probabilistic representation is given by the positive values and finite probability at zero that I(t) has to present. Parameters β and υ are estimated from the existing two streamflow time series. The temporal patterns found for the two series are rather similar, thus we assumed β and υ spatially invariant over the catchment (table 2).

Table 2: Net rainfall input distribution β and υ parameters.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

υ 1.05 1.19 1.21 1.39 1.48 1.82 2.07 1.66 1.28 1.09 1.04 1.03

β 5.29 2.65 1.52 1.43 6.33 0.73 0.30 0.86 2.85 2.69 1.77 3.63




Figure 6 shows, as an example, the generated streamflow probability distribution, at section B07 for august. Measured values during august 2003 and august 2004 correspond respectively to the 55 percentile and to the 98 percentile. Model performance in 2004 is generally poor because of particular climatic conditions occurred in that year.

101

0.001

0.003

0.01

0.02

0.05

0.10

0.25

0.50

0.75 0.90 0.96 0.99 0.999

river discharge l/sec

Pro

babi

lity

Aug

ust

2003

Aug

ust

2004

Figure 6: Generated streamflow probability distribution compared with

Similar results have been found at the remaining sections, with model adequacy being affected by the degree of anthropic impact and hydro-geological conditioning over the corresponding sub-basin.

5 Conclusion

In this paper we highlight the modeling difficulties that have to be faced when basin with particular features have to be investigated. The case study is the Bussento river basin, located in southern Italy, which is well know to hydrogeology and geomorphology scientists for its karst features, characterized by soils and rocks with highly different hydraulic permeability and above all an highly hydrogeological conditioning. In this case, a lumped methodology cannot guarantee satisfactory results but a semi-distributed formulation can be more appropriate. The methodology we have presented has some similarity with a few well known conceptually schemes, based on the existence of linear reservoirs and liner channel to describe the different components the streamflow can be decomposed in, but it join all together hydraulic, hydrological and geological data. In particular, the hydro-geological catchment’s features have been used to drive the identification of the monitoring sections, i.e. the extension of each sub-basin, and to assign a priori model parameters. Overall, the proposed modeling approach has a good performance, even though it appears in some sections inadequate, acknowledging the poor available database and the complexity of the system we have tried to model.




occurred values over two years.

The authors wish to thank the Regional Water Basin Authority, Sinistra Sele, and the CUGRI, Centro Universitario per la Previsione e Prevenzione dei Grandi Rischi, for their support. The research was partially supported by MIUR and MURST grant.

References

[1] Claps, P., Rossi, F. & Vitale, C., Conceptual- stochastic modelling of seasonal runoff using autoregressive moving average models at different scales of aggregation. Water Resources Research, 29(8), pp. 2545–2559, 1993.

[2] Celico, P. B., Schema idrogeologico dell’Appennino Meridionale; Mem e Note Ist. Geol. Appl., 19 Napoli, 1978.

[3] Anderson, M.P. & Woessner, W.W., Applied Ground Water Modelling. Academic Press, San Diego, 1992.

[4] Bonardi, G., Ciampo, G. & Perrone, V., La formazione di Albidona nell’Appennino calabro-lucano: ulteriori dati stratigrafici e relazioni con le unità esterne appenniniche. Boll. Soc. Geologica Italiana, 104, Roma , 1985.

[5] Bonardi, G., Ciampo, G. & Perrone, V., La formazione di Albidona nell'Appennino calabro-lucano: ulteriori dati stratigrafici e relazioni con le unità esterne appenniniche. Boll. Soc. Geologica Italiana, 104, Roma , 1988.

[6] Guida, D., Iaccarino, G. & Perrone, V., Nuovi dati sulla successione del Flysch del Cilento nell’area di M.te Centaurino: relazioni fra Unità Litostratigrafiche, Unità Litotecniche e principali Sistemi Franosi”. Mem. Soc. Geol., 41, 1988.

[7] Iaccarino, G., Guida, D. & Basso, C., Caratteristiche idrogeologiche della struttura carbonatica di Morigerati. Mem. Soc. Geologica Italiana, 41, 1065-1077, Roma, 1988.

[8] Shuster , E., T. & White, W., B., Seasonal fluctuations in the chemistry of limestone springs. A possible means for characterizing carbonate acquifers. J. of Hydrology, 14, 93-128, 1971.

[9] White, W. B., Conceptual models for limestone acquifers. Groundwater, 7 (3), 15-21, 1969.

[10] White, W. B., Conceptual model for karstic acquifers. Speleogenesis and Karstic Aquifers - The virtual Scientifical Journal, 1, 1-6, 2002.

[11] White, W. B., Conceptual models for carbonatee acquifers: revised. In Dilamarter, R. R. and Casallany, S. C. (Eds): Hydrologic Problems in Karst Terrain. Western Kentucky University, Bowling Green, KY, 176-187, 1977.

[12] White, W. B., Conceptual model for karstic acquifers. Speleogenesis and Karstic Aquifers - The virtual Scientifical Journal, 1, 1-6, 2003.




Acknowledgements

Hydrological modelling of snow cover in the large upper Po river basin: winter 2004 results and validation with snow cover estimation from satellite

D. Rabuffetti, A. Salandin & R. Cremonini

Arpa Piemonte, Italy

Abstract

The study of the hydrological budget of mountainous river basins requires the understanding of the annual cycle of snow accumulation and melting. In fact a deep knowledge of snow cover distribution and dynamic offers several possibilities to improve water resource management and exploitation. The implemented model reproduces an energetic budget of the snowy mantle at small scales using DEM resolution. The use of a distributed model accounts for the high time-space variability of meteorological factors, such as precipitation, air temperature and solar radiation, whose fields at soil level are reconstructed from spot measurements through interpolation procedures based on the topography of the river basin. The modelled variables are the snow water equivalent (SWE) and the discharge generated by the snowmelt while the mechanics of the snowy mantle like thickness and density are not considered. At the same time the simplified degree day method, that uses only air temperature as an index to melt is also implemented and compared. The propagation of melting water inside the snowy mantle is modelled through the conceptual linear reservoir approach; the outflow from snowy mantle is propagated as superficial run-off to the closing section of the river basin through the Muskingum-Cunge hydrological model. The basin of the river Po closed at the section of Ponte Becca is studied in the simulation; it covers a surface of approximately 38000 km2. The simulation ranges from winter 2004 to spring 2005. The modelled snow cover over the catchment is compared with estimation from satellite, derived using the Normalized Difference Snow Index (NDSI), concurring to the validation of the model. Keywords: snow, energy budget, degree day, satellite images, model validation.




doi:10.2495/GEO060301

1 Introduction

In hydrology, the study of accumulated water as snow and its successive release during melting season plays particular importance either in the long period budget and in short flood simulations. In the first case, the knowledge of SWE dynamics allows one to understand the availability of water resource stored in snow, and contributes to improve its management policy. For example, with regard to engineering applications, one can point out the useful information that snow models give for the management of hydroelectric or irrigation reservoirs. In the second case stored water in snowy mantle reduces the available volumes for floods and their peak discharge; flood wave estimation relative to meteoric phenomena in melting season are also improved. This work aims to study the dynamics of the snowy mantle exclusively from a hydrological point of view, with the hypothesis that the parameters of the constituent equations stay constant and that wind doesn’t affect significantly the distribution of snowy mantle in space. In particular a numerical tool for the calculation of SWE and the relative processes of melting and propagation of the outflow is implemented. The validation of the model is obtained through local verifications (qualitative comparison between the measure snowy height and SWE modelled) as well as through the comparison with MODIS satellite images of snow cover. An important feature of this work is linked to the space scale of interest. The large Po catchment and its varied characteristics with the high Alps surrounding the Padana plain represent a significant test case for the hydrological snow model.

2 The hydrological snow model

The application of distributed models is based on the solution of the equations describing the physical phenomena at local level in particular for each elementary unit in which the river basin is subdivided; this allows to describe the hydrological answer to meteorological variables either at local and at catchment scale. The model of the snowy mantle essentially concerns the following processes: accumulation of the snow on the ground, it’s melting and the propagation of the melting water inside of the snowy mantle. The snow model described has been then included into the hydrological distributed model FEST, Mancini et al. [1], for the application at catchment scale.

2.1 Accumulation model

Precipitation measured by rain-gauges can be distinguished by the use of the atmospheric temperature into rain and snow, Tarboton et al. [2].

P)(P

PP

Ps

Pl

αα

−==

1 (1)




where P is the total precipitation volume (millimeter), Pl and Ps are respectively liquid and solid precipitation, αP the fraction of liquid precipitation. αP value is calculated with the successive equations

supaIF

P

supainfIF

infsup

infaP

infaIF

P

TT

TTTTT

TTTT

≥→=

≤≤→−

−=

≤→=

1

0

α

α

α

(2)

where Ta represents air temperature, Tinf and Tsup are two threshold temperatures to be determined during model calibration.

2.2 Energy budget

The sources of energy that drive snowmelt include both shortwave and longwave net radiation, convection from the air (sensible energy), vapour condensation (latent energy), conduction from the ground, and energy income with rainfall; these fluxes are labelled Qsn , Qln , Qh , Qg , Qp respectively. The energy budget equation that describes the energy available for snowmelt is given in eqn (3); the total energy available for snowmelt is Qm (kJ/m2), ∆Qi is the increment of internal energy stored in the snow per unit area of snowpack

ipgehlnsnm QQQQQQQQ ∆−+++++= (3)

In this work the snowpack is considered single-layer with constant medium temperature equal to 0°C. This is equivalent to consider always null variations of the snowy mantle internal energy; this hypothesis is generally considered acceptable in the alpine environment where the thickness of the snowy mantle is elevated for long periods. Sensible and latent energy can be often ignored because quantitatively less important than radiative exchange; moreover this also avoids to estimate wind speed and humidity fields which are affected by high approximation at river basin scale. The amount of snowmelt M (mm of water equivalent) may be expressed by

B.

QM

w

m

ρ9334= (4)

B is the thermal quality of the snow (ratio of heat required to melt a unit weight of snow to that of ice at 0°C), 334.9 (kJ/kg) is latent heat of ice melting, and ρw is the density of water.

2.3 Degree day method

In order to carry out a correct estimation of the snowmelt with a full physical approach, as seen in 2.2, it is necessary to know the time and space variability of




the heat flows as well as of the characteristics of the snow. For large scale applications, the possible lack of accuracy in the necessary information has carried to the development of models that operate with only an index that is a variable used in order to explain a physical phenomenon in integral way. In the calculation of the energetic budget of the snow, the temperature carries out a predominant role and it’s often used like index of snowmelt. The use of the degree-day in the valuation of the energetic contribution to the snowmelt is originally due to Martinec [3]. The melting rate (m/sec) is proportional to the measured temperature described by the introduction of the coefficient Cm

)( bams TTCM −= (5)

where Tb and Ta are respectively base and air temperature. In this work, the value of Cm is kept constant during all the day and the entire simulation. Base temperature and Cm values are obtained from calibration. Generally coefficient Cm ranges from 4.8*10-8 to 6.9*10-8 m/(°C s) and base temperature is 0°C.

2.4 Propagation model

To avoid the complex solution of the equations that describe the motion of the water inside the snowy mantle one can decide to use a conceptual approach, i.e. the linear reservoir cascade. The propagation of liquid fraction of SWE in every cell is modelled as a linear tank; the outgoing flow is proportional, through the use of a constant of time, τ, to the volume of the liquid member in the snowy mantle. The value of the constant τ is determined from the value, commonly found in literature, of the average speed of the water inside of the snowy mantle, equal to 6 (m/h). The propagation of the water inside of the mantle is based on the solution of the mass budget

usinl QQQ

dtdh

−+= (6)

where hl the height of the liquid fraction of the tank, Qin is the flow coming from the upstream cells,

lu hQ τ= (7)

it’s the flow directed to the downstream cell and

SMPQ ls += (8)

it’s the source term that is the sum of the water produced from the snowmelt and the liquid part of the precipitation in that cell. When the propagated flow gets up to the limit of the snowy mantle, it is added to overland flow.




3 The Po river basin

3.1 Geographical data

In this study the Po river basin closed at Becca bridge covers a total area of 38000 km2 including the Aosta Valley and some areas in Liguria, Lombardia and Switzerland. It is situated on the Padana plain and bounded on three sides by mountain chains. Hydrologically speaking, the Upper and Mid Po catchment is varied. The mountainous areas of the Alps provide a complex hydrological regime. In the winter, the Alps are covered in snow and most of the precipitation is stored as snow and glacial depth. In the spring the snow melts, aided by rainfall, which can result in the spring floods. The adopted DEM mesh of 1 km subdivide the river basin in square cells with time-invariant characteristics of the territory such as elevation, aspect and inclination; moreover every cell is characterized with a value of precipitation, temperature and solar radiation obtained from the interpolation of site measurements.

3.2 Meteorological data

There are over 300 rainfall-temperature stations providing measure depths at 30 minute intervals. The model creates hourly rainfall depth surfaces which are constructed from the rainfall data interpolated using the I.W.D. (inverse weight distance) method. To avoid the well known problem of precipitation underestimation due to very low temperatures only gauges with elevation lower than 2000 m asl are considered. The temperature surface maps are constructed from the spot data with adjustment to consider its dependency on elevation. In particular spot measurements are all referred to the same level (1000 m asl) by a constant lapse rate (6.5°C/km) and then interpolated. Finally the temperature of the generic cell is calculated on its real altitude again through the use of the same lapse rate. The calculation of the short waves incident radiation on the single cell happens through a simplified model of the phenomena of transmission, absorption and atmospheric spread also consider the geographical characteristics of the territory, D. Rabuffetti et al. [4]. Over 60 snow-gauge stations provide a measure of height at 30 minute intervals and they are used for validation at local scale.

3.3 Satellite data

The MODerate resolution Imaging Spectroradiometer (MODIS) instrument provides high radiometric sensitivity (12 bit) in 36 discrete spectral bands ranging in wavelength from 0.4 µm to 14.4 µm, Salomonson and Toll [5]. A ±55-degree scanning pattern at the EOS orbit of 705 km achieves a 2330-km swath and provides global coverage every one to two days. Snow has strong visible reflectance and strong short-wave IR absorbing characteristics. The Normalized Difference Snow Index (NDSI), defined as




)TMBandTMBand()TMBandTMBand(NDSI

5252

+−

= (9)

is an effective way to distinguish snow from many other surface features. Pixels that are approximately 50% or greater covered by snow have been found to have NDSI values about 0.4,[6]. Snow cover over eight days is mapped as maximum snow extent in one SDS and as a chronology of observations in the other SDS stored in HDF-EOS format. Eight-day periods begin on the first day of the year and extend into the next year. The product can be produced with two to eight days of input. Using HDFlook-Modis utility, the MODIS/Terra snow products have been extracted and it have been remapped in UTM European Datum 50, to compare with ground observation. The images obtained have a resolution of approximately 400 m, different from the snow model resolution. For the comparison we report these images to model resolution of 1 km considering that each cell is covered by snow if it is so for more than 30% of its surface.

4 Results

The simulation starts from 1-10-2004 until 31-03-2005. Since the interest is to evaluate performance of the model for the annual cycle of snow accumulation and melting, it is assumed that, at the starting time, the snowy mantle is present nowhere. Two validation approaches are used: first the comparison with snow gauges, second with snow cover images obtained from MODIS satellite. It must be underlined that in both cases only a qualitative verification of the model is possible because either snow gauges and satellite provide just an indirect estimation of SWE. Anyway it is important to highlight the very different scales in which these two approaches allow investigating. Figure 1 shows the qualitative comparison in the station of Formazza, taken as an example, with both models. Simulation look fairly good picking the main periods of the snowy mantle accumulation and melting.

Figure 1: Comparison between local measure of snowy height and modelled SWE with the different melting models.




For a general overview, the correlation coefficients (ρ) between observed snowy height and predicted SWE at gauge site have been calculated for some representative stations all around the region.

Table 1: Local validation.

Energy budget model Degree day model Station ρ ρ

Alpe Veglia 0.91 0.48 Capanne Marcarolo 0.71 0.93

Colle Lombarda 0.80 0.60 Formazza 0.84 0.94

Formazza Bruggi 0.92 0.88 Lago Agnel 0.49 0.70 Larecchio 0.89 0.82 Monviso 0.04 0.78

Passo del Moro 0.89 0.93 In general correlation values show a good performance of both models, even if for few sites (Alpe Veglia, Lago Agnel, Monviso) important differences among the models are present, probably due to gauges problems. As far as satellite images are concerned the periods used for the validation are chosen on available data; in particular clear sky periods are considered. Data used covers autumn with high snow precipitation in the mountains, winter with very low precipitation and spring with focus on heavy snowy day in the lowlands. As an example, Figure 2 shows the result of the comparison between model and observation is expressed by means of a contingency matrix. Some classical indices based on this contingency matrix are calculated to assess model performances: threat score (TS) is the statistical measure of accuracy taking into account the numbers of missed and false predictions (the model accuracy improves as the threat score comes closer to 1); bias (B) measures prediction over/under-estimates of snow cover (overestimation for bias scores major than 1 and vice versa); hit rate (HR) represents model capability to predict snow covered areas (best for hit rat equals to 1), Murphy and Winkler [7]. Even though the energy budget model obtains local results as good as the degree day model, it proves to be less satisfactory at distributed level. Table 2 gathers the result for the different analysed periods and the energetic model budget model is better only for the period from 6/3/2005 to 13/3/2005, where a heavy snow fall was registered in the low lands. To allow a deeper understanding of model behaviour, statistics were calculated focusing on the mountain areas (i.e. elevation greater than 1000 m asl). Results resumed in table 3 show a good performance of both models with a better behaviour of the degree day model. This can be addressed to a high




melting rate showed by the energy budget model mainly in the north-eastern areas.

Period: 2005-01-25 – 2005-02-01 Energy budget Degree Day

Period: 2005-03-06 – 2005-03-13

Energy budget Degree Day

Figure 2: Contingency matrix obtained with two models.

Finally, comparing tables 2 and 3, one can clearly notice that model performance in mountain is definitely enhanced than in the whole catchment.




This highlights a poor capability to represent snow dynamics in the lowlands, where a general quick melting is observed in the results.

Table 2: Contingency tables statistics.

Energy budget model Degree day model Period (8 days) TS B HR TS B HR

01/01/2005 0.59 1.49 0.93 0.65 1.00 0.79 25/01/2005 0.63 1.53 0.97 0.70 1.27 0.93 02/02/2005 0.58 1.55 0.93 0.67 1.04 0.82 26/02/2005 0.65 1.32 0.92 0.65 1.36 0.93 06/03/2005 0.57 1.06 0.75 0.55 0.92 0.68 14/03/2005 0.56 0.96 0.71 0.67 1.02 0.81

Table 3: Contingency tables statistics for mountain.

Energy budget model Degree day model Period (8 days) TS B HR TS B HR

01/01/2005 0.78 1.24 0.98 0.74 1.07 0.88 25/01/2005 0.68 1.41 0.98 0.72 1.25 0.94 02/02/2005 0.62 1.47 0.95 0.69 1.04 0.83 26/02/2005 0.76 1.20 0.95 0.78 1.27 0.99 06/03/2005 0.72 1.13 0.89 0.76 1.29 0.99 14/03/2005 0.57 0.95 0.71 0.67 1.02 0.81

5 Conclusion

The application of hydrological snow model at distributed level on the large Po catchment give satisfactory results; nevertheless both the melting models adopted need improving. Probably the characteristics of the snowy mantle, i.e. the albedo dynamics for the energy budget model and coefficient Cm for the degree day model, change during the winter season and require a better description. Local comparison with snow gauges allow a sound validation of the model and shows that where meteorological forcing is known one can reproduce the snowy mantle dynamics with precision either with a physical description of the processes involved either with the degree day conceptual approach. The use of snow cover derived from satellite images offers a further and important means for validation and for understanding model behaviour. Looking at the entire catchment scale, the simplified degree day model picks in a slightly better way the complex dynamics of the snowy mantle in spite of the energetic budget model that needs much more parameters and data. Finally it is important to emphasize that the model results can be used itself for a better exploitation of satellite image. In cloud covered areas satellite cannot




give any information: a jointly usage of model and images can certainly produce better estimates of snow cover. Future developments will be addressed to the extension of the analysis to a longer period in order to obtain more significant statistics. An on-line prototype application will also provide a sound evaluation of practical usefulness of this kind of products.

References

[1] Mancini, M., Montaldo, N. & Rosso, R., La modellazione distribuita nella valutazione degli effetti di laminazione di un sistema d’invasi artificiali nel bacino del fiume Toce, L’acqua, pp. 31-42, 2000.

[2] Tarboton, D. G., Chowdhury, T. G. & Jackson, T. H., A spatially distributed energy balance snowmelt model, Utah Water Research Laboratory, 1994.

[3] Martinec, J., The degree-day factor for snowmelt runoff forecasting. Proc. of general assembly of Helsinki commission on surface waters, IASH publication n° 51, 1960.

[4] Rabuffetti, D., Salandin, A., Volontè, G. & Mancini, M., Modellazione idrologica del manto nevoso. Il caso del lago epiglaciale del ghiacciaio del Belvedere sul Monte Rosa. Atti del XXIX Convegno di Idraulica e costruzioni idrauliche, vol. 3, pp. 867-874, 2004.

[5] Salomonson, V.V., & Toll, D.L., The moderate resolution imaging spectrometer-radar (MODIS-N) facility instrument, Advances in Space Research, 11, pp 231-236, 1991.

[6] MODIS Snow Products, http://modis-snow-ice.gsfc.nasa.gov/sugkc2.html [7] Murphy, A. H. & Winkler, R. L., A general framework for forecast

verification, Monthly Weather Review, 115, pp. 1330-1338, 1997.




Section 8 Landscape analysis

Spatial correlograms and landscape metrics as indicators of land use changes

R. Aunap, E. Uuemaa, J. Roosaare & Ü. Mander Institute of Geography, University of Tartu, Estonia

Abstract

Land use changes over time can be analysed in several ways. We studied the spatial autocorrelation (Moran’s I) of raster format land use maps from three different time periods (1900, 1940, and 2000) in 13 study areas representing most of the landscape regions in Estonia. Human influence was taken into consideration in compiling a scale of the contrast between 10 land use groups. We introduce a simple characteristic based on spatial correlograms: a half-value distance lag, hI=0.5 – a distance where Moran’s I drops below 0.5. No significant change was detected in values of hI=0.5 over time. In addition, we did not detect a difference between lowlands and heights. In analysis of landscape metrics Edge Density (ED), Patch Density (PD), Contrast Weighted Edge Density (CWED), Mean Patch Area Distribution (AREA_MN), and Percentage of Like Adjacencies (PLADJ) showed significant changes comparing the year 2000 with 1900 and 1940. However, the results showed no significant change in landscape metrics between 1900 and 1940. ED, PD and CWED had higher values in 2000 than in 1900 and 1940. Therefore landscape heterogeneity has increased in recent decades. ED, PD, CWED, AREA_MN and PLADJ metrics also indicated a significant difference between lowlands and heights. It appeared that heights have a more heterogeneous landscape structure than lowlands. Generally, the heterogeneity of Estonia’s landscapes overall has changed within recent decades. Keywords: FRAGSTATS metrics, landscape pattern, landscape regions, land use change, Moran’s I, spatial autocorrelation.

1 Introduction

Studies on land use changes is the basic area in landscape research [1, 2] being one of the key issues in global environmental change [3]. Both natural and socio-




doi:10.2495/GEO060311

economic factors have been used in the analysis of land use changes [4, 5]. In the majority of them, the main problem is to correctly characterize the spatial pattern using various landscape metrics [6, 7]. Widely used means to describe landscape texture metrics can be calculated with the help of FRAGSTATS [8]. It has been shown [9, 10] that these metrics are scale-dependent, and not all indexes demonstrate a regular behaviour in relation to scale changes. Despite the great number of indexes, FRAGSTATS does not include measures of variography, e.g. different spatial structure functions such as correlograms and variograms, which are popular in geostatistics [11], and describe the dependence of variability on distance. In order to study landscape heterogeneity and spatial autocorrelation, correlograms are preferred to semivariograms, since – according to Legendre and Fortin [6] – they are standardized and make it possible to compare different landscapes. This method is used by Radeloff et al. [12] to study artificial landscapes with a regular pattern, in order to detect periodicity in their correlograms. The behaviour of a correlogram’s wavelength and amplitude within a specific range of spatial orders can be used as an indicator of spatial pattern [13]. A classical estimator of spatial dependence is Moran’s I, “associated with statistician P.A. Moran (1948)” and proposed as the spatial analogy of autocorrelation used in time series analysis [14]. Since the introduction of the autocorrelation index by Moran [15], spatial correlograms have been used for the spatial analysis of several natural and social phenomena. At present, various studies use Moran’s I correlograms to avoid systematic mistakes due to spatial autocorrelation in spatial analyses [6, 16, 17, 18], or using the correlograms for ecological and landscape analyses at different spatial scales. Koenig [19] showed the importance of the Moran effect and spatial autocorrelation (environmental synchrony) for the analysis of patterns of animal populations at continental and global scale. Likewise, Diniz-Filho et al. [20] calculated the Moran’s autocorrelation value when analysing avian populations at continental scale. Large-scale (100-1000 km) analysis of Moran’s I correlograms was carried out for the investigation of vegetation pattern dynamics in the Great Lakes region during the Holocene [21], for the prediction of deforestations in Saskatchewan [22], for the analysis of human land transformations in South African avian diversity [23], for the identification of operational units for conservation in continuous populations [24], and for the analysis of alien plant invasion in Catalonia [25]. At the local and landscape scale (10-1000 m), Moran’s I was used for analyses of the distribution pattern of several bird and mammal species [26], carabid beetles [27], Neotropical migrant songbirds [17], the impacts of logging in Amazonian forests [28], urban spatial features [29], the variability of soil properties in wetlands [30], and the spatial patterns of greenhouse gas emissions in tropical rainforests [31]. Surprisingly, we were able to find only two papers in which Moran’s I statistic has been used in connection with changes in land use/cover [32] or vegetation cover [33]. Read and Lam [32] have found that Moran’s I is effective




at detecting changes in land cover types in Costa Rica’s lowlands, whereas FRAGSTATS indexes were not sufficient in this sense. The main objectives of this study were: (1) to analyse three map series (from approximately 1900, 1940, and 2000) of selected landscape areas in Estonia concerning their differences in spatial autocorrelation and FRAGSTATS indexes; (2) to find out whether the Moran’s I characteristic and landscape indexes respond to the land cover changes.

2 Material and methods

2.1 Study areas

Thirteen study areas were selected on the basis of Estonian landscape regions so that they represent most Estonian landscape types (Fig. 1). The selection of study sites was based on Uuemaa et al [34].

Figure 1: Study areas and landscape regions of Estonia.

Of 13 study areas, there were sites dominated by agricultural land use, forests, bogs or urban areas. Study areas were formed on the basis of Estonian Basic Map Sheets. Each study area was 5*5 km.

2.2 Land use data

Land use data was derived from three maps from different time periods: 1:42,000 (from Russian topographic map sheets dating from 1886-1917; later referred to as “1900”), 1: 50,000 (topographic map sheets published by the Estonian




Military Topo-Hydrographic Department, in 1935-1939; later referred to as “1940”), and 1:20,000 (Estonian Basic Map sheets from 1998-2004; generalised to 1:50,000 scale; later referred to as “2000”) and converted into raster format using 10 m pixel size.

Figure 2: Scale of the contrast of main land use types.

For the generalisation of the 1:20,000 Estonian Basic Map sheets into the 1:50,000 scale, we used the MapInfo tools Polygon Area Thinning and Gap Removal. The minimum recognisable area was set to 0.4 ha. The maps of the test sites were scanned (except for the Estonian Basic Map, which is already in digital form), digitised, and rasterized. Ten land use types distinguished from all map series were reclassified so that new type numbers could be used as contrast indexes (Fig. 2). In the case of land use types Mi and Mj, their difference (|i-j|) shows the contrast between these types.

2.3 Moran’s I

In our analysis we applied the Idrisi Kilimanjaro software [35]. A module named AUTOCORR calculates the first-lag autocorrelation coefficient of an image. The following equation, which is similar to the usage in other software, is applied:

( )

−

−−⋅=

∑∑ ∑

∑∑

=≠

= =

n

iiji ij

n

i

n

jjiij

yw

yywnI

1

2

1 1

)(

))((

µ

µµ (1)

where n – number of values to be taken into account (in the case of a raster image, pixels); w – spatial weights: 1 in the directions up/down/left/right, 0.70711 (square root of 2) as a weight for the diagonal neighbouring pixels; yi/j – value of pixel i resp. j; µ – mean of values y [36]. In addition to Moran’s I, Idrisi also calculates several statistics including tests of significance under two null hypothesis assumptions. For raster images, the autocorrelation has been calculated with all appropriate pixels using so-called King’s case analysis [35].

1 2 3 4 5 6 7 8 9 10

Bogs, fens

Paludified

forests

Forests

Bushes

Grasslands

Arable lands

Clear-cut

forests

Mining areas

Urban areas

Water




Using auxiliary images (the CONTRACT module with so-called pixel thinning), we computed the 1st, 2nd, 3rd etc. lags of Moran’s I value. Accordingly, we found all necessary Ih, changing h as a multiple of pixel resolution. In our investigation, the size of the pixel side was 10 m, and for the test sites we calculated series of Ih, h=10, 20, 30, …, 100, 120, …,200, 300, 400, 500 and 1000 m. We used the results to construct the graphs of I(h), called correlograms, which ideally are monotonically decreasing curves. Since n is very large (tens and hundreds of thousands), all I(h), except for some I(1000), are statistically significant. We investigated the correlograms of the test areas based on 3 map series from different periods, and found these to be quite regular. In order to compare Moran’s I correlograms from different test areas and different map series and also with other landscape metrics derived from FRAGSTATS, we introduced a simple characteristic of half-value distances: hI=0.5 – the distance lag where Moran’s I drops below 0.5 [34].

2.4 Landscape metrics

Using FRAGSTATS 3.3, landscape metrics were calculated for all study sites. We calculated the following landscape metrics: 1) Edge Density (ED); Patch Density (PD); Mean Patch Area Distribution (AREA_MN); Mean Shape Index (SHAPE_MN), Contrast Weighted Edge Density (CWED); Percentage of Like Adjacencies (PLADJ); Contagion (CONTAG) and Shannon’s Diversity Index (SHDI). For details and metrics formulae see McGarigal and Marks [8].

2.5 Statistical analysis

According to the Kolmogorov-Smirnov test for normality, all of the variables under consideration were normally distributed. The homogeneity of variances was verified using the Cochran C and Levene tests. In correlogram analysis, when comparing different groups we used a one-way ANOVA (Tukey’s HSD test). In landscape metric analysis, most of the variables did not meet the analysis of variance assumptions (Levene and Cochran C tests). Therefore, the significance of differences was analysed using the non-parametric Kruskal-Wallis test. For the statistical analysis of all data, the computer program STATISTICA 7.1 was used. The level of significance of α = 0.05 was accepted in all cases.


3.1 Moran’s I correlograms

We did not detect statistically significant land use changes during the years 1886-2004. The overall significance of the model was >0.3, and also Tukey’s HSD test did not show significant differences between group means. Therefore we could say that landscapes overall have not changed significantly over the past 100 years in Estonia. However, Figure 3 shows that the mean of the hI=0.5




decreased from 250 in 1900 to 160 in 2000 (Fig. 3). Also, the variation is smaller in 2000. Therefore we could say that human influence has made landscapes slightly more heterogeneous. We could even claim that land use in some areas did change dramatically. In Alutaguse paludified lowland, for example, the forests and bogs were turned into mining areas during the last 50 years (Fig. 4). However, the heterogeneity of the landscape has decreased (Fig. 5). In 2000 the spatial autocorrelation is highest, and in 1940 lowest. In the case of West Estonia it is vice versa (Fig. 5). In recent decades the spatial autocorrelation has decreased, i.e. human influence has increased the heterogeneity of landscape. In the case of all heights except Vooremaa, the spatial autocorrelation had decreased in recent decades. Landscapes were more heterogeneous in 2000 than they were a hundred years ago.

1990 1940 20000

50

100

150

200

250

300

350

400

450

I h=0

,5

Mean Mean±SD

lowlands heights50

100

150

200

250

300

350

400

450

I h=0

,5

Mean Mean±SD

Figure 3: Average and standard deviation values of Moran’s Ih50 of 3 map

series (1900, 1940, 2000), and heights and lowlands over all 13 test sites.

1900 1940 2000

Figure 4: Land use change in the test area of the Alutaguse 2 paludified lowlands. For numbers of land use types, see Table 1.

We also found that there is no significant difference between Ih50 in heights and lowlands as could be expected based on the results obtained by Uuemaa et al. [34]. The Tukey HSD test did not show a statistically significant difference between heights and lowlands. Nevertheless, it can be seen from Figure 3 that the mean of the Ih50 is 155 in the case of heights and 250 in the case of lowlands. The variation of Ih50 is also smaller. This shows that the spatial autocorrelation of




heights is lower than the autocorrelation of lowlands, i.e. lowlands are more spatially homogenous. This can also be seen from Fig. 5, where the decrease of correlograms is more abrupt in the case of heights (Vooremaa and Otepää) than in the case of lowlands (Alutaguse 2 and West-Estonia).

Alutaguse 2

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000lag(m)

Mor

an's

I

190019402000

Vooremaa

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000lag(m)

Mor

an's

I

190019402000

West -Estonia

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000lag(m)

Mor

an's

I

190019402000

Otepää

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000lag(m)

Mor

an's

I

190019402000

Figure 5: Spatial correlograms of 3 map series (1900, 1940, 2000) from 4 test

areas.

Interestingly, we found that near 180m and 400m of lag distance, a “jump” appeared on almost each correlogram (Fig. 5). This phenomenon is probably related to the periodicity of Moran’s I correlograms, which was, however, only detected at larger scales [12].

3.2 Landscape metrics

Although no change in Ih50 values appeared in the analysis of correlograms, statistically significant changes were detected in several values of landscape metrics. The average values and standard deviations in the years 1900 and 1940 are very stable (Table 1). In 2000, remarkable changes can be detected in values of PD, ED, AREA_MN, CWED and PLADJ. These metrics also gave statistically significant changes in the Kruskal-Wallis test (Table 2). No change was detected between 1900 and 1940, which also confirms the results of Palang et al., [37]. ED, PD and CWED had increased in 2000 compared to earlier dates (Fig. 6). This shows that heterogeneity has increased in recent decades. AREA_MN and PLADJ had decreased in recent decades, which also indicates the increase in landscape heterogeneity. We also tried to identify differences between heights and uplands. The results showed that PD, ED and CWED values are significantly lower in the lowlands (Table 2 and Fig. 6), i.e. the landscape structure is more homogenous than in the




heights. Palang et al. [37] obtained similar results: in southern Estonia, where all the heights are located, test sites had higher heterogeneity. AREA_MN and PLADJ had lower values in heights (Table 2), i.e. landscape patches are smaller, and therefore the landscape has a more complex structure than in the lowlands.

Table 1: Average values and standard deviations of measured landscape metrics. *significant difference with 2000 (p<0,01).

1900 1940 2000

PD 2.66±1.78* 3.26±1.75* 15.03±7.15 ED 46.467±20.73* 51.94±19.75* 100.83±28.85 AREA_MN 56.97±42.56* 38.58±18.63* 8.20±3.81 SHAPE_MN 1.88±0.28 1.76±0.12 1.81±0.24 CWED 19.43±7.57* 22.47±7.65* 44.01±12.96 CONTAG 63.384±7.86 60.14±6.02 61.82±6.44 PLADJ 97.48±1.04* 97.20±0.99* 94.76±1.44 SHDI 1.3±0.30 1.39±0.22 1.28±0.22

Table 2: Changes in landscape metrics agewise and difference between lowlands and heights. Estimated by Kruskal-Wallis test (significance levels ***p< 0.001; **p<0.01; *p<0.05).

Agewise Lowlands and heights

PD 1900 and 1940<2000*** lowlands<heights* ED 1900 and 1940<2000** lowlands<heights** AREA_MN 1900 and 1940>2000*** lowlands>heights* SHAPE_MN no difference no difference CWED 1900 and 1940<2000** lowlands<heights* CONTAG no difference no difference PLADJ 1900 and 1940>2000** lowlands>heights** SHDI no difference no difference

Summarizing the results of this study, the heterogeneity of Estonian landscapes has changed over the last 50 years, according to landscape metrics analysis. However, the spatial autocorrelation of landscapes has not changed significantly, but it also showed decreasing trends in recent decades. The results of this analysis were not unequivocal, because in some study areas the heterogeneity had increased, and some study areas had become more homogeneous over time. Thus the overall change in Estonian landscapes is not so remarkable.




1900 1940 20000

20

40

60

80

100

120

140

160

ED

(m

/ha)

Median 25%-75% Min-Max

lowlands heights0

20

40

60

80

100

120

140

160

ED

(m/h

a)

Median 25%-75% Min-Max

Figure 6: Median, quartiles, maximum and minimum values of ED of 3 map

series (1900, 1940, 2000), and heights and lowlands over all 13 test sites.

4 Conclusions

The results of the study demonstrated that the average value of spatial autocorrelation in Estonian landscapes has not significantly changed over time. We were also unable to find a significant difference between spatial autocorrelation in heights and lowlands. We propose the distance (lag) of spatial correlograms at which the Moran’s I value reaches 50% of the maximal value (hI=0.5) as a new landscape metric for the characterization of landscape pattern. Its benefit is its simple interpretability and the independence of the scale. Thus this characteristic can effectively be used as an indicator in landscape planning and management. Although the analysis of correlograms did not show significant change over time, several landscape metrics indicated that landscapes are more heterogeneous in 2000 than they were in 1900 or 1940. There was also a statistically significant difference between heights and lowlands, latter being more homogeneous.

Acknowledgements

This study was supported by Estonian Science Foundation grant No. 6083 and Target Funding Project No. 0182534s03 of the Ministry of Education and Science of Estonia. We acknowledge Sten Mander for his help in digitizing maps.

References

[1] Forman, R.T.T., Godron, M., Landscape Ecology, John Wiley and Sons, New York, 619 p., 1986.

[2] Zonneveld, I., Land Ecology, SPB Academic Publishing, Amsterdam, 199 pp., 1995.

[3] Alcamo, J., Leemans, R., Kreileman. E., Global Change Scenarios of the 21st Century, Results from the IMAGE 2.1 Model, Elsevier Science Publishers, London, UK, 287 pp., 1998.




[4] Wear, D.N., Bolstad, P., Land-use changes in southern Appalachian landscapes: Spatial analysis and forest evaluation, Ecosystems, 1(6), 575-594, 1998.

[5] McDonald, R. I., Urban, D.L., Spatially varying rules of landscape change: lessons from a case study, Landscape Urban Plan. 74(1), 7-20, 2006.

[6] Legendre, P., Fortin, M.-J., Spatial pattern and ecological analysis, Vegetatio, 80(2):107–138, 1989.

[7] Gustafson, E.J., Quantifying landscape spatial pattern: What is the state of the art? Ecosystems 1(1), 143-156, 1998.

[8] McGarigal, K., Marks, B.J., FRAGSTATS: spatial pattern analysis program for quantifying landscape structure, USDA Forest Serv. Gen. Tech. Rep. PNW-351, 1995.

[9] Wu, J., Shen, W., Sun, W., Tueller, P. T., Empirical patterns of the effects of changing scale on landscape metrics, Landscape Ecol. 17(8), 761-782, 2002.

[10] Uuemaa, E., Roosaare, J., Mander, Ü., Scale dependence of landscape metrics and their indicatory value for nutrient and organic matter losses from catchments, Ecol. Indic. 5(4), 350-369, 2005.

[11] Cressie, N.A.C., Statistics for Spatial Data, Revised Edition. Wiley, New York, 1993.

[12] Radeloff, V.C., Miller, T.F., He, H.S., Mladenoff, D.J., Periodicity in spatial data and geostatistical models: autocorrelation between patches, Ecography 23(1), 81-91 2000.

[13] Chou, Y.H., Spatial pattern and spatial autocorrelation, Lect. Notes Comput. Sc. 988, 365-376, 1995.

[14] Taylor, P.J., Quantitative methods in geography, Houston Mifflin Company, Boston 1977.

[15] Moran, P.A., The interpretation of statistical maps, J. Royal Stat. Soc. B 10, 243–251 1948.

[16] Legendre, P., Spatial autocorrelation: Trouble or new paradigm? Ecology, 74(6): 1659-1673, 1993.

[17] Lichstein, J.W., Simons, T.R., Shriner, S.A., Franzreb, K.E., Spatial autocorrelation and autoregressive models in ecology, Ecol. Monogr. 72(3), 445-463, 2002.

[18] Overmars, K.P., de Koning, G.H.J., Veldkamp, A., Spatial autocorrelation in multi-scale land use models, Ecol. Model. 164(2-3), 257-270 2003.

[19] Koenig, W.D., Global patterns of environmental synchrony and the Moran effect, Ecography 25(3), 283-288, 2002.

[20] Diniz-Filho, J.A.F., Bini, L.M., Hawkins, B.A., Spatial autocorrelation and red herrings in geographical ecology, Global Ecol. Biogeogr. 12, 53-64, 2003.

[21] Graumlich, L.J., Davis, M.B., Holocene variation in spatial scales of vegetation pattern in the upper Great Lakes, Ecology 74(3), 826-839, 1993.




[22] Hobson, K.A., Bayne, E.M., van Wilgenburg, S.L., Large-scale conversion of forest to agrculture in the boreal plains of Saskatchewan, Conserv. Biol. 16(6), 1530-1541, 2002.

[23] Fairbanks, D.H.K., Kshatriya, M., van Jaarsveld, A.S., Underhill, L.G., Scales and consequences of human land transformation on South African avian diversity and structure, Anim. Conserv. 5, 61-73, 2002.

[24] Diniz, J.A.F., Telles, M.P.D., Spatial autocorrelation analysis and the identification of operational units for conservation in continuous populations, Conserv. Biol. 16(4), 924-935, 2002.

[25] Pino, J., Font, X., Carbo, J., Jove, M., Pallares, L., Large-scale correlates of alien plant invasion in Catalonia (NE of Spain), Biol. Conserv. 122(2), 339-350, 2005.

[26] Koenig, W.D., Knops, J.M.H., Testing for spatial autocorrelation in ecological studies, Ecography 21(4), 423-429, 1998.

[27] Judas, M., Dornieden, K., Strothmann, U., Distribution patterns of carabid beetle species at the landscape-level, J. Biogeogr. 29(4), 491-508, 2002.

[28] Read, J.M., Spatial analyses of logging impacts in Amazonia using remotely sensed data, Photogramm. Eng. Rem. S. 69(3), 275-282, 2003.

[29] Myint, S.W., Kam, N.S.-N., Tyler, J.M., Wavelets for urban spatial feature discrimination: Comparisons with fractal, spatial autocorrelation, and spatial co-occurrence approaches, Photogramm. Eng. Rem. S. 70(7), 803-812, 2004.

[30] Bruland, G.L., Richardson, C.J., Spatial variability of soil properties in created, restored, and paired natural wetlands, Soil Sci. Soc. Am. J. 69, 273-284, 2005.

[31] Ishizuka, S., Iswandi, A., Nakajima, Y., Yonemura, S., Sudo, S., Tsuruta, H., Muriyarso, D., Daniel Spatial patterns of greenhouse gas emission in a tropical rainforest in Indonesia, Nutr. Cycl. Agroecosys. 71(1), 55-62, 2005.

[32] Read, J.M., Lam, N.S.-N., Spatial methods for characterising land cover and detecting land-cover changes for the tropics, Int. J. Rem. S. 23(12), 2457-2474, 2002.

[33] Landis, A.G., Bailey, J.D., Reconstruction of age structure and spatial arrangement of piñon-juniper woodlands and savannas of Anderson Mesa, northern Arizona, Forest Ecol. Manag. 204(2-3), 221-236, 2005.

[34] Uuemaa, E., Roosaare, J., Kanal, A., Mander, Ü., Spatial correlograms of soil cover as an indicator of landscape heterogeneity, Ecol. Indic. Submitted, 200X.

[35] Eastman, R., IDRISI Kilimanjaro: Guide to GIS and Image Processing, Clark University, 2003.

[36] Single Raster Analysis Tools VII, http://www.sbg.ac.at/geo/idrisi/wwwtutor/s_tools7.htm [24.11.05], 1999.

[37] Palang, H., Mander, Ü., Luud, A., Landscape diversity changes in Estonia, Landscape Urban Plan. 41(3-4), 163-170, 1998.




The role of geosciences and landscape in the management of Natural Parks of Guadalajara (Central Spain): in search of suitable applications

A. García-Quintana, M. P. Abad, M. Aguilar, L. Alcalá, I. Barrera, M. Cebrián, M. C. Fernández de Villalta, J. F. García-Hidalgo, R. Giménez, A. E. Godfrey, J. A. González-Martín, A. Lucía, J. F. Martín-Duque, M. Martín-Loeches, E. Quijada, J. M. Rodríguez-Borreguero, R. Ruiz López de la Cova & A. Solís PANAGU Project, Complutense University, Spain

Abstract

University professors, researchers and professionals from different governmental bodies in Madrid and Guadalajara have formed a research group to investigate how to use geo-environmental information in the management of Spanish Natural Parks. Funding has been obtained from the Spanish Ministry of Science and Technology. The pilot area for the project is the Natural Parks of the province of Guadalajara in the Castilla – La Mancha Autonomous Region. This article describes two proposals to analyse the physical environment and the landscape of the natural parks of Guadalajara: analytical-parametric inventory, and synthetic-physiographic inventory. The objective is to provide useful information on how to manage these areas. Additionally, the article includes results of the study of active karstic processes in two of these natural parks, Río Dulce and Alto Tajo. Keywords: natural parks, landscape, environmental geology, karstic processes, Guadalajara.

1 Introduction

Landscape and geology are often the most important factors leading to the designation of an area as a national or natural park. This has been the case since the beginning of the national park movement in North America (Yellowstone




doi:10.2495/GEO060321

1872, Waterton Lakes 1885) and in Spain (Picos de Europa, Ordesa and Monte Perdido, 1918). However, once a park is established, geological and landscape value tends to be relegated to a secondary role in the daily management of such places. Growing interest in the preservation of fauna and flora during the second half of the 20th century, the ambiguity of the term “landscape”, and a lack of general knowledge about geology on the part of Mediterranean societies have all been obstacles to parks’ maximizing and managing these resources. Faced with this reality, studies on geology and landscape appear to be undervalued and underutilised in the management of natural places in many countries, among them Spain. This limits the chances that societies will protect, enjoy, and value these areas. Nevertheless, the information from such studies can provide: knowledge about active processes, conceptual tools, classifications, and territorial compositions, predictions, evaluations, etc. This perceived deficit is the starting point for the PANAGU project. Landscape is the “meeting point” of the project’s component parts, as it is the most conspicuous and highly valued element in most natural parks. We believe that analysing geology and geomorphology can provide a well-informed understanding of landscape. Landforms (bedrock, topography and processes) affect microclimate conditions, soils, vegetation, land use and the visual composition of many of the natural parks of the Iberian Peninsula. For this reason, although the research team is led by geologists and geomorphologists, other specialists on landscape are also involved.

2 Geological and geographical setting

The province of Guadalajara is situated on the Meseta, a region composed of wide barren plains and mountain chains. This morphological unit occupies the greater part of the Iberian Peninsula; it has an average altitude of 600 m.a.s.l. and constitutes the Peninsula’s dominant geographic element. Three types of landscape stand out: mountains, open plains, and gorges cut into the plains. The gorges are the result of the big difference in altitude between the central zone of the Meseta and sea level. The three natural parks of Guadalajara offer fine examples of these landscapes. The province of Guadalajara is shaped by the three large regional geological units that intersect there: the Central System, the Iberian Chain and the Tajo Basin. The Central System is a reactivation and uplift of the ancient Variscan Massif, composed predominantly of metamorphic and igneous rocks. The Tejera Negra is located in this Central System. It is quite small (1,600 ha) and was created to protect a relic beech tree (Fagus selvatica) forest, although its name is due to the presence of stands of yews (Taxus baccata, in Spanish tejeras). The park has relatively simple topography shape and a fairly uniform geological composition. It is characterised by sharp peaks and crests that rise above 2000 m.a.s.l., sculpted from quartzite and black slate of Ordovician age. It is poorly developed, far from large population centres, and has low human impact. At present, there is a proposal to considerably expand the park’s size due to the beauty and untouched nature of much of the surrounding area.




The Iberian Range is a foreland mountain chain, with a very thick sedimentary layer predominantly of carbonate rocks. Due to intense erosion during the Neogene, today it appears as a set of barren plains from which some residual mountain chains jut out. These mountains control the shape of the stream network. The Río Dulce and the Alto Tajo are situated in the Iberian Range. Both parks are characterised by narrow, deep canyons and river gorges which in places can drop to 500 m below the plains and plateaus, located at 1,100 m above sea level. The Río Dulce Natural Park (literally, “Sweetwater River”, because it does not have a high percentage of salts, unlike other rivers in this region which have been exploited for their salts since the days of the Roman Empire because they drain salty Triassic formations) is small in size (8,300 ha) and, although its main landscape types are gorges and plains, it has a more complex topographic structure. The plains have different shapes and land uses as they are developed on substrates of different lithologies and structures. The same can be said of the Rio Dulce and its tributaries. This has led to a diversity of landscapes. There has also been greater human impact, thanks to its location near the historic city of Sigüenza, today a magnet for cultural and gastronomic tourism. Since the Middle Ages, the area has been well travelled, first by two drovers’ routes and royal roads that passed close by. Today it is flanked by a national highway, a regional highway, and it can be traversed by a local road. The Alto Tajo natural park (literally, “High Valley” of the “Cut River”) is larger (106,000 ha). It constitutes one of the most unique natural areas on the Iberian Peninsula so its designation as a national park is under consideration. The park is famous for its deeply incised fluvial network. There are more than 100 km of spectacular canyons, deep gorges, and narrow valleys with steep slopes, some covered by thick forests. To this scientifically and aesthetically interesting geo-morphologic setting one must add geological, plant and ethnographic formations that make the park a valuable national heritage site: unique large sedimentary structures, juniper groves (Juniperus thurifera), and “pallozas” (primitive dwellings of pre-Roman technology).

Figure 1: Location of the three natural parks within the regional geological

units of Guadalajara and the Iberian Peninsula. DB, Duero Basin; IR, Iberian Range; CS, Central System; TB, Tajo Basin; EB, Ebro Basin; BR, Betic Range; GB, Guadalquivir Basin; HM, Hercynian (Variscan) Massif, CPR, Cantabrian Pyrenees Range.




3 The landscape study of the natural parks of Guadalajara

The primary and operative objective of the PANAGU Project is to carry out an analysis of the landscape of the protected natural areas of the province of Guadalajara (Castilla-La Mancha). This analysis is first and foremost cartographic and digital and is being performed following two concepts.

3.1 Analytical or parametric method

The analytical or parametric method examines a series of maps and thematic reports, whose partial results are integrated to establish the true relationship between different elements of the environment. 3.1.1 Lithology and tectonic structure In this research, rocks are studied not so much for their own sake but rather as to how they condition ecosystems, human uses and landscape. Therefore, using a standard 1/50,000 geological map, lithological maps were compiled that grouped rock units with a high degree of physical and chemical homogeneity, and whose ecological and visual response were also homogeneous. A subsequent evaluation of these units allowed the capabilities and limitations of use to be established as well as their relationships with other environmental factors. For each scenic landscape site, the size, orientation and form of each topographic element, as well as the surface area, intensity and type of land uses and the placement of human settlements and developments can be explained in the litho-structural configuration of the different plots in the territory. Each body of rock provides at least one distinguishing property to the landscape: colour, alteration, texture, permeability, forms of erosion, etc. The tectonic structure affects the park’s landscape on a larger scale, such as: surface forms, folded areas with irregular relief, valleys oriented along faults, etc. Detailed analysis of the parks’ territory by a large group of experienced professionals revealed geological formations that should be of aesthetic interest to the public, and of scientific interest as national treasures, but which at the moment are underappreciated. These include ancient valleys with hanging formations, karst systems cut by modern valleys, prehistoric landslides, etc. 3.1.2 Relief and topography The study of landforms plays a fundamental role in the study of landscape. From a strictly visual standpoint, landforms mould both the viewpoints and the view shed. Within the framework of the study of “natural scenarios”, it seems necessary to make an objective classification of the landforms and their characteristics if we want to be able to measure, correlate and study a landscape. In so doing we facilitate comparison of the results of the different studies. The way to accomplish this objective is through a mathematical manipulation of numerical data that can be computer processed. Therefore, we use the Digital Elevation Models (DEM), as an approximation of reality. In general, the DEM creates a representation of physical phenomena, in this case the ground surface, from a spatial, geometric point of view. This digital




representation is, by definition, only an approximation to reality through a finite sample of values and a set of functions. In order to achieve the best adaptation possible to the phenomenon it represents, the project used the geodesic reference system (ED50) and the UTM projection (huso 30), the official system of Spanish cartography. The DEM has a 25 m grid spacing, and the data are referenced to the Alicante sea level datum. Accordingly, the DEM fulfils two functions. On the one hand, it computes relief and its derivatives (slopes, orientations, curvatures, etc.). These characteristics constitute a Digital Terrain Models (DTM). On the other hand, the DEM serves as a basis for geo-referencing other thematic maps. The project will attempt, on the basis of this modelling of the landscape features of the three natural parks of Guadalajara, to acquire a fuller knowledge of the physical condition and construct a tool to improve management and planning of these protected natural places. All this information is being processed with the help of appropriate software, a Geographical Information System, whose purpose is to optimise and facilitate analysis of the data. 3.1.3 Hydrogeology Understanding groundwater is very important to the management of the Río Dulce and Alto Tajo natural parks. Most of both parks are underlain by karstic aquifers, highly susceptible to contamination. Hydrogeological research is focused on the precise localisation of the groundwater recharge areas, as these are the points where potentially contaminating agricultural or livestock activities can be regulated. In some cases recharge occurs outside the park boundaries, so that the mapping of these areas is necessary to justify their protection. The precision of this research includes carrying out classic tracer tests. The protection limits should also take the discharge points into account, some of which appear to be independent of the river basins. Also important in their implications for the management of these two parks are the hydro-geochemical processes in the Keuper facies that result in mineralized springs. The deposited minerals are essential to plant growth and attract certain animals in search of water and dietary salt. The metamorphic bedrock and mountainous relief which predominate in the Tejera Negra natural park prevent any significant infiltration of precipitation. However, we intend to monitor the permeability of the numerous debris flows, scree slopes and areas of surface alteration to determine their contribution to base flow and response times. 3.1.4 Soils Soils are one of the more dynamic components of any landscape. They are constantly being formed by physical, chemical, and biologic processes. They are essential to the maintenance of a healthy ecosystem, especially in these natural parks. Land management without adequate knowledge of soils and their correct handling have in many cases destroyed what they intended to conserve. Such actions have damaged the entire natural environment and increased erosive




activity on the remaining resources, with the consequent modification of the landscape. This is especially true in these areas of the Iberian Peninsula where the severity of current climatic favours intense soil destruction. Conserving the soil resources depends on a complex combination of economic factors, on the knowledge of conservation means and techniques, and on the possibilities within the social environment which favour it. For this reason, the soil surveys of the three natural parks in Guadalajara consist of a soils mapping and the evaluation of how vulnerable they are to physical and chemical degradation. 3.1.5. Vegetation The plant cover often hinders the observation of underlying geological elements. However, the specialised observation of the plant cover allows the geological characteristics hidden from sight to be deduced. The relationship and correspondences between geology and botany have long been dealt with and written about from very different viewpoints: from those who consider geology as a secondary factor [1] to those who put it first when justifying the discontinuities and even the vegetation typology [2]. The most obvious approximations, such as those in cases of azonal vegetation [3], or in the ridge crest-slope-valley floor models, are relatively easy, but more complex approaches can lead to difficulties. The discrepancies and difficulties are due above all to the web of complex relationships within the natural environment and also to the different rhythms of geological and vegetation processes. When geology is cited as the vegetation controller there are many factors involved, for example topography, lithology, chemical composition, pH, joints, soil development... Further, each factor can vary in importance depending on the case. Some medium-scale (1/50,000) studies carried out by drawing up independent thematic mappings have produced only partly satisfactory results [4]. The destruction of the plant cover and its spontaneous recovery tendency, are the origin of the mosaics of different plant communities existing within geologically uniform areas because of differing regeneration requirements. But in most cases, the botanical component of the landscape mirrors the history of land uses. To overcome these problems requires the use of mapping units which combine arrangement and dynamism [5], grouping complexes of plant communities linked to each other through dynamic relationships or “eucomplexes”, and separating them from the mosaics of communities without these relationships (the so-called “geo-complexes”). Within this framework, one of the aims of the PANAGU project is to discover the spatial and mapping coincidences between these units, eucomplexes, catenas and mosaics, and the synthetic units defined by the geo-edaphic elements which model the landscape, lithology, topography and processes. In this way the analysis of the “visual” landscape, where the plant cover is an important element, can be completed. 3.1.6 Visual landscape The visual elements of a landscape are probably the first and most noticed by the general public in the three natural parks of Guadalajara. Since it is the result of




all past geologic, climatic, and vegetative history, it is truly integrative. Although the landscape constitutes the integrating element of the whole project, its “specific” study is approached with four premises: a) to prioritise the correlation between the compositional and visual properties of the territory; b) to build a typological classification for the territory of the parks and their surroundings; c) to carry out a specific mapping of landscape types; and d) to analyze the visual scenarios and their relationship with geology (rocks, structure, morphogenetic processes). Given that the three natural parks have different characteristics, different aims have been set and different methodologies followed for each area. For the Río Dulce natural park a detailed mapping of landscape scenarios has already been carried out (Fig. 2), along with a catalogue and a genetic classification. Three main scenario groupings have been differentiated: (a) the more-or-less flat uplands; (b) the narrow valleys of the Río Dulce and some of its tributaries; and (c) slopes of transition territory between groups A and B. The geological substratum of the uplands conditions its partial erosion and the agricultural use. Thirty upland scenarios have been distinguished, belonging to ten landscape types. Nineteen valley scenarios have been distinguished, grouped in six landscape types. Thirty five transition area scenarios have been differentiated, belonging to six landscape types.

Figure 2: Example of landscape mapping in the Río Dulce natural park. 3.1.7 The artistic contribution Science, at its core, really means the scientific method, a way of observing or accessing nature from the senses and reasoning, from observation and the making of hypotheses, from verification and raising them to the category of scientific theories or relative interpretation models. It is never absolute, but rather subject to changes in time, in line with advances in other fields such as technology. Art, however, can be defined historically: we cannot confine it to a single conception. We can only say that art is the way artists interpret nature, so that there will always be as many “arts” as there are artists, although it is true that there are certain categories which repeat themselves throughout the history of art. They are driven by the search for expression, for harmony, for beauty, etc. Every artist sees the landscape according to his or her own sensitivity, and for this reason it is an open-ended way of looking, a poetic reading of nature,




receptive to suggestions and to the recognition of signs and symbols of every sort. The scientific reading of the landscape, on the other hand, although not without sensitivity, gives a reading in prose, demanding in its words, in the concepts and in the rigorous and systematic description of the medium. In this context, the artistic contribution to this project attempts to: (1) represent the different landscape typologies obtained from their classification; and (2) make abstractions from the different elements which make up the landscape (Fig. 3).

Figure 3: The artistic representation allows the abstraction of the landscape

components; left, geological skeleton; right, a delicate tapestry of vegetation provides the visible scenario (M. Cebrián).

3.2 Synthetic or physiographic method

The integrated study of the landscape, represented in maps of homogeneous territorial units, is being carried out on the basis of geologic and geomorphologic information, which are given an integrated content through the progressive and comprehensive inclusion of existing ecological relationships, and not simply by addition by the use of overlays of other elements, such as soils or vegetation. 3.2.1 The mapping of landscape elements in the Tejera Negra natural park The bedrock of this natural park is mostly a series of highly fractured Silurian and Ordovician black slates. These slates are exposed in numerous surface outcrops. In the higher areas there is discontinuous debris overburden while in the lower areas colluvium overburden can be found. Finally, the main rivers, the Lillas and the Sorbe, flow over an alluvial deposit of some depth. A detailed geomorphological mapping (scale 1:25,000) of the whole natural park is being carried out because of the importance it has for understanding soil development, and subsequently for vegetation. This mapping is of the synthetic or physiographic type, not the morphogenetic as it will look at and describe bedrock plus landform. This mapping of the landforms will then be used to analyse the relationship of the soils and finally the relationship of these soils with the existing vegetation types. The aim is to obtain a map of the homogeneous territorial units in detail, or “landscape elements”. We intend that the characterisation and evaluation of these units, both ecological and visual, will provide useful information for the management of this natural area.




4 Active karstic processes in the Guadalajara natural parks

The methodology proposed in the previous paragraph attempts to approach the study of the physical environment and the landscape of the natural parks in Guadalajara from an integral and methodological point of view. Another essential aim of the PANAGU project is the characterisation and quantification of the geologic and geomorphologic processes active in these karstic parks, in so far as these processes form part of the natural dynamic, with repercussions for ecological and human activity.

4.1 Tufa deposits in the Alto Tajo Natural Park

In the Alto Tajo Natural Park there are numerous karstic features, which are undoubtedly of landscape, environmental and scientific interest [6, 7, 8]). Among these, there are numerous outstanding tufa deposits. They range in age from Neogene to late Holocene times with some still currently active, although their growth is increasingly slow. From the paleoclimatological point of view, the tufa deposits are contemporary with the last karstification periods which occurred in the Iberian Chain throughout the Quaternary Age. They developed during the Marine Oxygen Isotopic Stages (OIS), especially the odd numbers 7, 5 and 1, in Mediterranean climates characterised by high humidity and by development of the plant covers [9, 10]). Within the Alto Tajo natural park, the following tufa typologies can be distinguished: a) Perched springline tufas developed from important karstic springs. They are the most numerous and spectacular tufa build-ups in the canyon of the Alto Tajo [10] due to the sizes. They reach volumes larger than a million m3. Their outcrop geometry is composed of tiers with subhorizontal tops leading to waterfalls of considerable height containing fossil or functional moss curtains. Their stratigraphy contains tufas with different petrological facies. There are also included fluvial deposits, periglacial lithoclast microbreccias, and paleosols which result from a varied paleoclimatic record [10]. b) Fluvial tufas. The narrow, deeply cut gorges carved out by the Tajo river controlled the genesis of the numerous phytoherm barrages of different palaeoclimatic stages. These barrages are frequently located close to knick-points associated with differential erosion of the underlying Mesozoic strata. The oldest generation of these barrages, lower Pleistocene, is placed some 200 m above the river bed of the Alto Tajo, while the most recent is Holocene and can be found at 5 to 10 m above the stream (Fig. 4) [10]. In the ponded areas upstream of the dams there are accumulations of abundant fine materials of tufaceous origin. In some non-contaminated stretches of the river in the Alto Tajo there are certain stromatolite groups currently growing laterally in the channel margins [11]. These are deposits of microbial origin and have not yet been investigated in spite of their rarity in Mediterranean river beds. The tufaceous deposits of the area key elements in establishing the most recent evolution of karstic landscapes. In fact, during the Holocene, numerous tufa barrages formed an authentic fluviolacustrine landscape (Fig. 4) in the Alto Tajo and in some of its tributaries [10]. After the Bronze Age, these systems




became degraded by human impact. As a result, in the Iron Age certain valleys record a change in their sedimentary dynamic caused by the removal of the plant cover for agriculture, cattle rearing or metal working activities. The numerous terrigenous deposits, which reached the river beds from the slopes, put an end to the precipitation stage of the carbonates and to the construction of the tufa barrages. The increasingly numerous detrital deposits gradually destroyed these vulnerable natural structures. At the present time, only one grouping of functional barrage survives in the Alto Tajo, which dams the Taravilla Lake.

Figure 4: Sketch showing one hypothetical Holocene barrage tufa in the Alto Tajo Valley (Bronze Age, 3,500 B.P.) (in [12] p. 160.)

4.2 Karstic erosion of the Río Dulce natural park

One of the most active geomorphologic processes in the Río Dulce natural park is karstic erosion. The Río Dulce runs along the floor of a narrow canyon cut in a Jurassic and Cretaceous limestone and dolostone plateau, where it is fed by various karstic springs. The recharge of these springs is from the plateau, which hardly shows any drainage. The precipitation filters through the soil and into the numerous fractures in the rock below. To determine the chemical dissolution rates within the carbonates, a protocol has been set up to periodically sample the waters of springs, for chemical analysis of dissolved ions (Figure 5, Table 1). Given the river’s mainly subterranean contribution, these values may be considered as representative of the Río Dulce. This means that if the quantities of dissolved calcite and dolomite are multiplied by the annual mean flow of the Río Dulce (27.5 hm3 at the stream gage at Aragosa, five km downstream from the La Cabrera spring) the following dissolution volumes are obtained: 1) 225.25 mg/L of calcite x 27.5 x 109 litres / year= 6,194.375 x 109 mg of calcite / 109 = 6,194,375 tons / 2.71 ton/m3 = 2285.74 m3 of CaCO3 dissolved per year. 2) 70.07 mg/L of dolomite x 27.5 x 109 litres / year= 1,926.925 x 109 mg of dolomite / 109 = 1,926.925 tons / 3 ton/m3 = 642.30 m3 of MgCO3 dissolved per year. To sum up, although these results are preliminary and perhaps only valid to give a general idea of the magnitude of the phenomenon, it can be estimated that a total of 2,928.04 m3 (c.3,000 m3) of rock is dissolved annually in the karstic system drained by the Río Dulce within the natural park.




Figure 5: The Río Dulce natural park and its hydrological setting: Dulce river watershed, springs, stream gage and bedrock lithology.

Table 1: The mean content of calcite and dolomite dissolved in the water from

the Río Dulce springs, for the analyses carried out to date (HCO3 total is 296 mg/l).

Ions Mg / L Ions mg / L Ca 90.2 Mg 20.2 HCO3 dependent (Ca) 135.05 HCO3 dependent (Mg) 49.86 CaCO3 225.25 MgCO3 70.07

5 Discussion and conclusions

By carrying out territorial inventories of the natural parks in Guadalajara using both analytic and synthetic procedures, an attempt is being made to compare the advantages and disadvantages of each of these procedures, as determined by the usefulness of the information obtained. Given that these inventories have not yet been concluded, it is still too early to know which kind of inventory is better adapted to the characteristics of each of the parks. As far as the tufaceous geo-systems are concerned, it is important to highlight the distinctly high value which their related landscapes offer in the Alto Tajo natural park. This is due to the exceptional dimensions of specific tufaceous deposits, to their “capricious shapes” and to their spectacular waterfalls and the humidity which impregnates the atmosphere of the systems which are still functional, almost all of them sited within the domain of the slopes where they are fed by the karstic springs. These geo-systems, however, are in grave danger: easily degraded and vulnerable to environmental changes. In addition their current growth rates are much slower than several centuries ago. Are recent climatic changes responsible for this? Is human activity the essential reason for this loss of functionality? To delimit the natural and human causes responsible for this deterioration is one of the aims of this project. Fortunately, this fragility is well known and reflected in legislation, thanks to our earlier research. Evidence of this is the protection and conservation framework which covers the tufa outcrops in the recent Law on Nature Conservation and Environmental Impact Assessment passed by the Castilla – La Mancha Government.




With respect to the volume of rock dissolved in the Río Dulce, 3,000 m3 may seem at first a significant volume, but in proportion to the area and volume occupied by the carbonate massif susceptible to karstification, this dissolution is tiny. Comparison between the annual dissolved volume and the volume of the carbonate rocks of the massif gives that the percentage of this total which is dissolved annually (2 x 10-5 %) is not much. This, in turn, allows an approximate calculation of how many years it will take to dissolve the total of the park down to river level (!) or, on the other hand, to carry out evolutionary interpretations of the formation of the canyon and of the park itself.

References

[1] Cain, S.A., Foundations of Plant Geography, Harper: New York, 1944. [2] Kruckeberg, A.R. Geology and Plant Life, University of Washington

Press: Seattle, 2002. [3] Walter, E. Vegetation of the Earth and Ecological Systems of the

Biosphere, Springer-Verlag: Berlin, 1979. [4] Forteza, J., Cruz, R., Goy, J., Barrera, I., et al., Soil representation in

landscape and geomorphology for the regulation of the Candelario Natural Reserve (Salamanca, Spain), GIS use and limitations. Proc. of the III Int. Congress of European Society for Soil Conservation, Valencia, 2000.

[5] Vigo, J. Some reflections on geobotany and vegetation mapping. Acta Bot. Barc, 45, pp. 535-556.

[6] González Amuchastegui, M.J., Parameras de Molina y el Cañón del Alto Tajo. Guía de los Espacios Naturales de Castilla – La Mancha, pp. 201-202, 1991.

[7] García Quintana, A., García Hidalgo, J.F., Martín Duque, J.F. et al., Geological factors of the Guadalajara Landscapes (Central Spain) and their relevance to landscape studies. Landscape & Urban Planning, 69, 417-435.

[8] García Quintana, A., Martín Duque, J.F., González Martín, J.F., et al. Geology and rural landscapes in central Spain (Guadalajara, Castilla – La Mancha). Environmental Geology, 47, pp. 782-794.

[9] Ordóñez, S. González, J.A., and García del Cura, M.A., Datación radiogénica (U-234/U-238 y Th-230/U-234) de sistemas travertínicos del Alto Tajo (Guadalajara), Geogaceta, 8, pp.53-56.

[10] González Amuchastegui, M.J. y González, J.A., Significado geomorfológico de las acumulaciones tobáceas del alto valle del río Tajo (sector Peñalen-Huertapelayo). El Cuaternario en España y Portugal, Vol. 1, ITGE, Madrid, pp. 99-109, 1993.

[11] Guerrero, I. y González, J.A., Características geomorfológicos del modelo de construcción tobáceo del Alto Tajo en su fondo de valle (Peralejos de las Truchas, Guadalajara), Geotemas, 1(3), pp. 375-378.

[12] González, J.A. y Rubio, V., Las transformaciones antrópicas del paisaje de los sistemas fluviales tobáceos del Centro de España. Bol. R. Soc. Esp. Hist. Nat. (Geología), 96, pp. 155-186.




Exploring the effect of demographic elements on the evaluation of the scenic beauty of various landforms – preliminary results

A. Tsouchlaraki Hellenic Open University, Greece

Abstract

This paper reports on a piece of research which attempts to take 32 different images of landforms and to show how different classes of people perceive or “value” the scenic beauty of these landforms. This is an interesting question and one which bears research. In this case the term “scenic beauty” refers to the public preference of various forms of the earth's relief. It is a figure which, even though it depends on various subjective factors, aims to quantify the general preference of the public for various landforms. A questionnaire survey takes place in order to investigate the public preference, using a sample of 221 persons in the area of Athens and Piraeus, and the city suburbs. This area concentrates a very large proportion of Greece's population and can ensure variety in terms of social and demographic status of the sampled population. The means used to demonstrate the various forms of relief are 32 digital relief images, created with the use of an algorithm developed for this purpose. The representative selection of the sample of digital images took place after the classification of the forms of relief in Lefka Ori mountain range in Crete. The questionnaire included 11 questions describing the person questioned in relation to the environment he/she has experienced or knows and his/her contact with the countryside. Each question corresponded to a factor (e.g. age, sex, education, income etc.) that had been generalised in categories. Following this, a primary statistical analysis of variance was carried out for each of the factors examined and some preliminary results are reported. Little research in this specific area has been done and it is interesting to explore further the way people with different social backgrounds react or perceive the various landforms. Keywords: scenic beauty, landscape aesthetic, perception, landforms, demographic elements.




doi:10.2495/GEO060331

1 Introduction

Landscape is the combined result of physiography, geological formations, vegetation, waters and the various cultural interventions that occur in a given area. This combination attributes shape, line, colour and texture to a landscape, while the aesthetic result is considered on the basis of the variety or the uniqueness offered and is usually classified into three main classes: 1) indistinctive, 2) common, 3) distinctive (USDA, [1]; USDA, [2]). This classification determines those landscapes which are most important and those which are of lesser value from the standpoint of scenic beauty. The classification is based on the premise that all landscapes have some scenic value, but those with the most variety or diversity have the greatest potential for high scenic beauty. The various approaches developed for determining landscape visual quality refer to the general output of the synthesis of all physical variables of the environment, physiography, soil, vegetation, hydrological elements, and not individually to each element. In these approaches, the visual quality of a landscape is evaluated either indirectly by thematic maps according to standardised criteria based on the experience of the scientists, or directly on the basis of psychometrical methods that quantify directly the public’s preference by demonstrating through a certain means the landscape under evaluation (Kaplan et al [3]; USDA [1]; Kaplan [4]; Daniel and Boster [5]; U.S. BLM [6]; Palmer [7]; Smardon et al [8], Hunziker and Kienast [9]). Relief constitutes, however, a part of these approaches and not the main objective. There are cases where it would be desirable to isolate the relief from the other elements which are making up the visual environment and to investigate its scenic beauty separately. Such cases include the technical works that cause major and permanent alterations to the relief, as well as the works, of which the spatial arrangement depends on the morphology of the ground. Slopes, distances, hypsometric difference values, viewsheds are elements of the relief that affect the visual quality of a landscape, but also the ability of a particular landscape to accept and absorb new activities. The experience of the scientists show that the more mountainous the form of the relief, the more distinct the landscape category offered (USDA [1]; U.S. BLM [6]). It is interesting in this case to explore the way various persons react, based on different demographic elements, to the different forms of the relief. The way each observer evaluates any given landscape is a very complicated issue, a matter difficult to predict. Many factors can possibly influence this evaluation, ranging from factors that can be registered, such as the usual demographic elements, to imponderable factors, such as the mood of the observers at the time of the evaluation. Besides, what “one likes or dislikes” does not remain constant with time. As a person matures, his/her attitude towards many things in life changes. Therefore it would be utopia to try to predict with precision the preference of an observer for a given landscape, but this is not the aim of this research. This paper presents some preliminary results of a research work that aims to investigate the way in which the various demographic elements of the persons questioned influence the preference in the case of the evaluation of the




scenic beauty of a landform. Whatever conclusion can be drawn, even based on the few factors that can be registered, would be useful in the interpretation of the phenomenon.

2 Investigation methodology

2.1 Digital relief visualisation

The investigation of the public preference to the various landforms can be achieved with the use of questionnaires, by utilising some means for presenting the different forms of relief. In this case, it was not possible to use actual photographs as a means to demonstrate these different forms, because they would provide simultaneously information on the vegetation, the soil and the land uses of a given area, something that could influence the preference of the persons questioned. Therefore, the survey was effected using digital visual representations as a means to demonstrate the various forms of relief. These are images created with the use of a Digital Terrain Model and the respective shading image. The creation of these images relied on an algorithm that had been developed for this purpose, producing the perspective image of the relief, as this would look like if it had been photographed from a known shooting point in relation to a given target, with the use of a photo camera of known geometry (Tsouchlaraki [10]). All shootings are strictly horizontal, considering the geometry of a 35 mm camera with a normal lens (f=50 mm) and a predetermined data analysis scheme. The study area was that of the Lefka Ori mountain range in Crete, owning to the variety of the forms of relief it includes. The IDRISI GIS package was used for the processing of information, as well as the creation of a DTM and of other derivative elements. The representative selection of the sample of the digital representations that would be included in the questionnaire took place following a classification of the relief forms, based on a method developed for this purpose, which is a modification of Hammond’s classification method (Tsouchlaraki [11,12]). The classification resulted in 32 relief categories. A horizontal position was selected for each category, for creating a digital representation. The 32 images created are representative of the different forms of relief in the study area and are depicted in Figure 1. As far as the scenic beauty is concerned, the persons questioned were asked to rate it on a 1-10 scale (1 representing a very small preference concerning the scenic beauty of the landform and 10 representing a high preference). This scale was chosen because the studies of landscape aesthetics which use photographic imagery (both actual and simulated) and 1-10 scales for response are common in use.

2.2 Questionnaire design and execution

There are many rules for the design and execution of a questionnaire and also many decisions that have to be taken (Damianou [13], Koutsopoulos [14]). The




1 2 3 4

5 6 7 8

9 10 11 12

13 14 15 16

17 18 19 20

21 22 23 24

25 26 27 28

29 30 31 32

Figure 1: The 32 images of landforms included in the questionnaire.




size of the sample, the sampling methodology, the form and type of questionnaire, the duration of a questionnaire are some of the matters that have to be examined. This subsection is devoted to these matters. The questionnaire was used in the area of Athens, Piraeus and the city suburbs, using a sample of 221 persons. This area, comprised of 52 municipalities and 5 communities, offered the following advantages: A) It is an area of Greece’s capital that concentrates a very large proportion of

the population and is suitable for ensuring variety in terms of the different social and demographic characteristics of the persons questioned.

B) The population of Greece’s capital comes mostly from different areas of Greece, and allows the inclusion of people from the provinces, with different experiences and mental representations of the various forms of the relief.

The total population of the study area is 3,020,562. The area includes both rich and poor quarters, both densely and scarcely populated, all of them being within the city plan limits so that they are mapped and facilitate the organisation and implementation of the questionnaire. A stratification of the municipalities of Athens was effected before the selection of the persons to be questioned. In order to ensure the representation of the entire population, two individual layers were used prior to adopting the final sampling. Upon the first layering, the municipalities were classified in three categories, according to their population: small (X<30,000 inhabitants), average (30,000<X<70,000) and big (X>70,000 inhabitants). Upon the second layering, we used the existing classification of the prefectures: Athens, Eastern Attica, Western Attica and Piraeus. We thus created 3x4=12 layers. In a simple random sampling, the sample size is calculated as follows (Damianou [13]): P [ | y - Y | < d ] > 1-a (1) where: Y : the population average of the requisite characteristic, which in this

case is the visual value; y : the population average estimate derived from the sample; d: the error margin or the desired measurement accuracy of the average; 1-a: the confidence coefficient. It is clear from equation (1) that the size of the “n” sample is determined by the following formula (Damianou [13]):

Nn

n0

0

1 +=n (2)

in which:

22a/0 )

dsz

(n = (3)




where: Ν: the total population; s2: the dispersion of the characteristic; za/2: the upper a/2-point of the normal distribution Ν (0,1). The term n0/N, when the value of N is very big, tends towards zero and is ignored. In the layered sampling analysis, the same reasoning is followed, with the exception being that the above formulae (2,3) are applied in each layer, using the respective dispersion of the characteristic which each layer exhibits. In case the dispersion is unknown, we can use its value from previous studies related to the subject matter (Damianou [13]). Because of the fact that there are no previous studies related to the certain subject matter, we used the results of an experimental implementation in 55 students of the 2nd and 8th semesters of the Department of Rural and Surveying Engineering of the National Technical University of Athens. This implementation was a simple questionnaire including the 32 images and asking students to evaluate them on a 1-10 scale, according to their preference. In this pilot study the previous experience of the scientists came true as the more mountainous the form of the relief, the higher the values that resulted from the students’ preference. The results of this implementation showed that the standard deviation differs among images and ranges from 1.64 to 2.75 units. In order to determine the final sample, we assumed that all layers present the same dispersion and took the worst case to be the standard deviation: s=2.75 units. The desired measurement accuracy of the mean was determined to be d=0.40, a value lower than the unit half. Therefore, for a 95% probability, the application of the above formulae shows that the size of the sample is:

1822)40.0

96.175.2(n ==x (4)

This figure was increased, for the sake of safety, by about 20 percent and the final sample was thus determined to be 221 persons. This sample was selected for two main advantages: 1) it was small in relation to the size and therefore useful for carrying out a further research; 2) it could assure the required reliability in order to draw some initial conclusions. The students’ responses were not used in this final sample or in the further analysis. The sample was distributed among the individual layers, based on the percentage of the population in each of the 12 layers in relation to the total population of the city. Following this, we proceeded to random sampling for the selection of the sample decided for each layer, using the files of the National Statistical Service of Greece. The questionnaire included 11 questions and was accompanied with the 32 coloured images. The questions had a multiple choice form and were selected in such a way as to describe the person questioned in relation to the environment he/she has lived in or is familiar with, during his/her contact with the countryside. Also basic demographic elements, such as sex, age, education,




profession, income are included in the questionnaire. Each question corresponds to a single factor, generalised in a maximum of three categories. Besides, the relatively small size of the sample would not have benefited the larger in number categories. The questionnaire is presented in the following Figure 2. The eighth question was devoted to the scenic beauty of the various landforms presented in the coloured images. The responder was asked to rate the landforms on a 1-10 scale according to his/her preference. With regard to the ninth question about the profession, the third category entitled “other” refers to unemployed, students, housewives and in general people who do not work. The technique used for the questionnaire was that of the interviews. The main advantage of this technique is the direct contact with the person questioned and therefore presents the highest participation rates. The time available for the interview depends mainly on the way it is conducted. In this case, we decided to visit the persons questioned at home, in order to allow for a 10-15 minutes interview.

2.3 Questionnaire results

Table 1 shows the descriptive elements of the results’ distribution, as derived for the scenic beauty of the various landforms. For the other questions the results are shown in Table 2. For the sake of brevity, all references hereinbelow will use the symbolism of factors shown in Table 2. It is clear from the frequency values that there is sufficient number of observations in each category. The sample includes individuals for all the categories of age, education, income, profession and also individuals that come from different places of Greece or abroad. As pertains to sex, women are the majority, however, the percentages are close to the respective percentages of the official census, in which men correspond to about 49% of the population and women to 51%; therefore these percentages are considered to be satisfactory for the balance between the two sexes. As pertains to the scenic beauty, all images were rated with values from 1-10. The average responses for each image range within 5.02-7.72, with standard deviations from 1.91 to 2.62. Therefore, the standard deviations are within the range of the values that had been observed from the experimental questionnaire to the students. When observing the images in a descending order as to the mean scenic beauty, we can easily conclude that low values correspond to more plane forms, while high values to mountainous forms. As it has been already mentioned, this is one of the criteria used in landscape analysis and evaluation, because mountainous forms present greater variety in the relief elements (slopes, curvatures, crest etc.) in relation to plane forms. An interesting observation may also derive, when we also look at the mean values and the respective deviations of the answers. Let’s consider that the standard deviation is the measurement of disagreement among the respondents, then we observe a greater disagreement of opinions in image 1, and a greater agreement in image 21. By observing the remaining images, it seems that




disagreement tends to increase as the relief’s visual value decreases. The respondents therefore seem to agree more on what is good rather than bad. This observation is not an object of this study, but we present it as an interesting issue for further investigation.

Figure 2: The questionnaire.

Scenic Beauty of Various Landforms Please give a single answer to the following questions. 1. Sex: 1. [ ] Male 2. [ ] Female 2. Age: 1. [ ] 18-35 years 2. [ ] 35-50 years 3. [ ] more than 50 years 3. You come from:

1. [ ] Athens 2. [ ] Outside Athens 3. [ ] Outside Greece 4. Where have you spent most part of your life?

1. [ ] Big city ( ………name of the city) 2. [ ] Small town (……… name of the town) 5. How often do you visit the countryside? 1. [ ] Very often, almost every month 2. [ ] Not very often, two or three times every year 6. Usually for what purposes do you visit the countryside? 1. [ ] Leisure 2. [ ] Other 7. Using a scale of 1 (low preference) to 10 (high preference) how much would you evaluate

the scenic beauty of the landform that is represented in each image? Image 1: ___ Image 2:___ Image 3:___ Image 4:___ Image 5:___ Image 6: ___ Image 7:___ Image 8:___ Image 9:___ Image 10:___ Image 11: ___ Image 12:___ Image 13:___ Image 14:___ Image 15:___ Image 16: ___ Image 17:___ Image 18:___ Image 19:___ Image 20:___ Image 21: ___ Image 22:___ Image 23:___ Image 24:___ Image 25:___ Image 26: ___ Image 27:___ Image 28:___ Image 29:___ Image 30:___ Image 31: ___ Image 32:___

8. Education: 1. [ ] Primary 2. [ ] Secondary 3. [ ] Higher 9. Profession 1. [ ] Employees 2. [ ] Self-employed 3. [ ] Other 10. Income yearly 1. [ ] Less than 5 millions GRD 2. [ ] More than 5 millions GRD 11. Usually how long do your vacation last? 1. [ ] Less than 15 days 2. [ ] 15-30 days 3. [ ] More than 30 days




Table 1: Descriptive elements of the distribution of results for the visual value of the landform of each image.

Image Mean

x

Standard error

s( x )

Standard deviation s

1 5.02 0.18 2.62 2 6.81 0.16 2.42 3 5.71 0.13 1.97 4 5.49 0.14 2.06 5 5.98 0.15 2.19 6 6.28 0.13 1.95 7 6.43 0.15 2.30 8 5.67 0.17 2.48 9 6.34 0.13 1.96

10 7.05 0.15 2.28 11 5.70 0.13 1.95 12 6.02 0.15 2.21 13 5.76 0.14 2.13 14 5.56 0.16 2.38 15 7.38 0.14 2.01 16 6.51 0.14 2.09 17 5.44 0.17 2.49 18 7.41 0.14 2.14 19 5.61 0.16 2.36 20 7.49 0.14 2.06 21 6.04 0.13 1.91 22 6.59 0.13 2.00 23 5.86 0.14 2.01 24 6.10 0.14 2.14 25 6.76 0.13 1.98 26 5.86 0.17 2.50 27 6.10 0.13 2.00 28 5.86 0.14 2.06 29 5.89 0.13 1.95 30 5.68 0.15 2.28 31 7.14 0.14 2.05 32 7.72 0.14 2.01

3 Analysis of variance – preliminary results

The results were proceeded with the analysis of variance for each individual factor (Bora-Senta and Moysiadis [15]). The dependent variable was the scenic beauty of landforms. For any one interviewee there is only one set of factors which are repeated against all 32 images. Repetition of characteristics in multivariate analysis has unpredictable results. In order to avoid this problem and get some preliminary results, instead of using all the ratings given by each interviewee we used the sum of ratings of all the images for each interviewee. In this manner the data matrix included 221 observations.




Table 2: Categories of factors.

Question/ Factor General Categories Occurrence Frequencies

Sex (SEX) - male - female

93 128

Age (AGE) - 18 - 35 - 35 - 50 - > 50

94 76 51

Origin (FROM) - Athens - outside Athens - outside Greece

90 115 16

Where have you spent most part of your life? (LP)

- big city - small town

177 44

How often do you visit the countryside? (CVI)

- > 1 / month - < 2-3 / year

65 156

For what purposes; (RCVI) - leisure - other

161 44

Education (EDU) - primary - secondary - higher

61 94 66

Profession (PROF) - Employees - Self-employed - Other

85 61 75

Income (FIN) - < 15,000 EURO - > 15,000 EURO

151 70

Time of vacation (HOL) - < 15 days - 15 - 30 days - >30 days

59 122 40

From the results gathered from this preliminary factor analysis of variance, we can draw some conclusions. In general, and in almost all of the factors, the mean values of the individual categories differ from one another with a high degree of reliability. The F-ratio level of significance ranges from 0.00 to 0.0022. With a probability of almost 100%, this means that the average values of the individual categories are not equal. The sum of the squares of errors between the groups, which the higher it is, the better the factor classifies the dependent variable, can become also a benchmark for the factors. Factor analysis extracts factors which maximise the variance explained in order of the most important first and so on. Considering the results of the analysis, the variation of the mean values in the categories of each factor leads to the following conclusions: 1. The persons who have spent most part of their lives in small towns or villages

give higher ratings than those who have lived in cities or in city suburbs. 2. The persons who originate from areas outside Greece give ratings higher than

those who come from areas outside Athens, and the latter give ratings higher than those who come from Athens.

3. Women give higher ratings than men. 4. The persons who spend more than 30 days annually for holidays give higher

ratings compared to persons who spend 15-30 days, and the latter give higher ratings than persons who spend less than 15 days.




5. The persons with a family income below EURO 15,000 give higher ratings compared to those with a family income of more than EURO 15,000.

6. The persons aged 50 years or more give higher ratings compared to those in the 35-50 age group and the latter give higher ratings compared to persons in the 18-35 age group.

7. The primary education graduates give higher ratings compared to the secondary education graduates, and the latter give higher ratings compared to university graduates.

8. Persons visiting the countryside for leisure purposes only give higher ratings than persons visiting the countryside for other reasons.

9. Unemployed persons give higher ratings than self-employed and self-employed give higher ratings than employees.

10. The persons who visit the countryside more often than once a month give higher ratings compared to the persons who visit the countryside from time to time, 2-3 times a year or do not visit it at all.

The way the respondents used the same scale of values is different from person to person, however it seems feasible to group and generalize their behaviour. This is a first conclusion and perhaps one we would expect. What is that makes certain groups of people use higher values in relation to other groups? An important generalisation drawn from the observation is the following: the persons who have or had in the past more chances to come to contact with a physical environment use higher ratings. This probably explains why the persons who have lived in or come from the countryside, together with the persons who visit the countryside more often, the persons with more freedom in their work time, or the more aged persons who had more chances to visit the countryside, are the ones who know the physical environment better and thus give higher ratings. It is through the high degrees of freedom that they express their preference for nature. Up to here, the previous observations derive from the whole sample of the respondents. However, the fact that women gave higher ratings than men is a matter of concern and shows that we should perform a further investigation for both sexes separately. This will be the main concern of the following phase of this research, in order to find out whether the factors that affect each sex remain the same or not.

4 Discussion

This study addresses the issue of the influence of the demographic data on the evaluation of the landforms scenic beauty. It is clear from the results that the factors examined might influence to a certain extent the scenic beauty of the landforms. Maybe there are many other factors, perhaps even more important than the ones examined. No relative research had been conducted in the past, so as to allow for a comparison. For example, perhaps the morphology of the place of origin of the respondent or the place where he/she has lived most of his/her life plays an important role in his/hers preferences and his evaluation, since each respondent is familiar, due to his experiences, to certain relief forms. The




investigation of such factors is not part of this study, however it is a very interesting issue for further research, in order to better understand and interpret the public’s preferences.

References

[1] USDA Forest Service, National Forest Landscape Management, Government Printing Office, Ag. Handbook 434, Washington, 1974.

[2] USDA Forest Service, Landscape Aesthetics, Government Printing Office, Ag. Handbook 701, Washington, 1995.

[3] Kaplan R., Kaplan S., Wendt J.S. Rated preference and complexity for natural and urban visual material, Perception and Psychophysics, 12(4):354-356, 1972.

[4] Kaplan S., Some methods and strategies in the prediction of preference, In Landscape Assessment – Values, Perceptions and Resources, edited by E.H. Zube, R.O. Brush, and J.A. Fabos, Stroudsburg, PA.:Dowden, Hutchinson and Ross, pp.118-119, 1975.

[5] Daniel C.T., Boster S.R., Measuring Landscape Aesthetics: The Scenic Beauty Estimation Method, USDA Forest Service, Research Paper RM-167, 1976.

[6] U.S. BLM, Visual resource management: Visual resource management program, U.S. Government Printing Office, USA, 1980.

[7] Palmer J., A visual character approach to the classification of backcountry trail environments, Landscape Journal, 2(1), USA, 1983.

[8] Smardon R., Palmer J., Felleman J., Foundations for Visual Project Analysis, John Willey & Sons, New York, 1986.

[9] Hunziker M., Kienast F., Potential impacts of changing agricultural activities on scenic beauty – A prototypical technique for automated rapid assessment, Landscape Ecology, 14(2), pp.161-176, 1999.

[10] Tsouchlaraki A., 1996. Digital Relief Visualisation in Landscape Analysis, Technika Chronika, Scientific edition of the Technical Chamber of Greece: I, 33, Athens, pp. 27-37, 1996 [in Greek with English extended summary].

[11] Tsouchlaraki A., A landform classification method with GIS for landscape visual analysis purposes, Proceedings of the Second International Conference on Sustainable Planning, 12-14 September 2005, Bologna, Italy.

[12] Tsouchlaraki A., A Methodology for the evaluation of the visual value of natural relief. Ph.D., Department of Rural and Surveying Engineering, NTUA, Athens, 1997.

[13] Damianou, Ch., Sampling methodology – Techniques and applications, Aithra, Athens 1992 [in Greek]

[14] Koutsopoulos K., Geography: Methodology and Spatial Analysis Methods, Symmetria editions, Athens, pp. 367-422, 1990 [in Greek].

[15] Bora-Senta E., Moysiadis X., Applied Statistics, ZITI editions, Thessaloniki, 1992 [in Greek].




An approach to the landscape analysis

B. Badiani Researcher of Urban and Regional Planning, University of Brescia, Italy

Abstract

The starting-point is the new conception of landscape proposed by the European Convention on Landscape (Florence, 2000): the landscape is a combination of elements, natural and not, that encircle us. In light of this new concept, the work proposes a reflection on the actual modalities of the description and interpretation of the landscape used as the first step in environmental planning. In 2002, the Lombardia region approved a law concerning the guide lines for the examination of the projects in relation with the landscape value. So it is very important to make clear what a landscape means. But the elements suggested in this law to describe the landscape do not seem to be useful in estimating its value. Taking into account the proposed methodologies of this regional law, an attempt has been made to devise one practical and flexible procedure for the evaluation of the characters of the landscape. The aim is to reach a fundamental knowledge of what a landscape is, taking into account its concrete and sensitive aspects and their possible descriptions. The methodology is based on a check-list of natural and human elements, implemented on a GIS. A database has been associated with each element. The information put into the database is organized in a way to facilitate the evaluation of the landscape, the communication of the result of the analysis and the participation of the citizens in the decisional phase. The methodology has been applied to a territory of a municipality in Northern Italy to test the good flexibility, the simplicity and the clarity of the description of the landscape image and the improvement of the quality and completeness of the check-lists. Keywords: landscape analysis, environmental planning, landscape planning, GIS.




doi:10.2495/GEO060341

1 Introduction

Safeguarding the landscape mainly entails governing its transformations due to the action of man or natural events. This translates into the fact that each landscape and territorial policy initiative must therefore originate from tools useful to govern the transformations. According to the National legislation referring to the Natural Heritage and Landscape Code, approved as Legislative Decree No. 42 dated 22nd January 2004 and that came into force on 1st May 2004, the urban municipal plan is assigned a special conclusive value in the process to construct the overall safeguarding system, as also expressly stated in the regional law of Lombardy No. 12 dated 15th March 2005 “Law to govern the territory”, requiring the foregoing planning tool to also include landscape-related aspects. The following aspects emerged from an analysis of the initial experiences of local landscape planning:

o the lack of a simple, effective, easily applied and above all shared operating methodology, with the consequent limited tangible use of the tool;

o a lack of clarity and precision in describing the criteria used to determine the classes of sensitivity of the local landscape;

o a limited or no participation and involvement of the local community in the preparatory phases of the Plans and when determining the classes of sensitivity of the landscape;

o no consideration has been given to shared tangible and schematic parameters or indicators, as a tool to analyse and evaluate the landscape.

The fundamental methodological aspects of the proposed method have been defined starting from this critical verification. The method must be:

- concise → a limited number of parameters to be assessed; - readily understandable and applicable → clear, well-defined and

practical; - general and flexible → applicable to different territorial environments

without too many difficulties; - easy to manage → based on readily obtainable data; - effective and useful → which enables tangible results to be attained.

However, the starting point remains the clarity of the concepts and the objectives which are to be achieved. Therefore, it is essential to specify what is meant by landscape, which of the infinite definitions and interpretations of the term is taken into consideration in relation with the scope of the methodology proposed. It was deemed appropriate to consider the concept of landscape proposed by the European Convention (Florence, 2000), since it has been formulated recently, incorporates various aspects addressed in previous years and expresses the complex nature of the concept in the best possible way. The landscape is everything that surrounds us and that we perceive with our eyes and with our emotions, and more in detail it is “a determined portion of territory which is perceived by man, the appearance of which is the result of the




action of human and natural factors and of their interrelations […] a shared asset, the basis of the cultural and local identity of populations, an essential component of the quality of life and expression of the richness and diversity of the cultural, ecological, social and economic heritage”. In this sense we can interpret the landscape as the series of signs left on the territory by natural and human history, read and interpreted by an observer. And it is at this point that the main problem to be addressed is encountered: if the landscape is “everything”, how can a practical procedure be found that takes into consideration each constituent feature? How can such an ambiguous concept be rendered concise and schematic? There is a need to deliberately limit the analysis of the landscape by considering the most significant and important aspects for its planning and management. But what is significant in a landscape? To make a selection among the features which surround us entails a choice that is undoubtedly arbitrary and subjective. Furthermore, in this way the understanding of the landscape will never be complete, because some factors or aspects that could be significant will always be excluded. A possible solution exists if one observes that complexity is a factor which cannot be eliminated in order to know the landscape, and it is pointless to try to understand everything. It is necessary to be aware of the fact that an “approximate knowledge” exists.

2 Presentation of the proposed procedure

The method that is illustrated below intends to foster the awareness of both the designer and the local populations as regards safeguarding the landscape and can be a useful help in landscape planning at the municipal level. The methodological path is characterised in stages:

1. a first stage to identity the objectives associated with the; 2. an analysis stage, based on collecting data and processing the data in a

reference list; 3. an assessment stage, which entails describing the aspects which

combine to attributing value to the features identified as being characterising features;

4. drawing up a landscape map, which is created by using GIS (Geographical Information System) software, which enables some alphanumeric information to be associated with the graphic elements represented, thanks to the use of a descriptive database that contains information regarding the features necessary to appreciate the landscape’s overall value.

3 Landscape analysis stage

The scale of the intervention is selected, which in this case is on a municipal scale, after having defined the objective of the landscape analysis, which could




be to identify the special local features in order to verify the sensitivity to changes associated with planning choices. The processing activity consists in selecting the data collected. The most important factors to be considered in the analysis stage are outlined in the following list, which can be enriched and extended:

1. physical and natural features; 2. anthropic features: these represent the landscape components that

derive from the action of man; 3. perception features: these represent the landscape components subject

to perception; 4. critical areas: these represent factors which identify particular

situations. The check-list structured in this way includes all the landscape features considered to be most significant and above all shared. This has been possible by involving the local community, also in the stage of preparing the checklist. Table 1 shows a number of the items adopted, for purposes of illustration. A value must be assigned after the landscape has been analysed and after all its constituent factors have been identified. This entails defining criteria to establish an opinion about the single feature. This represents an extremely subjective and delicate operation. In this session it was deemed more important to define the aspects of the feature that can contribute to establishing the value. An attempt was made to break down the feature into its fundamental aspects to appreciate fully all the aspects that can influence the value of a given feature. The concepts of shape, meaning and use were selected. First of all, a landscape feature is characterised by a shape, which represents its physical constitution, its aesthetic features such as size, colours, contours, etc. The value opinion in relation to the shape can depend on the presence of rare or unusual features, the diversity and variety of the features, their integration, etc. The meaning is the result of the individual's perception and is influenced by a series of interrelationships which are established among the object, the surrounding context and the person making the assessment. The relationship with the objective of the analyses must also be established to assess the meaning. The following criteria have been identified:

• Readability → meaning the quality that confers on a feature of the landscape a high probability of evoking a well identified, structured and functional environmental image in an observer;

• Use → has a dual meaning. To be understood as the possibility of using the feature in question or the possibility of using other features via this feature;

• Context → represents the set of features that surround the feature under examination and can be “natural”, or “man-made”;

• Visual → is understood as the type of view presented by observing the landscape feature in its context;

• Value → this term is associated with the social and cultural aspects and with the local history, as well as the importance of the landscape heritage;




• State → this term refers to the degree of naturalness and integrity of a given feature.

Table 1: Brief description of the landscape features and the data sources.

Feature Brief description Sources NATURAL FEATURES

Mountain, ridge surveys

Represent the watershed lines of the principal catchment

basin.

Technical regional map

Morphological factors Characteristic forms of the land due to past or current

morphological transformations.

Technical regional map

ANTHROPIC FEATURES Historical centres and

settlements Built-up areas which represent the oldest district of the urban

system.

Municipal cartography

Historical archives PERCEPTION FEATURES

Panoramic views Visual overviews of shared importance

Panoramic points Place from which it is possible to enjoy a panoramic

view Routes to enjoy the

landscape Road or footpaths, pedestrian-bicycle lanes and routes from which the landscape can be

enjoyed

Municipal and provincial

cartography Site visits

CRITICAL AREAS Degraded areas Areas subject to change Municipal

documents and cartography Site visits

The function is associated with the how the feature is used. The function aspect is associated with a value related to the following criteria, as in the case of the meaning assessment:

• Accessibility → this represents the possibility of reaching the landscape feature, therefore of enjoying it and exploiting it from the recreational or useful point of view;

• Use → this is understood as the use of the land referred to the area in which landscape feature under examination is located;

• Profitability → this refers to the possibility of generating income associated with the landscape feature as a tourist venue or as a source of commercial resources or again as a place for recreation.




The most obvious and important novelty introduced by the methodology proposed, consists in characterising the landscape features directly at the cartographic level: the features are not only represented by locating them on the map, but an assessment is also associated with them through a combination of symbols associated with the aspects which define their value. Some examples of the map detailing the shape, meaning and function aspects of a number of features that constitute the landscape are shown in the illustrations below.

Figure 1: Shape table.

LEGEND • source • eternal • precise feature • clear water

Figure 2: Meaning table.

LEGEND • panoramic point • significant view • strong readability • limited use • natural context • panoramic view • perceptive value • in good conditions

Figure 3: Function table.

LEGEND • hollow • wood • difficult access • object of speleological studies




4 Conclusions

An attempt has been made to create a practical and flexible procedure to assess the quality of the landscape, using on a knowledge database to manage the territorial transformations by referring to the methodologies and the operating methods proposed by the regional regulations which relate to landscape planning and managing landscape transformations, and their application to a number of municipal territorial contexts, (in particular the guidelines for landscape-related examination of projects, approved with Regional Council Resolution (D.G.R.) No. 7/11045 dated 8th November 2002). The starting point of this study is represented by the endeavour to establish an in-depth knowledge of the landscape features, to focus attention on the most tangible and sensitive aspects and on their possible representation. The greatest difficulty found was associated with the need to represent the landscape concisely, without reducing the significance and complexity. The landscape is analysed by specifying its shape, its meaning and its function and the methodology proposed enables its potential, values and criticalities to be interpreted. The procedure proposed has been applied to a case study: the Municipality of Polaveno, located in North Italy (Province of Brescia) and a number of aspects worthy of note have been highlighted. On the one hand, problems and difficulties emerged due to the limits of the form to represent the landscape that is necessarily partial. On the other hand new potential for new fields of application have emerged. The need to make significant choices regarding the features to be taken into consideration stems from the impossibility of representing “everything”. In relation to the objective that is to be achieved, the best and most effective criteria to be adopted in making this selection, is undoubtedly the criteria of interpreting the landscape based on the social perception of the persons that benefit directly from the landscape. Difficulties were experienced in retrieving all the useful information, both with reference to the location of the landscape features, as well as to determine their meaning. The available and up-to-date cartographic sources are very limited. The best solution is to verify the landscape features by site visits, and above all to become acquainted through the eyes of those that live there and benefit from the landscape directly every day. The methodology proposed does not impose any rigid schemes. Using flexible check-lists becomes a simple reference and a guide in the selection of the features which describe the landscape and for their assessment. It was found useful to exploit the advantages and the potential offered by the GIS (Geographical Information System) digital tool to be able to consider the continuous transformations of the built up landscape framework, in the best possible way. The data collected and summarised can be changed or consulted at any time, also solving the problem of updating the maps.




References

[1] R.Arnheim, La dinamica della forma architettonica, Feltrinelli, Milano, 1981 [2] B.Badiani, Movimento e qualità dello spazio urbano, Ed. Bios, Cosenza,

2004 [3] A.Clementi, Revisioni di paesaggio, Meltemi ed., 2002 [4] J. Gehl, Vita in Città: spazio urbano e relazioni sociali, Maggioli Editore,

Rimini, 1999 [5] C.Lévy-Leboyer, Psicologia dell’ambiente, Laterza, Bari, 1982 [6] A.Sestini, Il paesaggio, TCI, Milano, 1963 [7] F.Steiner, Costruire il paesaggio. Un approccio ecologico alla

pianificazione, McGraw-Hill, 2004 [8] L. Quaroni, Il volto della città, in Comunità n.25, 1954 [9] M.Tira (a cura di), Metropoli e mobilità: il caso di Brescia, vol.4, Sintesi

Ed., Brescia 1996




Evaluation and analysis criteria of the environmental risk factor of the anthropic perturbation in the infrastructure works

G. Gecchele & G. Pizzo Land, Environment and Geo-Engineering Department, Turin Polytechnic, Italy

Abstract

The present study represents a methodological example of the knowledge of the territory through the identification of the qualitative parameters that give technical information of the territory characteristics, of the environmental effects and changes due to anthropic perturbation. In this work some analysis and evaluation criteria of the environmental risk factors are presented, due to the infrastructure works, in order to supply the planning choices and to manage the execution phase, referring to great infrastructure works. In particular, this work, relating to the meaning of impact and environmental quality established by the literature, wants to present a check list for the environmental analysis, which is applied as an example to a phase of an infrastructure work and which represents an important scheme to identify, to analyse and finally to evaluate the necessary information to reach the environmental safety. Keywords: environmental analysis and evaluation, infrastructure works, environmental safety.

1 Introduction

The great civil infrastructure works need specific analyses aimed to identify, evaluate and manage the environmental impacts.

It is then necessary to establish which is the environmental state before the beginning of the works, in order to understand its possible evolution.




doi:10.2495/GEO060351

The study of the relation between the anthropic activities and the environment [1] needs to take into account not only the physical, chemical and biologic components of the environment self, but also the social and economic dynamics of the community, the relations, the structure, the energetic balances [2], in order to understand the whole induced impact.

Every action which is done on the territory produces an alteration of the environmental parameters. Such alteration normally presents positive and negative aspects and has to be compared to the induced benefits and to the environmental and social costs.

The present study wants to study the negative environmental effects produced by the great and continuative infrastructure works; it will be taken into account the only environmental system, leaving to the specific disciplines the complementary analyses of the social and economic dynamics and the relevant costs and benefits.

The criticities of the environmental alterations have to be analyzed at first in the planning phase, by the application of the study models of the ecosystem response deriving to the anthropic pressure levels. The consequent decreasing of the environmental state quality will produce a decreasing of the environmental safety, which is the condition of absence of danger and of safeguard of the population and ecosystem integrity. Such decreasing will be related to the check and control procedures, put into act to manage the impacts.

2 Environmental quality, environmental impact

The environmental impact, according to the definition of Malcevschi [3] is the consequence of the action of a source that, after a more or less complex sequence of events, generates pressures on environmental targets, which could also be altered.

In order to limit the impact due to an intervention on the territory, it is necessary to study in the planning phase the possible modifications induced on the environment and on the existing anthropic activities.

The dimension and the importance of the impacts is related to the possible increasing of the risk on the safety and health of the exposed population [4], to the variation of availability of the existing resources and to the impossibility to reach the aims of safeguard and of environmental quality expressed by the local administration and by the scientific and cultural community. The sum of such factors implies a decreasing of the environmental safety.

These evaluations have to be integrated with the principles of sustainable development, put into act by the local authorities and aimed to the integration of all the interested social, politic, economic, administrative and legal subjects [5].

If the local authorities can put into act the criteria for a local sustainable development, the real problem is the interaction with the global system (national or international) [6] whose dynamics of sustainable development could be different or incompatible with the local decisions.

The infrastructural networks are a typical example of action that often shows the contradiction between the local and global development models that,




although pursuing the same principles, however they don’t collaborate on the dynamics of environmental and social economic development.

The great infrastructure works show how the natural system factors involve in particular the local system, without interacting, sometimes not at all, with the global system.

If it is not possible to understand what will be the destiny of the global ecosystem, it is possible and necessary to study the different scenarios of the evolution deriving from the different strategies and politics of safeguard in the local system.

The analysis of the ecologic systems and their interaction with the social and economic systems, have moreover permitted to put into act adequate programming and control instruments aimed to pursuit a sustainable development of the territory.

In the next paragraph are underlined the correlation between some of these instruments, with particular attention to the evaluation and control instruments.

3 Correlation environmental state – environmental evaluation

The instruments for the environmental evaluation are aimed to analyze the criticities deriving from the perturbation of one or more elements of the environmental sphere, of the health and of the territory.

In order to apply such instruments, it is at first necessary to analyze the level of perturbation of the environment due to a specific pressure and to correlate such perturbation to the existing environmental state. It is then necessary to introduce some definitions that represent the modification of the environmental state.

The “persistence” describes the state in which an environmental system does not modify its inner structure, due to external interferences.

The “resilience” indicates the capacity of a system to maintain its own structure, due to external interferences.

The “inertia” is the capacity of a system to maintain its own dynamicity. The “collapse” is a fast modification of the system directed to a definitive

destruction. It is possible to represent the evolution of the environmental state, due to

pressures generated by the realization of an intervention on the territory, with different cases.

1) In the case 1, there is an event (construction of a work) on time t0. The environment component passes from a condition of persistence to a condition of resilience. If the system can support the intensity of the pressure, then the environmental state, when the work is finished (time tf), could come back to same beginning level and stay in a condition of persistence. The environmental safety is safeguarded.

The environmental impact is represented by the colored area between the straight line without intervention and the curve after the intervention (case1).




Figure 1: Effects of the impacts on the system.

2) If the pressure on the system is important (i.e. the great infrastructure works), after the phase of resilience, the environmental state can reach different levels of persistence, according to the entity of the impact, represented by the area between the straight line without the intervention and the curve after the intervention (case 2). In this case the system has been impacted, but it can come back to a condition of persistence.

3) If the system cannot support the pressure of an intervention, the resilience limit will be exceeded and when the work is finished, the environmental state will present a lower level respect to the beginning condition of persistence, and moreover it will decay (case 3). Such decay will reach the collapse in different times, depending on the inertia of the system.

Considering the complexity of the conflicts due to the interaction between the human activities and the environmental dynamics, the application field of the environmental evaluation is wide [7].

tf t0

Limit ofresilience

time

Environmental Strategic Assessment

Environmental Impact Assessment

Safety Plans Environmental Management System procedures

persistence

resilience inertia

collapse

CASE 1

CASE 2

CASE 3




From a general to a particular application, the evaluation can interest plan and territorial programs (Strategic Environmental Assessment, Ecologic Incidence Assessment), preliminary and definitive projects (Environmental Impact Assessment, Ecologic Incidence Assessment, Safety Plans), existing plants or executive projects (Integrated Environmental Prevention, Operative Safety Plans, Environmental Management System).

It is possible to find a relation between such evaluation instruments and the effects of the impacts on the examined system. The Strategic Environmental Assessment, that represents a preventive instrument of global evaluation, is directly related to the beginning state of the system, since it studies the phase of programming, before the work is planned.

In this programming phase, the Environmental Impact Assessment studies the system after the Strategic Environmental Assessment, but before the perturbation state of the system, since it evaluates the project.

The results of the planning choices after an environmental impact evaluation should make the system maintain its own structure, safeguarding not only the compatibility, but also the sustainability. The Operative Safety Plans and the Environmental System Management (ISO 14000, EMAS) can be acted in the executive phase and are aimed to avoid the exceeding of the resilience limit.

For that situations on the limit of the environmental collapse, the operation of control and mitigation of the impacts are more complex, since the system lost its elasticity.

Such instruments should then be able to manage the impacts from the first phase of execution of the infrastructure works, in order to make the system maintain its inner elasticity.

4 Criteria for the environmental analysis and evaluation

The environmental evaluation of a project or the construction phase of an infrastructure work is not a rigid scheme, but it is an dynamic instrument that depends on the meaning of evaluation. In every case the evaluation is aimed to check the relation between the intervention and the safeguard of the environment, according to the principles of the environmental sustainability.

In order to make detailed evaluations of an area interested by a project, it is at first necessary to make a descriptive analysis of the characteristics of the project self. After a first descriptive analysis there should be a phase of qualitative evaluation and, when it is possible to get data from available sources or direct measurements, the evaluation can be quantitative. Such evaluation, through the use of adequate instruments, can give and estimation of the alterations due to the realization of the project.

The descriptive analysis can be done by check lists, yet used by the control authorities [8], which are a selected list of environmental factors, structured on the different target to be reached (description of the impact related to the project, knowledge of the possible alternatives of the project, indication of the environmental components with indicators).




The present study wants to show a check list which is structured beginning from the description of the activities indicated by the project to the identification of the risk, in order to represent a reference guide for the environmental evaluation.

In tab. 1 there is an example of check list for the environmental analysis in the construction of streets, applied to a classical macro phase, expressed by the literature [9], of preparation, boundary and cleaning of the area interested by the infrastructure work.

The activities are divided into phases and subphases and are linked to the environmental aspects, which are defined by the EMAS as the elements of the activities that have an interaction with the environment [10]. For each environmental aspect the interested environmental component and the consequent risk for the environmental safety have been identified. The risk for the environmental safety is the possible modification of the environmental state, of safety and health of the population that can be represented by pollution, by consumption of raw material and energy, by an annoyance or an environmental disaster.

In the check list specific indicators have been also identified. Such indicators, which can be used in the following evaluation phase, represent and measure particular environmental impacts. Moreover this checklist links the impact to the interested phase of the activity (correlation source – effect).

The items in the check lists have to be identified according to the specific activity and territory which is studied and have to be separated in sub items in order to permit a complete evaluation of the environmental impacts.

In order to apply the indicators to the specific areas which have to be studied, they should be able to represent the measures of the environmental monitoring. It is important to take into account not only the possible damage conditions, but also the annoyance ones. The indicator should then represent the beginning environmental state and measure the variation of the examined parameter. The comparison with the existing situation permits to evaluate if the environmental condition is more or less critical (disturbance – annoyance – damage).

It is then necessary to get information on the environmental system concerning the area interested by a project. Such information, with the analysis of the environmental aspects of the project, make possible to identify and evaluate, through specific indicators, every impact of an infrastructure work.

If the beginning environmental state is good, the impact of an infrastructure work can not be critic, but its effect will be a decreasing of the environmental quality. Moreover if the beginning environmental state is not good, a little decrease of the environmental quality could generate the collapse of the system, that means a condition of damage. The indicators should then be related to the examined system in order to give important information for a correct environmental evaluation [11].




Table 1: Example of check list for the environmental evaluation.

Detail of the operative

phases

Environmental aspects

Environmental components

Identification of the risk for

the environmenta

l safety

Indicators for the

environmental evaluation

MACRO PHASE: PREPARATION, BOUNDARY AND CLEANING OF THE AREA

Antroposphere, biosphere

Local acoustical pollution

Sound level Modification of the vehicular traffic

Atmosphere Local emission

Concentration of combustion products

Boundary

Installation of boundary structures

Antroposphere Visibile impact

Reduction of visible cones

Installation of mobile and stationary equipment

- - - -

Lithosphere Increasing of land vulerability

Concentration of chemical substances

Elimination of loam soil, disafforestation

Hydrosphere Increasing of land porosity

Concentration of chemical substances

Ground leveling Atmosphere Dust production

Concentration of dust

Mucking removal

Lithosphere Waste production

Quantity of waste

Atmosphere Release of combustion products

Concentration of combustion products

Elimination of loam soil, disafforestation, ground leveling

Movement of dumpers and equipment


Acoustical pollution

Sound level

Conveyance and movement of surface waters

Modification of watercourse


Reduction/increment of flow rate

Variation of flow rate




5 Conclusions

The meaning of environment as the sum of the physical, chemical and biologic conditions is reductive, a linear relation cannot take into account every interested part [12], nor represent the relations of an Ecologic, Social and Economic System. The best definition of environment is a whole of systems, with their inner dynamics, relations, structures and energetic balances.

It is important to choice adequate indicators to make a correct evaluation, in order to establish administrative and planning procedures aimed to face the environmental impacts, due to the realization of a project inside a territory.

Such indicators should find a correlation in the different level of schemes for the environmental evaluation, from the Environmental Strategic Assessment, to the Environmental Impact Assessment arriving to the Operative Safety Plans, in the case of the infrastructure works.

The final target is to pursuit a condition of environmental safety, safeguarding the health and integrity of the population and the ecosystem quality, operating then not only for the control of the pollution, but for the environmental, economic and social sustainable development.

The only compatibility is not enough to assure a development finalized to maintain a high ecosystem quality level. It is necessary then to be able to evaluate the capacity of a territory to support the impacts, in order to establish management procedures for the environmental safety.

To reach this target, the methodological instrument for the knowledge of the territory, which is necessary to get information on the effects and modification induced on the environment, should use the analysis and evaluation criteria of the environmental risk described in this study. The operative steps to get information for the environmental safety in the case of an infrastructure work, should be following:

- analysis of the documents. This phase is necessary to understand the legislation and the technical procedures which should be applied to an infrastructure work. It is then necessary to identify and analyze the compulsory and voluntary norms for the environmental, safety and health control and safeguard. Moreover it is necessary to analyze the documents belonging to the local administration and to the company concerning the risk for the environmental safety.

- environmental analysis. The pollution sources are identified and related to their effects. The environmental state is evaluated by the use of appropriate indicators.

- environmental evaluation. This phase wants to underline the environmental criticities of the infrastructure work, by evaluating the environmental impacts, through the studying of an evaluation criteria.

The application of such operative scheme to the infrastructure works will permit to calibrate the analysis methodology indicated by the present study, and will be object of thorough examination.




References

[1] Taylor, William M., The vital landscape : nature and the built environment in nineteenth-century Britain, William M. Taylor. - Aldershot ; Burlington : Ashgate, 2004, 252 pp.

[2] Afgan, Naim H. Sustainable development of energy, water and environment systems : proceedings of the conference on sustainable development of energy, water and environment systems, 2-7 June 2002, Dubrovnik, Croatia, Naim H. Afgan, Zeljko Bogdan, Neven Duic'. - Lisse : Balkema, 2004, 367 pp.

[3] Malcevschi S., Qualità ed impatto ambientale, Etaslibri, Milano, 1991, 355 pp.

[4] Gerrard, Simon, Environmental risk planning and management, Simon Gerrard, R. Kerry Turner and Ian J. Bateman. - Cheltenham : Elgar, 2001, 615 pp.

[5] Clerico M., Pizzo G., Environmental issues for sustainability development in underground works, IABSE Symposium ‘Towards a better built environment – innovation, sustainability, information technology’, Melbourne 11-13 September 2002; p. 106,107; CD p.1-8

[6] Naveh Z., Lieberman A., Landscape Ecology, Springer-Verlag, New York, 1994

[7] Ravetz, Joe, City - Region 2020 : Integrated planning for a sustainable environment, Joe Ravetz. - London : Earthscan, 2000, 307 pp.

[8] Arpa Piemonte, Sostenibilità ambientale dello sviluppo, Arpa Piemonte, Torino 2002, 367 pp.

[9] Comitato Paritetico Territoriale Per La Prevenzione Infortuni, L’igiene E L’ambiente Di Lavoro Di Torino E Provincia, Conoscere per prevenire n.12, Valutazione dei rischi nel settore delle costruzioni, vol.1, CPT, Torino, 2005, 802 pp.

[10] Regulation (EC) No 761/2001 of the European parliament and of the council of 19 March 2001 allowing voluntary participation by organisations in a Community eco-management and audit scheme (EMAS), Official Journal of the European Communities L 114 Vol.44, pag. 1 on 24 April 2001

[11] Australian And New Zeland Environment And Conservation Council 1996. Canberra : Dept. of the Environment and Heritage, 1996

[12] Bottero M., Mondini G., Valle M., Verso una pianificazione ambientale compatibile: i piani territoriali di coordinamento, GEAM rivista della Associazione Georisorse e Ambiente, No. 104, December 2001, pp 265-269.




Section 9 Natural hazards and risks

M3 (Monitoring, Management and Mapping) –

Z. Boukalová, V. Beneš & P. Kořán Cross Czech, G-Impuls, Czech republic

Abstract

Natural disasters are typical examples of people living in conflict with the environment. Vulnerability of populated areas to natural disaster is partly a consequence of spatial planning policies that failed to take account of hazards and risks in land use zoning/development decisions. Thus it is important to combine knowledge, technology, M3 and actors in the field of risk assessment and land use zoning to achieve effective natural disaster prevention and mitigation. Understanding geologic processes is essential to research fields, such as engineering, environmental management, land preservation and restoration, urban environment, soil and water pollution, soil erosion and landscaping. Keywords: monitoring, management, mapping, natural disasters, maintenance of dikes, geophysical methods, dipole electromagnetic profiling, sustainable hazard mitigation, multi-hazard risk assessment.

1 Monitoring, management and mapping

In recent years, due to more and more frequently occurring weather effects of extreme nature which cause disastrous floods, increased attention has been paid to two main issues, in the context of the conflict “people and environment”:

- Inspection and maintenance of dikes and embankments. - Risk mapping of natural hazards (floods) and spatial planning policies

definition.




doi:10.2495/GEO060361

tool for the solution of the conflict: “people and environment”

1.1 Monitoring, management and maintenance of dikes

Inspection, monitoring, management and maintenance or remediation of dikes and embankments is very important task, in recent Europe. Our experience gained in the Czech Republic (projects IMPACT, FLOODsite, ARMONIA) shows that inadequate attention has so far been paid to the documentation of dike breaches and failures after extensive floods. Basic data on the reasons for, and the extent and course of dike breaches are missing in the majority of the cases. Exact data are seldom known, even from the recent disastrous floods in central Europe that occurred in 1997, 1998 and 2002. The data are often incomplete and of insufficient authenticity. However, it is evident that analyses of such information, followed by appropriate adjustments and repairs of the dikes, may significantly reduce the risk of occurrence of new dike breaches and failures. We particularly talk about those dike segments where the reasons for destruction were, for example, inappropriate dike structure, inappropriate material or reduced stream channel capacity due to clogging. Furthermore, after analyzing a database, it often turned out that dike breaches in these sections had occurred repeatedly. Statistical analysis of dike breach parameters may also allow some important generalizations related to the causes and characteristics of breach in specific river basins (catchment areas). For example, it turns out that the prevailing reason for dike breach occurrences in Slovakia is liquefaction caused by seepages in the underlying beds. The main reason for dike failures in Hungary is overtopping. Entirely different mechanisms of dike breach occurrences of course require different types of preventive dike modifications. At present, dike maintenance and preventive repairs are based on a system of visual inspection complemented by analyses of airborne or satellite photographs, or sometimes on slow intrusive methods (boreholes drilling etc.) results. Only rarely is the project documentation of dikes and embankments complemented by detailed information on structures and material properties, i.e. information acquired by engineering-geological investigation, drilling, laboratory tests of soils, etc. The reason for this is the considerable cost of such investigation and the large extent of the dikes. However, we believe that information on the nature of materials and basic dike structure is essential for efficient failure prevention. This particularly applies to old dikes for which construction documentation is missing. Furthermore, in some countries (for example, developing countries or countries of former East Europe) we may expect low quality of construction work that may contribute to dike breach when stressed (see Fig. 1). It is in this area that a package of geophysical methods can be of particular value. Geophysical methods investigation and monitoring provide a continuous image of physical properties of a dike body and, furthermore, this type of investigation is relatively inexpensive. Last years, we concentrated on testing the possibilities of application of the following geophysical methods:

- Geo-electric methods resistivity profiling (RP), self potential method (SP), multielectrode method (MEM), electromagnetic frequency method (EFM).




Figure 1: Example of inappropriate material in the core of a damaged dike.

- Seismic methods shallow seismic method (SSM), seismic tomography (ST), multi-channel analysis of seismic waves (MASW).

- Microgravimetric method. - GPR method. - Geomagnetic survey, gamma-ray spectrometric survey.

Potential now exists to apply an innovation: a Geophysical investigation tool, which is based on generating electro-magnetic (EM) fields and mapping their propagation through soils and structures. The method finds perturbations in the EM fields arising from concealed boundaries or changed materials. Previous tools and approaches have shown limited effectiveness and poor data interpretation accuracy compared with conventional intrusive geotechnical investigation and description methods. Speed has been affected by the need to repeatedly remount transmitters and receivers. The new technology, so-called Geophysical Methods Suite – GMS (using GEM-2 tool), could provide a breakthrough in an area of science in which infrastructure managers and engineers have been highly skeptical. Experience gained in the Czech Republic shows that the new technology GMS could support national government asset owners across Europe. The GMS is proposed to enable rapid, economic and repeatable identification of non-homogeneities (possible points of disruption) in line embankments and is especially promising for the identification of problem or weak spots in line embankments for water management and flood defense, where embankments could be subject to rupture under extreme hydrological conditions. Thanks to fairly fast and economical




monitoring procedures it should also be possible to use the technology to facilitate the maintenance of line embankments in a state which should prevent leakage and ruptures. More, this technology should make it possible to check fairly long sections of line embankments in relatively short time (about 10 km per day), whilst maintaining adequate levels of precision. What is innovative about the proposed GMS? Having used the new tool to define the hot spots of an existing embankment system, detailed investigation, maintenance and renewal efforts can be concentrated in a cost-effective way on the critical parts of the embankments. The core part of the GMS is Dipole Electromagnetic Profiling with the unique brand new and innovative apparatus the GEM-2 (multi-frequency device). In order to incorporate the geophysical methods into a complex of dike prevention and maintenance, we first have to identify the effects that can be monitored by these methods. Figure 2 illustrates an approach to incorporation of geophysical methods into a dike and embankments maintenance program. From the viewpoint of dike maintenance – dike breach, timing of the action is of central importance. The breach formation itself takes place at a time scale of hours, max. a few days. A hazardous segment is evident, application of geophysical measurements is not assumed here. However, except for overtopping, the remaining defects mostly show somewhat hidden PRE-breach formation stage (for example, seepage through the underlying beds, repeated seepage at an increased water level, structure defects, etc.) which predisposes the point of future dike breach. This stage often lasts for even tens of years and is our area of interest for the application of geophysical methods. The database of quick testing measurements, which is the basic component of the monitoring system (see Fig.2), provides a basic description of dike materials and structures, division of dikes into quasi-homogeneous blocks (i.e. dike segments showing similar geotechnical and physical properties). Productivity of measurement is rather high, based on the dike character ranging between 10 and 20 km of a dike per day. From the viewpoint of dike maintenance, these data are an appropriate complement to a visual inspection, allowing us to assess relative permeability of the dike material and its homogeneity and to detect subsurface distribution systems reaching a dike, etc. This allows us to more precisely identify problematic dike segments that are disturbed and weakened inside. All types of dike-testing measurements should be linked to GPS. The GMS system: is composed of 3 basic building blocks: Quick testing measurement – fast and cheap measurement for basic evaluation of the dike condition and homogenity within the whole river-basin. This method is also the core for repeated (monitoring) measurement. As a method for this purpose we suggest DEMP using multi-frequency tool (for example GEM-2). Diagnostic measurement – detailed measurement of the eroded (non-homogeneous) sections aimed at finding hidden defects of the dikes. The method is based on the application of the set of geo-electric methods, especially multi-probe resistance method MEM complemented by another independent method based on the type of the defect searched for accordingly.




Figure 2: A diagram of incorporation of geophysical methods into dike maintenance.

Measurement of geotechnical conditions – geophysical measurement to monitor geomechanical conditions of eroded dike sections. For the analysis of dikes geomechanical characteristics; especially seismic methods and micro-gravimetry will be used. The GMS asset lies in the possibility of objective evaluation of dike homogeneity and condition. Geophysical methods are suitable supplement for current methods of checks (visual check, aerial and satellite pictures analysis). Monitoring function of GMS lies in the analysis of relative changes of geophysical parameters. GMS database construction requires so called initial stage and so called following check stages.

input data

- better fit Large-scale,

laboratory and mathematical

simulation

better understandi

ng of breach

Time

YEAR MONTH DAY HOUR

Breach Formation

OvertoppingPiping

Slope deformation

Seepages through or below the dike

Dike structure defects

Repeated high water attacks

Actual dike condition

Internal erosion

P R E -effective and timely maintenanceprevents the defects from occurring

Geophysical Monitoring Systemnewly included in inspection and maintenance of thedikes (together with airborne photographs analyses,visual inspection, etc.)

Database of quick testing easurements

Monitoring of geomechanical

properties Diagnostics of

problematic segments




Initial stage of GMS is based on quick testing measurement of dikes within the whole area of the river-basin and following diagnostics measurement of selected problem sections. These checks result in complex evaluation of dikes condition in the river-basin including the suggestion of the necessary repairs. Special selected sections may be checked using measurement of geotechnical condition. Check stages are planned for the dikes which are constructed to protect some place against high water level (flood protection). In Europe, the check stage should be carried out after 3 years at the latest without the reference to flood conditions. Check stages may of course be carried out based on agreement at whatever times according to the needs of the dike owner/caretaker, e.g. in limited time setting during floods. The check stage includes repeated quick testing measurement and comparison of the acquired data with the data from previous stages (when using the results to eliminate the influence of climatic conditions). Thanks to analysis of repeated measurements we are able to locate time unstable anomaly areas which often coincide with the places where the dike ruptures occur. The check stage can be supplemented with diagnostic measurement if needed. The GMS system success is largely based on narrow co-operation between geophysics specialists and dikes caretakers. They have large quantity of information which can help in making the geophysical measurements interpretation much more precise. Without mutual trust and communication the GMS database program has no meaning.

1.2 Mapping of natural risks and hazards

The implementation of flood risk assessment procedures for spatial planning in order to prevent and mitigate floods in urban areas is very important issue. From October 2004, the project ARMONIA is trying to find right solution regarding the implementation of multi-risk assessment procedures for spatial planning in order to prevent and mitigate natural disasters in an urban environment. The main objective of the project is to provide guidelines for EU standards on the harmonisation of data, methodologies and maps related to all main natural hazards acting on urban areas, incorporating consideration of climate change impacts. ARMONIA seeks to achieve outcomes that can mitigate the adverse effects of natural phenomena through joint effort of the scientific community, technology experts and users. The target is not only a scientific output, but a measurable impact on policies/practices for disaster mitigation initiated within the period of the project, which fits with Europe's goals regarding sustainable development in supporting environmental and security policies by facilitating and fostering the timely provision of quality data, information and knowledge, developing tools and improving management practices. Disaster management in Europe currently suffers from following main reasons: There is no common shared strategy at European level for the




prevention and mitigation of natural disasters, mainly when dealing with the integrated and combined impacts of natural hazards on modern society, including secondary social effects. Main important initiatives are on regional or national levels. As disasters are often affecting several countries with heavy transboundary effects (as for example the floods 2002), there is important to support common multidisciplinary EU activities focused on prevention and mitigation of natural disasters. There is the need for a modern “disaster science” which can better deal with the complexity of reality in Europe (taking into consideration dynamic systematic interactions). The common accepted “natural and induced technological disaster management cycle”, including stages of prevention, mitigation, preparedness, response and recovery is a theoretical view that is usually marginally working in practice: the synergic interaction and implication of various hazards and risks are missing. A new approach for the harmonization and further development of all component of the disaster cycle is needed. The effective use of available technological tools in the field of disaster management by end users is not common in many European countries, especially in some of the new member states. More, the sustainable hazard mitigation programmes addressing both the short and long term consequences of their implementation in a holistic manner have not been adequately included in the agenda of European research, end users and stakeholders. The project ARMONIA should bring the important outputs and solve the issues as harmonisation of different risk mapping processes for standardizing data collection/analysis, monitoring, outputs and terminology for end users and optimisation of methodologies for hazard/risk assessment for different types of disastrous events. More, the project will suggest the design of a harmonised decision-making tool for applying hazard and risk mitigation in spatial planning and optimisation of a guideline on natural hazard mitigation in the context of the EU Environmental Assessment Directive (2001/42/EC), till end of year 2007. Human and economic losses due to natural disasters continue to increase world-wide. The high importance of disaster reduction policies was stressed at the UN-World Conference on Disaster Reduction (January 2005, Kobe, Japan). As a consequence there is a need to integrate knowledge, technology and actors and to update state of the art into a disaster management approach that reflects the complexity of the modern society in a realistic way.

2 Conclusions

There is an urgent need for better disaster and vulnerability reduction support actions, in the catchment scale. In order to achieve overall management of water resources, at river basin level in particular, it is of prime importance for decision-makers (Directors of River Basin Organizations and Administrations, Basin Committee members, representatives of Local Authorities and associations of users) to have easy access to comprehensive, representative and reliable information, at all relevant levels. The M3 (Monitoring, Management and Mapping) tool, adapted on the modern society would be the best start to find




realistic and end users friendly “disaster and vulnerability reduction support actions” in the context of the conflict: “people and environment”.

References

[1] ARMONIA: Applied multi Risk Mapping of Natural Hazards for Impact Assessment, FP6 project n. 511208; web: www. armoniaproject.net/html4

[2] Boukalova Z., Beneš V.(2004): Case studies and geophysical methods. Association of State Dam Safety Officials: Dam Safety 2004. Phoenix, Arizona, USA, September 2004.

[3] Boukalova Z., Beneš V. (2005): Long-term monitoring of geotechnical state of flood protection dikes using non-destructive geophysical methods, report. Cross Czech a.s., Prague, 2005.

[4] Margottini C. (2006): Natural hazards and economic impact in Europe: some concepts for the 7FP. International Symposium on Climate Challenges. Brussels, 2-3 February 2006.




Gas hazard: an often neglected natural risk in volcanic areas

W. D’Alessandro Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, Italy

Abstract

Volcanic areas release huge amounts of gases, which apart from having important influences on the global climate could have strong impact on human health. Gases have both acute and chronic effects. Carbon Dioxide and Sulphur gases are the main gases responsible for acute mortality due to their asphyxiating and/or toxic properties. On the contrary Mercury and Radon have important chronic effects respectively for its toxicity and radioactivity. The problem has long been neglected until the “Lake Nyos” catastrophe in 1986, in which about 1700 people were killed by a volcanic CO2 emission, attracted the worldwide attention of the mass media. In this paper we present some studies on gas hazards in three different volcanic systems chosen for their different activity status: Mt. Etna (Italy), characterised by frequent activity with a mean CO2 emission of about 450 kg s-1; Pantelleria island (Italy) at present in quiescent status and a CO2 emission of about 12 kg s-1; and Sousaki (Greece) a recent (Quaternary) but now extinct volcano with a CO2 emission of about 0.6 kg s-1. In all three systems the main problems arise from CO2 emissions while secondary problems are due to SO2 and Hg (Etna), H2S (Sousaki) and Rn (Pantelleria). Keywords: gas hazard, Carbon Dioxide, Sulphur gases, Radon, Mercury.

1 Introduction

Pliny the Elder, who perished during the 79 A.D. eruption of Vesuvius in Italy, is possibly the most famous known victim of volcanic gases. The description of his death, made by his nephew Pliny the Younger (Letter 16, 6), clearly points to cardio-respiratory collapse of a person with chronic respiratory disease.




doi:10.2495/GEO060371

Although it is unclear if the fatal causes were volcanic gases or fine suspended ash or even a heart attack, the description closely resembles that reported for one of the victims of volcanic gases at Aso volcano in Japan in recent times, who suffered from pulmonary emphysema [1]. Gas manifestations have been recognized since olden times and often sanctuaries were built near them since the Neolithic to get protection against them from divinities. In these sanctuaries sometimes oracles gave their responses under the narcotic effect of elevated CO2 concentrations. An important sanctuary for example was placed at the Palici Lake in Sicily (5th cent. B.C.). The abundant CO2 released from its shores and trough its waters was used to test the truth of persons [2]. Standing on the shore of the lake the person had to swear in the name of the sanctuary’s gods looking to the sky, and then had to touch the waters with his hands. If he had no consequences he was saying the truth. Probably in wind free days there was a higher proportion of ‘liars’. Nowadays fluxes in the area are still very high [3] but no signs of the sanctuary is still present, the lake was dried up and CO2 is industrially exploited for gas addition to beverages. On February 20th 1979 at the onset of eruptive activity of Sinila crater on Dieng Plateau, Indonesia, 142 persons, fleeing from the nearby Batur village, died asphyxiated by a cloud of CO2-rich volcanic gas rolling down the flank of the volcano [4]. But it was not until 1986 that a wide public knew about the risk of volcanic gases. On August 21st of that year about 1700 people were killed and 850 injured by a massive gas release from Lake Nyos, Cameroon [5]. The gas release of this volcanic crater lake, composed almost exclusively of CO2, was not related to any paroxysmal volcanic activity. The release occurred on a calm night and the gas cloud, heavier than air, flowed in a valley that was more than 250 m below the surface level of the lake. Recently Witham [6] published a new database of volcanic disasters and incidents of the 20th century. Of the 491 reported events 11% were referred to volcanic gases, accounting for 2016 (2% of the total) of the killed and 2860 (18%) of the injured people. As the author himself states, this database has to be improved and completed and indeed only one of the many incidents occurred in Italy has been reported. But it is equally impressive to learn that the Lake Nyos disaster ranks at the 8th place for the number of dead and at the 5th place for the number of injured in this database. An important research project aiming at the study of diffuse degassing of the whole Italian peninsula financed by Italian Civil Protection is currently in progress [7]. The new awareness of the importance of determining the risk connected with natural gases mainly of volcanic origin probably derives from the expansion of Rome’s metropolitan area well within the area of Albani Hills. This recent, possibly still active, volcanic area is characterised by anomalous gas fluxes, which led to some fatal outcomes in the last years [8].

Sometimes the risk associated to gases is enhanced by improper human operations. For example an area within the urban area of Rome is naturally protected from high gas fluxes from the soil by shallow impermeable sediments. But excavation or drilling operations sometimes perforate these impermeable layers leading to massive gas outflows [8]. One of these episodes evidenced that




both of the medical and fire brigade structures were (and probably still are) unprepared to face this kind of emergency. The doctors in fact didn’t recognize the CO2 intoxication of 7 persons that lived close to the drilled well, and attributed ailments to collective hallucination because no clinical signs were still present at the emergency station. Returning to their homes these persons escaped death only by chance and were eventually hospitalised. On the other hand the remedy applied by fire brigade, to drill additional wells trying to exhaust the gas source, had as the only consequence to exacerbate the problem that was fixed only sealing all the drilled wells with the injection of special cements [8].

2 Hazardous gases

2.1 Carbon dioxide

Carbon dioxide is the more abundant gas, after water vapour, released by volcanic activity. Total CO2 output of the different volcanic systems span over a wide range of values, from about 450 kg s-1 of Mt. Etna (Italy) and Popocatepetl (Mexico) down to values of less than 0.1 kg s-1 (Iwoyama, Japan) [9]. Open conduit volcanoes, like Mt. Etna and Popocatepetl, emit this gas mainly through the craters. On the contrary volcanoes with closed conduit, even if characterised by intense fumarolic activity, release CO2 almost exclusively through diffuse soil degassing [9]. Carbon dioxide annually released by volcanoes at the global scale was estimated in about 300 Tg, which represents only a small fraction (1%) of the CO2 released by human activities [10]. The geographical distribution of natural CO2 release corresponds prevailingly to areas of active or recent (< 10 Ma) volcanism and seismicity [10]. Carbon dioxide has different origins: biogenic (respiration), hydrocarbon oxidation, thermal or chemical breakdown of limestone and mantle degassing. The first two processes very rarely create dangerous natural CO2 accumulations. Carbonate minerals in the crust can release huge quantities of CO2 both through thermal breakdown due to high geothermal gradients and through reaction with acid hydrothermal fluids. Magma ascent from the earth’s mantle is a very efficient CO2 degassing process. Mantle-generated magmas contain up to 1.5% by weight of CO2, but its low solubility leads to early (deep) gas separation during ascent of magmas. Considering that CO2-depleted magmas could eventually solidify within the crust, some volcanic system, like Stromboli Island (Italy), emit on long time average by weight more gas than lava. Normal CO2 concentration in the atmosphere at sea level is about 350 ppm (by volume) but its concentration can rise if production exceeds consumption and dissipation. Being heavier than air, in high flux areas, CO2 can accumulate in topographic depressions and enclosures reaching concentrations as high as 100%. Carbon dioxide concentrations higher than 10% can be lethal to humans and animals, and at concentrations above 20-30% even a few breaths can very quickly lead to unconsciousness and death from acute hypoxia, severe acidosis and respiratory paralysis [11]. Therefore, poorly ventilated places below and




immediately above ground such as caves, galleries, cellars, water wells, etc. can be very dangerous in areas of anomalous CO2 emissions.

2.2 Sulphur gases

Sulphur gases follow carbon dioxide in order of abundance in volcanic gases. The most important species are SO2 and H2S and their relative abundance is fixed by the thermodynamic parameters within the volcanic system (pressure, temperature, oxygen fugacity). Low temperatures and the presence of a hydrothermal system favour H2S. Open conduit volcanoes display high SO2/H2S ratios (> 20) while on the other hand volcanoes with low temperature fumarolic activity display low ratios (< 0.1). Sulfur dioxide is a highly irritant gas. In healthy persons 5 to 10 ppm SO2 cause eye, nose and throat irritation while 30 to 40 ppm can lead to respiratory failure. In individuals with bronchial asthma or other chronic lung diseases exposure to much lower levels (0.25 to 0.5 ppm) can be life threatening [1]. Concentrations of up to a few ppm can easily be achieved close to active volcanic craters. Indeed six fatalities due to SO2 have been documented at Mt. Aso volcano, Japan in the period 1989-1997, while other 59 persons had to be hospitalized in the period 1980-1995 [1]. Five of the dead and 29 of the injured suffered from chronic lung diseases, evidencing the higher risk for this class of people. In 1996 a monitoring system for the measurement of SO2 in air was set up in the crater area. No one could enter the area when SO2 level was exceeding 5 ppm, while an advise was given when it was exceeding 0.2 ppm for more than 5 min discouraging the visit of the crater. People suffering of respiratory and cardiovascular diseases were asked to vacate the area when SO2 was exceeding 2.5 ppm for more then 5 min. But such precautions were not able to save the lives of two persons on 23 November 1997. So it was decided to apply a more rigorous criterion, forbidding the access if SO2 was exceeding 0.2 ppm for more than 1 min or in the presence of instantaneous peaks exceeding 5 ppm [1]. Hydrogen sulphide is both an irritant and asphyxiant gas. Levels of up to 20 ppm have generally no effects on healthy people while for asthmatic persons this level have to be reduced to 2 ppm [12]. Concentrations above 20 ppm cause irritative effects on eyes and respiratory tract, above 50-100 ppm neurotoxic effects appear and 500-1000 ppm are considered of immediate life danger [12]. Although the human odour threshold is very low (0.02 ppm) the warning signal is lost above 150 ppm because of olfactory nerve paralysis by H2S itself. Like CO2 also H2S, being heavier than air, tend to accumulate in closed and/or depressed areas. Some of the fatal incidents with volcanic gases were attributed to the effect of H2S released by low temperature fumarolic vents or by gas bubbling through thermal springs. Furthermore, although not conclusive, some studies evidenced also health effects from chronic exposure to H2S (0.1 – 2 ppm) in the city of Rotorua (New Zealand), which is built on a geothermal field [13].

2.3 Radon

Radon is a natural radioactive gas being and intermediate product of the radioactive decay series of Uranium and Thorium. Radon can easily enter the




human body by inhalation. Radioactivity of Radon and of its decay products has been linked to an increase in the risk of developing lung cancer. Only smoking is a greater risk factor for lung cancer with respect to Radon inhalation [14]. The concentration of Radon in the indoor air and in soil atmosphere depends from many factors. One the most important is the content of parent elements (Radium, Uranium, Thorium etc.) in the rocks of the subsoil or of the building material. Furthermore, release and transport of radon are controlled by nature and alteration of the containing minerals, moisture content, and by the nature (i.e. carbon dioxide gas or groundwater) and flux of a carrier fluid [14]. In volcanic and geothermal areas higher risk of Radon accumulation occur where differentiated magmas, enriched in parent elements, are involved, where geothermal alteration is widespread and where anomalous gas fluxes are present.

2.4 Mercury

Mercury is a highly volatile, bioaccumulating toxic trace metal strongly enriched in volcanic and geothermal emanations. The contribution of volcanic activity (about 700 Mg yr-1), although highly debated, represents 20-40% of the global natural emission [15]. Very high mercury concentrations in air (up to 40 µg m-3) have been measured in volcanic and geothermal areas of Hawaii and of Iceland [16], well above the guideline value of 1 µg m-3 recommended by WHO for general population exposition and sometimes also above the occupational long-term exposure limit of 25 µg m-3. Although mercury accumulation in professionally exposed persons (volcanologists, guides, employed of the geothermal industry) has been demonstrated [17], no study on the neurotoxic effects of volcanic-derived mercury in this population has been until now performed.

3 Gas hazard at selected volcanoes

3.1 Mt. Etna

Mount Etna, located in eastern Sicily, is the largest strato-volcano in Europe (3300 m a.s.l.; base, 60-40 km) and one of the most active in the world. It grew in proximity to the collision boundary of the African and Eurasian continental plates, from repeated eruptions of alkali basalts-hawaiites over the last 200 ka [18]. Its recent activity is characterized by permanent open-conduit passive degassing, interrupted by paroxysmal activity (effusive to moderately explosive) at the summit craters and/or newly formed flank craters. At present, Etna’s central conduit feeds four summit craters called Voragine, Bocca Nuova, South-East and North-East. Degassing at the summit craters has continued without interruption in the last few decades, although only rarely have all the craters been degassing contemporaneously. Mt. Etna is considered, on long time average, the greatest point source of many volatile compounds to the atmosphere [18].




3.1.1 Lower flanks Etna emits yearly about 1 Tg of CO2 diffusely through the soils of its flanks. This huge amount represents about 10% of the total CO2 output, the greatest part being released through the summit craters. Spatial distribution of diffuse degassing is strongly controlled by the tectonic setting of the volcanic system and two anomalous degassing areas on the SW and on the E lower flanks have been recognized [18]. In these areas very high values of CO2 fluxes from the soil (up to 500 g m-2 d-1) and of CO2 partial pressure in groundwaters (up to 5 atm) are measured. These areas are also densely settled and intensively cultivated due to the high fertility of the soils and to their huge groundwater resources. Groundwater is exploited through wells and drainage galleries. Until recent times wells were excavated with large diameters (generally 2 m) up to depths of 400 m. These wells, still diffused and in use, in the two anomalous degassing areas display often dangerous accumulations of CO2 on their bottoms. People in the area know the problem very well and all operations inside the wells are generally made by expert persons under steady air-pumping, in worse cases with two or more independent pumping systems connected to autonomous electric generators. Such precautions warrant a reasonable safety and no fatal accidents have been registered at least in the last 30 years period. But in the last years an increasing trend of land abandonment left many of these wells unattended so that unprepared persons could get inside unaware of the high risk. For the mitigation of this risk it would be very important: i) to take census of all the old wells, at least, in the two anomalous degassing areas and ii) regularly control their state and if abandoned close all possible access and evidence the danger with adequate warning notices. Groundwaters are also extracted from the flanks of Mt. Etna with drainage galleries (or horizontal wells). A few tens of these galleries have length of more than 1000 m and water yields of 0.1-1 m3 s-1. Some of these galleries, built in the anomalous degassing areas, emit together with the groundwater also huge quantities of CO2, which represents a big problem not only inside the gallery but also outside its exit. The gas, in fact, tend to follow the water flowing in the canal at the exit of the gallery and in wind free days maintain dangerous concentrations of some % up to distances of some tens of meters. In May 1993 one fatal incident was registered at the gallery known as ‘Ponteferro’ near the village of S. Venerina on the eastern flank of the volcano. There was no eye-witness of the episode but the reconstruction points to the huge gas release as one of the main causes. The victim was an old man that kneeing down to collect water from the canal with a bottle, loosed consciousness due to the high CO2 concentration and eventually drowned falling in the canal. The access to the area, where many people come to collect drinking water, was immediately restricted to supervised periods and the canal was covered with a metal grid.

3.1.2 Summit crater area The access to the summit area is generally unrestricted in periods of low volcanic activity. On the crater rims in the downwind direction lethal gas concentrations




can easily been reached. But only volcanologists have sometimes the necessity to go inside the volcanic gas plume to collect samples and in this case they have to wear efficient gas masks and leave the area as soon as possible. Strong winds and water vapour condensation that evidences the plume help all other people to avoid areas with dangerous gas concentrations. Dangerous conditions could arise in rare cases, when wind blows very mildly or changes rapidly direction, but fortunately only healthy people reach the crater rims after a hard climb and they can stand higher concentrations. But there’re two areas close to the summit craters, Pizzi Deneri (2800 m a.s.l. about 2 km NE from the summit craters) and Torre del Filosofo (2900 m a.s.l. about 1 km S from the summit craters), where tourists are transported with all-wheel drive busses. Often between these tourists there’re elderly people that possibly suffer of chronic lung disease. No danger advice is given to these people of the possible effects of volcanic gases. In my opinion it is only by chance than until now there have been no serious consequences. The few measurements made in these two areas, in fact, gave monthly average values between 0.2 and 0.5 ppm [19], which are life threatening for persons affected by chronic lung disease [1]. For this reason it would be desirable to set up a warning and monitoring system like that of Mt. Aso in Japan [1]. Torre del Filosofo is also an area of anomalous soil degassing connected to old eruptive fractures. There is also an abandoned hut that was used for long time as shelter for volcanological surveillance instruments. One of its rooms, partially below ground level, used in winter times as emergency toilet, suffered for long time of CO2 accumulation but fortunately ventilation was always enough to prevent dangerous concentrations. In November 2003, during an eruption the hut was completely covered by volcanic ash. In the following summer the hut was partially uncovered to show its rests to the tourists, creating an access to some of its rooms. But these rooms that could represent an attractive shelter for excursionists during winter, in recent times (summer 2005) displayed very dangerous CO2 concentrations up to some percent. Recent measurements evidenced also high concentrations of mercury in the summit crater area (E. Bagnato and S. Giammanco – personal communication) with values of some µg m-3 in the volcanic plume and some hundreds of ng m-3 in the atmosphere in the Torre del Filosofo area. In this case the most exposed persons are not the tourists but the guides that spend a lot of time in the area.

3.2 Pantelleria Island

The island of Pantelleria represents the top of a large active volcano developed on the African continental crust. The volcano is located between Sicily and Tunisia, within NW-SE trending tectonic depressions related to the opening of a rift system developed during the Neogene–Quaternary [20]. The island is dominantly composed by volcanics showing compositional variations ranging from basalts to peralkaline rhyolites (pantellerites) ranging in age from 320 ka to the present. The volcanic activity of the island is mainly characterized by violent explosive eruptions and secondarily by basaltic effusive activity [20]. Its last eruption occurred in 1891 about 5 km NW off its coast. The island is at present




characterized by widespread occurrence of surface hydrothermal manifestations (fumaroles and thermal springs) [21]. Also Pantelleria island displays two anomalous degassing areas with cold (mofettes, bubbling gases) and hot (fumaroles) gas manifestations and areas of strong CO2 fluxes from the soil [21]. One of these mofettes is placed in a topographically depressed area near Lake Specchio di Venere. In wind-free days CO2 accumulates near the ground being a lethal trap for small animals. Such conditions could be very dangerous also for human that would lay down resting on the shores of the lake. Fortunately in recent times warning signals have been set up, but it would probably be safer also to fence the most dangerous area (some tens of m2). Furthermore in the inner of the island there’re many natural saunas exploiting the heath of fumaroles. Almost all of them are fed by fumarolic gases composed exclusively by water vapour and atmospheric air, but a few, within one of the anomalous degassing areas, contain also dangerous concentrations of CO2. These improvised saunas are built by foreign people that live on the island only on holidays and are not aware of this type of risk. Recently Radon measurements in soils and in dwellings of the island evidenced for most of the island very high values [22]. The highest values were detected in central and southern part of the island where the most radionuclide-rich rocks crop out. The higher natural radiation doses absorbed by the local population has been considered an additional risk factor in developing some kind of cancer and indeed a statistical study confirmed their higher incidence [Brai M. personal communication].

3.3 Sousaki

The Sousaki area is located in Greece, about 65 km west from Athens, near the Isthmus of Corinth and represents the NW end of the active Aegean volcanic arc. Here, sparse outcrops of dacitic rocks are the remnants of late-Pliocene to Quaternary volcanic activity (4.0 – 2.3 Ma [23]), while widespread fumarolic alteration and warm (35 – 45°C) gas emissions are still recognizable. Drilling exploration assessed the presence of a low enthalpy geothermal field. In the area showing the highest hydrothermal alteration, located along a narrow valley, several small caves were dug in the past century to extract hydrothermal alteration minerals (alunite, magnesite, sulfur). Some of these caves display at present hydrothermal gas emission from their bottoms. The gases, being denser than atmospheric air, flow on the grounds of the caves and eventually spill out from the mouth of the caves dispersing in the atmosphere after descending the flanks of the valley. The gases are composed of more than 90% of CO2 and have also high concentrations of H2S (1000 – 6000 ppm). Hazardous concentrations of CO2 and H2S are measured inside the caves were dead small animals are always found. But H2S dispersing in the surrounding atmosphere displays concentrations of some ppm that produces surely annoying smell and represents also a potential chronic health impact for the nearby living persons.




4 Conclusions

In the 20th century more than 2000 persons died and nearly 3000 were injured by volcanic gases. The most dangerous gas species is CO2, responsible of more than 90% of the victims and of the worst episodes (Lake Nyos and Lake Monoun, Cameroon and Dieng Plateau, Indonesia), but lethal episodes are also attributed to sulphur gases (SO2 and H2S). Furthermore also Mercury and Radon have dangerous chronic effects on human health due to their toxicity and radioactivity. Gas hazard is often disregarded because it is almost always connected to low or absent volcanic activity when attention is low. But we know that volcanic gases and especially CO2 can be released by volcanic systems up to some million years after volcanic activity ended. It is therefore important not to underestimate potential risks and the effort of the Italian scientific community under the patronage of Civil Protection for the understanding and mitigation of this natural risk goes in the right direction.

References

[1] Ng’Walali, P.M., Koreeda, A., Kibayashi, K., Tsunenari, S., Fatalities by inhalation of volcanic gas at Mt. Aso crater in Kumamoto, Japan, Legal Medicine 1, 180–184, 1999.

[2] Ferrara, F., Memoria sopra il lago de Palici ora lago Naftia in Sicilia, Reale Stamperia, Palermo, 1805.

[3] De Gregorio, S., Diliberto, I.S., Giammanco, S., Gurrieri, S. & Valenza, M., Tectonic control over large-scale diffuse degassing in eastern Sicily (Italy), Geofluids 2, 273-284, 2002.

[4] Le Guern, F., Tazieff, H. & Faivre–Pierret, R., An example of health hazard: people killed by gas during a phreatic eruption: Dieng Plateau (Java, Indonesia), February 20th 1979. Bulletin Volcanologique 45, 153–156, 1982.

[5] Baxter, P.J., Kapila, M. & Mfonfu, D., Lake Nyos disaster, Cameroon, 1986: the medical effects of large scale emission of carbon dioxide? British Medical Journal 298, 1437-1441, 1989.

[6] Witham, C.S., Volcanic disasters and incidents: A new database, Journal of Volcanology and Geothermal Research 148, 191–233, 2005.

[7] http://www.ingv.it/progettiSV/Progetti/Vulcanologici/vulcanologici_con_frame.htm

[8] Carapezza, M.L., Ranaldi, M. & Tarchini, L., Gas hazard in the Roman area: soil CO2 discharge and accidents induced by uncontrolled shallow boreholes, Abstract Book of the 8th International Conference on Gas Geochemistry ICGG 8, Palermo and Milazzo, Italy, 2-8 October 2005, p. 12, 2005.

[9] Pecoraino, G., Brusca, L., D’Alessandro, W., Giammanco, S., Inguaggiato, S. & Longo, M., Total CO2 output from Ischia Island volcano (Italy). Geochemical Journal 39, 451-458, 2005.




[10] Mörner, N.A. & Etiope, G., Carbon degassing from the lithosphere, Global and Planetary Change 33, 185-203, 2002.

[11] Henderson, Y. & Haggard, H.W., Noxious gases, Reinhold Pub. Co, 1943. [12] World Health Organization, Hydrogen Sulfide: Human health aspects,

Concise International Chemical Assessment Document 53, 26 pp., 2003. [13] Bates, M.N., Garret, N. & Shoemack, P., Investigation of health effects of

Hydrogen Sulfide from a geothermal source, Archives of Environmental Health 57, 405-411, 2002.

[14] Appleton, J.D., Radon in air and water, in: Essentials of Medical Geology (Selinus, O., Alloway, B. Centeno, J.A., Finkelman, R.B., Fuge, R., Lindh, U. & Smedley, P. eds.), Elsevier, 2005.

[15] Pyle, D.M. & Mather, T.A., The importance of volcanic emissions for the global atmospheric mercury cycle, Atmospheric Environment 37, 5115-5124, 2003.

[16] Siegel, S.M. & Siegel, B.Z., Geothermal hazard. Mercury emission, Environmental Science and Technology 9, 473-474, 1975.

[17] Siegel B.Z., & Siegel, S.M., Mercury in human hair: uncertainties in the meaning and significance of ‘unexposed’ and ‘exposed’ in sample population, Water Air and Soil Pollution 26, 191-199, 1985.

[18] Calvari, S., Bonaccorso, A., Coltelli, M., Del Negro, C. & Falsaperla, S. (eds.) Etna Volcano Laboratory, Geophysical Monography Series, AGU, 143, 2004.

[19] Aiuppa, A., D’Alessandro, W., Federico, C., Ferm, M. & Valenza M., Volcanic plume monitoring at Mount Etna by diffusive (passive) sampling. Journal of Geophysical Research 109(D21), D21308, 2004.

[20] Civetta, L., Cornette, Y., Gillot, P.Y. & Orsi, G., The eruptive history of Pantelleria (Sicily Channel) in the last 50 ka, Bulletin of Volcanology 50, 47-57.

[21] Favara, R., Giammanco, S., Inguaggiato, S. & Pecoraino, G., Preliminary estimate of CO2 output from Pantelleria island volcano (Sicily, Italy): evidence of active mantle degassing, Applied Geochemistry 16, 883-894, 2000.

[22] Cinti, D., Pizzino, L., Galli, G., Quattrocchi, F. & Voltattorni, N., Radon and thoron soil survey in the Pantelleria island: first results. Abstract book of the Annual Assembly of Italian National Group for Volcanology, Naples, 20-22 December 2004.

[23] Pe-Piper, G. & Hatzipanagiotou, K., The Pliocene volcanic rocks of Crommyonia, western Greece and their implications for the early evolution of the South Aegean arc, Geological Magazine 134, 55-66, 1997.




Section 10 Remote sensing

Identification of As-bearing minerals associated with mine wastes from former metalliferous mines in France using laboratory reflectance spectra

V. Carrère Laboratoire de Planétologie et Géodynamique, UMR CNRS 6112, Université de Nantes, France

Abstract

Arsenopyrite-rich wastes from former metalliferous mines in the Massif Central Region, France, contribute, through intensive leaching, to the formation of thin layers of As-Fe crusts on the tailing surfaces acting as a cement. When subject to rainfall, acid mine drainage develops, As is remobilized and trapped secondarily by oxyhydroxides or sulphates such as goethite, hematite, jarosite, etc., depending on water pH. The aim of this study is to assess the potential of hyperspectral remote sensing in this particular environment (small outcrops, important leaching process, vegetated environment). Samples were collected in the field from tailings and stream bottoms in various sites. Reflectance spectra of the samples were acquired in the laboratory using a field spectrometer under artificial illumination (0.4–2.5 µm spectral range). Mineral composition was estimated by comparing reflectance spectra to reference spectral libraries. This was done after continuum removal, using two different techniques, in order to minimize the influence of factors such as grain size or moisture content. Geochemical and SEM-EDS analyses were performed to gather information on the mineralogy. Preliminary results from spectral analyses show clear evidence of jarosite, goethite or limonite, schwertmannite and scheelite in various proportions, depending on crust types, confirmed by SEM-EDS analyses. Keywords: surface spectral reflectance, mine wastes, continuum removal, iron oxyhydroxides.




doi:10.2495/GEO060381

1 Introduction

Mining activities generate a multiform pollution due to sulphides oxidation present in the tailings. Their contact with atmospheric conditions can lead to Acid Mine Drainage (AMD) (Johnson and Thornton [1]; Webster et al. [2]), a remobilisation of toxic elements (As, S, Pb, …) which could be immobilized in the sediments, the soils, by a more or less stable trapping. Some of these elements can be temporarily fixed in different ways: precipitation as oxides, hydroxides or sulphates (Herbert [3]), and absorption onto different materials (Karathanasis and Thompson [4]). Due to its presence as a major element accompanying many types of ore deposits, As is present in numerous mining areas. Because of its toxicity, its solubility has been extensively studied (e.g. Azcue et al. [5]; Hunt and Howard [6]; Roussel et al. [7]). Richardson and Vaughan [8] studied the early stages of arsenopyrite oxidation using spectroscopic investigations. They showed that in conditions of low pH, such as those found in tailings dumps, arsenopyrite oxidizes rapidly and forms highly soluble As compounds as well as Fe(II) or Fe(III) arsenites or arsenates. According to local chemical conditions, As, Fe and S may be exported in drainage waters (Nesbitt et al. [9]; Savage et al. [10]). These elements could precipitate as authigenic minerals (Scott [11]; Jambor [12]) or amorphous material or can be adsorbed onto Fe oxides, hydroxides, or suspended material (Pierce and Moore [13]; Tsung-Hui et al. [14]; Manning et al. [15]).

In the French Massif Central, mining activities were greatly developed in the twentieth century. Thus, before the 1980’s abundant mining tailings were generated and laid in dumps without any remediation. The present study focuses on one site, the site of Enguialès, located in the Aveyron Department, France (Figure 1), which was slightly remediated. Courtin-Nomade et al. [16] provided a detailed description of mineralogical and geochemical conditions of the tailings which is used as a reference for analyzing the spectral information. As-rich Fe cements are well developed as weathering crusts within the dumps. The size of the site (roughly 1 km2) is well suited for estimating the efficiency of hyperspectral reflectance data, a non-destructive approach, to identify and map the various minerals that could potentially be used as markers for As contamination. Previous studies on other sites clearly showed (see for example Swayze et al. [17]) that hyperspectral reflectance data allow to identify mineral phases onto which As can be adsorbed (iron oxyhydroxides, clays) based on diagnostic absorption features. Moreover, mineral assemblages can be used as indicator of acidity or chemical conditions that can control As mobility in soils and rivers.

We present preliminary results on the potential of hyperspectral surface reflectance measurements to identify mineralogical phases susceptible to have high As concentration and/or mineral associations indicative of soil acidity in such a specific context. More classical approaches such as mineralogical, chemical and SEM-EDS analyses were also used to validate these identifications.




1.1 Site description

This tungstiferous deposit is located in the “La Châtaigneraie” district, in the south of the French Massif Central (Figure 1). Mineralized veins are mainly composed of wolframite, scheelite, pyrite and arsenopyrite. The mining activities ceased in 1979. The tailings represent 300 000 t directly spread out in a steep slope (~35°) on about 22 000 m2. The tailings are disposed as a pile ravined by AMD and meteoric waters. The remediation that was done on this site consisted of closing galleries access and gathering all the waters in only one dewatering point, which induced AMD phenomenon. Currently only the meteoric waters circulate across the tailings. Nevertheless, the Eh-pH conditions are oxidant and always very acidic with pH = 2.8 +/- 0.3. The oxidation of the tailings is expressed by local iron oxyhydroxides crusts, especially well developed in the gullies. Grains of ore gangue (quartz, muscovite and tourmaline) mainly constitute these iron crusts. These primary minerals are bound by a Fe-As rich cement. The richest As product corresponds to an amorphous iron “arsenate” which is mainly observed in the ochre cement rather than in the red cement. In the current Eh-pH conditions in Enguialès, the estimated solubilities show a high potentiality to release As from the identified amorphous arsenate. On the contrary, the most crystallized products as goethite or jarosite, in which As is also trapped (Courtin-Nomade et al. [18]), appear as stable minerals in the long term (Stahl et al. [19]).

Figure 1: The Enguialès mine test site: General view and site location (after

Courtin-Nomade et al. [16]).




1.2 Sample collection

The various surface types that can be observed at the tailings surface were sampled in the field in December 2002. Only the surface was scraped in order to be representative of what can be observed by field or airborne hyperspectral instruments. Samples collection was based on surficial visual characteristics (color and aspect). When compared to descriptions published in the literature for this site (Courtin-Nomade et al. [16]; Courtin-Nomade et al. [18]), they could be attributed to 5 main categories: (1) micaschists (in place bedrock); (2) grey sand; (3) ochre cement; (4) red cement and (5) dark-reddish cement.

2 Spectral analyses

Spectral absorption features observed in the visible to short-wave infrared wavelength region result from several distinct processes. In the spectral range from 0.4 to 1.2 µm, absorption features are produced mainly by energy level changes in the valence of transition metals (crystal field transitions), by paired excitations of metal cations, or by charge transfer between metal cations and their associated ligands (Hunt and Ashley [20]; Sherman and Waite [21]; Burns [22]; Clark [23]). Absorption features in the SWIR region (1.3–2.5 µm) are generated by molecular vibration processes: for example, the 1.4 and 1.9 µm absorptions are related to the first overtones of the water O-H stretching fundamentals and the combination water O-H stretching with H-O-H bending vibrations, respectively (Hunt et al. [24]).

Laboratory spectral measurements in the VIS-SWIR spectral range of samples of the five surface types were recorded using a Spectralon TM white reference standard and an ASD FieldSpec FR spectrometer under artificial illumination. Figure 2 shows the resulting reflectance spectra. The five spectral signatures are rather different and show most of their distinctive absorption features in the visible-near infrared (VNIR). The drop-off toward the U.V. corresponds to charge transfer between ferric cations and adjacent oxygen anions. The strong absorption edge between 0.4 and 0.6 µm is related to paired excitations between magnetically coupled ferric cations (Sherman & Waite [21], Crowley et al. [25]), while the broad absorption bands around 0.9 µm can be attributed to crystal field absorption bands in ferric (0.75–0.95 µm) or ferrous (0.9–1.1 µm) iron bearing minerals. Features observed around 2.20 µm are attributable to the Al-O-H combination bands while those observed between 2.25 and 2.38 µm are due to Fe-O-H and Mg-O-H combination bands (Crowley et al. [25], Clark [23]). Based on these observations, it is clear that the different surfaces can be discriminated from one another. The next step consists of identifying the various minerals that contribute to their composition using the wavelength position of the absorption features. Comparison with spectra of pure reference minerals was performed on “raw” spectra and after continuum removal.




Figure 2: Reflectance spectra of typical surface types.

2.1 Continuum removal

The overall shape of a reflectance spectrum (or continuum) is related to physical properties of the surface (e.g. grain size, roughness, moisture content, local slope, etc.) which may change from place to place despite a homogeneous composition. In order to better identify a spectral feature by its wavelength position, it must be isolated from these other effects. The first step is continuum definition and removal. Following the definition of Clark and Roush [26], the continuum needs tie points at local maximum reflectance but there might be no stable tie points toward the blue or the infrared for natural surfaces. Therefore, the continuum can be defined in different ways: (1) using straight lines in wavelength, linking local maxima between absorption features, including starting and ending wavelength; this can be called the “straight line” approach (Figure 3A); (2) using a gaussian decomposition for the absorption features and a straight line in wavenumber for the continuum to overcome the instabilities at both end of the spectrum; this is done by the Modified Gaussian Model (MGM) developed by Sunshine et al. [27] (Figure 3B). The spectral deconvolution by MGM uses a non-linear least-squares analysis. The measured spectrum reflectance at a given wavelength is fitted by a superposition of n Gaussian distributions (each is defined by three parameters: central wavelength, half-width and negative strength) and a continuum curve (slope and intercept for a straight line in wavenumber). The inversion process must start with an initial set of parameters that includes absorption band centers and full width at half maximum. These parameters are chosen by observing the spectral shapes and following a priori knowledge about spectral characteristics (Figure 3B).




Once the continuum line is established, the continuum-removed spectra are calculated by dividing the original reflectance values by the corresponding values of the continuum line. As the tie points can be different for the various surfaces depending in their mineralogical composition, the spectra were treated globally.

Figure 3: A - Continuum removal using the “straight line” approach; solid line:

original spectrum (bottom) and after continuum removal (top), dashed line: continuum. B - Principle of the spectral modelling using the MGM; solid line: original spectrum (bottom) and after continuum removal (top), dashed line: continuum, dotted line: gaussian functions used to model the original spectrum.




2.2 Mineral discrimination based on spectral features

A reference spectral database, published by Crowley et al. [25] containing the main minerals that can be encountered in sulphide-rich mine waste piles was used. Only minerals supposed to be stable in the environmental conditions of our test site (exposed settings, dry conditions), were selected for comparison. Primary ore minerals (muscovite, scheelite or dravite) were also included but not quartz or feldspar since these minerals, although present, do not show any absorption features in the wavelength range considered here. A list of reference minerals and formula is presented in Table 1. The 1.4 and 1.9 µm absorption features attributable to H2O were excluded of the identification procedure since they are present in most minerals and therefore not discriminant.

Table 1: Reference database mineral names and formula.

Mineral Formula Dravite (Tourmaline) NaMg3(Al,Fe)6Si6O18(BO3)3(OH)4 Ferrihydrite 5Fe2

3+O3*9H2O Fibroferrite Fe3+(SO4)(OH)*5H2O Goethite α-Fe3+O(OH) Hematite α-Fe2

3+O3 Jarosite KFe3

3+(SO4)2(OH)6 Limonite Mixture of hydrated iron oxides Muscovite KAl2(AlSi3)O10(OH)2 Paracoquimbite Fe2

3+(SO4)3*9H2O Rozenite Fe2+SO4*4H2O Scheelite CaWO4 Schwertmannite Fe16

3+O16(OH)12(SO4)2 Szomolnokite Fe2+SO4*H2O

2.2.1 “Raw” spectra Precise location of absorption features is more difficult to define on raw reflectance spectra, particularly in the case of broad absorptions. In a first approximation, jarosite was identified in grey sand, micaschists and dark-reddish cement. Goethite is found in ochre and red cement, schwertmannite in micaschists and possibly in dark-reddish cement. Scheelite is present in grey sand, micaschists and ochre cement and muscovite in dark-reddish cement.

2.2.2 “Straight line” continuum removal Removing the continuum helps locating more precisely the absorption positions. Results show that it is more likely that hematite rather than goethite is present, or more probably a mixture of hydrated iron oxides such as limonite, based on the position of the broad absorption feature around 0.9 µm. The presence of jarosite in micaschists and grey cement, and of schwertmannite in dark-reddish cement is confirmed. Dravite is clearly present in micaschists and possibly in grey sand and ochre cement. Exact match with scheelite and muscovite is difficult probably




because of mixing effects, and of the close positions of their absorption features around 2.20 µm.

2.2.3 Using MGM gaussian function positions Deconvolution of spectra using the MGM produces more gaussian functions than “realistic” absorption features. Therefore, it is harder to decipher the results. This technique appears also highly sensitive to input parameters which require some a priori knowledge. Several runs were necessary to adjust the parameters and reach a reasonable fit. Although the wavelength positions of the gaussian function did not exactly match the absorption positions determined in the previous steps, the presence of scheelite in grey sand and of hematite rather than goethite was confirmed. This approach requires further improvement in order to be operational on spectra of natural surfaces. A thorough sensitivity analysis of the deconvolution to input parameters is also required.

3 SEM-EDS and XRD data

XRD, performed on a limited number of samples due to time constraints, confirmed the presence of jarosite, dravite and muscovite in the ochre cement. Indications of quartz and chlorite or serpentine were also found but not for clays. SEM-EDS analyses were performed on minerals appearing to be statistically representative of the samples. These analyses confirmed the presence of primary ore minerals such as muscovite, quartz, feldspar and garnet as well as pyrite, ferberite or wolframite [(Fe, Mn)WO4]. Hydroxides, hydrated arsenates and phosphates, also present, were harder to identify since (OH) or (H2O) quantities cannot be estimated accurately with this technique.

4 Conclusion

These preliminary results show that identifying specific minerals present in cements developed at the surface of tailings can be complicated. In principle, many ferric oxide and sulphate minerals should be distinguishable by using field spectral measurements, as well as existing hyperspectral sensors. In practice however the ability to discern individual iron sulphate or oxide species is limited by the occurrence of the minerals in complex mixtures with each other and with other surficial materials. Mineral mixing, grain size and mineral composition are important factors influencing the spectral signature, making direct comparison between reference and measured reflectance spectra difficult. Selecting the proper endmembers to perform spectral unmixing is rather complicated because of the different sources of changes previously mentioned. This study showed that removing the continuum improved identification as it allows one to focus on absorption wavelength positions. It might be necessary to work on specific absorptions individually to improve the results. Using the MGM technique should give access to more quantitative information such as absorption depth and area that can be related to concentration or fractional composition but this approach needs to be more thoroughly evaluated when applied to natural




surfaces. Its sensitivity to input parameters needs to be characterized. Finally, detailed analysis of the chemical and mineralogical composition of the samples is necessary to better understand the optical properties of such surfaces.

References

[1] Johnson, C.A. & Thornton, I., Hydrological and chemical factors controlling the concentrations of Fe, Cu, Zn and As in a river system contaminated by acid mine drainage. Water Res., 21, pp. 359-365, 1987.

[2] Webster, J.G., Nordstrom, D.K. & Smith, K.S., Transport and natural attenuation of Cu, Zn, As, and Fe in the acid mine drainage of Leviathan and Bryant Creeks, Environmental Geochemistry of Sulfide Oxidation, eds. Alpers, C.N., Blowes, D.W., Am. Chem. Soc. Symp. Series, 550, pp. 244-260, 1994.

[3] Herbert, R.B., Properties of goethite and jarosite precipitated from acidic groundwater, Dalarna, Sweden. Clays and Clay Minerals, 45, pp. 261-273, 1997.

[4] Karathanasis, A.D. & Thompson, Y.L., Mineralogy of iron precipitates in a constructed acid mine drainage wetland. Soil Sci. Soc. Am. J., 59, pp. 1773-1781, 1995.

[5] Azcue, J., Muroch, A., Rosa, F. & Hall, G., Effects of abandoned gold mine tailings on the arsenic concentrations in water and sediments of Jack of Clubs Lake, B.C. Environ. Technol., 15, pp. 669-678, 1994.

[6] Hunt, L.E. & Howard, A.G., Arsenic speciation and distribution in the Carnon Estuary following the acute discharge of contaminated water from a disused mine. Marine Poll. Bull., 28, pp. 33-38, 1994.

[7] Roussel, C., Bril, H. & Fernandez A., Arsenic speciation: Involvement in evalution of environmental impact caused by mine wastes. J. Environ. Qual., 29, pp. 182-188, 2000.

[8] Richardson, S. & Vaughan D.J., Arsenopyrite: a spectroscopic investigation of altered surfaces. Mineral. Mag. 53, pp. 223-229, 1989.

[9] Nesbitt, H.W., Muir, I.J. & Pratt, A.R., Oxidation of arsenopyrite by air and air-saturated, distilled water, and implications for mechanism of oxidation. Geochim. Cosmochim. Acta, 59, pp. 1773-1786, 1995.

[10] Savage, K.S., Tingle, T.N., O’Day, P., Waychunas, G.A. & Bird, D.K., Arsenic speciation in pyrite and secondary weathering phases, Mother Lode Gold District, Tuolumne County, California. Appl. Geochem., 15, pp. 1219-1244, 2000.

[11] Scott, K.M., Solid solution in, and classification of, gossan-derived members of the alunite-jarosite family, northwest Queensland, Australia. Am. Mineral., 72, pp. 178-187, 1987.

[12] Jambor, J.L., Nomenclature of the alunite supergroup. Can. Mineral, 37, pp. 1323-1341, 1999.

[13] Pierce, M. & Moore, C., Adsorption of arsenite and arsenate on amorphous iron hydroxide. Water Res., 16, pp. 1247-1253, 1982.




[14] Tsung-Hui, H., Shang-Lien, L., Cheng-Fang, L. & Dar-Yuan, L., Characterization of arsenate adsorption on hydrous iron oxide using chemical and physical methods. Colloids and Surfaces A: Physiochem. Eng. Aspects, 85, pp. 1-7, 1994.

[15] Manning, B.A., Fendorf, S.E. & Goldberg, S., Surface structures and stability of arsenic (III) on goethite: spectroscopic evidence for inner-sphere complexes. Environ. Sci. Technol., 32, pp. 2383-2388, 1998.

[16] Courtin-Nomade, A., Bril, H., Néel, C. & Lenain, J.F, Arsenic in iron cements developed within tailings of a former metalliferous mine – Enguialès, Aveyron, France. Applied Geochem., 18, pp. 395-408, 2003.

[17] Swayze, G.A., Smith, K.S., Clark, R.N., Sutley, S.J., Pearson, R.M., Vance, J.S., Hageman, P.L., Briggs, P.H., Meier, A.L., Singleton, M.J. & Roth, S., Using imaging spectroscopy to map acidic mine waste. Environmental Science and Technology, 34, pp. 47-54, 2000.

[18] Courtin-Nomade, A., Néel, C., Bril H. & Davranche, M., Trapping and mobilisation of arsenic and lead in former mine tailings – Environmental conditions effects. Bull. Soc. Geol. France, 173, pp. 479-485, 2002.

[19] Stahl, R.S., Fanning, D.S. & James, B.R., Goethite and jarosite precipitation from ferrous sulfate solutions. Soil Sci. Soc. Am., 57, pp. 280-282, 1993.

[20] Hunt, G.R. & Ashley R.P., Spectra of altered rocks in the visible and near infrared. Economic Geol., 74, pp. 1613-1629, 1979.

[21] Sherman, D.M. & Waite, T.D., Electronic spectra of Fe3+ oxides and oxide hydroxides in the near IR to near UV, American Mineralogist, 70, pp. 1262-1269, 1985.

[22] Burns, R.G., Mineralogical application of crystal field theory. Cambridge University Press, Cambridge, 1993.

[23] Clark, R.N., Spectroscopy of rocks and minerals and principles of spectroscopy (Chapter 1). Manual of Remote Sensing, ed. A.N. Rencz, John Wiley and Sons, New York, pp. 3-58, 1999.

[24] Hunt, G.R., Salisbury, J.W. & Lenhoff, C.J., Visible and near-infrared spectra of minerals and rocks: IV. Sulphides and sulphates, Modern Geol., 3, pp. 1-14, 1971.

[25] Crowley, J.K., Williams, D.E., Hammarstrom, J.M., Piatak, N., I-Ming Chou & Mars, J.C., Spectral reflectance properties (0.4 - 2.5 µm) of secondary Fe-oxide, Fe-hydroxide, and Fe-sulphate-hydrate minerals associated with sulphide-bearing mine wastes, Geochemistry: Exploration, Environ., Analysis, 3, pp. 219-228, 2003.

[26] Clark, R.N. & Roush, T., Reflectance spectroscopy: quantitative analysis techniques for remote sensing applications. J. Geoph. Res., 89, pp. 6329-6340, 1984.

[27] Sunshine J., Pieters, C.M. & Pratt, S.F., Deconvolution of mineral absorption bands: an improved approach, J. Geoph. Res., 93, pp. 6955-6966, 1990.




Rapid mapping with remote sensing data during flooding 2005 in Switzerland by object-based methods: a case study

Y. A. Buehler, T. W. Kellenberger, D. Small & K. I. Itten Remote Sensing Laboratories (RSL)/NPOC, Department of Geography, University of Zurich UZH, Switzerland

Abstract

Rapid mapping and monitoring with remote sensing techniques is an important source of information for decision-makers faced with large scaled disasters. In August 2005 several regions in central Switzerland were affected by severe flooding. The “National Emergency Operations Centre” (NEOC) of Switzerland invoked the International Charter on “Space and Major Disasters”, requesting support by remote-sensing data for disaster-management. In this paper, a SPOT-5 and a RADARSAT-1 satellite-scene acquired a few days after the flood-peak are described and processed to map the affected area with object-based methods. The main difficulties with optical images are the small extent of the Swiss landscape, the large height-differences in the relief, shadows, clouds and snow covered areas. Comparing different approaches, it was found that beside the NIR and SWIR band the application of a DEM and a VISNIR water index showed the most promising results for a fast discrimination of the affected areas. The object-based classification is mainly dependent on threshold-values and Boolean operators. As investigations show, the spatial resolution of the acquired radar dataset in this case is not sufficient to map the affected areas. The results of the classification based on SPOT-5 scenes are up-to-date maps of flooded areas. We show that maps for decision-makers can be produced using auxiliary topographical and land use data. Aerial infrared images acquired from SWISSTOPO after the flooding were used to test the accuracy of the result. Procedures for rapid mapping applications in future disaster cases are discussed. Keywords: rapid mapping, flooding, disaster response, international charter on space and major disasters, natural hazards, object-based classification.




doi:10.2495/GEO060391

1 Introduction

Disaster mapping with remote sensing data has been investigated for several years. In the past, the poor availability of appropriate sensors, low temporal resolution, time-consuming data acquisition and distribution, limited image analysis techniques and computer resources – and in addition, the lack of missing awareness about the existence of products limited widespread application Iglseder et al. [1]. Today, major advances reduce these obstructions, enhancing possibilities for a useful disaster-management tool. In general, two types of rapid mapping products are required by end-users: overview-maps of affected areas and damage-maps combined with additional information (e.g. land-use type of flooded area, destroyed traffic routes, changes in flood-levels) Allenbach et al. [2]. This paper focuses on the detection of flooded areas and a raw differentiation of water depth. The creation of overview-information is not required due to the area-wide availability of 1:25’000 topographical maps. Organisations active in the domain of rapid mapping associated with the International Charter on “Space and Major Disasters” in Europe today are for example UNOSAT in Geneva Switzerland, SERTIT in Strasbourg France and the ZKI at the DLR in Oberpfaffenhofen Germany. The majority of past charter calls concerned areas of a large extent in regions of mainly even topography. Examples are the Elbe-River flooding in autumn 2002 and the destructive Indian Ocean Tsunami in December 2004. In contrast to these disasters types, the flooded areas investigated in this paper are of small spatial extent in the rough topography of the Swiss Alps. An overview of mapping natural hazards in alpine region is provided in Metternich et al. [3]. In August 2005 a 5b-cyclone carried large quantities of humid air from the Adriatic Sea to the central Alps. The rainfall that ensued reached intensities above 300 l/m2 over two days, which had never been measured in situ before. Due to a snow line over 3000 meters above sea level and wet conditions in the days before the disaster, the soils were already saturated and could not absorb the additional intense rain Frei [4]. The levels of rivers and lakes rose and flooded farmland, roads and urban areas. The steep terrain of the central Alps caused landslides and mudflows so that many people had to be evacuated. An overview of the main affected regions in Switzerland and the study area is shown in Figure 1. Especially the large extent of affected areas is exceptional for a flood event in Switzerland. Three days after the rainfalls the “National Emergency Operations Centre” (NEOC) invoked the International Charter on “Space and Major Disasters” requesting support by remote-sensing data for disaster-management. The “National Point of Contact” (NPOC) of the “Federal Office of Topography” (SWISSTOPO) as project manager produced rapid mapping products and handed them out to support the decision-makers. Due to the federal organization of the Swiss Civil Protection, several different organisations e.g. cantonal crisis committees, federal offices, the police, fire departments and the army are all possible end-users of rapid mapping products Federal Office for Civil Protection [5]. This paper focuses on the optimisation of rapid mapping products in this particular case.




To acquire useful satellite-data, a very short timeframe is available in which the affected area is only viewed by a small number of sensors. Additionally, the Swiss Alpine area presents specific challenges to the use of satellite data. The small extent of the landscape requires a high spatial resolution of better than 15 meters, which is only available from a few optical sensors (e.g. SPOT, IKONOS). The large height-differences in the relief cause shadows in optical images and radar-shadow, layover and foreshortening in radar images. In optical images, clouds, cloud-shadow and snow complicate the classification of flooded areas. Due to these difficulties, there are presently only a few datasets available for a rapid classification of the flooded areas in Switzerland.

Figure 1: Overview of the main affected areas (ovals) in Switzerland and the

study area (dashed rectangle).

2 Remote sensing data

2.1 RADARSAT-1

The “Canada Centre for Remote Sensing” (CCRS) submitted in response to the Charter invocation a RADARSAT-1 standard mode C-band dataset with a spatial resolution of 28 meters. A small subimage covering the Lake Brienz region is shown in Figure 2. The advantage of radar data compared to optical sensors is its capability to acquire data during the night and through clouds, which is of great advantage for flood monitoring. Investigations with a pre-flood and a post-flood RADARSAT-1 image showed that the spatial resolution is not sufficient to detect small flooded areas of several square meters in the Swiss landscape. The fine mode with a spatial resolution of 9 meters was not acquired but would have been




a preferable alternative. Further problems such as radar shadow, layover and foreshortening, increased by the rough topography complicate the classification. By the time the Charter was activated and the RADARSAT1 image planed and acquired, much of the flooding had receded, leaving only residual ponds. Additionally, the archive of images acquired with the same geometry for a particular region is small and it is difficult to find pre-disaster scenes for change detection. Further information about geometric and radiometric correction of radar data in mountainous terrain is provided in Small et al. [6]. Due to difficulties mentioned above, the RADARSAT-1 data was not used for the classification of affected areas in this investigation.

Figure 2: Radarsat-1 geocoded terrain corrected standard mode image in full resolution, acquired on August 28th 2005. The oval highlights a known flooded area.

2.2 SPOT5

SPOTIMAGE delivered two SPOT5-scenes, acquired on August 26th and 30th 2005. As base image for change detection, an image from the data archive acquired on June 12th 2003 was applied. SPOT5 carries two sensors HRG providing two panchromatic bands (spatial resolution 5 m), two visible bands, a near infrared band (spatial resolution 10 m) and a short wave infrared band (spatial resolution 20 m). The image swath is 60 km x 60 to 80 km depending on the acquisition angle Spotimage [7]. The utility of optical images is highly dependent on characteristics like the revisit time, the repetition rate, the acquisition angle and the scene’s cloud-cover Campbell [8]. After the heavy rainfall, clouds obscured more than 80% of the scene acquired on August 26th 2005. By chance, the main parts of the affected valleys were fortunately cloud-free. The scene from August 30th 2005 has nearly no cloud-cover but was recorded more than a week after the peak of the flood-event. Figure 3 shows quick-looks of the two post-flood images.




Figure 3: Quicklooks of the Spot5 HRG images acquired after the flooding.

3 Methodology

3.1 Data preprocessing

Usually the first step in the preprocessing chain for rapid mapping is the geometric calibration. It is the foundation for change detection methods and visualisation by overlaying the images with auxiliary geographical data. The required accuracy lies below the extent of one pixel (SPOT5 10 m, RADARSAT1 28 m). The rough terrain of the Swiss Alps makes accurate geometric calibration both more important and more difficult. Use of a digital elevation model (DEM) improves the accuracy remarkably Meier [9], Kellenberger [10]. Correction of atmospheric and illumination effects is not always possible within the fast preprocessing required in rapid mapping. Due to the inhomogeneity within a single scene, the lack of atmospheric parameters and time consuming processing, these corrections were not conducted in this investigation. To normalize the datasets, the pixel values are transformed from digital numbers to physical units (at sensor radiance) applying the calibration coefficients in equation (1) Schowengerdt [11]. Ls = C0 + C1 * DN (1) Ls = at sensor radiance C0 = calibration offset or bias C1 = calibration gain DN = digital numbers

3.2 Segmentation

Traditionally, classification of remote sensing data has been conducted using a pixel-based approach, where the spectral response of one pixel and possibly its very close surroundings is used for classification. Object oriented approaches




first merge adjacent pixels of high homogeneity in spectral reflectance or shape to image objects, dependent on a chosen scale factor. Classification on the basis of image-polygons allows use of the spectral statistics in addition to the integration of intrinsic, topological and contextual information. Furthermore, the approach improves the possibilities to include knowledge from sources other than remote sensing data Benz et al. [12]. A disadvantage in the field of rapid mapping is the large amount of required processing power and memory, slowing down the segmentation and classification process. The aim of segmentation is to create image objects that accurately represent the shapes and sizes of the water-covered areas in the SPOT5 pre-disaster and post-disaster scenes. Investigations showed that best results are achieved based on a VISNIR water index (2), the near infrared band NIR and the Swiss digital elevation model DHM25 on a low scale. According to the normalized difference vegetation index NDVI Campbell [8], the VISNIR water index is based on the band containing high spectral response (green) and the band that contains low spectral response (NIR) for water. Figure 4 shows part of the calculated water index band from the pre-disaster scene superimposed with the computed segments.

VISNIR water index = green − NIR

green + NIR (2)

Figure 4: Computed water index of the pre-disaster scene superimposed with the generated image-objects.

3.3 Classification

Based on the computed segments, the first classification is performed on the pre-disaster scene. The aim is to generate a precise representation of the areas




covered by water before the flood. The classification is based mainly on the VISNIR water index (2), the NIR band and the standard deviation of the height of the pixels merged in one polygon, which is an estimate of the flatness of a segment. The second classification, performed in the post-disaster scene, highlights the areas that are covered by water now but were not in the pre-disaster scene. The detection of flooded areas is based on the water index (1), the SWIR band, the NIR band, the standard deviation of the height of the pixels merged in one polygon and Boolean combinations. In addition to the flooded areas, clouds, shadow and snow are separated to avoid intersections with the flood classes. To save time, the classes are defined by thresholds and not by samples. An advantage of this method is that later images from the same sensor can be classified with the same or slightly modified class definitions. The dataset from August 30th 2005 (Figure 3) was acquired under different atmospheric conditions and has a larger offnadir view angle, leading to differences in the measured spectral response. A classification with the same class definitions applied in the first post-disaster scene shows accurate results for flooded areas. This approach allows an uncomplicated monitoring of the affected areas, and production of dynamic flood maps. The same approach could be applied during future flood events, helping to speed up the production of rapid mapping products. More research has to be done to check the potential for adapting the established class definitions to further scenes from the same or different sensors. To control the achieved accuracy of the object-based classification, infrared aerial images flown on August 29th 2005 by SWISSTOPO were used. The scanned aerial infrared images have an average scale of 1:10’000 and have been resampled to a pixel size of 0.25m to accomplish the accuracy assessment. Flooded areas of different water levels are perfectly identifiable on these images, which makes them the best groundtruth available to assess the accuracy of the classification from the SPOT5 scene. Table 1 shows the results of this evaluation.

Table 1: Accuracy assessment for the two flood classes.

Flooded area low water level Flooded area high water level Producer’s accuracy 0.907 0.929

User’s accuracy 0.985 0.995 Kappa Index per class 0.864 0.816

Overall accuracy 0.921 Kappa overall 0.839

Especially the high Kappa coefficient value of 0.839 verifies the good results of the classification that a visual evaluation promises. Minor misclassifications are mainly caused by the time difference between the acquire dates of the satellite scene and the aerial images. Additionally, flooded areas of a very small extent or with a very high fraction of debris cannot be detected in the Spot5 image.




4 Results

Rapid mapping products aim to be quickly available and easy to read for end users. The polygons classified as flooded are combined with topographical maps to generate a layout that the decision makers are familiar with, and allow them to add further information. This procedure enables extraction of information about the type of affected area (e.g. affected traffic routes or flooded settlements). Figure 5 shows an example of a topographical map with flooded areas superimposed, as detected within the SPOT5 post-disaster image. To produce overview maps on a larger scale, the flood-polygons could be plotted over the original satellite data or over satellite data combined with topographical maps. These products can be valuable tools when planning rescue efforts shortly after the disaster.

Figure 5: Rapid mapping product to support the disaster management, topographical map ©SWISSTOPO (BD053189).

5 Conclusions

The results from this investigation show that object-based classification of remotely sensed data can produce valuable base data for rapid mapping products, supporting decision makers in time of crisis. An estimation of the time required for the production steps after the arrival of the remotely sensed raw data is given in Table 2. The extracted objects representing the flooded areas can be exported to a GIS and intersected with land-use or population density maps to estimate




economic damage and other valuable information. These possibilities can provide an important support for disaster management.

Table 2: (for a single scene).

Basic work Geometric calibration: ca. 1 h Segmentation ca. 1 h Adaptation of the classification for

the pre-disaster scene ca. 1 h

Adaptation of the classification for the post-disaster scene:

ca. 1 h

Validation of the classification result:

ca. 1 h

Map production Map generation: ca. 1/2 h per map Map layout: ca. 1/2 h per map Total ca. 6 h

The temporal availability of high spatial resolution satellite data has improved within the last decade and this trend is still going on. The “International Charter on Space and Major Disasters” provides the infrastructure to acquire the appropriate data a short time after a disaster and enables the support with rapid mapping products. To improve the international success of this institution it is important that the regions, which are affected by a disaster, invoke the charter immediately after the event and that more satellite operators join the charter to raise the probability of image availability. To enable the detection of small flooded areas in rough terrain, the spatial resolution ought to be better than 15 meters. The acquisition date should be as close as possible to the disaster event so that most of the affected areas are still covered by water. Common cloud cover in the time after flood-events triggered by heavy rainfall is the main problem in rapid-detecting affected areas in optical satellite data. The spectral appearance of urban areas under cloud shadow is very similar to the appearance of water in SPOT5 data and reduces the accuracy of the classification. In a further step of research a fast automatic extraction of clouds and shadow should be developed to improve the speed and quality of the classification. Application of radar data can bypass the problems of clouds and cloud shadow, but is limited by layover, foreshortening and radar-shadow in mountainous areas. Although these limitations exist, there is a need for research on object-based classification in rapid mapping with combined data from radar and optical sensors.

The authors like to thank swisstopo for data-support and the geometric calibration, SPOTIMAGE and the Canadian Space Agency for the satellite data.




Estimation of the time needed for the main steps of rapid mapping

Acknowledgements

References

[1] Iglseder, H., Arensfischer, W., Wolfensberger, W., Small Satellite Constellations for Disaster Detection and Monitoring, Natural Hazards, volume 15, issue 11, pp. 79-85, 1995

[2] Allenbach, B. et al., Rapid EO Disaster Mapping Service: Added value, feedback and perspectives after 4 years of Charter actions, SERTIT, 2005

[3] Metternich, G., Hurni, L. & Gogu, R., Remote Sensing of landslides: An analysis of the potential contribution to geo-spatial systems for hazard assessment in mountainous environments, Remote Sensing of Environment, volume 98, issues 2-3, pp. 284-303, 15 October 2005

[4] Frei, C., August-Hochwasser 2005: Analyse der Niederschlagsverteilung, Bundesamt für Meteorologie und Klimatologie (MeteoSchweiz), 2005

[5] Federal Office for Civil Protection (Bundesamt für Bevölkerungsschutz,) http://www.bevoelkerungsschutz.admin.ch (access December 12th 2005)

[6] Small, D. Meier, E. Nüesch, D., Robust Radiometric Terrain Correction for SAR Image Comparison, Geoscience and Remote Sensing Symposium IGARSS '04. Proceedings, IEEE International, volume 3, pp. 1730 - 1733, 2004

[7] Spotimage, http://www.spotimage.fr (access December 12th 2005) [8] Campbell, J. B., Introduction to Remote Sensing third edition, Taylor &

Francis London, pp. 272-285, 465-467, 2002 [9] Meier, E., Geometrische Korrektur von Bildern orbitgestützter SAR-

Systeme, Remote Sensing Series volume 15, 1989 [10] Kellenberger, T. W., Erfassung der Waldflächen in der Schweiz mit

multispektralen Satellitenbilddaten, PhD Thesis, RSL Department of Geography, UZH University of Zurich, pp. 142-164, 1996

[11] Schowengerdt, R. A., Remote Sensing – models and methods for image processing, Academic Press, San Diego, pp. 311-313, 1997

[12] Benz, U. et al., Multi-resolution, object-oriented fuzzy analysis of remote sensing data for GIS-ready information, ISPRS Journal of Photogrammetry & Remote Sensing volume 58, pp. 239 – 258, 2004




Using spatial technology for analyzing disturbed areas and potential site selection in Chihuahua, Mexico

V. M. Tena1, A. C. Pinedo2, A. H. Rubio1, P. de L. G. Barragán3, A. A. Pinedo3, M. V. Hernandez3 & C. Velez2 1Investigadores del INIFAP, Chihuahua, Chihuahua, México 2Profesores-Investigadores, UACh, Chihuahua, México 3Estudiantes de Doctorado, Chihuahua, Chihuahua, México

Abstract

The objectives were to: 1) analyze the physical-chemical variables of a site in order to characterize the disturbed area’s suitability for planting; 2) detect disturbed areas through spectral transformation techniques; and 3) generate cartographic maps with the location of the disturbed areas suitable for planting. The field data, Landsat TM 7 images, and digital elevation models were analyzed with IDRISI and ArcView. Multivariate Cluster Analysis, and Principal Components Analyses were applied to valuate the biometric and physical-chemical variables of the soil, and detection of disturbed areas. Physical-chemical analyses showed similar characteristics for the whole area in the study; only soil depths are considered important for the establishment of forest plantations. The combination of bands 3, 4 and 5 allowed detection of, in a preliminary way, the disturbed areas. The Principal Component Analysis showed that the first component reduced the dimensionality of the data while the second component detected the disturbed areas. Keywords: reforestation, disturbed areas, detection, spatial, Chihuahua, México.

1 Introduction

Worldwide, deforestation is associated with increasing demands for forest products due to increased human populations [1]. Mexico has a deforestation rate estimated at 600,000 hectares per year [2]. This is one of the highest rates of




doi:10.2495/GEO060401

deforestation in the world with the presence of associated problems like erosion, desertification, less biodiversity, and lost productivity of the land. This problem hurts the economy of the forested regions of the State of Chihuahua, where ecological damage includes irregularities in weather, water, soil, flora and wildlife among others. In addition, protection measures and commercial tree planting activities showed a series of limitations and obstacles to success, because they do not result in clear and well defined objectives due to the lack of reliable and up to date information related for possible areas to be planted. To establish successful plantations that provide wood sources for the forest industry and to diminish pressure in the natural forests, environmental factors must be considered such as the biophysical and chemical features of the planting sites. Field data analysis combined with satellite imagery and digital elevation models (DEM) of high spatial resolution allow the generation of up to date and reliable information to make decision making easier [3,4]. This helps in making adequate planning and execution of recovery programs in disturbed areas, looking to reactivate their productivity as well as conserving and preserving the associated natural resources. The objectives of this study were to analyze the physical and chemical variables of a site to characterize the disturbed areas for suitability for planting, to detect the disturbed areas through spectral transformation techniques, and to generate cartographic maps with the location of the suitable areas to be planted.

2 Regional setting

The study was conducted in four areas: (1) San Juanito, (2) Bocoyna, and (3) San Ignacio de Arareco, all in the Municipality of Bocoyna, and (4) the commonly owned area called Cusarare in the Municipality of Guachochi, Chihuahua, located in the Sierra Tarahumara. Geographically, these areas are located in the UTM coordinates 228369 E, 3046635 N and 265236 E, 3105332 N.

2.1 Climate

The study areas are humid and temperate (C(E)W2) with a mean annual precipitation of about 600 mm. The mean annual temperature is 12.9 °C with a minimum of -9.9 °C and a maximum of 26.7 °C. Extreme temperatures varied from -18 °C in December of 2003 to 40 °C in June of 1992.

2.2 Geology

The area has outcrops of extrusive igneous rocks dominated by the acid types of rhyolite and tuff of the superior tertiary Ts (Igea). Rough topography of the study area includes a great diversity of topographic shapes such as plateaus, canyons, mountains, and hills. Most of the surveyed sites where on plateaus with elevation ranges from 2,212 meters to 2,572 meters.




2.3 Physiography

Rough topography of the study area includes a great diversity of topographic shapes such as plateaus, canyons, mountains, and hills. Most of the surveyed sites where on plateaus with elevation ranges from 2,212 meters to 2,572 meters.

2.4 Soils

The soils of the municipality of Bocoyna are the eutric regosol type and those in the Cusarare in the Municipality of Guachochi are haplic feozem type. In general, the soils have shallow depths fluctuating from 35 through 250 cm with textures that range from sandy clay loamy soils to sandy loams that are medium high in organic matter content and nitrogen, and deficient in phosphorous, with a highly acid pH and a hydraulic conductivity from moderate to moderately fast.

2.5 Vegetation

The study areas are characterized by a mixture of tree species that are dominated by Pinus arizonica, Pinus duranguensis, and Pinus engelmannii and associated with Quercus spp., Populus spp., Juniperus depeana, Pinus leyophila, Pinus chihuahuana, Pinus ayacahuite, and Picea chihuahuana.

3 Analysis

Information was analyzed with IDRISI and ARC/VIEW 8.x programs. There were four data sources: (1) Landsat TM7 January 2003 images, (2) four digital elevation models (DEM) in 1:50,000 scale, (3) edaphology, topography, and soil use thematic charts, and (4) field data collected for variables of interest. A radiometric analysis procedure was applied to the bands used from Landsat TM7 with geometric correction to fix possible noise effects, pollutants attenuation, relative humidity, and cloudiness levels [5]. The sub-scene of interest was adjusted according to the analysis of the study areas for better details in the analysis of the variable. Quadrangles of the DEM in 1:50,000 scale were segmented in sub-scenes according to the location of each site. Once the DEM was generated, two conversion algorithms were applied to generate the Slope Digital Model (SDM) and the Exposure Digital Model (EDM). Through the Boolean operator (AND), the slope was reclassified in the next range: 0–2% commercial plantations, 3–5% forest recovery, and >5% for protection of other existing resources. For a preliminary detection of disturbed areas, a monoband analysis of landsat TM was explored, and a band combination was selected according to the suggested procedures of Pinedo [6] to better discriminate the disturbed areas. Once the areas were detected, the SEEDTOOLS module from ERDAS was used, interpolated by the ARCVIEW program to generate in an automatic way the masks of the areas, the same that were overlain in the sub-scenes derived from landsat TM as well as from the DEM, SDM and EDM. Under a stratified random sample, data from the variables of interest were collected using three sampling techniques: (1) a 1 square meter plot for location




(GPS), elevation (altimeter), exposure (Silva compass), slope (clinometer), and soil variables, (2) 12.56 square meters (2 meters radius) for the arboreal vegetation regeneration data, and (3) 200 square meters plots (4 meters radius) where forest variables included plant species, total and commercial tree heights, normal diameters, product distributions, ages, annual increments, and year pass). Statistical analyses were conducted with a multivariate technique of Cluster Analysis of the MINITAB program. To know the relation of each of the landsat TM bands and reduce the possible dimensionality of the spectral data, a principal component analysis was applied beside the newly derived images which allowed the detection of disturbed areas suitable of being planted [7].

4 Results

4.1 Disturbed areas analysis

The correlation matrix of the principal component analysis (PCA) showed, with exception of band 4, a high degree of association with the rest of the spectral bands of Landsat TM (Table 1), in a way that the combination of any of the three bands can generate a cover type map with an easy detection of the disturbed areas. Although, band 4 was chosen as the least correlated and combined with an infrared and one visible band, it was used to generate a composition of false color using band 3 in the blue channel, band 4 in the green channel, and band 5 in the red channel. This process was useful in this study since it allowed the generation of actual land use cartography to discriminate in a preliminary way the disturbed areas, locate the sample sites, and then, quantify the polygons of areas suitable for planting. Besides attenuating the effects of the differences of variability between bands and diminishing the dimensionality of the data between them, the PCA allowed the generation of new images, making as many components as bands that were introduced in the analysis. The first component synthesized 94.37% of the total variability while the second component explained 3.55% of the total variability. The two components represented 98% of the total variability of the original variables, while the rest of the components synthesized 2.08% of the total variability. Even when the first component reduced the dimensionality of the data preserving almost all the information, the second component showed clearly the disturbed areas of detection.

4.2 Cluster analysis

4.2.1 Distribution and grouping of the sites Variables from the 40 sites were grouped among the four resulting clusters of the multivariate analysis. (Clusters analysis) based on the similarity in values of the studied variables, in a way that cluster 1 grouped 6 sites located in the lands of San Juanito and San Ignacio de Arareco, cluster 2 grouped 9 sites in San Juanito,




San Ignacio de Arareco and Cusarare, cluster 3 grouped 10 sites of the four study areas while cluster 4 grouped 15 sites from the four areas (Table 2).

Table 1: Matrix of correlation between bands of the sub scenes of landsat tm7 image for the study area.

Scene TM1 TM2 TM3 TM4 TM5 TM7 band_1 1.0000 band_2 0.9814 1.0000 band_3 0.9725 0.9877 1.0000 band_4 0.7520 0.8112 0.7833 1.0000 band_5 0.9162 0.9421 0.9467 0.8408 1.0000 band_7 0.9307 0.9468 0.9591 0.7661 0.9830 1.0000

Table 2: Created clusters based on the interest variables.

Clusters Sites 1 9,10,13,14,15,17 2 1,2,5,8,16,18,36,38,40 3 4,11,12,26,27,30,31,33,37,39 4 3,6,7,19,20,21,22,23,24,25,28,29,32,34,35

4.2.2 Physical and chemical analysis of the sites The grouped variables in the four clusters showed similar physical and chemical qualities for the whole area under study, with a relative difference shown by the variable “soil depth”, which values are higher in the sites grouped in clusters 1 and 2 with regard to the sites grouped in clusters 3 and 4. Because of this, they showed better conditions for plantation establishment. (Table 3).

4.2.3 Cartography of the planting susceptible areas The masks of the disturbed areas that originated in component 2 of the PCA were overlaid for their visualization and analysis in the composite image. This cartographic product represents the base map that in a thematic and digital format contains the necessary information to cover the necessities of other variables that were not analyzed in this study as the estimation of the disturbed areas, commercial type of affected threes, severity degree of the disturbance, and ecological impact analysis, among other variables. The planting areas have a total of 5,820 hectares (Table 4) that are a consequence of diverse factors that determine the conditions of the existing forest. In all the studied areas, condition is associated with management history and includes over-harvesting, illegal harvesting, forest fires, and changes in soil use.

4.2.4 Surfaces analysis Besides elevation, the processing and analysis of the DEM gave data for various slopes, which is important in commercial planting and environmental restoration




programs. Slopes greater than 5% are present in approximately 88% of the areas with gradients from 62% in at San Juanito to 94% in Cusarare. This physiographic situation is associated in an inverted way to the class 0-2 % where San Juanito has the highest percentage of plain or nearly flat land (23%) while Cusarare has only 1% of this class. Diverse authors [8,9,10] assure that the selection of the right species, slope, soil, and climate among other factors are variables that determine the success of plantations since the intended commercial plantations are meant to be established in preferred nearly flat soils, while the areas dedicated to protection are meant to be planted on steep slopes.

Table 3: Values of interest variables and their grouping in the four resulting clusters.

Cluster Variable 1 2 3 4

Elevation 2390.00 2410.28 2456.66 2318.41 Soil Depth (cm)

34.5000 33.7143 26.6000 25.0000

Mould (cm)

0.9167 0.4429 1.0800 0.6667

ApDens. 0.6220 0.6624 0.7226 0.8039 E. Poros (%)

63.9307 64.6811 62.0351 60.4064

C. E. (mmhos /cm)

0.2200 Normal

0.1743 Normal

0.1867 Normal

0.2000 Normal

pH 4.18 Highly acid

4.32 Highly acid

4.36 Highly acid

4.20 Highly acid

Org. Mat.(%)

1.7183 M. high

1.6671 M. high

1.5407 M. high

1.1325 M. high

RAS 9.51 Normal 7.27 Normal 7.58 Normal 7.27 Normal N-NO (Kgm/ha)

404.69 M. high

469.29 M. high

235.75 M. high

230.62 M. high

P (Kgm/ha)

12.78 Deficient

14.88 Deficient

11.84 Deficient

9.43 Deficient

K (ppm) 770.83 Excessive

278.57 High 210.83 High 275.00 High

C. H. (cm/hr)

6.473 MF 6.68 MF 4.44 M 8.62 MF

Sand (%) 52.7280 45.1337 57.6853 60.9947 Clay (%) 23.1520 25.0720 20.0800 20.2787 Bare soil (%)

39.0667 49.5571 27.3333 41.7250




Table 4: Planting susceptible surfaces based on the spectral analysis.

Property Area (ha) Planting area (ha) San Juanito 5229 1877.83 Bocoyna 6055 956.46 San Ignacio de Arareco 21051 1370.04 Cusárare 29914 1616.09 TOTAL 62249 5820.42

5 Discussion

The creation of preliminary maps was very useful because it allowed the generation of actual use of soil cartography to discriminate the disturbed areas, locate the sampling sites, and then quantify the polygons of the areas suitable for planting. The careful selection of these three bands was made according to the suggestions of Richards [11] and Beaubien [12] who recommended the reduction in the number of bands and use of those whose combination maintained the level of discrimination for the types of cover of interest. From a statistical point of view, the PCA easily made a first interpretation over the variability axles of the image and allowed identification of those features that showed up in almost all of the bands and those others that are specific to some group of bands [7,13,14]. Although Muchoney and Haack [15] referred to the temporary changes and disturbance in vegetation that are observed in the last components of Landsat TM, in this study the component 2 allowed discrimination in a clear way for the disturbed areas which indicates the importance of using this multivariate tool to ease the processes for this type of study. Variables of interest among the 40 sites were grouped in the four resulting clusters from the multivariate analysis (Cluster analysis) based on the similarity of the values of the studied variables that allowed the distribution and grouping of the sites. Johnson [16] noted that this multivariate technique involves techniques that produce classifications from data that initially are not classified and must not be confused with the discriminating analysis in which from the beginning it is known how many groups exist, and it has data that comes from each one of these groups. The site analysis showed similar physical and chemical features for the whole study area, with a relative difference expressed by the variable soil depth whose values are greater in the grouped sites in clusters 1 and 2 (34.5 and 33.7 cm.), in respect to sites grouped in clusters 3 and 4 (26.0 and 25.0 cm.), for which they showed the best conditions for plantation establishment. This matches another study [17] where soil depth was one of the edaphic variables of greatest impact in determining the quality of sites for plantations, since with greater depth the water retention increases in the soil and makes it easier to incorporate organic matter from the decomposition of logs, branches, and twigs among others. This permits an adequate radicular development for the establishment of crops [18]. In respect to this, Narváez and Armendáriz [19]




noted that verticiles in plantations of Pinus arizonica, Pinus duranguensis, and Pinus engelmanii depend on the total depth of the soil. The planting area on the assessed properties, which together make a total of 5,820 hectares, is a consequence of diverse factors that determine the actual forest conditions, which is associated with the management history of the properties related to factors like over-harvesting, illegal harvesting, forest fires, and changes in land use. Besides elevation, other processes and analysis gave important features of the terrain, mainly in the derivation of slope, which is an important variable in commercial plantations and environmental restoration programs. Diverse authors [8,9,10] pointed out that the selection of adequate species, slope, soil, and climate among other factors, are variables that determine the success of plantations.

References

[1] Rotmans, J. and R. J. Swart. 1991. Modeling tropical deforestation and its consequences for global climate. Ecol. Modeling. 58: 217-247.

[2] CONAFOR, 2005. Comisión Nacional Forestal. Inventario Nacional Forestal. Versión Preliminar s/p.

[3] Franklin, J. 1994. Thematic Mapper Analysis of Coniferous Forest Structure and Composition. In. J. Remote Sensing of Environment. p. 35.

[4] Francois, 1999. Aplicación de Imágenes de Satélite para Análisis de Recursos Forestales Deforestación y Fragmentación Forestal en la región de la Laguna de Términos, Campeche: un análisis del período 1974-91. Centro EPOMEX Universidad Autónoma de Campeche. Campeche, Camp. México. p 12.

[5] Lillesand, T. M. y R. W. Kiefer. 1987. Remote Sensing and Image Interpretation. John Wiley and Sons. New York.

[6] Pinedo, A. A. 2004. Análisis multitemporal de áreas deforestadas en la región centro-norte de la sierra occidental, Chihuahua, México. Tesis de Maestría. Facultad de Facultad de Zootecnia. UACH. p.53.

[7] Ferrero, B.S., M.G. Palacio, y R.O. Campanella, 2004. Análisis de Componentes Principales en Teledetección. Consideraciones estadísticas para optimizar su interpretación (Parte II). Universidad Nacional de Río Cuarto, Córdoba, Argentina. p. 9.

[8] Campos, R. D., Caráves, S. I. y Varela O. R. 1998. Normas para el Trabajo Técnico No 53. p 54.

[9] Moreno, S. R., Moreno, S. F. y Cruz, B.G. 1994. Determinación de áreas potenciales para plantaciones forestales. IV Reunión Nacional de Plantaciones Forestales. SARH. pp. 180 – 186.

[10] Escárpita, H. A. 2002. Situación actual de los bosques en Chihuahua. Madera y Bosques 8(1), 2002:3-18.

[11] Richard, J. A. 1986. Remote Sensing Digital Image Analysis. And introduction, Ed. Springer-verlang.




[12] Beaubien, J. 1994. Landsat TM Satellite Images of Forests: From Enhancement to Classification. Canadian Journal of Remote Sensing. Vol. 11. No Québec, Canadá. p. 17-26. Bosques 8(1), 2002:3-18.

[13] Robin, M. 1995. La télédéction. Nathan Universite. Des satellites aux systémés d´ information géographiques. Edition Nathan.

[14] SPIPR2. 1992. Sistema Personal Interactivo de Percepción Remota. Versión 2.0. Aguascalientes, Ags. 1992.

[15] Muchoney, D.M. and Haack, B.N. (1994). Change Detection for Monitoring Forest Defoliation. Photogrammetric Engineering & Remote Sensing. Vol. 60, No. 10. pp 1243-1251.

[16] Johnson, E. D. 1998. Métodos Multivariados Aplicados al Análisis de Datos.Kansas State University. p. 319.

[17] Carmean, W. H. 1975. Forest site cuality evaluation in the United States. Advances in Agronomy. 27:209-269. USA.

[18] Davel, M. y Ortega, A, M. 2002. Predicting site index from environmental variables in Douglas-fir plantations in the Patagonian Andes, Argentina. En línea Disponible: [email protected]. Accesado: 6 de Noviembre del 2005.

[19] Narváez, F. R. y Armendáriz ,O. R. (2005). El suelo y clima en relación con la calidad de sitio de las plantaciones forestales en los municipios de Bocoyna, Guerrero y Madera del estado de Chihuahua. En Memoria del VII Congreso Mexicano de Recursos Forestales. Chihuahua, México. pp 94-95.




Section 11 Soil and rock properties

The pedoecologial conditions of natural and opencast peat fields in Estonia

M. Noormets1, T. Köster1, T. Tõnutare1, K. Kauer1, R. Kõlli1, T. Paal2 & M. Oder1 1Department of Soil Science and Agrochemistry, Estonian University of Life Sciences, Estonia 2Forest Research Institute, Estonian University of Life Sciences, Estonia

Abstract

Peat soils in five different bog areas were studied during 2004. The study areas were chosen according to their exploitation type; natural, cultivated and milled peatlands. The peat soil nutrient content, plant associations and their nutrient content were examined. The peat soils were analysed for the CHWE, total N, P, K, Ca, Mg, C, for ash content and for plant-available P, K, Ca, and Mg. For the distribution of nutrients in the peat soils, samples were taken from 0-5, 5-10 and 10-15 cm depths. The dominating plant species and the type of plant associations were determined in areas where it was presented. The plant samples were analysed for N, P, K, Ca and Mg. The study results showed that in natural peat areas the variability in the ash content was high; the same was found for the Ca and Mg (%) content. With increasing depth, the nutrient content decreased, but this depth relationship was not significant for every parameter examined. The plant cover in cultivated peat areas had the highest nutrient contents. Key words: Fibri Dystric Histosol, natural peat bog, milled peatland, cultivated peat soils, plant association, total nutrients, available nutrients, pH, C, N, C:N ratio, CHWE, ash, P, K, Ca, Mg.

1 Introduction

Nowadays we consider the peatlands as a valuable ecological biotope that should be protected and rationally used because it is essentially a nonrenewable resource. Currently the peatlands are used for growing wild berry species like: Rubus chamaemorus, Oxycoccus palustris and Vaccinium species [1, 2, 3]; they




doi:10.2495/GEO060411

are used for forestry [4, 5], and as a growing media source for horticulture [6]. Historically and presently, peatlands have been heavily utilized and degraded by their agricultural and mining use. After drainage, the self-regulation system of mires is changed and the natural peat formation process is destroyed and then the dominating process is peat mineralization [7, 8]. Due to various agricultural activities in peat lands, the peat loss as a result of mineralization could reach 10 to 15 tons of organic material per hectare annually [9]. One possible strategy to stop or slow the mineralization is to cultivate wild berries (Vaccinium species). However, even with this strategy, mineralization and surface deflation are fast compared with peat formation. The preservation of peatland resources is critical because they are reservoirs of clean water, important sinks for carbon sequestration, important for maintenance of the hydrologic regime, and lastly they are valuable biotopes.

The Northern mires represent the largest store of carbon in the terrestrial ecosystem. Therefore, these areas play an important role in the Earth’s gas emission budget that affects the atmospheric CO2 concentration. According to Kreshtapova and Maslov [10] the CO2 storage in boreal peat deposits is estimated at 455 Pg. Moreover, Horn et al. [11] report that 30% of global reserves of soil carbon are preserved on natural bog areas. These areas can contribute up to 7% of the global annual emission of the greenhouse gas methane. If the natural condition of mires is destroyed by drainage, peat harvest, or agricultural use, they will change from being a carbon sink, to an important CO2 emission source.

Currently the annual peat production of different types of peat is approximately 4.5 – 5 million m3 in Estonia. Of the produced peat, 2.5 – 3 million m3 is comprised of the weakly decomposed peat used as a growing media in various horticultural production operations. The peat for growth substrate is produced mechanically by tractor drawn vacuum machines or by block cut harvesting. Most of the well decomposed peat (1.5 – 1.8 million m3) is used for energy consumption. Exports from milled peatlands are in excess of 90% for plant growing media, 80% for briquettes, and approximately 60% for block cut peat [12].

In this article will identify the peat soil properties as found under ‘natural’ conditions, milled peatland and under the cultivation in order to find out their characteristics and suitability for natural revegetation. An additional purpose of this study was to compare the natural, milled and cultivated peat soils agrochemical properties.

2 Study area

The study areas are located in the County of the Tartu, in the southern part of Estonia. According to the WRB soil classification system, the soil of the experimental fields belongs subgroups of Fibri–Dystric Histosols - HSdy(fi). The plant associations of the study areas were determined according to the classification system developed by J. Paal [13]. The study areas Sangla I and




Sangla III are used currently for peat mining, therefore no plant cover is presented.

Table 1: The type of study area use and plant associations or dominating

Number of study area

Name of study site

Type of use Plant association or dominating plant species

1 Sangla I

Milled peat land Without plant cover.

2 Sangla II

Technogenic stripe

No clear plant association is presented

3 Sangla III

Milled peat land Without plant cover.

4 Ilmatsalu I

Natural peat bog Ledo-Pinetum, Ledum pine heath moor

5 Ilmatsalu II

Cultivated peat bog area.

V. angustifolium, V. angustifolium X V. corymbosum

Sangla I. The peat processing started in this particular area in 1962 and currently the mining area is increased up to 590 ha. The Sangla mire peat deposit thickness is 6.5-9.5 m [5, 14]. Before the natural plant cover was removed for mining in Sangla I, the area was characterized as the heath moor growth site type. The main processing commodities produced from this site are horticultural peat and briquettes. For the package peat with the following criteria is used: 100% of poorly decomposed Sphagnum sp. peat (white peat); the moisture content is between 35% and 55%; the peat is free of weed seeds and radioactive or chemical substances. The Von Post decomposition index is H1…H3, the ash content does not exceed 5%; the bulk density is 130 - 200 kg/m3; and the pH is 3.0 - 4.0. Sangla II is a small area surrounding the main peat milling ground where the natural revegetation is occurring and for the purposes of this paper, it is described as a technogenic stripe. The whole area is influenced by drainage and in the past during the planning of the milling ground the plant cover was removed and pushed by machines to the sides. Nowadays there is a succession plant cover that could potentially function as the donor area for wild plant species during the rehabilitation work. The dominating plant species are Pinus sylvestris, Betula sp., and Eriophorum vaginatum. In the grass canopy, the following were identified; Carex cespitosa, C. acuta. C. elata, C. nigra, C. lasiocarpa, Phragmites australis, Phalaris arundinacea, Filipendula ulmaria. Sangla III. The mining area is located in the Sangla mire (same as the Sangla I peat deposit) and in the Laugesoo mire. The main commodities are fuel peat, peat briquettes and peat extraction. Ilmatsalu I is part of Sangla mire, named for the region where it is situated. This peat deposit is 6.5 – 9.5 m thick (same as Ilmatsalu II) and the bog is classified as heath moor. The somewhat more important plant association of this bog is




plant species.

Ledo-Pinetum. The main tree species is Pinus sylvestris with accompanying species of Ledum palustre, Calluna vulgaris, Andromeda polifolia, Empetrum nigrum, Oxycoccus palustri, Vaccinium uliginosum and Vaccinium myrtillus. The following bryophytes were represented Polytrichum commune, P. strictum, Sphagnum sp. In the grass canopy Rubus chamaemorus and Carex sp is presented. Ilmatsalu II. In, 1994 the peat milling on current area was finished. Before the peat excavation the growth type was heath moors, and important plant association was Ledo-Pinetum. Currently in opencast peat field have the dominant species Vaccinium angustifolium and in addition V. angustifolium X V. corymbosum cultivar Northblue was planted. Randomly among cultivated species appeared Pinus sp., Betula sp. and Eriophorum vaginatum. In the bryophyte canopy Marchantia polymorpha, Ceratodon purpureus, Polytrichum strictu, Funaria hygrometrica are randomly scattered. Drosera rotundifolia was randomly found. In early spring the plantation received 50 kg ha-1 of (11:11:22) N:P2O5:K20 fertilizers.


For the current investigation the soil sampling was carried out during the vegetation period in 2004. The plant association description was made during the summers of 2004 – 2005. Soils were described on the basis of test pits and the samples for analysis were taken from different depths (0-5, 5-10 and 10-15 cm). The peat soil samples were taken in four replicates and analysed in laboratory in triplicate. Peat samples were analysed for organic C according to the Tjurin method [15] and total N according to Kjeldahl [16]. For the direct estimation of the organic matter content the loss-on-ignition (LOI) method was used. Available P, K, Ca and Mg were analysed according to the Mehlich-3 method [18]. The pH was measured from the soil suspension with 1M KCl (1:5 w/v). For the total phosphorus, potassium, sodium, calcium and magnesium in peat and plant determination samples were destroyed by wet digestion with sulphuric acid. Total phosphorus was determined spectrophotometrically by vanadomolybdophosphoric acid method. K, Ca and Mg were determined with atomic absorption spectrometer [17]. Hot water extractable organic C from the soil-water extract (1:100, w/v) was measured [16]. STATISTICA 7 [19] was used for the statistical analysis and the standard error (± SE) is presented on the figures. In the tables the standard deviation (± SD) is presented.


The study area peat pH values varied from 2.4 to 5.2. In Sangla III the pH was relatively high (pH 5.2) it might be hypothesized that the higher values for pH and ash are indicative of the increased importance of fen deposit. However, the nutrient content is low at this location. Spiers [20] is found that in low acidic soils at pH 3.5, the plant growth and yield is decreased. In the cultivated area




Ilmatsalu II despite to fertilizers application the soil nutrient content for total and available nutrients was low when compared to other areas (Table 2).

Table 2: Peat soils agrochemical properties in the study areas (n=12).

Number of study area Soil

characteristic 1 2 3 4 5 pHKCl 4.0 ± 0.7 4.0 ± 0.7

5.2 ±0.1

2.4± 0.1

3.1± 0.9

Ctot. % 38.9± 2.4 38.3± 3.3 38.3± 3.1 38.4± 4.4 41.0± 2.9 Ntot. % 1.3± 0. 3 1.6± 0.2 2.1± 0.1 1.0± 0.3 1.0± 0.1

C:N 31.5± 10.2 24.4 ± 2.6 18.3±2.3 40.7±12.9 42.3±7.6 CWE. % 0.62± 0.16 1.75± 0.44 0.5± 0.1 1.1± 0.4 1.0± 0.3 Ash % 3.3± 1.6 27.1± 18.6 10.7± 4.7 3.3± 2.9 1.3± 0.6 Ptot. % 0.03± 0.01 0.05± 0.01 0.03± 0.02 0.03± 0.02 0.02± 0.01 Ktot. % 0.19± 0.05 0.64± 0.24 0.14± 0.13 0.29± 0.25 0.11± 0.08 Catot. % 0.7± 0.2 0.7± 0.2 1.2± 0.4 0.2± 0.1 0.4 ± 0.3 Mgtot. % 0.18± 0.05 0.26± 0.10 0.23± 0.03 0.06± 0.01 0.12± 0.08

Pavb. mg kg-1 23.5± 10.4 85.6± 36.9 8.9 ± 4.8 88.3 ± 85.8 34.8± 31.7 Kavb. mg kg-1 224.8± 86.4 207.2± 56.9 120.2± 19.7 297.4± 129.5 190.8± 34.5 Caavb. mg kg-1 9258± 1375 9070± 2410 12889± 2319 3530± 959 5551± 1988 Mgavb. mg kg-1 1598± 328 1840± 381 1733± 217 766± 248 1253± 567

*tot- total content of nutrients, *avb- available content of nutrients. *The numbers in tables are presented as average of different depths (upper 15 cm).

According to the soil nutrient assessment valuation done at Estonian Agricultural Research Centre the content of plant available P is in Sangla I and Ilmatsalu II low (23.5 and 34.8 mg kg-1, respectively), in Sangla III very low (8.9 mg kg-1). Unexpectedly, the content of available Ca and Mg was very high in all study areas peat soils.

The plant nutrient content was higher in the cultivated area than on natural area grown plants (Figure 1). Although, the agrochemical parameters of the cultivated area were in the same range, or slightly lower, the plant nutrient content was higher. It is reported that the availability of P could be enhanced while the yield is decreased [21]. But Holmes [22] has indicated that this situation could be improved by additional phosphorous application. Fertilization of plants grown in peat soils is important [23, 24, 25] because of the low mineral nutrient availability in soil. From Noormets et al. [26] in a study using Vaccinium species have found that plant growth and yield formation on opencast peat field (Fibri Dystric Histosol) was influenced by the fertilisation rates and nutrient balance at pH level 2.8-4.0. Significantly different content for total N and Ca, also available Mg was found in milled peatland when compared to the milled and cultivated peatlands (Figure 1). The natural peat soils were found to have higher levels of total K and plant-available P and K when compared to mined locations. The content of ash depended upon the decomposition level of peat material itself, and it varied from 3.3% to 27.1% in the areas examined. The higher content of ash in technogenic stripe might be explained by the mineral material that is carried from the road




nearby. Some studies have found that the liming of peat soils could lead to the pH increase but in natural areas, like an exhausted peat soil, it could increase the possibility for the wild plant species expansion from the surrounding areas [27]. The revegetation of milled peat lands areas in Estonia with trees and bushes occurs rapidly, but the restoration of natural Sphagnum mosses would seem to be very complicated. According to Salonen [28] the plant succession in opencast peat fields differs from the natural sites. As demonstrated in trials in the Cacouna Station (Canada), the revegetation in some plots by trees and bushes occurred very quickly. The scarcity of Sphagnum and other mosses indicates that the bog is not returning to the functional peatland ecosystem [29]. The revegetation depends on the peat harvesting technology, where the block-cut mined bogs were rapidly recolonized (less than 5 years) by ericaceous shrubs, but the vacuum-mined peatlands were covered by plant cover much more slowly (approximately 25 years) and not by typical peatland species. The study has pointed to the two main problems occurring by revegetation: firstly, large areas are without any plant cover; and secondly, the moisture deficiency in the top peat layers. Both of these conditions contribute to the wind erosion of the peatland [29].

milled peatland natural cultivatedType of use

-0,20,00,20,40,60,81,01,21,41,61,82,0

N, % P, % K, % Mg, % Ca, %

milled peatland natural cultivated

Type of use

-2000

200400600800

100012001400160018002000

P, mg kg-1

K, mg kg-1

Mg, mg kg-1

A. B.

natural cultivatedType of use

-0,5

0,0

0,5

1,0

1,5

2,0

2,5

3,0 N, % P, % K, % Mg, % Ca, %

milled peatland natural cultivated

Type of use

0

5

10

15

20

25

30

35

40

45 C, % Ash, %

C. D.

Figure 1: The nutrient content of different investigation areas according to the type of exploitation (M±SE). A- content of total nutrients. B- content of available nutrients. C- content of total nutrients in plants. D- content of total carbon and ash.




5 10 15Depth, cm

0,0

0,5

1,0

1,5

2,0 N, % P, % K, % Mg, % Ca, %

5 10 15Depth, cm

0

500

1000

1500

2000

2500 P, mg kg-1

K, mg kg-1

Mg, mg kg-1

A. B.

5 10 15Depth, cm

0

5

10

15

20

25

30

35

40

45

50

Valu

es

C, % Ash, %

C.

Figure 2: The nutrient content in different depths (M±SE). A- the content of total nutrients. B- the content of available nutrients. C- content of total carbon and ash.

The mean content of some nutrients (Ntot, Catot, Mgavb) decreased with increasing depth, but as the variability was high, this decrease was not significant (Figure 2). However, the decrease of some nutrient by depth like Ktot and Mgtot was observed. Also, there was found to be a significant decrease in the content of Pavb.

5 Conclusions

Currently, in Estonia there is a tendency to expand the working milling areas of peatlands and then the rehabilitation work is completed afterward. This study has demonstrated that the content of available P was highly variable, ranging from 10-88.3 mg kg-1, and the variability was highest in natural bog areas. In the technogenic stripe areas, the ash content (27.1%) was significantly higher than in other study areas where it ranged 1.3-3.3%. The content of available Ca and Mg was also very high. The nutrient content in the plant cover was higher than in natural area, likely due to fertilization. The studied depths for nutrient content in the peat deposit showed significance difference for total K and available P, for the rest of parameters the variability found to be very high.




Acknowledgements

We thank colleagues from the Department of Soil Science and Agrochemistry EMU for their collaboration with the laboratory analyses, especially Mrs. Imbi Albre. We thank the farmers and companies for their kind permission to use their property for the research. This study was supported by the Estonian Ministry of Education and Research, No 0172613s03. The conference participation fee was supported by RAK and by the Estonian Science Foundation (Project 4726).

References

[1] Noormets, M., Köster, T., Karp, K., Paal, T. & Tõnutare, T., The recultivation of opencast peatland in Estonia. Proc. of the 12th International Peat Congress. Wise Use of Peatland. ed. J. Päivänen, Finland, pp. 1195-1201, 2004.

[2] Jaadla, T., Cranberry production on peatlands. Estonian Peat, 1, pp. 19, 1994.

[3] Paal, T., On cultivating European cranberry in Estonia. Journal of Small Fruit and Viticulture. 1(1), pp. 59 - 61, 1992.

[4] Hartman, M., Karisto, M. & S. Kaunisto. Carbon dynamics for a drained peatland forest with drainage and nutrition gradients. Ecohydrological Processes in Northern Wetlands. pp. 209-214, 2003.

[5] Valk, U., Estonian mires. Valgus, Tallinn-Estonia, pp. 270, 1988. [6] Carlile, W.R., Horticultural developments, environmental challenges and

growing media. A global perspective. Proc. Int. Conf. Peat in Horticulture, Quality and Environmental Callenges, eds. G. Shmilewski. & L. Rochefort, Estonia, pp. 17-23, 2002.

[7] Ilomets, M., The peat mining on account of what? Estonian Nature, 2/3, pp. 20-24, 2003.

[8] Armentano, T.V. & Menges, E.S., Patterns of change in the carbon balance of organic soil wetlands of the temperate zone. Journal of Ecology, 74, pp. 755-774, 1986.

[9] Tomberg, U., Breakdown of peat as a result to drainage, Saku, Estonia, pp. 32, 1992.

[10] Kreshtapova, V., N. & Maslov, B. S., Contents of carbon compounds in reclaimed peat soils as a function of the properties of peat organic matter. Proc. of the 12th Int. Peat Congr. Wise Use of Peatland. ed. J. Päivänen. Finland, 2, pp. 988-992, 2004.

[11] Horn, M.A., Matthies, C., Küsel, K., Schramm, A. & Drake, H.L. Hydrogenotrophic methanogenesis by moderately acid-tolerant methanogens of a methane-emitting acidic peat. Applied and Environmental Microbiology, 69, pp. 74-83, 2003.

[12] Estonian Competition Board, 2002. http://www.konkurentsiamet.ee/ dokumendid/ko200272.htm.

[13] Paal. J., Classification of Estonian vegetation site types. Tartu-Tallinn. pp. 297, 1997.




[14] Orru. M., Širkova. M. & Veldre. M., Estonian peat resources. Estonian Geological Survey. Tallinn, pp. 146, 1992.

[15] Vorobjova, L.A., Chemical analysis of soils. Textbook, Moscow University Press, pp. 272, 1998.

[16] Procedures for soil analysis, ed. L.P. van Reeuwijk Wagening: ISRIC, pp. 112, 1995.

[17] Methods of soil analysis. Part 3. SSSA Book Series: 5. Madison. Wisconsin. USA. pp. 1390, 1996.

[18] Handbook on Reference Methods for Soil Analysis. Soil and Plant Analysis Council. Inc. pp. 202, 1992.

[19] Statsoft. Statistica 7,0. Copyright 1984-2005. [20] Spiers, J.M., Influence of lime and sulfur soil additions on growth, yield

and leaf nutrient content of rabbiteye blueberry. Journal of the American Society of Horticultural Science, 109, pp. 559-562, 1984.

[21] Bishko, A.J. & Fisher, P.R., The pH-Response of a Peat-based Medium to Application of Acid-reaction Chemicals. Horticultural Science, 38(1), pp. 26-31, 2003.

[22] Holmes, R. S., Effect of phosphorous and ph on iron chlorosis of the blueberry in water culture. Soil Science, 90, pp. 374-379, 1960.

[23] Noormets. M. & Karp. K., The influence of fertilization to vegetative growth of the lowbush blueberry (Vaccinium angustifolium Ait.) in a young and cropping plantation in peat bog. Transactions of the Estonian Agricultural University, 212, pp. 149 – 154, 2001.

[24] Noormets, M., Karp, K., Starast, M. & Paal, T., The influence of fertilization on the production of lowbush blueberry (Vaccinium angustifolium Ait.) seedlings on opencast peat pits. Journal of Agricultural Sciences, 5(8), pp. 293-303, 2002.

[25] Noormets. M., Karp. K. & Paal. T., Recultivation of opencast peat pits with Vaccinium culture in Estonia. Ecosystems and Sustainable Development IV. eds. E. Tiezzi. & C.A. Brebbia. Wessex Institute of Technology. UK and J-L. USO. Universitat Jaume I. Spain. 2, pp. 584, 2003.

[26] Noormets, M., Karp, K., Kelt, K.,Tõnutare, T. & Paal, T., Fertilizers effects on the lowbush blueberry (Vaccinium angustifolium Ait.) plants and berry chemical composition, grown in opencast peat pits. Journal of Agricultural and Food Science, (submitted), 2006.

[27] Paal, T., Starast, M. & Karp, K., The influence of liming on growth of lowbush blueberry on exhausted peat fields. Uprawne rosliny wrzosowate, ed. T. Ligocka, Skiernewice, Poland, pp. 71-76, 2003.

[28] Salonen. V., Penttinen. A. & Särkkä. A., Plant colonization of a bare peat surface: population changes and spatial patterns. Journal of Vegetation Science 3, pp. 113-118, 1992.

[29] Lavoie. C. & Rochefort. L., The natural revegetation of a harvested peatland in southern Quebec: A spatial and dendroecological analysis. Ecoscience 3(1), pp. 101-111, 1996.




[30] Naucke, W., Heathwaite, A.L., Eggelsmann, R. & Schuch, M., Mire chemistry. Mires. Process, exploitation and conservation. ed. Heathwaite, A.L. pp. 263-310, 1993.




Monitoring programme for underground rock characterization facility

K. Lehto & J. Lahdenperä Posiva Oy, Finland

Abstract

Posiva is a Finnish nuclear waste management company with the duty to take care of the high-level nuclear waste produced in Finland. Pursuant to the Decision-in-Principle of 2001 Posiva has moved to the phase of underground characterisation of the repository site. The main objective of this phase is to confirm the suitability of the Olkiluoto site by investigations conducted underground. A programme of monitoring has been launched as part of the investigations and to follow the changes occurring within the site due to the construction. The programme of monitoring started in the year 2004, one year before the construction work. Baseline monitoring was carried out prior to the construction phase and the results gained from monitoring during construction phase are compared to the baseline data. Rock mechanics, hydrology, geochemistry, environment, and the use of foreign materials are included the monitoring programme. Keywords: monitoring, high-level nuclear waste, hydrology, rock mechanics, geochemistry, environment, foreign materials, underground rock characterization facility.

1 Introduction

Posiva Oy is a Finnish high-level nuclear waste management company owned by the two Finnish nuclear companies Teollisuuden Voima Oy and Fortum Power & Heat Oyj. Posiva’s duty is to take care of final disposal of spend nuclear fuel. Posiva is established in the year 1996.

In the year 2001, the Finnish Parliament ratified the Decision-in-Principle on the disposal of spent fuel from the Finnish nuclear power reactors. According to the decision, Posiva may concentrate the investigation on one site, Olkiluoto, in




doi:10.2495/GEO060421

the municipality of Eurajoki in the western part of Finland. The Decision-in-Principle also means that the repository would be located at Olkiluoto and the disposal would be based on the KBS-3 concept.

On the year 2004 Posiva started to construct an underground characterization facility called ONKALO, which will be a part of actual high-level waste repository later on. The repository is planned to be in use in the year 2020.

Before Posiva started to construct ONKALO a baseline monitoring was done. The monitoring results gained during the construction and use of ONKALO will be compared to the data gained during the baseline monitoring.

Figure 1: The design of the underground rock characterisation facility at the main drawings stage [6].

2 ONKALO: underground rock characterization facility

The construction of the underground rock characterization facility called ONKALO was started in July 2004. By the end of the year 2005 about 990 m of tunnel has been excavated and the first underground investigations program is implemented. The target depth of the ONKALO is 520 m below the sea level. The main characterization level is at the depth of 420 m below the sea level and the lower characterisation level in depth of 520 m below the sea level, fig. 1. The total length of the tunnels will be about 9 km. The entire facility is planed to be

ACCESS TUNNEL

LOWER CHARACTERISATION LEVEL

MAIN CHARACTERISATION LEVEL

VENTILATION SHAFT




ready in the year 2010. ONKALO is planned to be a part of the actual high-level waste repository, which is planned to start its operation in the year 2020 [6].

The construction of the ONKALO and the high-level waste repository will affect the surrounding rock mass and the groundwater flow system as well as the chemical environment on the surface but especially at the greater depths. To determine the magnitude and the extent of effects the programme of monitoring was established. The programme of monitoring at Olkiluoto during construction and operation of underground rock facility was established prior to the construction phase [7].

3 Programme of monitoring

Before the programme of monitoring during construction of ONKALO begun a baseline condition of Olkiluoto was reported [5]. The main purpose of the baseline report was to establish a reference point for the following phases of the Finnish spent nuclear fuel disposal programme. The focus of the baseline report was to define the current surface and the underground conditions at the site, to establish the natural fluctuation of properties that are potentially disturbed and to provide the reference data to the models and modellers.

During the construction of ONKALO rock mechanical, hydrological, chemical and environmental monitoring is carried out and the quality and the quantity of the foreign materials used in the construction and investigations is measured and followed [7]. The monitoring is carried out both on the surface and underground. In the Baseline report the current understanding of the investigation site before the constructions of ONKALO started was established. The results gained from monitoring have been and will be compared to the baseline data.

3.1 Hydrology

The following parameters have been included in the hydrological monitoring programme:

o level of groundwater and seawater table o hydraulic head o flow conditions o in situ salinity o precipitation o surface runoff o infiltration o soil frost o water balance in the tunnels

The hydrological monitoring is carried out from the both on surface and underground. Both the open and the packed-off boreholes are used in the surface based monitoring. The flow conditions, in situ salinities and the hydraulic heads are measured from the open boreholes by Posiva Flow Log measurement device and Posiva’s hydraulic testing unit (HTU).




Figure 2: Map of deep drillholes (KR) and multilevel piezometers (EP) in

Olkiluoto.

The hydraulic heads and the level of the ground water table are measured from the open and packed off boreholes as well as from the multilevel piezometers, fig 2. The level of water table is measured manually from the open boreholes once a week or month and the hydraulic head automatically from the packed-off boreholes once every ten minute.

The rest of the parameters listed above form a background data for the studies and the analysis of the actual monitoring results [1].

3.2 Rock mechanics

A local GPS-network was established at Olkiluoto in the year 1995. It consists of 14 stations one of which belongs to the Finnish permanent GPS network FinnRef. Four of the stations are located outside the island. The purpose of the measurements is to observe long-term large-scale rock movements and land uplift.

Because the GPS-measurements are not sensitive enough, precise levelling is carried out to measure the annual land uplift at the Olkiluoto.

A microseismic network of Olkiluoto was built in 2002. The network consists of 12 stations. Four of the stations are located outside Olkiluoto. In the beginning, the network monitored tectonic earthquakes in order to characterise the baseline of seismicity of the Olkiluoto. When the construction started it also monitors excavation-induced seismicity [9].




Later on when construction of ONKALO progresses more measurements such as convergence measurements and acoustic emission will be done and extensometers installed [7].

3.3 Geochemistry

The geochemical monitoring programme includes mapping of the changes in the groundwater chemistry caused by the construction and mapping of the influences caused by the engineering and the foreign materials entering ONKALO [3]. Most of the geochemical measurements have taken place from the surface based holes, but some underground sampling is also done.

Groundwater samples have been taken both from the overburden and deep bedrock by different methods. The chemical properties of the water used in the ONKALO and pumped out from there have been analysed.

Possible movement of the saline waters in the bedrock is studied by Gefinex 400s measurements.

3.4 Environment

Environmental monitoring has been carried out for a long time: the environmental monitoring performed by Teollisuuden Voima Oy (TVO) started in the 1970’s. TVO is obliged to carry out monitoring of marine environment and also of radioactivity of the environment around Olkiluoto, both in the marine and terrestrial systems and human food chains. Noise, dust and water quality of the private wells are monitored in campaigns. In addition, aerial photographing and characterisation of the flora and fauna is included in the programme [7].

The data gathered through the monitoring is used as input for the biosphere modelling in the long-term safety analysis as well as for the broader environmental impact analysis [2].

3.5 Foreign materials

During the construction and operation of the ONKALO several types of foreign materials are introduced to the underground facility, mainly for the engineering purposes. The amount of the materials, such as cement, rock bolts, labelled water etc. is monitored and recorded.

4 Results

The monitoring at Olkiluoto is part of the site investigations. The monitoring activities are done continuously or regularly and always in the same way. The programme of monitoring will be updated at the latest in 2012, but maybe earlier.

4.1 Hydrology

The results of hydrological monitoring are studied and analysed to determine the effects of the construction of ONKALO on the water balance of the reservoir.




The total leakage water inflow into ONKALO is about 16 l/min when length of tunnel in 990 m. Because of the fairly small inflow into ONKALO no long-term hydrological changes have occurred so far, fig 3.

Shallow boreholes (L, PP, PA, PR)

9909408867920 15 71 733 175 253 307 342 426 502 572 621 688

-2

-1

0

1

2

3

4

5

6

7

8

9

10

11

12

1.9.2004 10.12.2004 20.3.2005 28.6.2005 6.10.2005 14.1.2006time

grou

ndw

ater

tabl

e le

vel (

m.a

.s.l)

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

prec

ipita

tion

(mm

)

PP1 PP2 PP3 PP4 PP5 PP6 PP7PP8 PP9 PP10 PP31 PP32 PP34 PP35PP36 PP37 PP38 PP39 PA1/1 PA1/2 PA1/3PA2/1 PA2/3 PA3 PA4/1 PA4/2 PA4/3 PA5L1 L2/1 L2/2 L2/3 L3/1 L3/2 L3/3L4/1 L4/2 L4/3 L5 L7 L8/1 L8/2L8/3 L9 L13/1 L13/2 L13/3 L14/1 L14/2L14/3 L15/1 L15/2 L15/3 L16/1 L16/2 L16/3L26 L27 PR1 PR2 PR3 PR4 H7Sea level ONKALO chainage KR3 L4 KR3 L8 Pori Kuuminainen Daily precipitation

Figure 3: Level of the groundwater table in the shallow boreholes.

4.2 Rock Mechanics

The annual land uplift at the Olkiluoto is 6 mm/year [5]. Precise levelling shows that two areas rise faster than the rest of the island. The reason to this is that lots of the rock masses are moved away from these two areas.

Five of the microseismic events are classified uncertain or unusual. After Saari [10] it is very unlikely that they are microseismic events, but they are not typical examples of an artificial seismic signals, either, table 1.

No other rock mechanical changes have occurred in Olkiluoto since the construction started.

4.3 Geochemistry

Some changes in the main ions can be noticed in the groundwater samples taken from the surface. All the noticed changes are mainly minor and cannot be caused by the ONKALO activities because most of the samples have been collected before the construction took place [3].

No other changes in the geochemistry have been observed.

4.4 Environment

No unexplainable deviations from the reference data can be seen so far [2].




Table 1: Events of uncertain origin 2002-2004. Loc. Err = location error, Mag Loc = local magnitude, Size = estimated size of the seismic source, ESEP = Energy of S-wave/energy of P-wave and No of Acc = number of recordings used in analysis [10].

No Date Origin Time

(UTC)

X (m)

Y (m)

Z (m)

1 19.4.2002 09:41:36.5 6792766.0 1525527.0 -152.8 2 23.4.2002 17:55:47.4 6792785.5 1525545.9 -143.4 3 8.5.2002 00:10:04.2 6792017.5 1525408.3 -81.4 4 24.5.2002 19:59:35.7 6791939.0 1525934.5 -6030.0 5 24.8.2004 14:45:40.2 6791924.0 1526224.5 -364.8

Loc. Err. (m)

Mag Loc

Seismic Mom. (log10)

Radiated Energy (log10)

(J)

Size

(m)

ESEP

No. of

Acc.

No

0.0 -2.9 6.5E+00 -3.1E+00 10.2 2.4E-01 1 1 0.0 -3.1 6.1E+00 -3.0E+00 9.1 3.3E+01 1 2 0.0 -3.2 6.1E+00 -3.4E+00 9.1 3.8E+00 1 3 0.9 -0.4 9.0E+00 2.5E+00 4.1 0 5 4 3.4 -0.4 9.2E+00 2.3E+00 16.7 4.7E+00 4 5

4.5 Foreign materials

The use of foreign materials is restricted and regulated in the ONKALO and a list of appropriate materials that can be used in the ONKALO is established and continuously updated. Altogether 46 different materials have so far been authorised.

During the first 990 m of tunnel length a total of about 321 tons cement has been used. This compares well with the estimations made before the construction of the ONKALO begun.

5 Conclusions

Many of the possible changes caused by the construction of ONKALO are driven by the drawdown of the groundwater table, which is a consequence of the water inflow into the tunnel. Because of heavy use of cement in the upper parts of ONKALO, the water inflow is only about 16 l/min and because of that no drawdown has occurred.

Another reason why there have not been any significant changes is a fairly short time since the construction started.




References

[1] Ahokas, H., Klockars, J. & Lahdenperä, A-M. Results of Monitoring at Olkiluoto in 2003-2004. Hydrology. Working report 2005-28. Posiva Oy, Eurajoki, 2005.

[2] Haapanen, R. Results of Monitoring at Olkiluoto in 2004. Environment. Working report 2005-31. Posiva oy, Eurajoki, 2005.

[3] Hirvonen, H. & Mäntynen, M. (ed.). Results of Monitoring at Olkiluoto in 2004. Geochemistry. Working report 2005-29. Posiva Oy, Eurajoki, 2005.

[4] Juhola, P. Results of Monitoring at Olkiluoto in 2004. Foreign materials. Juhola, P. Working report 2005-27. Posiva Oy, Eurajoki, 2005.

[5] Posiva Oy. Baseline conditions at Olkiluoto. Posiva 2003-02. Posiva Oy, Eurajoki, 2003.

[6] Posiva Oy. ONKALO Underground Rock Characterization Facility - main drawings stage. Working report 2003-26. Posiva Oy, Eurajoki, 2003.

[7] Posiva Oy. Programme of Monitoring at Olkiluoto During Construction and Operation of the ONKALO. Posiva 2003-05. Posiva Oy, Eurajoki, 2003.

[8] Posiva Oy. TKS-2006. Nuclear waste management of the Olkiluoto and Loviisa power plants: Programme for Research, Development and Technical Design for 2007-2009 (in prep).

[9] Riikonen, S. (ed.) Results of Monitoring at Olkiluoto in 2004. Rock mechanics. Working report 2005-30. Posiva Oy, Eurajoki, 2005.

[10] Saari, J. Seismic Network at the Olkiluoto Site. Working Report 2003-37. Posiva Oy, Eurajoki, 2003.




Geotechnical evaluation of Stabilized Dredged Material (SDM) from the New York/New Jersey Harbor

A. Maher1, A. Sarmad2 & M. Jafari1

1Department of Civil and Environmental Engineering, Rutgers University, USA 2DMJM Harris Inc., New York, USA

Abstract

As a result of the 1997 ban on ocean dumping of dredged sediments, the States of New York and New Jersey have embarked on a rigorous program of seeking environmentally friendly solutions to the handling of dredged material, including the beneficial use of stabilized dredged material (SDM) in roadway applications. A pilot study was initiated in 1998 to construct two embankments on a site in Elizabeth, NJ, where SDM was successfully used as a cover for more than 100 acres of commercial development area. The pilot study included a laboratory phase for geotechnical evaluation of SDM, and a field phase for monitoring and evaluation of the construction process, as well as the performance of the fills following construction. The results of the laboratory study, as reported in this paper, demonstrate that SDM satisfies most of the geotechnical criteria for fill construction, except those for durability, requiring proper coverage and protection similar to those provided for fills constructed on cohesive soils. Keywords: dredged material; contaminated sediments; beneficial use; embankments.

1 Introduction

The Port of New York and New Jersey is the largest Port on the East coast of the United States, situated in the metropolitan center of the Hudson Raritan Estuary complex. The New York / New Jersey Harbor complex is naturally shallow, with an average depth of 19 feet at low tide. Due to the Port’s strategic position in regional and international trade, the U.S. Army Corps of Engineers has




doi:10.2495/GEO060431

provided some 250 miles of engineered waterways at depths ranging from 20 to 45 feet. Plans are underway to deepen the main channels to 53 feet during this decade. Maintenance of these waterways, crucial to safe navigation, requires dredging 4-6 million yd3 of sediment, or “dredged material”, annually. Unfortunately, at least half of the material scheduled for removal is contaminated with industrial chemicals and trace metals from historical and ongoing sources, making management of the material challenging. Historically, dredged materials from the channels and berths in the Port have been relocated to other parts of the Harbor, used to fill in shallows, or dumped in the ocean. Following the London Convention, the United States Environmental Protection Agency (USEPA) directed consignees to evaluate dredged material for its potential environmental impact prior to dredging. Materials found suitable for open water disposal were to be placed in one or more designated sites. In the case of the NY/NJ Harbor, this meant placing the material at a 2.2 square mile area off Sandy Hook, NJ, known locally as the “Mud Dump”. Starting in 1991, further modifications to the ocean disposal testing requirements resulted in strict restrictions on disposal at the site. In 1993, environmental groups began legally challenging even the most recent regulations, eventually resulting in an outright ban of disposal of dredged materials at the site by 1997. Today, only material considered to be completely free of potential to cause environmental harm is placed at the site, doubling as a cap of older, more contaminated materials. Unfortunately, these new regulations did nothing to slow the rate of sedimentation in the Harbor complex. Berths and channels in this heavily trafficked system require nearly continuous maintenance to ensure safe passage of commercial vessels. The Port community was unprepared for the loss of management options for dredged material. Managers were forced to either delay dredging or pay sums 15-20 times higher than usual. Dredging has all but ceased in the Port, threatening the maritime industry. In response the States of New Jersey and New York, the U.S. Army Corps of Engineers (Corps) and the Port Authority of New York and New Jersey (PANYNJ) created teams to find alternative methods for management of contaminated dredged material. One of the alternatives considered was to seek beneficial use of stabilized dredged material (SDM) in upland disposal sites. This entails the stabilization of dredged material with pozzolanic admixtures to create structural and non-structural fills for various applications, including those in brownfield development projects and transportation infrastructure systems. The beneficial use of SDM as a fill has been demonstrated to be cost effective for high volume usage. For example, approximately 600,000 cubic yards of SDM were successfully used as structural fill for the construction of parking areas for the Jersey Gardens Mall in Elizabeth, NJ. In this project, the developer utilized dredged material amended with Portland cement for the grading, filling and capping required for the remediation of the landfill. Amending dredged material with Portland cement yields three benefits: it binds contaminants to the sediment particles, it removes excess water and it improves the structural characteristics of the silt and clay particles.




2 Objective

During the course of the Jersey Gardens development project, the Office of Maritime Resources of the New Jersey Department of Transportation (NJDOT) initiated a pilot study to evaluate the feasibility of SDM as a fill material for roadway embankments. Two embankments were constructed on existing municipal solid waste fills at the Jersey Gardens Mall site using SDM as the fill material. The project had two phases: a laboratory phase (phase I) consisting of a comprehensive geotechnical evaluation of SDM for beneficial re-use applications, and a field phase (phase II) consisting of performance evaluation of embankments following construction. This paper summarizes the first phase of the study.

3 Geotechnical properties of SDM

The controlling parameters for the laboratory investigation were the type and the content of admixtures (cement and fly ash) that were used in the field phase, as well as the sequence of mixing, curing and placement activities specific to the project. The preparation of SDM in the field was conducted on the Jersey Gardens site using a pugmill system. After preparation, the stabilized dredge material (SDM) was placed on various locations at the site for stabilization for curing. Unlike typical soil-cement mixtures in which the soil and cement are mixed and then immediately compacted, the SDM due to its high initial water content was placed on holding sites while it dried and cured, and the final site preparations were made. Once the SDM had cured, it was moved to the embankment sites for final placement, molding and compaction. As a result, a direct comparison between the SDM used in this project and typical soil-cement materials could not be made. However, soil-cement properties are used as point of reference for the evaluation of laboratory results. Three different mixtures were prepared for the laboratory evaluation; each using raw dredged material (RDM), Portland cement and fly ash. The recipes were all mixed on a wet-weight basis. The three recipes used were: 1) RDM with 4% Portland cement, 2) RDM with 8% Portland cement, and 3) RDM with 8% Portland cement and 10% fly ash. The following tests were conducted to characterize each mixture: Unified Soil Classification ASTM D-1140, and D-422 Shear Strength (tri-axial), ASTM D-4767, 2850-87 Swell Pressure ASTM D-4546 Consolidation Test ASTM D-2435 Resilient Modulus AASHTO T274 Hydraulic Conductivity (Permeability) ASTM D-5084 Compaction Test ASTM D-1557 Durability ASTM D-559 Cement Content Determination ASTM D-806-96




3.1 Classification

The dredged material tested in this investigation is mostly silt with low percentages of fine sand and clay. Sediments dredged from navigational channels do not naturally contain coarse or medium sand (although incidental pieces of gravel were found in some samples), because sand will settle before it reaches still waters. In addition, these sediments cannot contain high percentages of clay, because clay particles will stay in suspension. However, deepening dredging in undisturbed areas might result in the generation of material containing significant amounts of gravel and rock mixed with fine material. This study did not address this type of material. The SDM samples tested consisted, on average, of 66% silt, 14% clay and 16% fine and medium sand (12.1% fine, 3.9% medium). Gravel content was negligible except for one sample, which contained 6.5% gravel. The percentage of clay size particles was higher for those SDM samples that had been mixed with fly ash, presumably due to the fine nature of fly ash particles. The organic content of the raw dredge material was determined to be around 8%, according to ASTM D2974.The effect of increased curing time on particle size distribution was minimal. Any variation in particle size is attributable to size variation in the source material. In addition to the gradation test, SDM samples were also tested for plasticity index. Based on the Atterberg Limits, all the samples tested are below the A-line and to the right of the LL=50 line on the Plasticity Chart. Therefore, the SDM could be classified as Elastic Silt (MH).

3.2 Moisture-density relationship

According to the test results, maximum dry densities ranged from 76.6 pcf to 78.8 pcf (1.23 to 1.26 Mg/m3), and optimum moisture contents ranged from 26% to 31.5%. A slight reduction in maximum dry density was observed when the percentage of cement and the curing time were increased prior to compaction of the material. This is similar to findings made by Kezdi [4], where the maximum dry densities of cement-treated silts were found to decrease slightly with increasing cement content.

3.3 Consolidation

Laboratory consolidation tests were conducted according to the ASTM D-2435 method. The samples were prepared using RDM amended with 4% Portland cement, 8% Portland cement, and 8% Portland cement with 10% fly ash. The SDM mix was remolded into a consolidometer with different compaction efforts applied. To determine the level of compaction achieved with each sample, a compaction test conforming to ASTM D-1557 was conducted for each recipe. According to the test results, samples were compacted to varying degrees ranging from 59% to 90% of their maximum dry density. The moisture contents used when the test samples were remolded were chosen to represent the site’s average and approved layers that did not meet the 85% Modified Proctor criteria. Samples were tested after one and six months of




curing. The energy applied for remolding the sample prior to the test played a major role in the consolidation behavior of the material. The test results indicate pre-consolidation stresses (Pc) as high as 8.7 tsf (833 kPa) once the sample is compacted to 87% of its modified maximum dry density. This means that the compacted material will compress before experiencing 8.7 tsf (833 kPa) of overburden (equivalent to approximately 170 feet ( 52 m) of SDM, unit weight of 100 pcf (1.6 Mg/m3), or 133 feet (40 m) of compacted granular fill unit weight of 130 pcf – 2.08 Mg/m3). However, Pc as low as 1.32 tsf (126.4 kPa) was recorded for a sample compacted to 86% of its modified maximum dry density. The average value of Pc, for samples compacted from 81% to 90% of their modified maximum dry density, is higher than 5 tsf (478 kPa). The compression index (Cc) values range from 0.22 to 0.9. Both of these values were recorded for SDM with 8% Portland cement. In general, for all recipes tested, once compaction reaches 81%, the compression index will not exceed 0.5. In that case, a Pc of 2 tsf or more should be expected. The compression ratio (CR =Cc/1+e0) varied from 0.085 to 0.24. This value did not exceed 0.19 for samples compacted to 83% or above. The results also show that based on consolidation settlement estimates, SDM embankments could be constructed to a height of 50 feet (15 m) with negligible settlement taking place within the SDM fill. This conclusion is supported by the results of the field settlement program. In the case of the two embankments in this study, and in similar cases where construction is proposed on marginal foundation soils, settlement is primarily a function of the foundation soil and its consolidation characteristics.

3.4 Permeability ASTM D-5084

Twenty-four samples were prepared and tested for permeability (hydraulic conductivity). Three different recipes for amending RDM were used in the sample preparation: 4% Portland cement, 8% Portland cement, and 8% Portland cement with 10% fly ash. The three different recipes were sampled at one month and at six months. Half of the samples were compacted to 85% and the other half were compacted to 90% of their maximum dry density, as determined by Modified Proctor (ASTM D-1557). The permeability results ranged from 1.25x10-6 cm/sec to 4.3x10 –7 cm/sec. The lowest values were recorded for samples of RDM amended with 8% Portland cement and 10% fly ash. Also, samples amended with 4% Portland cement generally had lower permeability than did samples amended with 8% Portland cement. This may be due to the apparent effect of cementation on imposing a flocculated fabric arrangement in SDM. In general, tests results indicate that SDM could be considered for use as a low permeability layer in landfill cap applications. For roadway applications, however, building roadways on SDM would be similar to building on compacted fine-grained sub-grades, such as those used in arid regions like Arizona, Texas, etc. For roadway construction, proper coverage must be provided using an appropriate base or sub-base materials.




3.5 Shear strength

The strength parameters of SDM were evaluated for feasibility of SDM as a fill material, and specifically for the slope stability of the pilot embankments. The consolidated undrained (CU) shear condition was determined to best reflect the realistic field conditions both during construction and post-construction periods. Both one and six-month-old samples of the three different recipes for SDM were tested for shear strength characteristics under CU conditions (ASTM D-2850-87). The samples were compacted to 85% and 90% of their modified maximum dry density and total and effective strength parameters were determined for stability analysis. The effective C and φ or (C′ and φ′) were calculated after the Mohr circles for effective stresses were plotted. As expected, the effective friction angle values were generally larger than the total values for SDM. No significant change or trend in the magnitude of the frictional angle, and, with the addition of cement and fly ash could be observed. This is similar to previous findings by Balmer [1], Clough, et al. [2] and Van Riessen and Hansen [6]; where different soil types, amended with varying cement contents, were extensively tested and showed no significant change in frictional angle as a function of the varying amount of cement. In general, an average angle of 34o can be estimated for long-term stability analysis of embankments constructed with SDM. On average, there is an 8° increase in the effective friction angle compared with the total friction angle. Cohesion, however, decreases as the friction angle increases. The test results also showed that compaction plays a significant role in the magnitude of strength parameters. For all the samples tested, a 5% increase in dry density resulted in increased strength. On average, the un-drained C values increased by 35%. Moreover, the average increases in φ′ and C′ were 1 % and 50%, respectively. On this basis, it can be concluded that compaction is the most important physical stabilizer of SDM with respect to strength parameters. A general comparison of SDM with typical soil-cement and cement-modified soils shows that with the same percentage of added cement, and similar compaction efforts (90% of optimum for SDM, and optimum for soil-cement) cement-modified soils are denser than SDM, have slightly higher friction angles, and have a much higher cohesion intercept under triaxial shear conditions.

3.6 Resilient modulus (AASHTO TP46-94)

The resilient modulus is a dynamic soil property used in the mechanistic design of pavements. The test provides a means of characterizing base, sub-base and sub-grade materials under simulated field loading conditions and is the basis for a deterministic approach to pavement design. In the resilient modulus test, the materials are tested under a variety of conditions, some of which include stress state, moisture content, temperature, gradation and density. A detailed description of the test and sub grade resilient properties of NJ soils is given by Maher, et al [5].




Sample Type

Stockpiling Period

(months)

Compaction Effort (%)

Resilient Modulus

(Psi)

Resilient Modulus

(MPa) 85 4827.5 33.28 4% PC 1 90 7720.2 53.22 85 5167.9 35.62 4% PC 6 90 8752 60.34 85 11911 82.12 8% PC 1 90 12326 84.98 85 8432 58.13 8% PC 6 90 8945 61.67 85 5610 38.68 8%PC + 10% FA 1 90 9254 63.80 85 1498 10.32 8%PC + 10% FA 6 90 6601 45.51

Rt. 23 in NJ (medium to fine sand)

max dry density 9633 66.42

Rt. 295 in NJ (medium to fine silty sand)


Rt. 206 in NJ (silt with fine sand)


For this study, the resilient properties of SDM were determined for all the mixture types used. Table 1 summarizes the resultant resilient modulus values for SDM mixtures and those for three New Jersey sub-grade soils that currently underlie roadways in New Jersey. According to the table, SDM compares favourably to the soil taken from various subgrades in NJ, indicating sufficient resiliency under dynamic loads.

Table 1: Comparison of resilient modulus values between SDM and typical NJ base materials.

3.7 Swell potential

Samples of SDM were also tested for swell pressure in order to determine if SDM could be used in applications where the material would be in contact with structures sensitive to swell pressures and excessive deformations. For example, if SDM were used as a base material in roadways, excessive swell pressures and deformations will be detrimental to the integrity of the pavement. For this study, samples of RDM were mixed with 4% Portland cement, 8% Portland cement, and 8% Portland cement plus 10% fly ash. Samples were cured in the laboratory for one month and for six months. These samples were then compacted to different densities in order to determine at what point the density level and moisture content would become critical in generating excessive swell pressure and deformation. Swell tests were performed in accordance with ASTM D-4546. The laboratory data indicate several trends. The strain or percent swell was not significant for any of the samples tested. The strain values ranged from 0.1




to 1.2 percent, with an average of 0.6. The maximum strain belonged to the sample amended with 8% Portland cement plus10% fly ash (1.2%). This magnitude of volume change is considered low and, therefore, not detrimental to adjacent structures. The swell pressure, however, was high for samples compacted to 94% or higher of their maximum dry density with moisture contents on the dry side of optimum. For these samples, the overall average swell pressure was 1.005 tsf (96.25 kPa). The average for one-month old samples was slightly higher at 1.34 tsf (128.32 kPa), with an average strain of 1.1%. Although strains were not high for any of the samples tested, the swell pressure generated was moderate. For SDM that was mixed with 8% Portland cement and compacted to 95% of its maximum dry density, the swell pressure was measured as high as 1.96 tsf (187.69 kPa). However, considering low associated strains, SDM would not have any detrimental effect on adjacent structures. For samples compacted on the wet side of their optimum moisture content, much lower swell pressures and strains were measured. The average swell pressure for those samples was 0.14 tsf (13.41 kPa), and the average strain was 0.3%. This is due to the fact that fine-grained soils have a flocculated structure at low moisture contents (below optimum moisture content). At moisture contents above optimum, the structure of the soil particles becomes more dispersed and layered. For dispersed structures, additional moisture does not result in significant volume changes.

3.8 Durability

The major durability concerns regarding SDM include potential strength loss due to freeze-thaw cycles and moisture variation. The freeze-thaw test simulates the internal expansive forces that result from the moisture in fine-grained soils. During freeze-thaw cycles, SDM experiences an increase in volume and a loss in strength. Some soil-cement mixtures have the ability to regain strength under certain conditions; specifically, the availability of reactive Calcium Oxide, adequate temperature and a high pH environment. For SDM, these conditions do not exist; therefore, any strength loss will be permanent. In order to study the effects of freeze-thaw cycles on SDM, samples were prepared from the three different recipes. The testing was performed in accordance with ASTM D560. Samples were compacted to 85 and 90% of their maximum dry density, as determined by Modified Proctor. To provide a point of reference, a natural clay sample was also tested for its behavior during freeze-thaw cycles. According to the test results, none of the samples could withstand more than three freeze-thaw cycles before failing. Significant volume change (ranging from 1.8% to 58%) was experienced during testing. Considering that the average volume change for the natural clay sample was 2%, it may be concluded that the freeze-thaw effect is several times more severe for SDM than it is for natural clay. As a result, all SDM should be protected against frost in order to maintain the cement contents within the percentages used for this project. Frost depth in New Jersey is approximately 2.5 to 3 feet (0.75 to 0.9 m). Under these




conditions, SDM should be kept at least three feet below the surface. This should apply to both pavements and embankment slopes. Wet-dry tests are conducted to simulate shrinkage forces in cement-modified or soil-cement specimens. Wet-dry cycle tests were conducted on the three different recipes of SDM. Tests were conducted according to ASTM D-559. All of the samples with the exception of one (8% PC @ 90% Modified Proctor) collapsed before experiencing 12 wet-dry cycles. Volume changes were in the range of 10% to 48% of the original volume. Therefore, SDM should be protected against frequent wet-dry cycles with placement of proper coverage for roadway applications, or low permeability layers in general fill applications. Furthermore, if SDM is compacted at moisture contents below the shrinkage limit, the potential for the development of tensile cracks and a consequent loss in strength could be minimized.

4 Conclusions and recommendations

Beneficial use of stabilized dredge material (SDM) has been shown to be a practical option for the management of navigational dredged material in the Port of NY and NJ. The laboratory study described in this paper evaluated the geotechnical properties of stabilized dredge material (SDM) from the NY/NJ Harbor for potential high volume applications in roadway construction. The study was the first phase of a two-phase pilot project sponsored by the New Jersey Department of Transportation for finding alternative methods for beneficial use of the 2-4 million yd3 of contaminated sediments dredged annually, to maintain the maritime the transportation system that serves the Port. The results of the laboratory study demonstrate that stabilized dredge material (SDM) satisfies most of the geotechnical criteria for construction of fills and embankments, except those for durability: freeze-thaw and wet-dry cycles. Proper coverage and protection need to be provided for SDM fills to address the durability problem, similar to those addressed in the construction fills with cohesive soils. A summary of the test results as described in the paper is as follows:

1. The raw dredged material from the NY/NJ Harbor is mostly silt with low percentages of fine sand and clay. The dredged material samples tested in this study consisted of 66% silt, 14% clay and 16% fine and medium sand (12.1% fine, 3.9% medium). The percentage of clay size particles was higher for those stabilized samples that had been mixed with fly ash. This is due to the fine nature of fly ash particles. The organic content of the raw dredge material was determined to be around 8% according to ASTM D2974. Based on the Atterberg Limits, all the samples tested are below the A-line and to the right of the LL=50 line on the Plasticity Chart, classifying SDM as Elastic Silt (MH).

2. The maximum dry densities for the different mixes tested ranged from 76.6 pcf to 78.8 pcf (1.23 to 1.26 Mg/m3), and optimum moisture contents ranged from 26% to 31.5%. A slight reduction in maximum




dry density was observed when the percentage of cement and the curing time were increased prior to compaction of the material.

3. The compression index (Cc) values for SDM ranged from 0.22 to 0.9. and did not exceed 0.5 for any of the samples, once the samples had been compacted to 81% of their maximum dry density. Therefore, a Pc of 2 tsf (191.52 kPa) or more should be expected. The compression ratio (CR =Cc/1+e0) varied from 0.085 to 0.24. It can be concluded that SDM embankments up to 50 feet (15 m) in height could be constructed with only minimal settlement within the SDM fill.

4. The hydraulic conductivity ( permeability) results ranged from 1.25x10-

6 cm/sec to 4.3x10 –7 cm/sec. SDM could, therefore, be considered for use as a low permeability layer in landfill cap applications. In roadway applications, however, building on SDM fills would be similar to construction on compacted fine-grained sub-grades, such as those in arid regions like Arizona, Texas, etc. Proper coverage must be provided using appropriate base or sub-base materials.

5. The addition of admixtures produced no significant change or trend in the frictional properties of SDM. In comparison to soil-cement and cement-modified soils, SDM has lower friction angle and much lower cohesion intercept under triaxial shear conditions mainly due to the sequence of sample preparation used in this study which followed the field operations. Temperature had a major effect on the curing process of SDM at temperatures below 40°F; it is recommended that SDM be placed during warm seasons (e.g., April through October in New Jersey).

6. The resilient modulus values for all of the samples tested compared well with three sub-grade soils that are currently under New Jersey roadways.

7. The strain or swell percentage was not significant for any of the samples tested. The strain values ranged from 0.1% to 1.2%, with an average of 0.6%. This magnitude of volume change is considered to be low and, therefore, not detrimental to adjacent structures.

8. The results from durability tests indicate that SDM is susceptible to frost action (several times more susceptible than natural clay) and should be placed below frost line. Based on the wet-dry tests, proper soil cover needs to be provided at all times to minimize strength loss due to erosion. Compacting SDM at moisture contents below the shrinkage limit would minimize the potential for tensile cracks and thereby minimize any further strength loss in the material.

References

[1] Balmar, G. G. (1958). “Shear strength and elastic properties of soil-cement mixtures under triaxial loading.” PCA, Bulletin, D32.

[2] Clough, G. W., Sitar, N., Bachus, R. C., and Rad, N. (1981). “Cemented sand under static loading.” JGED, ASCE, Vol. 107, No. GT6, 799-817.




[3] Cotton, M. D. (1962). “ Soil-cement technology – a resume.” Journal of the PCA Research and Development Laboratories, Vol.4, No.1, 13-21.

[4] Kezdi, A. (1970). “Handbook of soil mechanic,” VEB Verlag fur Bawsen, Berlin.

[5] Maher, A., Bennert, T., Papp, W., (2000). “Resilient properties of New Jersey sub-grade materials,” Report submitted to NJDOT.

[6] Van Riessen, G. J., and Hansen, K. (1992). “Cement-stabilized soil for coal retaining berms.” ASCE, Geotechnical STP 30, 981-992.




The impact of soil quality on cocoa yield in Nigeria

O. A. Amusan1 & F. O. Amusan2

1University of Bonn, Institute of Agriculture, Water Engineering and Land Improvement, Bonn, Germany 2Alcon Labs, UK

Abstract

The Southwestern region of Nigeria is the largest administrative region in Nigeria; it occupies about 30 percent of Nigeria and has an estimated population of about 40 million. Over 50 percent of the area is categorised as rural areas of which about 90 percent of the people depend on farming for their livelihood. Over 90 percent of Nigerian cash crop cocoa is produced in the cocoa belt of the Southwestern region, but both cash and food crops have consistently declined in the last few years. This phenomenon constitutes a threat to food security and calls for efforts to explain the downward trend and make recommendations for improvement. The objectives of this study were to evaluate the soils of some areas in Southwestern Nigeria for cocoa on the one hand, and identify factors affecting cocoa yield on the other. A novel technique that combines soil survey with socio-economic analyses was adopted. Socio-economic surveys covered resource quality and constraints to agricultural production, whereas soil sampling and analyses were carried out to assess the contribution of soil to yield. Three locations having similar agro-ecological features were selected, namely Ibadan, Ife and Akure. Cocoa farmers were randomly selected and interviewed on their farms using standardised questionnaires to elicit information on factors affecting crop yield. Relationships between cocoa yield and variables presumed to influence yield were determined using linear multiple regressions. Soil organic C, Age of farm soil, and ECEC were identified as the major constraints to yield. Other variables are related to biophysical and management factors. It is recommended that emphasis should be placed on soil management techniques that conserve organic matter and enhance the nutrient and water holding capacity of the soils. Policies that would enhance sustainability of agricultural land use and crop marketing are also required. Keywords: cocoa yield, food security, land use, linear multiple regressions, soil organic C, soil quality.




doi:10.2495/GEO060441

1 Background and aim of the study

The main soil factor affecting cocoa in Nigerian conditions is the clay content. This is important chiefly because a relatively high amount of clay in the soil gives it a fairly high capacity for retaining moisture, which remains available to the trees during the dry season. Soils in Southwestern Nigeria exhibit varying degrees of fertility. They are derived from hard crystalline rocks of the “Basement Complex” (gneisses, granites, schists, quartzites, and amphibolites), which have been eroded in varying degrees in different parts of the main cocoa belt. Fertile soils are found next to areas with poor soils and contiguous tracts of fertile cocoa soil are rare. Pockets of fertile soils can be found around Ife, Ibadan and Akure (Figure 1).

Figure 1: Map of the cocoa growing regions of Nigeria.

Cocoa grows best in soils with the following properties (Rehm and Epsig [1]):

1. A dark, grey brown, crumby humid surface layer 3.5–7.5 cm in thickness consisting of medium texture and free from coarse sand, grit or small gravel.

2. About 30cm of medium-textured earth, brown or reddish-brown in colour.

3. A layer of friable red clay 60–90 cm in thickness 4. A layer of friable clay mottled whitish, orange, red and pale yellow, and

60–90 cm thick which grades into 5. Rotten, disintegrated rock, preferably of a granite nature or derived

from dark-coloured igneous or volcanic material.

Cocoa-growing year, distinguished from cocoa-marketing year begins when moist air (the Equatorial Maritime air mass) sweeps in from the coast at the end




of April or the beginning of May. From May to July and in September and October, rainfall averages at least 150mm per month in nearly all cocoa-growing areas. The “little dry season” in August when the climate remains humid but rainfall is much less is characteristic of Southwestern Nigeria. If it is unduly long or dry, the cocoa crop may suffer severely. The crop may also suffer severely if the rains between July and October when the pods are ripening are unduly heavy. The northern limit of cocoa belt is marked partly by the line where rainfall approaches the lower limit of 1100mm/ year and partly, in the area of Ondo provinces where the rainfall is well above that limit. In this zone the influence of the dry Harmattan wind in the months from December to March is very strong and cocoa, which cannot stand long drought, seems unable to survive. Temperatures are lowest and relative humidity highest in the rainy months (the growing season) and plant growth is then very rapid. In the dry season, temperatures are high during the day and relative humidity falls very low during the afternoons. Conditions are then unfavourable to growth. The period during which even some of the hardiest crops cannot be successfully planted without constant watering is rather over three months in most of the main cocoa belt. Even at the highest elevations in the cocoa-growing areas temperature never fall so low as to be adverse to growth. Farmers therefore face replanting problems in old cocoa farms and high seedling failure in new plantings in marginal soils. This situation is affecting the future planting plans of cocoa farmers in Oyo, Ogun and Ondo States of Southwestern Nigeria. Although contiguous land masses suitable for cocoa growing exist outside the main cocoa belt (particularly in Ikom and Baissa of South-eastern Nigeria), the exceptional productivity arising from the fertile soil conditions has been eroded by the high incidence of Phytophthora (black pod) disease due to high rainfall. The disease is more prevalent in the high rainfall areas of the Southeastern state of Cross River than in the relatively drier cocoa areas of Southwestern Nigeria. This study is aimed at assessing the impact of soil quality on Cocoa yield in the study areas in Southwestern Nigeria. The specific objectives include:

1. To evaluate the soils of some areas in South-western Nigeria for cocoa 2. To identify factors affecting cocoa yield 3. To determine the biophysical and socio-economic constraints to quality

cocoa production

2 Methodology

2.1 Data set

The analysis presented in this paper is based on primary data, which were collected by surveys of farm households, extensionists, farm mangers, etc. The Cocoa farm household samples (classified by farm size) were selected from three locations having similar agro-ecological features, namely Ibadan, Ife and Akure. The primary data from the socio-economic survey were used to elicit information on farm–farmers’ characteristics vis-à-vis constraints to cocoa. Soil sampling and analysis were carried out to examine the role of soil properties in crop yield.




Socio-economic and soil data were thereafter integrated to study their influence on crop yield.

2.2 Socioeconomic survey

A survey was carried out among cocoa growers in South-western Nigeria in the year 2002. Three locations having similar agro-ecological features were selected, namely Ibadan, Ife and Akure. Thirty cocoa farmers were randomly selected and interviewed on their farms using standardised questionnaires. Farmers interviewed in Ibadan belonged to three different groups of cooperative societies or farmers’ organisation. The farmers interviewed at Ife and Akure belonged to the respective cooperative multipurpose unions (CPMU) in these towns. These cooperatives are profit -making organisations, however, they offer more than just marketing services for their members. Access to the farmers and their farmlands was obtained under the auspices of the Association of Nigerian Cooperative Exporters Limited (ANCE) and ongoing Project for Improvement of Cocoa Marketing and Trade in Nigeria (ICMT) sponsored by the International Cocoa Organisation (ICCO). While cocoa farmers in Ife and Akure are fully integrated into the ICMT project, the cocoa farmers in Ibadan are non-participants. The selection along the line of participation or non-participation on the project moreover enabled the evaluation of the project’s overall influence on cocoa yield. The interview comprised of qualitative and quantitative components, covering aspects of agronomic activities, resource quality and availability, problems perceived and objectives set both at individual and cooperative levels. Additional information was collected in key-persons interviews from cocoa exporters, cocoa processors, cocoa researchers and representatives of governmental and non-governmental institutions. Official statistics and other secondary data served as background information.

2.3 Soil sampling and analyses

Biophysical data was obtained by analysis of soil samples taken from cocoa farms. Soils were analysed for chemical analyses. The soils were analysed for basic cations (determined in 1N NH4OAc), total N (Kjedahl method), available P (Bray P method) organic C (Walkey-Black wet oxidation method) and pH (0.1M CaCl2).


3.1 Soil properties, land use history and crop yield

The topography of the land at all study farms ranges from flat to gentle slope (2-8%). Soil analyses (Tables 1and 2) showed that the soils are fairly high in total N (mean = 0.20%), organic C (mean = 1.89%), available P (mean = 3.19 ppm), but low in Na. The pH is close to neutral (mean 6.5). In terms of variability, pH is the least variable (cv = 7%) while available P is the most variable property (cv = 69%). Statistics of soil properties at the sampled




locations are presented in Tables 5 and 6. Soil pH is the least variable across locations (CV=3-9%). Highest CV (100%) among Ibadan soils occur for EA, while the highest CV of 119% occur for P for Ife soils while Ca and K manifests the most variable soil property amongst Akure soils (C=55%). This suggests that cations are the most variable properties among Akure soils. Test of mean differences show that none of the properties are significant across locations (p>0.05). This is because the soils have developed from similar soil parent materials derived from Precambrian rocks.

Table 1: General Statistics of soil properties (N=16). (EA= Exchangeable Acidity, ECEC= Effective cation exchange capacity.)

Ca (cmol /kg)

Mg (cmol/kg)

Na (cmol /kg)

K (cmol/kg)

EA a (cmol /kg)

ECEC a

(cmol /kg)

Total N

(%)

Availabl-e P

(ppm)

Orga-nic C (%)

pH

Mean 4.34 1.11 0.07 0.38 0.02 5.90 0.20 3.19 1.89 6.46 Std 2.08 0.48 0.02 0.21 0.02 2.55 0.05 2.19 0.61 0.44 Cv (%) 48 43 29 25 100 43 27 69 32 7 min 1.49 0.30 0.04 0.06 0.00 2.15 0.14 0.99 1.16 5.70 max 7.51 1.22 0.10 0.76 0.07 9.77 0.33 7.05 3.33 7.30

3.2 Relationship between land use history and crop yield

Yields obtained differ depending on the variety of cocoa planted, age of the trees and farm maintenance. Higher yield was observed on farms planted predominantly with Amazon cocoa in comparison with those planted with Amelonado. Older farms have lesser yield than younger farms and farms are better maintained in regards to cultural operations were more productive. Table 18 presents the yield recorded on the farms during the 2001/02 cocoa seasons.

Table 2: Location statistics of soil properties (N=16).

Location No of samples

Statistics Ca Mg Na K EA ECEC N P Org.C pH

Mean 4.46 1.09 0.07 0.33 0.02 5.96 0.19 3.48 1.84 6.45

Std. dev 2.06 0.44 0.02 0.23 0.02 2.39 0.05 2.53 0.57 0.49

Ibadan 8

CV (%) 46 40 29 70 100 40 26 73 31 8

Mean 3.89 0.97 0.06 0.42 0.02 5.36 0.21 5.5 2.11 6.56

Std. dev 2.37 0.47 0.01 0.12 0.01 2.74 0.08 6.56 0.88 0.57

Ife 4

CV (%) 61 48 17 29 50 51 38 119 42 9

Mean 4.73 1.31 0.08 0.44 0.02 6.58 0.2 2.33 1.79 6.38

Std. dev 2.61 0.64 0.02 0.24 0.01 3.44 0.04 0.93 0.45 0.22

Akure 4

CV (%) 55 49 25 55 50 52 20 40 25 3




Table 3:

Location No of farmers Mean Std. Deviation CV (%) Ibadan 10 245 86 35 Ife 10 347 166 48 Akure 10 434 189 44

Table 3 shows that average yield per hectare is highest in Akure and lowest in Ibadan. However Ife has the highest variability in yield. Mean yield among the sampled locations and coefficient of variation may largely reflect variability in management practices. Test of mean differences in yield at the locations is presented in Table 4.

Table 4:

Location (1) Location (2) Difference in mean yield (Location 1-Location2) in

kg/ha

Significant Probability

Ife Ibadan 102 0.15 Akure Ibadan 189 0.01** Akure Ife 87 0.22

Table 4 shows that yield differences in Ibadan and Akure are highly significant (p<0.01), while yield differences at other pairs of locations are not significant (p>0.05). Ife and Akure farmers participate in the on-going ICMT project of the International Cocoa Organisation (ICCO) and therefore benefit from the expertise knowledge received at the farm management training sessions. Statistical relationship between land use history and crop yield were studied in terms of recorded yield, and three other variables: age of farm, proportion of dormant trees, and proportion of trees replaced. Figure 2 shows that yield of cocoa generally increases with age of farm up to 35 years, and then declines. Age of trees explains 29% of variability in cocoa yield. This phenomenon may be related to the physiology of the crop, crop varieties and the fact that the input management strategy in terms of fertilizer application is generally low. Yield decreases with the proportion of dormant trees on farmers’ field (Figure 3). Highest yields (650-750 kg/ha) were obtained in fields with no dormant trees. As dormancy is a function of several factors such as age, disease and pest attack etc, the need for proper control of diseases and succession planting cannot be over-emphasised. Dormant trees explain 81% of variability in cocoa yield. Figure 4 shows that yield of cocoa is strongly related to the proportion of replaced trees. This management option is particularly appealing to farmers given the fact that the application of farm inputs like fertilizers and pesticides is generally low. Proportion of replaced trees explains 89% of variability in cocoa yield. Figure 5 shows that the tendency of cocoa yield to be low at extreme values of soil organic C (that is, organic C below 1.5% and above 2.5%). The same observation holds for total N where yields are observed to be low for N values




Yield statistics (kg/ha) on surveyed farms in the 2001/ 02 season.

Least Significant Difference (LSD) test of yields.

below 0.2% and above 0.25% (Figure 6). This result suggests that accumulation of organic materials in the soil is not necessarily favourable to the growth of cocoa. Figure 7 also suggests that pH below 6.0 and above 6.5 adversely affects yield. Yield is maximum at about a pH of 6.3 and attempt should be made to maintain soil pH around this level.

Figure 2: Relationship between yield and age of farm.

Figure 3: Relationship between yield and proportion of dormant trees.

Figure 4: Relationship between yield and proportion of replaced trees.

y = 9 . 9 6 x 2 - 1 3 8 . 3 6 x + 6 3 2 . 1R 2 = 0 . 8 1

0

2 0 0

4 0 0

6 0 0

8 0 0

0 2 4 6 8 1 0 1 2

d o r m a n t t r e e s ( % )

yiel

d (k

g/ha

)

y = -0 .0 7 x3 + 4 .7 5 x2 - 6 1 .1 0 x - 7 5 .0 2R 2 = 0 .2 9

0

2 0 0

4 0 0

6 0 0

8 0 0

2 0 2 5 3 0 3 5 4 0 4 5 5 0

A g e (y e a rs )

yiel

d (k

g/ha

)

y = 0 .0 7 x 2 + 7 .0 0 x + 1 7 1 .2 6R 2 = 0 .8 9

0

2 0 0

4 0 0

6 0 0

8 0 0

0 1 0 2 0 3 0 4 0 5 0 6 0

r e p la c e d t r e e s (% )

yiel

d (k

g/ha

)




Figure 5: Relationship between yield and soil organic carbon.

Figure 6: Relationship between yield and soil total nitrogen source.

Figure 7: Relationship between yield and soil acidity.

Thus proper monitoring of acidity in the soils is crucial. Relationships between some soil variables are presented in Figures 8 and 9. Figure 8 shows the tendency of pH to increase with the organic C content of the soil. This has the potential to decrease yield (see Figure 5). The strong relationship between

y = -1 2 7 .1 5 x 2 + 4 7 2 .8 8 x - 2 3 .9 3R 2 = 0 .1 0

0

2 0 0

4 0 0

6 0 0

8 0 0

0 .5 1 1 .5 2 2 .5 3 3 .5

o r g a n ic C (% )

yiel

d (k

g/ha

)

y = 4 0 0 .2 7 x3 - 8 2 1 5 .8 x2 + 5 5 8 3 7 x - 1 2 5 3 0 8R 2 = 0 .2 2

0

2 0 0

4 0 0

6 0 0

8 0 0

5 5 .5 6 6 .5 7 7 .5 8

p H

yiel

d (k

g/ha

)

y = -24016x2 + 10421x - 687 .66R 2 = 0 .13

0

200

400

600

800

0 .1 0 .15 0 .2 0 .25 0 .3 0 .35

to ta l N (% )

yiel

d (k

g/ha

)




organic C and total N (figure 9) suggests that management of soil organic C is crucial to maintaining adequate levels of N in the soil.

Figure 8: Relationship between pH and organic carbon.

Figure 9: Relationship between total N and Organic C.

3.3 Effects of biophysical and management variables on crop yield

Linear multiple regression was used to relate crop yield to biophysical and socio-economic data. Owing to high level of multicollinearity, independent variables were reduced to six. Among the variables in the model, two (Organic C and Age of farm) are negatively related to cocoa yield, whereas other variables are positively related to cocoa yield. However, soil variables are not significant to the model (p>0.05), whereas three management variables (plant density, proportion of dormant plants replaced and crop variety) are significant (p<0.1). All the variables explain 97% of the variability of yield and the model can be used to predict yield at 99% confidence level. Figure 10 further shows the relative importance of the variables as measured by the standardized β coefficients. Proportion of dormant plants replaced is the most significant management variable affecting yield, followed by crop variety (that is, F3 Amazon).

y = 0.31x3 - 1.84x2 + 3 .85x + 3.69R 2 = 0.44

5

6

7

8

0.5 1 1.5 2 2.5 3 3.5

organic C (%)

pH

y = 0 .0 8 x + 0 .0 4R 2 = 0 .9 3

0

0 .1

0 .2

0 .3

0 .4

0 .5 1 1 .5 2 2 .5 3 3 .5

o rg a n ic C (% )

tota

l N (%

)




Figure 10: Relative importance of variables in the multiple regression model.

4 Conclusions and recommendation

Seventy percent of the cocoa farmers in the study areas are more than 60 years of age and over 50% of the farms are older than 45 years. More than half of the farmers are illiterates who keep no management records. Furthermore, yields on over 50% percent of the farms are below 300 kg/ha. These figures pointedly show a relationship between old age of farmers / farms and poor yield on the one hand and poor or no education on the other. The world market is increasingly flooded with average or low quality cocoa from new hybrid varieties. As long as the market is flooded with relatively large amounts of cocoa, buyers are not forced to pay a premium for cocoa of whatever origin. To the contrary, they can enforce large discounts for lower quality. They are able to select and mix according to their needs, all at relatively low cost. On the other hand, a small producer experiencing severe problems to expand production, or quantity, will probably find it more attractive to concentrate on improving quality. Once the market tightens up, the potential quality premium can even be a valuable asset and provide a competitive advantage. Achieving this improvement in quality will however depend, in the first place, on the physical potential of the product. Farmers who apply chemical inputs, must consider the aspect of profit making and quality of cocoa due to chemical residues in the product, coupled with the problem of environmental pollution through chemicals. Farmers should be encouraged to practise conservation and rehabilitation agriculture. This involves all steps which negate the processes of degeneration on cocoa farms, for example through consequent replacement of dormant trees, control of tree population for effective ground cover, properly managed canopy, integrated management systems that utilise resistant cultivars and application practices that emphasise proper timing and effective use of fertilizers at lower application rates. This would help not only to improve yield but also has the advantages of profitability, product quality and environmental protection. Given the situation in the international market in and the natural limitations of the cocoa farmers in Nigeria, the development of a market niche for quality was suggested. Further research will need to find out whether new breed varieties have the physical potential to achieve the said improvement in quality. The traditional Amelonado

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

Organic C Age offarm

Plantdensity

ProportionReplaced

CropVariety

ECEC




s t a

n d

a r

d i z

e d

b

e t a

cocoa has been proved to be superior in every physical aspect (shell content, nib yield, fat yield and moisture content) when compared to beans derived from other varieties. The Amazon variety which is now very common in Nigeria, in addition to the suitable cocoa flavour characteristics, are more resistant to diseases in the environment and give better yields.

References

[1] Rehm, S. & Epsig, G., The Cultivated Plants of the Tropics and Subtropics, Verlag - Josef Magraf: Germany, 1991

[2] Amusan, O.A., Amusan, F.O., Braimoh, K.A. & Oguntunde, P.G., Quality Management Practices in Cocoa Production in South - Western Nigeria: Deutscher Tropentag 2005, ID–29 Online. www.tropentag.de

[3] Adegbola, M.O.K., Rehabilitating cocoa swollen shoot virus infected Cocoa Farms- the Offa-Igbo experience. Proc. 5th Int. cocoa Res. Conf. Ibadan, Nigeria 1975: 182-8, 1977

[4] Adenikinju, S.A., E.B. Esan and A.A. Adeyemi, Nursery Techniques, Propagation and Management of Cacao, Kola, Coffee, Cashew and Tea Progress in Tree Crop research (Second Edition) A Commemorative Book to mark the 25th Anniversary of CRIN, 1989

[5] Ajayi and Okoruwa Managing Uncertainties and Risks in Cocoa Production and Marketing in Nigeria, 1996

[6] Anon., Technical Consultation on Agricultural Methods and Techniques for Cocoa. Report of a meeting held in Itabuna, Bahia, Brazil. Cocoa Producers Alliance: Lagos, 1977

[7] Are, L., Methods of rehabilitating cocoa farms in Nigeria. Proc. 2nd Int. Cocoa Res. Conf. Salvador, Brazil 1967: 383-7, 1969a

[8] Atanda, O.A. T. Quarcoo and M.O. Osundolire, Nursery techniques for tree crops in Nigeria. In Progress in tree Crop Research in Nigeria. A commemorative book, CRIN, 96-104, 1977

[9] Baiden, E. and A. Asare-Nyako (1986) Report of the Cacao Research Institute of Ghana 1982/83–1983/84, 205, 1986

[10] Odegbaro, O.A., Prospects of rehabilitating Amelonado cocoa with improved cocoa varieties in Nigeria without complete planting. Proc. 5th Int. Cocoa Res. Conf. Ibadan, Nigeria 1975: 259-64, 1977

[11] Cocoa/Chocolate and Confectionary Alliance. Cocoa beans: Chocolate Manufacturers quality requirements. The Cocoa, Chocolate and Confectionary Alliance, 11 Green Street, London, 3rd Ed. Jan. 1984. 19pp, 1984

[12] Prior, C., Cocoa replanting practices and the possible dangers from root-rots. Cocoa Industry Board of Papua New Guinea Newsletter 1, 2, 1981




Section 12 Vulnerability studies

Influence of ground water extraction in the seismic hazard of Mexico City

J. Avilés1, L. E. Pérez-Rocha2 & H. R. Aguilar3 1Instituto Mexicano de Tecnología del Agua, Mexico 2Instituto de Investigaciones Electricas, Mexico 3Centro de Investigación Sísmica, FJBS, Mexico

Abstract

The influence of ground water extraction in the seismic hazard of Mexico City is examined. Available information on settlements of the land surface is used to evaluate the subsidence effects on the predominant ground period. Microzoning maps, as a function of this relevant site parameter, are constructed for the present and future geotechnical conditions. Based on these maps, site-specific design spectra throughout the whole city are determined by applying seismic code provisions. It is found that the regional subsidence will generally be beneficial for structures with fundamental period longer that the current period of the site, but detrimental if the structure period is shorter than the site period. Keywords: design spectra, ground water extraction, microzoning map, regional subsidence, seismic hazard.

1 Introduction

Extensive ground-water exploitation has been identified as the main cause of land subsidence in Mexico City. The phenomenon has caused large settlements in the past, up to 10 m at some areas of the city. Subsidence of the city was studied for the first time by Carrillo [1], showing a clear dependence with the amount and rate of water extraction from artesian wells. Since then, many efforts have been made to understand the phenomenon and mitigate its consequences. A review of these efforts can be found in Ovando-Shelley et al. [2]. These authors have studied the consolidation process in the central part of the city using the well-known Terzaghi’s consolidation theory. They showed that exploitation of the aquifers under the lacustrine clays is reflected not only on the gradual




doi:10.2495/GEO060451

reduction of their thickness, but also on the modification of their mechanical properties.

In Mexico City, ground motion amplification during earthquakes is the most important factor associated to the subsoil characteristics. The sedimentary basin measures approximately 30×70 km and the thickness of compressible clay deposits may exceed 70 m. With shear wave velocities ranging from around 50 to 100 m/s, the predominant ground period can be as long as 5 s. In view of these geotechnical conditions, the resonant spectral acceleration for sites with dominant period around 2 s may reach up to one g when intense subduction earthquakes occur as far as 300 km. This was observed during the great 1985 Michoacan earthquake. Any variation in the present geotechnical conditions, however, will result in a change in the seismic hazard in the future.

In this work, an empirical method to predict the effects of regional subsidence on the seismic hazard in Mexico City for the coming years is presented. Using available information regarding precise leveling of benchmarks and geotechnical soil profiles, Aguilar et al. [3] have found correlations of the subsidence rate and the thickness of sediments with the predominant ground period. With the use of these data in a year-by-year incremental procedure, the evolution of the microzoning map and the corresponding design spectra specified by the building code, both in terms of the site period, is predicted for an exposure period of 50 years. This would allow the designer to evaluate the effects of regional subsidence during the life span of a structure.

2 Available information

The subsidence of Mexico City is studied here by using an empirical method based on extrapolating available data for predicting future trends. With spectral amplification functions for about 100 instrumented soft sites at Mexico City, complemented with around 500 microtremor measurements, predominant ground periods sT were computed for a rectangular grid of 80x80 points covering most part of the city, using the interpolation technique devised by Pérez-Rocha et al [4]. This resulted in the microzoning map displayed in fig. 1, which will be referred hereafter as the 2000 version. The isoperiod curves for =sT 0.5 and 1 s roughly mark the separations between both the firm and transition zones as well as the transition and soft zones, respectively.

For about 360 sites, the thickness of sediments sH is also known. Fig. 2 left-hand side shows the correlation of this parameter with the site period obtained by Aguilar et al. [3]. This relationship is expressed by

[ ] sTsTmH sss 1for 5.0)(31)( 21 ≥−= (1)

Note that, for a given site, the shear wave velocity sV is defined indirectly by eqn. (1) since, according to the one-dimensional wave propagation theory, the expression sss VHT 4= applies.




Figure 1: Curves of predominant ground period (s) in Mexico City.

For monitoring subsidence in Mexico City, a network of more than 2200

benchmarks distributed along the city has been installed. This network was releveled in 1983 for the last time. By collecting elevation measurements at land surface during the time period 1983-1998, Aguilar et al. [3] estimated the amount of subsidence at soft sites. Then, the mean subsidence rate was correlated with the site period by the expression

[ ] sTsTyearcm sss 1for )(2)( 9.1 ≥=υ (2)

Fig. 2 right-hand side illustrates that the subsidence rate declines as the site period shortens which in turn occurs when the depth of sediments shrinks (see Fig. 2 left-hand side). This is consistent with the consolidation process in the lakebed zone of the city. It should be mentioned that the observed subsidence include not only the effect of ground-water extraction, but also the consolidation from the own weight of buildings.




Figure 2: Left-hand side: thickness of compressible clay deposits in Mexico

City’s lakebed zone. Right-hand side: subsidence rate in Mexico City’s lakebed zone.

3 Predicting procedure

Pumping operations in Mexico City have varied over the years, making difficult the prediction of subsidence effects. Nevertheless, assuming the present pumping conditions will be maintained in the near future, useful estimations may be made as to changes expected in the existing microzoning map and the corresponding design spectra. With the information that has been presented, a year-by-year incremental procedure can be implemented for predicting the evolution of site period, as follows: ● For a given site, the initial dominant period 1

sT is taken from fig. 1.

● By application of eqn. (1), calculate the initial thickness of sediments 1sH

corresponding to 1sT .

Then, ● Calculate the subsidence rate i

sυ by use of eqn. (2).

● Calculate the amount of compaction as tH is ∆υ=∆ , which for =∆t 1 year

represents the annual subsidence. ● Calculate the change in thickness of sediments as HHH i

sis ∆−=+1 .

● From eqn. (1), calculate the change in site period as 5.0)31( 211 += ++ is

is HT .

● Replace i by i+1 and repeat the process for successive time steps until the target exposure period is reached.

Let us now illustrate the application of the proposed procedure for estimating

the changes expected in the microzoning map of the city for the next 50 years. Fig. 3 exhibits the modified isoperiod curves by the effects of regional subsidence. The main variations are observed at the airport, Xochimilco and




Tlahuac areas, where long site periods undergo drastic reductions. In contrast, the border between the transition and soft zones experiences little modification. The evolution of site period can be appreciated in fig. 4 for sites with current values of =sT 1 to 5 s.

Figure 3: Modified isoperiod curves by the effects of regional subsidence for a

time period running from 2000 to 2050.

4 Evolution of seismic hazard

The Valley of Mexico is affected by earthquakes having different causes. They have been divided in four groups, namely: 1) local earthquakes, 2) continental plate earthquakes, 3) intermediate depth earthquakes and 4) subduction earthquakes. For the contribution of all these events, the expected Fourier amplitude spectrum at firm ground for a 125-year return period has been estimated with the use of a standard probabilistic approach. This result is shown in fig. 5.




Figure 4: Evolution of site period in Mexico City’s lakebed zone for a time

period running from the 2000 to 2050 year.

The site effects due to local soil conditions have been explicitly considered in the Mexican Building Code [5] after the great 1985 Michoacan earthquake. The resulting site-specific design spectra are defined by the following expressions:

aeae

a TTcTTgS <+

= if,4

31 , (3)

beaa TTT,cgS ≤≤= if , (4)

be

r

e

ba TT,c

TT

gS >

= if . (5)

where aS is the spectral acceleration expressed as a fraction of gravity, =r 1 for the lakebed zone of Mexico City, aT and bT are the lower and upper periods of the flat part of the spectrum, respectively, and c is the seismic coefficient. The flat part of the spectrum is specified by the limiting periods

sTsTmaxT ssa 1for),64.0,35.0( >= , (6)

sb T.T 21= . (7) while the seismic coefficient is specified as

244

s

s

TT

c+

= (8)




These spectra are intended to cover not only the resonant peak response associated with the first soil period, but that with the second period as well.

Figure 5: Fourier amplitude spectrum at firm ground in Mexico City for a 125-

year return period, including the effect of earthquakes from different origins.

To illustrate the evolution of code design spectra, a characteristic site with

current value of =sT 3.5 s was selected. The shortened site period resulting for an exposure period of 50 years is seen from fig. 4. A comparison between the 2000 and 2050 design spectra is made in fig. 6. Also shown are the corresponding response spectra obtained by using as input motion the Fourier amplitude spectrum of fig. 5. The structural response was computed by applying the random vibration theory [6] to a simplified model consisting of a one-story structure placed on a uniform stratum under vertically incident shear waves, corrected empirically to account for the bedrock flexibility [7]. Although the representation is not perfect, the design spectrum intends to reproduce the general trends observed in the response spectrum, for both present and future geotechnical conditions. It is seen that regional subsidence will have either favorable or unfavorable effects on the seismic safety of existing buildings, depending primarily on the period ratio of structure and site.

The peak structural response can be conveniently represented by means of spectral contours of acceleration expressed in terms of the structure and site periods. In this way, the maximum response of any building at any location may be readily estimated. We computed the 2000 and 2050 spectral contours, the comparison of which is shown in fig. 7. They are constructed from site-specific design spectra specified by the building code throughout the whole city. Each site spectrum can be recovered by drawing a section along the site period of interest.




Figure 6: Comparison between the 2000 (solid line) and 2050 (dashed line)

spectra for a site with presently dominant period Ts = 3.5 s; design spectra (thick line) versus response spectra (thin line).

Figure 7: Comparison between the 2000 (thick line) and 2050 (thin line)

spectral contours of acceleration expressed as a fraction of gravity.

Looking at the current site spectra, notice that the plateau width is an increasing function of the site period. This is to cover the influence not only of the first mode of vibration of the soil, but of the second mode as well (see fig. 6).




In fact, the peak response associated to the latter may be as large as that associated to the former for long site periods, say, >sT 3.5 s. Also, it is apparent that the plateau height increases with the site period for <sT 2 s, but decreases for >sT 2 s. Thus, the most vulnerable buildings would be those with roughly 20 stories, assuming a fundamental period of 0.1 s per story. Nevertheless, this is true for the time being but not for the coming years.

The difference between the 2000 and 2050 spectral contours reflects the effects of regional subsidence. The 2050 spectral contours tend to rotate counter-clockwise and elongate with respect to those for the 2000 year. Note that sites with >sT 2 s will migrate in the future to the region in which spectral ordinates reach their peak values.

Figure 8: Subsidence impact on seismic base-shear coefficient: ratio of the

2050 to 2000 spectral contours given in fig. 7.

To have a whole scenario for the changes expected in the seismic hazard of the city, the ratio )(S)(S aa 20002050=α was computed. Fig. 8 shows a general view of the subsidence impact on base-shear coefficient. Both detrimental ( >α 1) and beneficial ( <α 1) effects are observed, depending on the period ratio of structure and site. For any structure and site configuration, the subsidence impact may be assessed directly from this figure, entering with the corresponding structure and site periods. It can be seen that, in general, regional subsidence will affect the structural seismic safety adversely for se TT < and positively for se TT > .




5 Conclusions

In this prospective study, the effects of regional subsidence on the seismic hazard for Mexico City were examined. For an exposure period of 50 years, changes expected in the existing microzoning map and the corresponding design spectra specified by the building code were assessed. In view of complexity for making an accurate prediction, only approximations were established about the magnitude of subsidence effects. After some numerical evaluations, the following main conclusions can be drawn: a) Long site periods will undergo large reductions (e.g., from =sT 5 to

≈ 3.25 s), whereas short site periods will undergo small reductions (e.g., from =sT 1 to ≈ 0.95 s).

b) Sites with dominant period >sT 2 s will migrate in the future to the region in which spectral ordinates reach their peak values.

c) In general, regional subsidence will affect the structural seismic safety adversely ( >α 1) for se TT < and positively ( <α 1) for se TT > . The assessment of subsidence effects was made in an environment of high

uncertainty, where data may change rapidly from time to time due to local pumping conditions. Although the results are location specific, the predicting procedure may be applicable to other geographic locations facing similar problems of land subsidence due to ground water extraction.

References

[1] Carrillo N., Influence of artesian wells in the sinking of Mexico City, Proc. 2nd Int. Conf. on Soil Mechanics and Foundation Engineering, Rotterdam, Holland, 1948.

[2] Ovando-Shelley E., Romo M. P., Contreras N. and Giralt A., Effects on soil properties of future settlements in downtown Mexico City due to ground water extraction, Geofísica Internacional, 42, pp. 185-204, 2003.

[3] Aguilar H. R., Galicia M., Pérez-Rocha L. E., Avilés J., Vieitez L. and Salazar M., Effect of regional subsidence on dynamic soil properties, Proc. 12th Panam. Conf. on Soil Mechanics and Geotechnical Engineering, Boston, USA, 2003.

[4] Pérez-Rocha L. E., Ordaz M. and Sánchez-Sesma F. J., Spatial interpolation of seismic data: the case of the Valley of Mexico, Proc. 10th Panam. Conf. on Soil Mechanics and Foundation Engineering, Guadalajara, Mexico, 1995.

[5] MBC, Complementary Technical Norms for Earthquake Resistant Design, Mexico Building Code, Federal District Government, 1987.

[6] Boore D. M. and Joyner W. B., A note on the use of random vibration theory to predict peak amplitudes of transient signals, Bulletin of the Seismological Society of America, 74, pp. 2035-2039, 1984.

[7] Avilés J. and Pérez-Rocha L. E., Site effects and soil-structure interaction in the Valley of Mexico, Soil Dynamics and Earthquake Engineering, 17, pp. 29-39, 1998.




Predicting favourable areas for landsliding through GIS modelling in Aparados da Serra (Brazil)

A. J. Strieder1, S. A. Buffon1, T. F. P. de Quadros2 & H. R. Oliveira3 1Lab. Modelagem Geológica e Ambiental, MODELAGE-UFRGS, Brazil 2Fund. Estadual de Proteção Ambiental, FEPAM – SEMA-RS, Brazil 3Transportadora Gasoduto Bolívia-Brasil S.A., TBG-Sul

Abstract

This paper presents the results of GIS modelling for predicting areas where natural landslides can occur in the Aparados da Serra region (Southern Brazil). The scarp is mainly developed upon basalts and dacites (Serra Geral Fm.) and sandstones (Botucatu Fm.) of the Paraná Basin. The 1000 m high scarp has been geomorphologically evolving since the break-up of the Gondwana Supercontinent and the opening of the South Atlantic Ocean. Geologic, geomorphologic and geotechnical mapping were conducted in order to acquire field data and to define a conceptual geomorphologic-geotechnical model for GIS data modelling. These mappings were aided with ASTER image processing and aerial photograph analysis. ASTER images do permit a high resolution and accurate DEM. The prediction of areas able to develop natural landslides along the scarp was based upon Factor of Safety (FS) algorithms. The GIS modelling results for Aparados da Serra region were grouped into four categories. The areas with the lowermost FS number were defined as highly susceptible to developing natural landslides. Field investigations upon some existing landslide structures in the region do confirm such predictions. However, large scale FS determination through GIS modelling in order to predict susceptible areas for landslides must be interpreted in a relative manner. FS modelling can, then, be applied to landslide susceptibility mapping in areas of poor historical records, since supported by adequate geological and geotechnical investigations. Keywords: landslides, susceptibility maps, GIS modelling, factor of safety.




doi:10.2495/GEO060461

1 Introduction

Landslides are a natural phenomenon which include a wide variety of material, such as debris, soils, rocks, organic matter, and constructions, moving down-slope [1]. Landslides describe many types of down-slope mass movement, ranging from rapidly moving catastrophic rock avalanches and debris flows, to slowly moving earth slides. Landslides concerned here relate to slow rotational and translational slides of debris and soils [2]. The gravity is the main driving force for these slides, but a number of factors can influence them. The increase in water content, for example, contributes to loading of the debris and soil materials in the slope. In the same way, the shear strength of the slope material is decreased while the pore water pressure is increased. This seems to be the second main driving force for landslides in tropical countries, such as Aparados da Serra region in Southern Brazil. Landslide susceptibility maps are produced by a number of methods by means of GIS. Quantitative methods may include inventory, slope angle, terrain aspect, geology, vegetation, land use maps overlaid through the named landslide susceptibility matrix technique [3]. Multivariate analysis of factors controlling landslides are also developed [4]. GIS development enhanced landslide susceptibility analysis in different ways (e.g. [5,6,7]). It seems that an inventory of historical data is a key factor in landslide susceptibility map construction. However, how can a landslide susceptibility map be assessed in areas displaying poor landslides records? This paper aims to present a landslide susceptibility map for Aparados da Serra area (Southern Brazil, fig. 1) based on a rational deterministic approach [8] through GIS modelling. The prediction of areas able to develop natural landslides along the Aparados da Serra scarp was based upon Factor of Safety (FS) algorithms. The factor of safety (FS) was determined for high resolution areas (15 m) according the discussion by Wu and Abdel-Latif [9]. The results were tested looking at some existing, shallow landslides at Aparados da Serra scarp.

2 Geological and geotechnical investigations: a landslide model for Aparados da Serra (Southern Brazil)

The Aparados da Serra region (Southern Brazil) shows a 1000 m high scarp developed since the break-up of the Gondwana Supercontinent and the opening of the South Atlantic Ocean. It mainly exposes Botucatu Fm. (sandstones) at the base, and basalts and dacites of Serra Geral Fm. up-ward. The base of scarps has thick Tertiary talus deposits, cemented by iron oxides and hydroxides. ASTER image processing through AsterDTM tool of ENVI software (RSI) do permit a high resolution (15 m) and accurate digital terrain model. The coordinate system was set to SAD69/96. ASTER and LANDSAT images processing, DTM, aerial photographs analysis and fieldworks were performed to




produce the geologic, geomorphologic and geotechnical maps. These data were all introduced in ArcGIS software (ESRI) for additional modeling.

Figure 1: Regional map locating Aparados da Serra region, Southern Brazil. Geological mapping could distinguish 29 volcanic flows composing the Serra Geral Fm. 15 of these flows are made of basic volcanic rocks (basalts), while 14 of them are acid volcanic rocks (dacites). The limit of basic and acid volcanic rocks is set at 800 m of altitude. The lower contact of Serra Geral volcanic rocks is close to 245 m of altitude. Geomorphologic mapping defined a step-platform structure for Aparados da Serra scarp (Fig. 2). Under these structural conditions, recent weathering and mass wasting processes accumulated debris and soils as triangular prisms of colluviums over the platforms. The soils are mainly composed of clays, and some of them show latteritic features. Geotechnical and geomorphologic mapping recognized a vertical differentiation for colluviums prisms. In the higher scarps (> 800 m of altitude), the step height is close to or greater than platform length (Fig. 2). On the other hand, in the middle and lower part of the Aparados da Serra scarp (< 800 m of altitude), the platform length is much greater than step height. It was also noticed that soils have low thickness (< 5 m) in the upper scarp, while they show up to 30-40 m in some places of the lower scarp. These differentiations explain why some block falls are reported in the upper scarp, and why some shallow, translational slides are reported in the lower scarp.




The shallow (1-3 m depth), translational slides in the Aparados da Serra region (Southern, Brazil) shows clear elliptical shapes. They wide vary from 8 to 25 m, and the length is greater than 25-30 m. Laboratory geomechanic essays of colluviums in the Aparados da Serra region were done by Silveira [10] (Table 1).

Figure 2: Geologic-geomorphologic structure of the Aparados da Serra region (Southern Brazil).

Table 1: Geomechanic properties of colluviums in the Aparados da Serra region (Southern Brazil).

Geomechanic parameters Mean Minimum Maximum ρw (water density) 9,806 9,806 9,806 ρs (natural soil density) 18 15,5 20,4 ρw/ρs ratio 0,544778 0,632645 0,480686 ρt (saturated soil density) 28,15 27,7 28,6 ρ (dry soil density) 11,7 10,5 12,9 Φ peak (peak shear strength) 29,5 28 31 Φ residual (residual shear strength) 16 10 18 β (slope) 20 2 70 tanΦpeak/tanβ ratio 1,554448 15,22617 0,218695 tanΦres/tanβ ratio 0,787826 5,049344 0,118261 Triaxial cohesion 18 12,5 25 Undeformed direct shear cohesion 27 25 28




3 Predicting favourable areas for landslide through FS

Predicting favourable areas for landslides through FS determination in GIS environment requires an evaluation on existing landslide dimensions. In this way, the reported landslide in the Aparados da Serra region covers 1-2 pixels in wide and 3-5 pixels in length. The dimensions of reported landslides enable classification procedure through filtering in ENVI software. The main variables to be considered in mathematical modelling are:

- W (mean wide of landslide structures); - L (mean length of landslide structures); - D (colluviums thickness); - H (water table height); - φ (effective shear strength), - β (slope).

These characteristics enable some simplification in FS computing through GIS modelling. Mean W and L for reported landslide structures are greater than pixel resolution (15x15 m). Colluviums thickness can be computed from DTM pixel and lower volcanic flow platform. The computed colluviums thickness map shows lateral and longitudinal slope variations. The deterministic FS model can, in this way, consider that forces acting parallel to slope are in equilibrium and that the slope is infinite. These conditions can be applied chiefly to middle and lower scarps of the Aparados da Serra region. The FS determination can take the equation presented by Haneberg [8]:

FS = {1 – [(ρw / ρs) . (H / D)]} . (tan φ / tan ß) (1) However, equation 1 applies to low cohesion, unconsolidated materials. A more realistic FS determination must then consider cohesion of colluviums deposited over the volcanic basement. The equation presented by Wu and Abdel-Latif [9] takes cohesion into account and was developed for planar failure and infinite slope:

FSp = {cf + c’ + [ρ(D – H) + (ρt – ρw)H] cos2ß tan φ} / {[ρ(D – H) + ρtH] sin ß cos ß} (2)

where c’ is cohesion and cf is cohesion increase for forested areas. The FS is clearly dependent on variations of physical properties. But, water table is known to influence slope failure. The water table is not parallel to groundsurface and varies according thickness, composition and permeability of soils. It is also dependent on precipitation: it is well known that slope failure occur during on immediately after raining periods, when water table is high enough to decrease colluviums shear strength. Taking into account that water table (piezometers) close to a monitored landslide in Aparados da Serra (Southern Brazil) is 1-2 m below the groundsurface during dry periods. Then, it is possible to consider H = D (conservative approach) and determine FS for extreme situations. Cohesion increase due to vegetation cover was not considered in this FS determination. The dense Atlantic forest cover surely contributes to cohesion and to FS value. The influence of Atlantic forest in cohesion increase for colluviums is being investigated. It is interesting to note, in this respect, that shallow




landslides reported in Aparados da Serra region occur in deforested areas. Then, to partially overcome this uncertainty, it was taken into account undeformed direct shear cohesion in FS determination.

4 Landslide susceptibility map for Aparados da Serra region

The FS is the ratio between resisting to driving forces acting in slope failure. It is usually said that FS > 1 represents stable conditions, while FS < 1 represents unstable conditions. A FS ~1.25 or somewhat higher is generally accepted for slope stability. However, this determination is valid for a specific slope, where all the geomechanic parameters are known, controlled and modeled. Table 2 shows that FS is very sensitive to slope angle and colluviums thickness. The colluviums in the Aparados da Serra region show small variations in cohesion, shear strength, and densities when regional analysis is performed. However, FS rapidly decrease to values lower than 1 when slope angle goes to 25 (D = 5), and to 17 (D = 10). It is also realized that FS decreased 1/3 as colluviums thickness increased by 2.

Table 2: Sensitivity analysis of Aparados da Serra colluviums considering FS equation presented by Wu and Abdel-Latif [9]. FS was calculated for different slope angles and two colluviums thickness (D). Water table was fixed to be 1 m below the groundsurface.

D = 5 D = 10 β H = 4 H = 9

FSp5 FSp10 2 12,02876 8,528288 4 6,018449 4,264291 6 4,016854 2,843039 8 3,01748 2,132492

10 2,419042 1,706247 12 2,021122 1,422175 16 1,526263 1,067379 20 1,232421 0,854974 24 1,039572 0,713977 28 0,905004 0,614038 32 0,807534 0,540058 40 0,682551 0,44025 50 0,611113 0,371193 60 0,618595 0,348311

The determination of FS in large regions using a GIS environment must overcome large scale parameters variability. This introduces uncertainties in GIS modelling. The FS values determined in GIS environment varies from 0.2 to up




to 200. Taking into account the sensibility of FS equation (2) for geomechanic physical parameters, the predicted map of FS was computed for two different cases: i) residual shear strength (φ = 16o), that is the worst condition attained during and after raining periods, and ii) peak shear strength (φ = 29o). The predicted FS maps are presented in figure 3A and 3B, respectively. Based on the locations of some recorded shallow landslides in the Aparados da Serra region (Southern Brazil), it was possible to define five classes of susceptibility based in FS values (Table 3).

Figure 3: Landslide susceptibility map for Aparados da Serra region (Southern Brazil), based in FS equation presented by Wu & Abdel-Latif [9]. A) Predicted FS classes using residual shear strength (φ = 16o). B) Predicted FS classes using peak shear strength (φ = 29o).




Table 3: Susceptibility classes for landslides in the Aparados da Serra region (Southern Brazil), based in FS equation presented by Wu & Abdel-Latif [9]. Susceptibility classes defined for residual shear strength (φ = 16o) and for peak shear strength (φ = 29o).

Predicted FS Susceptibility class

FSp

Areas highly favourable for landslides 1 < 0.5 Areas favourable for landslides 2 > 0.5 < 1.0 Areas that can develop landslides 3 > 1.0 < 1.5 Areas not favourable for landslides 4 > 1.5 < 2.5 Stable areas under natural conditions 5 > 2.5

5 Discussion

The Factor of Safety determination using GIS modelling can be a useful technique for assessing landslide susceptibility map in areas of poor historical records on hillslopes of tropical regions, such as Aparados da Serra (Southern Brazil). The methods based on intense data records seek on statistical and multivariate capabilities for an adequate susceptibility map assess in Aparados da Serra region. In this way, careful geological, geotechnical and geomorphologic investigations in existing records were applied to assess the main causes of shallow, translational and rotational landslides in Aparados da Serra region. The geological and geotechnical model for Aparados da Serra region, the accurate and high resolution DTM (15x15 m), and the GIS capabilities do enable FS determination. The FS determination, however, introduces uncertainties in GIS modelling due to spatial variability of parameters. The value determined for each cell (15x15 m pixel) must be evaluated in a relative way. Then, a careful examination of existing records surely helps on establishing the limits between different classes. It is to be kept in mind that absolute FS value is an important parameter for construction decision in a given, specific investigation site, where a large number of geotechnical soil properties are well controlled. Landslide susceptibility map derived form FS determination in Aparados da Serra (Southern Brazil) helps, in this way, in showing “where” the areas more favourable to develop translational and rotational slides are. Ongoing investigation on Atlantic forest contribution on soils and colluviums cohesion can put some clues on questions regarding “when” and “how big” these landslides can be.

The authors want to thank FINEP, PETROBRAS and TBG (Proc. No. 0682/01) for research funds that supported this investigation.




Acknowledgements

References

[1] Cruden, D.M., A simple definition of a landslide, Bull. IAEG, 43, pp. 27-29, 1991.

[2] Cruden, D.M. & Varnes, D.J., Landslide Types and Processes. Landslides Investigation and Mitigation, A.K. Turner & R.L. Shuster Eds.. Transportation Research Board, Special Report 247, Washington DC, p.36-75, 1996.

[3] DeGraff, J.V. & Romesburg, H.C., Regional landslide-susceptibility assessment for wildland management: a matrix approach. Thresholds in Geomorphology, D.R. Coates & J. Vitek Eds., George Alien & Unwin Publ., pp. 401-414, 1980.

[4] Carrara, A., Multivariate models for landslide hazard evaluation, Mathematical Geology, 15(3), 403–426 pp., 1983

[5] Carrara, A., Cardinali, M., Guzzetti, F., & Reichenbach, P., GIS technology in mapping landslide hazard, Geographical Information Systems in Assessing Natural Hazards, A. Carrara & F. Guzzetti Eds, Kluwer Academic Publishers, pp. 135–175, 1995.

[6] Santacana, N., Baeza, B., Corominas, J., Paz, A. & Marturia, J., A GIS-based multivariate statistical analysis for shallow landslide susceptibility mapping in La Pobla de Lillet area (Eastern Pyrenees, Spain), Natural Hazards 30: 281–295, 2003.

[7] Tangestani, M.H., Landslide susceptibility mapping using the fuzzy gamma approach in a GIS, Kakan catchment area, southwest Iran, Australian Journal of Earth Sciences, 51(3), pp. 439-450, 2004.

[8] Haneberg, W.C., Deterministic and probabilistic approaches to geologic harzad assessment, Environmental & Engineering Geoscience, 6(3), pp. 209-226, 2000.

[9] Wu, T.H. & Abdel-Latif, M.A., Prediction and mapping of landslide hazard, Can. Geotech. J., 37, pp. 781-795, 2000.

[10] Silveira, R.M., Propriedades geotécnicas dos solos coluvionares do gasoduto Bolívia-Brasil em Timbé do Sul (SC), Dissertação de Mestrado em Engenharia Civil, Escola de Engenharia/UFRGS, Porto Alegre, 131 pp, 2003




Author Index

Abad M. P................................ 317 Aguilar H. R. ........................... 457 Aguilar M. ............................... 317 Alcalá L. .................................. 317 Amusan F. O............................ 443 Amusan O. A. .......................... 443 Aunap R................................... 305 Avilés J. ................................... 457 Badiani B. ................................ 341 Bahena A. ................................ 263 Barragán P. de L. G. ................ 401 Barrera I. .................................. 317 Bautista M. R. .......................... 135 Beneš V.................................... 361 Boukalová Z. ........................... 361 Bridge D. McC......................... 187 Brissette F. ............................... 273 Broderick M. A.......................... 15 Buehler Y. A............................ 391 Buffon S. A.............................. 467 Capra A.................................... 175 Carrère V. ................................ 381 Cebrián M. ............................... 317 Cecioni A..................................... 7 Champagne P. .................. 115, 125 Chang J.-H. .............................. 143 Cheng S. F. .............................. 105 Chopra M................................. 215 Clerico M................................... 33 Cremonini R. ........................... 293 Czarnecki A. ............................ 241 D’Alessandro W. ..................... 369 de Quadros T. F. P. .................. 467 Durning B. ................................. 15 Emmanouloudis D. .................. 253 Fernández de Villalta M. C...... 317 Férnández S. ............................ 263 Frey J. ........................................ 69

Frey T. ....................................... 69 Fyfe W. S..................................... 1 García-Hidalgo J. F.................. 317 García-Quintana A................... 317 Gecchele G. ............................. 349

Giménez R. .............................. 317 Godfrey A. E. .......................... 317 González-Martín J. A. ............. 317 Guida D. .................................. 283 Gutiérrez A. ............................. 263 Hall P. J. .................................. 151 Hernandez M. V. ..................... 401 Holguin C. ............................... 135 Hsiao L. S. ............................... 105 Huang C. Y. ............................. 105 Huang C.-P. ............................. 143 Itten K. I................................... 391 Jafari M.................................... 431 Jiménez J. ................................ 135 Kaikis M. ................................. 253 Kask P........................................ 69 Kauer K. .................................. 413 Kellenberger T. W. .................. 391 Kelly J...................................... 115 Khalekuzzaman M................... 125 Khalili M. ................................ 273 Kõiv M....................................... 93 Kõlli R. .................................... 413 Koodhathinkal B...................... 215 Kořán P.................................... 361 Köster T. .................................. 413 Kriipsalu M................................ 93 Lafragua J. ............................... 263 Lahdenperä J............................ 423 Leconte R................................. 273 Lehtinen K. ................................ 25


Giles J. R. A............................. 187

Lehto K. ................................... 423 Lewandowska-Czarnecka A. ... 241 Li P. P. ..................................... 225 Liu C. Q. .................................. 225 Longobardi A........................... 283 Lucía A. ................................... 317 Luukkanen O. ............................ 55 Maher A................................... 431 Mander Ü..................... 69, 93, 305 Martín-Duque J. F.................... 317 Martín-Loeches M. .................. 317 Michel F................................... 115 Mirzaeian M. ........................... 151 Möller A. ................................... 43 Moriel G. ................................. 263 Morris B. L. ............................. 187 Negreiros J. ................................ 79 Nieminen J. ................................ 55 Nishihama Y. ........................... 231 Noormets M. ............................ 413 Oder M..................................... 413 Oliveira H. R............................ 467 Oliver T...................................... 55 Ortega-Guerrero M. A. ............ 205 Paal T....................................... 413 Painho M.................................... 79 Pajuste K.................................... 69 Pérez-Rocha L. E. .................... 457 Pineda V. ..................................... 7 Pinedo A. A. ............................ 401 Pinedo A. C.............................. 401 Pizzo G. ................................... 349 Porto P. .................................... 175 Qiang Z. ................................... 143 Quijada E. ................................ 317

Rabuffetti D. ............................ 293 Ranke U. .................................... 43 Reinhart D. .............................. 215 Rodríguez-Borreguero J. M. .... 317 Roosaare J................................ 305 Rubio A. H............................... 401 Rubio H. O............................... 135 Ruiz López de la Cova R. ........ 317 Salandin A. .............................. 293 Sarala P...................................... 25 Sarmad A. ................................ 431 Saucedo T. R. .......................... 135 Scicolone B.............................. 175 Small D.................................... 391 Soffredini G. .............................. 33 Solís A. .................................... 317

Strieder A. J. ............................ 467 Tamura R. ................................ 231 Tena V. M................................ 401 Tokola T. ................................... 55 Tõnutare T. .............................. 413 Treier K. .................................... 69 Tsouchlaraki A. ....................... 329 Uuemaa E. ............................... 305 Vajirkar M. .............................. 215 Velez C. ................................... 401 Villani P................................... 283 Wang B. L. .............................. 225 Williams D. D.......................... 163 Wilson K. P. ............................ 163 Wood K. .................................. 135 Wu Y. Y................................... 225 Yanai S. ................................... 231 Zhou P. ...................................... 55


Stone Jr B. ............................... 195

WITPressAshurst Lodge, Ashurst, Southampton,SO40 7AA, UK.Tel: 44 (0) 238 029 3223Fax: 44 (0) 238 029 2853E-Mail: [email protected]

Modelling, Monitoring andManagement of AirPollution XIVEdited by: C. A. BREBBIA, Wessex Instituteof Technology, UK

Pollution is widespread throughout the worldand elimination of the risks to human healthis of the utmost importance. This bookexamines the development of experimentalas well as computational techniques to achievea better understanding of air pollutionproblems and seek their solution. The bookbrings together work from researchers whoare active in the study of air contaminantsand contains papers from the FourteenthAnnual International Conference on theModelling, Monitoring and Management ofAir Pollution. Papers featuring case studiesparticularly those discussing the evaluationof proposed emission techniques and strategiesare featured, as well as papers of a moretheoretical nature dealing with advancedmathematical and computational methods.

ISBN: 1-84564-165-5 2006 apx 450ppapx £165.00/US$295.00/€247.50

Environmental ToxicologyEdited by: A. G. KUNGOLOS, University ofThessaly, Greece, C. A. BREBBIA, WessexInstitute of Technology, UK,C. P. SAMARAS, TEI of West Macedonia,Greece, V. POPOV, Wessex Institute ofTechnology, UK

This book addresses the need for theexchange of scientific information amongexperts on issues related to environmentaltoxicology, toxicity assessment andhazardous waste management. Publishingpapers from the First InternationalConference on Environmental Toxicology,the text will be of interest to biologists,environmental engineers, chemists,environmental scientists, microbiologists,medical doctors and all academics,professionals, policy makers andpractitioners involved in the wide range ofdisciplines associated with environmentaltoxicology and hazardous waste management.The text encompasses themes such as: Acuteand Chronic Bioassays; Tests for EndocrineDisruptors and DNA Damage; InteractiveEffects of Chemicals; Bioaccumulation ofChemicals; Assessment of EcotoxicologicalProperties of Hazardous Wastes; HazardousWaste Management Techniques; LegislationRegarding Environmental Effects ofChemicals; Hazardous Waste Reduction andRecycling Techniques; Biodegradation andBioremediation; Monitoring of HazardousWaste Environmental Effects; LaboratoryTechniques and Field Validation; EffluentToxicity, Microbiotests; On-line ToxicityMonitoring; Forensic Toxicology;Genotoxicity/Mutagenicity; ExposurePathways; Risk Assessment; Biotesting andEnvironmental Control Strategy; Hot Spotsand Accidental Spills.ISBN: 1-84564-045-4 2006 apx 400ppapx £145.00/US$265.00/€217.50

Water Pollution VIIIModelling, Monitoring andManagement of Water PollutionEdited by: C. A. BREBBIA, WessexInstitute of Technology, UK,J. S. ANTUNES DO CARMO, Universityof Coimbra, Portugal

This book publishes the proceedings of theEighth International Conference onModelling, Measuring and Prediction ofWater Pollution. Water pollution is a subjectof growing public concern. The scientificcommunity has responded very rapidly tothe need for studies capable of relating thepollutant discharge with changes in the waterquality. The results of these studies arepermitting industries to employ moreefficient methods of controlling andtreating waste loads, and water authoritiesto enforce stricter regulations regarding thismatter. Bringing together papers from worldrenowned experts in this field, the textencompasses themes such as: Groundwaterand Aquifer Contamination; WastewaterTreatment; Re-use of Water; Lakes, Riversand Wetlands; Coastal Areas and Seas;Biological Effects; Agricultural Pollution;Oil Spills; Mathematical and PhysicalModeling; Experimental and LaboratoryWork; Surveying Techniques, Monitoringand Remote Sensing; Remediation Studies;Health Risk; Social and Economic Issues;Pollution Prevention; GIS and RemoteSensing Applications; EnvironmentalManagement and Decision Analysis;Environmental Impact Assessment.

ISBN: 1-84564-042-X 2006 apx 400ppapx £145.00/US$265.00/€217.50

Water ResourcesManagement IIIEdited by: M. DA CONCEICAO CUNHA,ISEC, Portugal, C. A. BREBBIA, WessexInstitute of Technology, UK

Today water resources are under extremepressure all over the world. This book bringstogether the work of researchers andprofessionals currently involved in thedevelopment of technical and scientific skillsthat contribute to achieving water resourcessustainability. Over 65 contributionsoriginally presented at the ThirdInternational Conference on WaterResources Management are included.The wide variety of subjects covered are asfollows: Hydrological Modelling;Groundwater Flow and Remediation;Reservoirs and Lakes; Coastal and EstuarialProblems; Water Quality; PollutionControl; River Basin Management; WaterManagement and Planning; WastewaterTreatment and Management; IrrigationProblems; Residential Water Management;Sustainable Water Use; and Arid RegionProblems.

Series: Progress in Water ResourcesVol 11

ISBN: 1-84564-007-1 2005 712pp£249.00/US$398.00/€373.50

All prices correct at time of going to pressbut subject to change.

WIT Press books are available throughyour bookseller or direct from

the publisher.

Geo-Environment and Landscape Evolution II

Documents

Transcript of Geo-Environment and Landscape Evolution II