334 Chapter 77

37. Koepf, H.H., 1989. The Biodynamic Farm. Anthroposophic Press, Hudson. New York.38. Capra, F., 1996. Het Levensweb. Levende Organismen en Systemen: Verbluffend Nieuw Inzícht in de

Grote Samenhang. Kosmos-Z& K Uítgevers, Utrecht. The Netherlands (Dutch translation of 'The Web ofLife·).

39. Sheldrake, R., 1989. The Presence of the Past. Morphic Resonance and the Habits of Nature. VintageBooks, New York.

40. Evans, L.T., 1993. Crop Evolutíon, Adaptation and Yield.Cambridge University Press.41. Schroevers, P., 1984. (Op.cit.).

Van Asseldonk, J.S.O., 1987. Holisme en reductionisme in de landbouwwetenschap: Samen meer dande som der delen. Landbouwkundig Tijdschrift. 99 (9).

42. Koningsveld, H.• 1986. Landbouwwetenschap ais systeemtheorie? Landbouwkundig Tüdschríft, 98(12).

43. Konlngsveld, H., 1987. Fragmentatie en systeembenadering in de landbouwwetenschappen. In:Koníngsveld, et aI. Landbouw. Landbouwwetenschap en Samenleving: Filosofische Opstellen.Mededelingen van de Vakgroepen VoorSocíologíe, 20. Landbouwuniversiteit. Wageningen.

44. Pretty, J. & R. Chambers, 1994. (Op.cit.).45. Van Eijk, A.M., 1988. (Op.cit.).46. Maharlshl Mahesh Yogl, 1968. Transcendental Meditation. Signet Book, New York (Original title: The

Science of Being and Art of Living, first published in 1963).Maharishl Mahesh Yogl, 1969. On the Bhagavad-Gita. A new translation and commentary, Chapters1-6. Penguin, UK(firstpublished: 1967).

11.2 THE GIS AND REMOTE SENSING CONTRIBUTION TO THE ElABORATIONOF SYSTEM HIERARCHIES IN F~

Evaristo Miranda

_Increasingly. the challenge of relating agriculture and farming to different land uses or landscape cate-gories can be attacked using GISand RS in both research and management.

11.2.1 Introduction

Land use and changes in land use are criticalelements in FSR. Rural areas are markedly het-erogeneous and many factors drive land usechanges in both time and space! . In the agricul-tural frontier areas of the developing countrieschange is particularly intense ando in addition,one type of land use (annual crops, permanentcrops, grazing lands, forestry) may be a featureof several different production systems withinthe landscape. This spatial and temporal díver-sity often precludes the application of the sametechnologies across the land-use type as awhole. To add a further dimension. technologieswhich are useful on an individual farm may beharmful when used more widely.

As a result, three important questions arisefor FSR:

• How to establish the spatial distribution ofproduction systems' at different scales?

• How to characterize the link between indi-vidual farming systems and different uses ofland?

• How to evaluate the sustainability of thediversity of farming systems and the interac-tions between them in the landscape?

Historically, FSR has concentrated at the farmlevei of what is in fact a hierarchy of systems.Analysis and evaluation have widened recently,helped by new research tools. GIS and remotesensing (RS) techniques are proving vaIuable inestablishing and monitoring systems hierarchiesin which the farm is one important leveI. Thiscontribution summarizes recent developmentsin GIS and RS methods, using examples fromapplications in the FSR-based systems hierarchyin Brazil. The methods were developed in aresearch programme executed by the NGOECO-FORCE (Research and Development @http://www.ecof.org.br) with the technicaland scientific collaboration of the

2000PI-APC2000SP-03.00616

2000

The GIS and remate sensingSP-03.00616

11111111111111111111111111111111111857 -1

At the Cutting Edge 335

Environmental Monitoring Center (NMA-EMBRAPA,http://www.nma.embrapa.br) in thecounty of Campinas, in São Paulo State, Brazil.

11.2.2 Complexity and systems hierarchies

It has been acknowledged, historically, thatfarms are complex systems. FSR has tried tomanage this complexity through its conceptualframeworks and models. This understanding ofcomplexity is related to many factors; the non-linearity and asymmetry of the relationshipamong the components and its structure; thedlfferent levels of organization and constraints;the simultaneous exístence of functional andstructural boundaries; the levei of uncertaintyof systems indicators; the permanent state ofevolution ín systems' components and interac-tions between them; varied sources of perturba-tions which destabilize a range of parameters;the spatial diversity of the landscape and theinteractions between the social. economic andecological systems.

While the concept of systems hierarchies íshelpful ín understanding these sources of com-plexity,the establishment of hierarchicallevels ínFSR cannot be an arbltrary processo Researchand rural development programmes- demon-strate a continuum of levels, their interactionsand linkages. Scale is a central concern to thosemodelling dynamic multivariate structures thatrespond to the interactions between many levelsof organization 3. Generally speaking, FSRhíerar-chical levels are associated with spatial scales ororganizationallevels in agrícultural productíon",ln FSRthe early notion of hierarchy involveddis-crete levels: fíeld, farm, watershed, valley,county.Hart, ín 1985, presented a series of concepts onagroecosystems, based on this kind of hierarchi-cal analysís". His study led to many FSR applica-tions ín Latin America. Many researchers haveconsidered the farming system hierarchy asnested, which requíres that upper levels contaínlower levels ín a continuum of structures.However, research results can rarely be general-ized, aggregated or disaggregated, from one leveito another. ln fact, in general. hierarchies whichinclude farm systems are non-nested, withstrong interactive tendencies. The levels are fre-quently a more convenient focus for researchthan the hierarchy, despite its explanatory power.

Computers are able to dísplay structures inhierarchicallevels without human value judge-ments", ln exploríng typologies of farming sys-tems, a process using clusters and multivariateanalysls, employing only numerical criteriagiven by data sets is now well known. Someauthors, such as Simon ín 1962, suggest thathierarchical structure itself ís a consequence ofhuman observatlons/. Without raising ques-tions of ontological realíty, ít seems fundamen-tal for gíven levels in FSR to take account ofhierarchical continuity and cohesiveness. Scaleis one continuously varying function that candescribe the continuum of levels and theirínteractions". The available methods try to elu-cidate the process and the critical parametersfor the different spatial scales and landscapeunits (land, property, hydrographíc basín, com-muníty mlcroregions, regions and county)using GlS and RS tools".

This contribution shows how GlS and RS areincreasingly used ín a complementary way tosimulate farming systems strategies at themicro levei; technologies for agriculture pro-duction systems, and at macro levels; publicpolicy, politics and land use. Ongoing develop-ments are improving the resolution available atseveral hierarchical levels. lncreasingly thechallenge of relating agriculture and farming todifferent land uses or landscape categories canbe attacked using GlS and RS in both researchand management!",

11.2.3 eis and RS: new techniques in FSR

GlS and RS techniques are proving their worthin helping to establish and monítor systemshíerarchíes ín FSR. GlS offers sophlstícated spa-tial analyses of the numerical descriptors of thefarming system. Whereas FSR developed com-plex numerical models, GlS spatial analysís ofproduction systems ís límited to farm fields. Thespatialization of productívíty variables, of farmsystem typology parameters, or of the system'senvironmental impact on a given resource atseveral hierarchical levels (e.g. field, farm,groups of farms or regíon) widens our under-standing and opens new horizons for FSR. GlSallows area, perimeter and volume calculationsand a series of basic operations for quantifyingthe spatial expressíon of variables. GlS also

336 Chapter 77

allows qualitative spatial analysís, such asdiversity. proper or improper land use. the símu-lation of alternative uses. the interactionsbetween different uses and the probable impactof new agricultural technologies on the enví-ronment-".

Spatial analysis of systems can occur at díf-ferent hierarchical levels. however, it oftendemands information that is not readily avail-able and sometimes does not even exist.Recently. the evolution of RS has made avaíl-able series of spatial data on existing praductionsystems and land use which have helped fillmany of these data gaps. In the developingcountries RS is frequently the only way to getthese data due to the lack of census data 12andthe difficulty in reaching some rural areasl '.Satellite imagery gives the researcher the meansto evaluate land use and changes in use!". Italso allows the researcher to detect the use ofsome technologies. particularly in soil conser-vatíon+! ando lmportantly, to relate land usesand vegetation behavíour!",

The terrestrial monitoring satellites repre-sent an efficient instrument to characterize landuse. measuring the spatial distribution of farmsand land use ín a very precise way17. Somefarming systems can be identified Irom orbitalimages and the uses of RS in FSR have beenincreasing with the development of new ímag-ing softwares and more sensitive sensors andsatellites. In 1996 satellites can already observedetail smaller than 50m2.

11.2.4 Remote sensing and FSR

The first LANDSAT satellite, originally calledERTS-l. was developed and launched by NASAin [uly 1972. Today about 300 satellites areavailable to monitor terrestrial ecosystems.agriculture and changes in land use. The ínter-est in the use of RS in FSR is linked to threepraperties of orbital images: spatial resolution.temporal resolution and radiometric or spectralresolution.

Spatial resolutionThis is important to the study of farming sys-tem based hierarchies. The orbital digital data'splasticity allows works at different spatialscales-", Agriculture can be analysed in díífer-

ent perception or hierarchical levels (local.microregíonal, regional. national). and eachperception levei can be at least partially assoei-ated with a cartographic scale ln spatial terms.Local studies range Irom 1: 1000 to1 : 10.000. microregional studies from1 : 25.000 to 1: 100.000. Regional studiesgenerally work wíth scales ranging frorn1 : 100.000 to 1 : 250.000 and national ormacraregional studies sometimes use spatialscales smaller than 1 : 1.000.000. The sameimage can be analysed frorn 1 : 1.000.000 to1 : 50.000. Recently there has been remarkabledevelopment in the spatial resolution and scalesof 1: 25.000 and 1: 10.000 can now beobtaíned, for example. from the IRS-C (India).SPOT 4 and 5 (France) and ORBVÍEW (USA)satellites.

The LANDSATTM ímage, used since 1985.covers an area of 34.000 km2 approximatelyand has a 30-m-pixel resolution. Agriculturecan be studied hierarchically frorn scales basedon the LANDSAT images which extend from1 : 1.000.000 to 1 : 50.000. The French satel-lite SPOT 3 has a 10-m resolution and theIndian satellite IRS-IC has a 6-m resolution.Both these satellites offer stereoscopic viewsand their images are already available. Thenext generation of satellites wilI be even moreaccurate, praviding resolution between 10 and100 times better than the existing commercialsatellites. formerly available merely as expen-sive aerial photos. As an example, the panchro-matic sensor of the satellites QuickBird andOrbView wilI have a 1-m resolution at nadirand the multí-colour sensor wilI have a 4-mresolutíon!",

The impraving spatial resolution of theimagery aIlows ever better sampling plans ínFSR. The distribution of land owners, the landuses and their localization can be mapped a pri-ori. This is particularly important in areaswhere censuses are insufficient or non-exístent.It is aiso vital in expanding agricultural fron-tiers such as the Amazon, or in areas whereagriculture has a strong spatial dynamíc, fre-quently expanding and contractíng-".

Temporal resolutionThis defines the frequency of repetition ín theimage's coverage at a same point: 16 days forthe LANDSAT.23 for the SPOT.Remote sensing

--- ----------------------------------

At the Cutting Edge 337

satellites are able to provide a monthly monítor-ing of agricultural and land use systems. Ifdifferent orbital systems are combined, forexample, LANDSAT,SPOT and IRS, a weeklymonitoring can be obtained. At a finer levei theNOAA/AVHRRsatellites provide information ontemperature-", drought, fires and burníngs-ê,soils moísture-! and vegetation activity at leastfour times a day. Thus orbital images can helpmonitor nutritional stress in vegetation, irriga-tion efficiency and even pest attacks.

In the last few years, the time between imageacquisition by the satellite and availability tothe user has been reduced to 1 or 2 months.The next generation of commercial hígh-reso-lution satellites will reduce this time still furtheras the images will be made available throughelectronic networks within hours. Severalorbital systems have been working since the1970s. The images obtained are preserved onfilesand made available through networks. Thisallows the reconstitution of land use evolutionand the monitoring of dynamic phenomena,like deforestatíon-+, erosion, the expansion ofthe cultivated area and salinization.

Spectra/ reso/utionThis is defined by on-board instrument bands.The instruments' spectral range includes thepanchromatic (PAN), the visible and the nearinfrared (VNIR), the short wave infrared(SWIR), the multiband thermal infrared (TIR)and the synthetic aperture radar (SAR).Satellites do not take pictures, but generateimages. Those images are digital and can beprocessed digitally. Each part of the spectralrange 'recognlzes' different surface elementssuch as soil, humidity, vegetation, dust, etc. Thecombination of the several spectral bandsthrough mathematical and statistical algo-rithms allows the identification and qualifica-tion of diverse cultures and the different kindsof land use2S•

The discriminatory power of the images isgreatly superior to that of the human eye.While the eye distinguishes an average of 20grey tones, hundreds of grey tones can be íden-tified on a satellite image. This, for example,makes it possible to identify irregularities inphotosynthetical activity that would be invisibleto the human eye. Several vegetative stages canbe identified on the same kind of plantation.

Phytomass and productivity levels can be evalu-ated, on pastures, sugar cane and cereal fields.Different soybean varieties have been dístín-guished on orbital images, due to their differ-ences in height and the insertion angle of theleaves. In the microwave field, the radar sensorsallow imaging during the night and under anyweather condítíons-". In humid tropical regionswith frequent cloud cover the radar images areof great help.

11.2.5 The use of GIS in FSR

The agricultural production cycle rarely corre-sponds to the time scales for the evaluation ofenvironmental phenomena (pedogenesis, mor-phogenesis, loss of fertility, land compactness,acidification, biodiversity reduction, riverobstruction). There has been little integration ofenvironmental phenomena in the reconstitu-tion of the history of production systems, or inthe modelling of new ones. Strict numericalmodels are inadequate to show spatial realities.The spatial and temporal dynamics of land use,either on the farm or at regional leveI, are goodexamples of crucial hierarchical issues that aredifficult to resolve without the use of carto-graphic methods. In FSR a spatial view can beacquired through the use of GIS.

GIS was born as a way to digitize cartogra-phy. Linked to the numeric data bank, a GIS isan efficient tool to characterize the spatial divi-sion of a great number of phenomena and theirdynamics. GIS allows spatial analysis to belinked to maps according to rules archives/files,equations or logical sequences-? and can thencreate new maps showing agricultural produc-tion cycles and the environmental impactsrelated to those cycles-".

GIS have a different mathematical structurefrom the satellite images treatment systems, butthere are interfaces between both. Cartographicdata can be digitally confronted with the orbitalimages and vice versa. A great deal of softwareuses GIS in a variety of applications, the the-matic and cartographic precisions vary for dif-ferent studies. Many GISare in the public domainand some versions can be very expensive to use.A GIS surrounding is convenient enough toanalyse land-use maps of different periods andarticulate them with the production systems.

338 Chapter 11

Several routines of agricultural zoning are cur-rentiy operational in GIS29. Land use plannersare increasingly using it and, in FSR, it is con-tributing to sample planning, extrapolation ofdata and the multivariate and multilocationalanalysis of productíonê".

11.2.6 GIS, RS and system hierarchies:an example

The work described in this contribution wascarried out in an area of approximately 800km2 in Campinas county, São Paulo State,Brazil. The data were obtained during a multi-disciplinary research project which included thesurvey of a sample of 100 small farms. Theenvironmental and land use characterizationwas supported by RS and the integration ofthese cartographic results with the FSR surveydata on the production systems was accom-plished through GIS. The project used the 2.4version of the GIS (GIS 2.4) of the NationalInstitute of Spatial Researches (INPE). Fourmain methods based on GIS and RS use weredeveloped and validated; mapping land-usecapacity, characterization of present land-usesystems, relationships between farming systemsand land use and a hierarchical evaluation ofagriculture environmental impacts.

Over 100 thematic and synthetic maps andorbital images were analysed and treated.Several themes were analysed at different hier-archic levels: farm, farm groups, watershed andcounty. It is impossible to reproduce these mapsand images here, but the main methodologicaland operational resuIts are discussed. Theseresults were the object of several publicationsand are available on the Internet (at the URL,http://www.nma.embrapa/projetos/cmp/gis.htmil).

Mapping land use capacityThere are several analogue methods of calculat-ing land use capacity. The method used by FAOwas adapted for GIS. The main steps developedand validated in the process were:

• Digitalization of the county limito• Digitalization of the contour curve map.• Digital generation of the hypsometric map

by the GIS.• Digital elevation model (DEM)generation.

• Declivity map digital generation, by the GIS.• Hydrographic map digitalization.• Basins and sub-basins map generation and

digitalization.• Generation and adjusts of the pedological

map, by fieldwork.• Pedological map digitalization.• Digital generation of the erodibility map, by

the GIS.• Digital generation of the hydric availability

map, by the GIS.• Digital generation of soils' phosphorus fixa-

tion capacity, by the GIS.• Digital generation of free aluminum toxicity,

using the pedology.• Digital generation of the interchangeable

bases availability map, by GIS.• Digital generation of a chemical fertility

map, using the pedology base.• Constitution of an integration programme

for generation of a land use capacity map,• Digital generation of the agricultural land

use capacity map.

Characterizatian af present land useThe characterization of present land use wascarried out using satellite multispectral irnages(LANDSAT TM 5 and SPOT) and IBGE's(Brazilian Institute of Geography and Statistics)topographic charts ín combination with field-work. The categories of land use were defined inthe county. In the Campinas case 17 categorieswere identified.

1. Urban areas.2. Water (lakes, irrigation dams and rivers).3. Natural forests and woodlands.4. Riparian forests.5. Savannas.6. Capoeiras (deforested areas, secondary

vegetation) .7. Pinus plantations.8. Eucalyptus plantations.9. Sugar cane.

10. Citrus.11. Coffee.12. Fruit plantations.13. Annual crops.14. Natural pastures.15. Artificial pastures.16. Vegetable gardens.17. Others (roads, rock, mines).

Digital classification methods were used tointerpret the satellite images and the resultsincorporated in GIS. This information can beextracted at several hierarchic levels: farrn, farmgroups, basin or sub-basin, region or county.The main methodological steps developed andvalidated were:

• Preliminary definitions of the agriculturaland non-agrícultural land use categories inCampinas county.

• Images acquisition of the SPOT and LAND-SATTM satellites.

• Digitalization of Campinas county bound-aries in GIS. and extraction of its area fromthe images.

• Migration of the corresponding digitalrecords.

• Preliminary digital treatment of the satelliteimages to differentiate and limit the mainland use categories.

• Terrestrial verifications in a diffused andconcentrated way.

• Map making of agricultural land uses in adefinite way.

• Vectoring the land use map and entry to GIS.

The re/atianships between farming systemsand /and use

Each type of land use can contain one or moreproduction or farming systems. The existing pro-duction systems were identified and their techní-cal coefficients quantified and the relationshipsbetween the use and production systems wereestablished at a variety of hierarchical levels:campo Iarrn, farms groups. basin, region and soon. Every farm was geocoded in GIS and a databank was created, articulating the land use andthe production systems. The main methodologí-cal steps were:

• Land use inventory and elaboration ofhypothesis about the existing variabilitybetween land uses and production systems.

• Preliminary identification of the main pro-duction systems, in relation to agriculturalland use.

• Acquisition of a IOO-farm sample usingaleatory stratified sampling techniques.

• Hypotheses verification about the relatíon-ships between the production systems andthe uses from the camp survey.

• For applications including more than one

production system, definition of a comple-mentary farm sample.

• Preliminary map making of spatial divisionof the main production systems interactionwith land use.

• Farming systems technical coefficient quan-tification from the IOO-farm survey.

• Final evaluation of the variability of mainland uses across production systems.

• Technical coefficient quantification andassessment of the possible environmentalimpacts of the production systems.

• Data bank constitutíon, for I ha of each typeof use identified in the fieldwork comple-mented with bibliographical data and dís-cussions with researchers.

• Creation of a data bank of technical coeffí-cients. production systems and possible cur-rent environmental impacts.

• Adequate polygon labelling of the presentland use map to allow their association withthe data bank (GIS).

Hierarchica/ eva/uatian oi agricultura Ienviranmental impacts

The environmental impact map of the agricul-tural activities. based on GIS. was made by link-ing the survey data from the farming systemsand the cartographic bases. The goal was to eval-uate the impact of agricultural activities on theland, aír, surface waters. fauna and natural vege-tation. The impact of agricultural inputs was alsoevaluated. A final synthesis was drawn up for theecosystems. The more recent data linking rou-tines via GIS enables the evaluations for severalhierarchic levels: farm, farm groups. basín, sub-basins. regions. counties or any other desiredarea. Pírst, through GIS. the following maps areproduced at several hierarchicallevels:

• Nitrogen use (kg N ha-1 year+).• Herbicide use (Iha-1 year'").• Pesticide (Iha' year+).• Synthesis map about chemical inputs impacto• The use of burning for agriculture.• Soi!compactness.• Run-off map.• Soi!loss.• Land use stability.• Land exposure period per year or bareness.

Second, the following synthesis maps are pro-duced at several hierarchic levels:

340 Chapter 11

• Farm system environmental impact on theland.

• Farm system environmental impact on thewater.

• Farm system environmental impact on thevegetation.

• Farrn system environmental impact on theair quality.

• Farm system environmental impact on thenon-biotic systems.

• Farm system environmental impact on thebiotic systems.

11.2.7 ConcJusion

FSR's recent history shows new researchthemes and methods emerging. System hier-archies are a useful operational device to helpdeal with the complexity of farming systems.They can be addressed through GIS and RS

which can be used independently or togetherin FSR. The results of the ECOFORCEandNMA/EMBRAPAapplication in Brazil demon-strate that GIS and RS can contribute to theimproved elaboration of system hierarchies inFSR. reducing the costs and time required.The nature of the RS data allows the scaletheme to be treated as a continuum acrossseveral hierarchical levels. At the same timethematic complexity gets adequate spatialtreatment through the GIS. GIS and RS cancontribute to FSR. from sampling to theextrapolation of results across space. Newsensors are providing unprecedented data.with detail to 1 m and wide spectral resolu-tíon, and will widen GIS's application to FSR.The great challenge is not in GIS and RSdevelopment anymore but perhaps on theresearchers' willingness to incorporate thesetechniques into their FSR toolbag.

REFERENCES1. Nassauer, J.I.. 1995. Culture and changing landscape structure. Landscape Ecology, 10(4), 229-37.2. Berdegué, J.A. & E. Ramíres, 1995. Investigación con Enfoque de Sistemas en Ia Agricultura y el

Desarrollo Rural. RIMISP.Santiago, Chile.Berdeguê, J.A. & E.E. de Míranda, 1991. Assesment of sustainable land-use systems research in SouthAmerica. In: Sustainable Land Use Systems Research, 12-16 Feb. 1990, New Delhí, India. Proceedings.India: Rodale Institute-USDA-India Council of Agricultural Research. pp. 201-18.

3. Allen, T.F.H.& T.B. Starr, 1982. Hierarchy: Perspectives for Ecological Complexity.University of ChicagoPress, Chicago.

4. de Míranda, E.E.• 1985. De Diferenciação camponesa e tipologias de produtores (Município de Euc\idesda Cunha). EMATER-BA/Embrapa-CPATSA.Salvador, BA.

5. Hart, R.D., 1985. Agroecosistemas; Conceptos básicos. 2nd edn. CATIE.Turrialba. Costa Rica.6. Gauch, H.G. & R.N. Whittaker. 1972. Comparison of ordination techniques. Ecology, 53, 868-75.7. Slmon, H.A.• 1962. The architecture of complexity. Proceedings of the American Philosophical Society,

106.467-82.8. Forman, R.T.T.&M. Godron.1986. Landscape Ecology.[ohn Wiley,NewYork.9. Davídson, D.A.• S.P. Theocharopoulos & R.J. Bloksma, 1996. A land evaluation project in Greece

using GIS based on Boolean and fuzzi set methodologies. International [ournal of· GeographicalInformation Systerns, 10(8), 369-84.

10. de Miranda, E.E.• C.O. Martos & J.A.C. Mangabelra, 1995. Na força das idéias. Indicadores de sus-tentabilidade agrícola na Amazônia. Embrapa-NMA, Campinas, SP.Brasil, p. 86.

11. Mltasova, H.• J. Hofíerka, M. Zlocha & L.R. Iverson, 1996. Modelling topographic potential for ero-sion and deposition using GIS. International Journal of Geographical Information Systems, 10(5),629-41.Xíang, W-N. & W.L. Stratton, 1996. The b-function and variable stream buffcr mapping: a note on 1\GIS method for riparian water quality buffcr generation'. International [ournal of GeographicalInformation Systems. 10(4).499-510.

12. GIS Aids EL Salvador Land Redistribution. 1996. GISWorld, 9(8).13.13. Petrte, G. & A.E. Níweírt, 1996. The applicability of space imagery to the small-scale topographic map-

ping of developing countries: a case study - the Sudan. ISPRIS [ournal of Photogrammetry and RemoteSensing, 51(4),209-15.

14. Lenney, P.M.• C.E. Woodcock. J.B. Collins & H. Hamdl, 1996. The status of agriculturallands inEgypt: the use of multitemporal NDVIfeatures derived from LANDSATTM. Remote Sense of Enviroment,56(1),8-20.

At the Cutting Edge 341

15. Rogowisk, A.S., 1996. Quantifying soil variability in GISapplications: Il. Spatial distribution of soil prop-erties. International [ournal of Geographical Information Systerns, 10(4),455-75.

16. Chuvieco, E. & J. Salas, 1996. Mapping the spatial distribution of forest danger using GIS. International[ournal of Geographical Inforrnatíon Systems, 10(3), 333-45.

17. Benson, RJ. & M.D. Mackenzie, 1995. Effects of sensor spatial resolution on landscape struture para-meters. Landscape Ecology,10(2), 113-20.

18. Moody, A. & C.E. Woodcock, 1995. The influence of scale and the spatial characteristics of landscapeson land-cover mapping using remote sensing. Landscape Ecology,10(6),363-79.

19. Corbley, K.P., 1996. One-rneter satellites: practical applications by spatial data users. Part '3. Geo InfoSystems, 6 (5),39-43.

20. de Miranda, E.E. & C. Mattos, 1992. Brazilian raln forest colonization and biodiversity. In: AgricultureEcosystems and Environment. Elsevíer,Amsterdam, 40,275-96.

21. Petkov, L., M. Píeri, F. Maselli & G. Maracchi, 1992. Study and modelling of temperature spatial varí-ability by NDAA-AVHRRthermal imagery. ISPRIS [ournal of Photogrammetry and Remote Sensing,47(1),1-42.

22. de Miranda, E.E., A. Setzer & A. Takeda, 1994. Monitoramento Orbital das Queimadas no Brasil.Embrapa-NMA, Campinas, SP,Brasil, p. 149.Pereira, J.R. & A.W. Setzer, 1996. Comparison of the fire detection in savannas using AVHRR'schannel3 and TM images. International [ournal of Remote Sensing, 17(10), 1925-37.

23. Tateishi, R. & C.H. Ahn, 1996. Mapping evapotranspiration and water balance for global land surfaces.ISPRIS[ournal of Photogrammetry and Remote Sensing, 51(4), 209-15.

24. Foody, G.M., G. Palubinskas, R.M. Lucas, P.J. Curran & M. Honzak, 1996. Identifying terrestrialcarbon sinks: classífícatíon of successional stage in regenerating tropical forest from LANDSA'fTM Data.Remote Sense of Enviroment, 55(3),205-16.

25. Cherrlll, A. & C. MacClean, 1995. An investigation of uncertainty in field habitat mapping and theimplications for detecting land cover change. Landscape Ecology,10(1), 5-21.

26. Taconet, O., C. Vldal-madjar, C.H. Emblanch & M, Normand, 1995. Taking into account vegetationeffects to estimate soil moisture from C-band radar measurements. Remote Sense of Enviroment, 56(1),52-6.

27. Zheng, D., E.R. Hunt & S.W. Running, 1996. Comparison of available soil water capacity estimatedfrom topography and soil series information. Landscape Ecology,11(1), 3-14.

28. HalIett, S.H., R.J.A. [ones & C.A. Keay, 1996. Environmental information systems developments forplanning sustainable land use. International [ournal of Geographical Information Systems, 10(1),47-64.

29. Arze, J., 1992. Sistemas expertos y fase de díseüo de Ia metodología para Ia investigación y desarroIlo dealternativas tecnológicas. In: International Meeting of RIMISP 5. RIMISP,Proceedings Santiago, Chile, 20.de Miranda, E.E., 1992. Posibles aportes de los Sistemas de Información Geográfica ai díseào de sistemasde producción. In: lnternational meeting of RIMISP,5. RlMISP,Santiago, Chile, p. 13.Merten, G.H., J.H. Kaviglioni, C.C. Glacomini, R.L. Rufino, D. Saintrand, C.A. Kessler, G.C. Ribas& R.O. Dedecek, 1995. Uso de SIG e do modelo USLEpara determinar a erosão potencial e atual nasmicrobacias pilotos de Água Grande e Córrego do Pensamento, Maborê, Paraná. In: Encontro da SociedadeBrasileira de Sistemas de Produção, 2Anais. SBS-IAPAR,Londrina, pp. 273-74.

30. Zonneveld, I.S. & R.T.T. Forman (Eds), 1990. Changing Landscape; an Ecological Perspective.Sprínger-Verlag, New York.