Post on 28-Apr-2018
Remote Sensing of forest fires
(focus on Mediterranean ecosystems)
Aristotle University of Thessaloniki
Greece
Laboratory of Forest Management and Remote SensingSchool of Forestry & Natural Environment
http://fmrs.web.auth.gr/
Ioannis Z. Gitas - Associate Professor
Tel: +30 2310 992699, Fax: +30 2310 992677E-mail: igitas@for.auth.gr
Presentation Structure
Mediterranean ecosystems and their characteristics
Forest fires in the Mediterranean
Forest Fire Management related Remote Sensing (and GIS) applications
Mediterranean ecosystems are limited to five relatively small areas around the planet
The Mediterranean forests, woodlands, and scrubbiome can be found in the world's fiveMediterranean climate zones, on the west coastof continents in the mid-latitudes.
Mediterranean ecosystems
The climate is characterized by hot, drysummers and mild, wet winters.
Vegetation types can range from forests towoodlands, savannas, shrublands, andgrasslands.
High biodiversity (uniquely adapted animaland plant species) due to the limited extentand isolation of the Mediterranean climateregions.
Of the most highly altered ecosystems on theplanet due to high human impact (habitation,agriculture, recreation).
Mediterranean ecosystems(characteristics)
› Natural fires are an integral part ofmany terrestrial ecosystemsincluding the Mediterranean.
› However, from the 1960s untiltoday, the general trend in thenumber of fires and surfaceburnings in the EuropeanMediterranean areas has increasedexponentially.
› This increase is mainly due to: (a)changes in traditional land uses, theconsequence of which is higher fuelaccumulation, and (b) globalclimatic warming.
Forest fire statistics for Greece (period 1980-2007)
Forest Fires in the European Mediterranean
Lebanon 2007 Greece 2007 Turkey 2006
Portugal 2005 France 2003 Spain 2006
Italy 2007
Fire, both natural and human-caused, has played a major role inshaping the landscape in the Mediterranean.
Forest Fires (examples from the Mediterranean region)
Forest Fires in Greece(the Peloponnese - August 2007)
Weather patterns Human activities
Topography VegetationSource: Emilio Chuvieco
Forest Fire Factors
A complete fire managementprogramme includes thefollowing:
o pre-fire planning (e.g.creation of fire-breaks,fuel type mapping,topography)
o fire detection andmonitoring (e.g. FUEGOsystem, observatories)
o fire suppression (e.g. useof retardants), and
o post-fire evaluation (e.g.burned area mapping,regeneration mappingand monitoring).
Forest Fire Management
Remotely sensed data and GIS analysis may be used in all phases of afire management programme.
Pre-fire Operations
Post-fire Operations
During fire Operations
Remote Sensing and GIS in Forest Fire Management
Forest Fire Management related Remote Sensing (and GIS) applications
Pre-fire planning
o Vegetation dryness mapping and monitoring
o Fire tower visibility analysis etc.
o Fuel type mapping (eg. ArcFuel)
During fire (assist operations)
o Fire Suppression
Post-fire impact assessment
o Burned area mapping
o Operational Burned Area Mapping at National Level
o Burn Severity mapping
o Short and long term damage assessment
Pre-fire Planning
Vegetation monitoring at
approx. weekly basis
Change detection map
Alexandridis, T.K., Katagis, T., Gitas, I.Z., Silleos, N.G., Eskridge, K.M., Gritzas, G. Investigation of aggregation effects in vegetation condition monitoring at anational scale (2010) International Journal of Geographical Information Science, 24 (4), pp. 507-521.DOI: 10.1080/13658810902798107
Pre-fire planning
Vegetation dryness mapping (MODIS – 250 m)
produced for the period 2002 – 2013 on a weekly basis
Pre-fire planning
Vegetation dryness monitoring (interannual comparison)
Alexandridis, T.K., Gitas, I.Z., Silleos, N.G. An estimation of the optimum temporal resolution for monitoring vegetation condition on a nationwidescale using MODIS/Terra data (2008) International Journal of Remote Sensing, 29 (12), pp. 3589-3607.DOI: 10.1080/01431160701564618
Pre-fire planning
Road network
Map of existing firetowers
Land coverDEM
Visibility map
Fire tower visibility analysis
Pre-fire planning
Estimation of defensible space
Pre-fire planning
› Classification of fuel is a difficult and complexprocess that requires good knowledge ofecology and fire behaviour.
› Fuel type classification is traditionally basedon the use of the, so called, fuel models suchas the:
o Northern Forest Fire Laboratory (NFFL) Fuel Model
o Prometheus fuel model (Mediterranean Basin specific)
o Other (Greek fuel model - Dimitrakopoulos 2001)
› Recent fuel type mapping methods rely on thecombined use of exisitng datasets (e.g. LPIS,CORINE) and remote sensing derivedinformation.
Fuel type classification
Fuel type mapping
PROMETHEUS model - IKONOS imagery
Giakoumakis M., I.Z. Gitas, J. San-Miguel (2002) Object-oriented classification modelling for fuel type mapping in the Mediterranean, using LANDSAT TM and IKONOS imagery- preliminary results, in Viegas, D.X. (ed.), IV International Conference on Forest Fire Research: Forest Fire Research & Wildland Fire Safety, Millpress, Rotterdam, ISBN 90-77017-72-0, p. 66
Resolution 60 cm
Output ready to be used in firebehaviour models such as FARSITE
Mallinis, G., Mitsopoulos, I.D., Dimitrakopoulos, A.P., Gitas, I.Z., Karteris, M. Local-scale fuel-type mapping and fire behavior prediction byemploying high-resolution satellite imagery (2008). IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 1(4), pp. 230-239. DOI: 10.1109/JSTARS.2008.2011298
Fuel type mapping
Using Quickbird imagery –Taxiarchis University Forest
A comparative analysis of Hyperion, Quickbird and Landsat TM and imagery for fuel type mapping of a
typical Mediterranean landscape
G. Mallinis, G. Galidaki, I. Gitas, A Comparative Analysis of EO-1 Hyperion, Quickbird and Landsat TM Imagery for Fuel Type Mapping of a Typical Mediterranean Landscape, Remote Sensing 6(2), 1684-1704, 2014.
Fuel type mapping
Fuel type mapping
• ArcFuel is an example of a multi-resolution multi-source mappingapproach.
• The method was recently developed within the frame of the ArcFUELLIFE project (http://www.arcfuel.eu/index.php/en/).
The ArcFuel model
Fuel type mapping
DMC su
mmer
DMC winter
LUCAS points
LPIS-ILOTS
• The NOFFi fuel map is developed on the experience gained fromthe ArcFUEL project.
• The map is produced on behalf of the Greek Forest Service who isfunding the project.
The National Observatory of Forest Fires (NOFFi) fuel map
Fuel type mapping
• Three pilot areas (Chalkidiki, Attiki, Xanthi) are currently used forthe development of the methodology.
• Once the method is checked for its accuracy then will beimplemented in the whole country.
The NOFFi fuel map (current status)
During fire
Fire Suppression
Satellite image-map for fire-fighting resources allocation
Fire Suppression
1. Estimation of vegetation density using image analysis
2. GIS analysis for locating the ‘line of defence’
Planning the ‘line of defence’
Post-fire mapping and monitoring
Burned area mapping
Traditional ways of mapping forest fires
(Greek Forest Service)
Burned area mapping
The starting point for remote sensing research in the field of burned areamapping is the study of the spectral properties of the burned areas, andtheir comparison to other major classes of land cover existing on theimage.
0
10
20
30
40
50
60
0.3
952
0.4
814
0.5
698
0.6
591
0.7
48
0.8
353
0.9
197
0.9
998
1.2
506
1.5
089
1.7
608
2.0
092
2.2
569
2.5
071
Re
fle
cta
nce
l
pine burnt100
Burned Area Mapping with Remote Sensing
Burned area mapping
Gitas et al. (2007) Wildfires & Remote Sensing - Fast Mapping Results Provide Deeper Insights, GEOInformatics, October/November 2007,Vol 10, 16-19.
Burned Area mapping using VHR imagery (Parnitha 2007 - IKONOS)
Burned area mapping
Gitas, I. Z., Polychronaki, A., Katagis, T. and Mallinis, G. (2008) Contribution of remote sensing to disaster management activities: A casestudy of the large fires in the Peloponnese, Greece, International Journal of Remote Sensing, 29:6, 1847 – 1853
Burned area mapping and rapid damage assessment
An FMRS - DMC ii - WWF Hellas partnership
Burned area mapping
Mitri, G.H., Gitas, I.Z. A semi-automated object-oriented model for burned area mapping in the Mediterranean region using Landsat-TM imagery (2004) International Journal of Wildland Fire, 13 (3), pp. 367-376. Cited 26 times. DOI: 10.1071/WF03079
Geographic Object-based Image Analysis (GEOBIA)
Operational Burned area mapping
In house developed stand-alone application for the implementation of advanced classification algorithms
Advanced classification techniques in BA (NOFFi)
Dragozi E, Gitas IZ, Stavrakoudis DG, Theocharis JB. Burned Area Mapping Using Support Vector Machines and the FuzCoC FeatureSelection Method on VHR IKONOS Imagery. Remote Sensing. 2014; 6(12):12005-12036.
Burned area mapping
Dragozi E, Gitas IZ, Stavrakoudis DG, Theocharis JB. Burned Area Mapping Using Support Vector Machines and the FuzCoC FeatureSelection Method on VHR IKONOS Imagery. Remote Sensing. 2014; 6(12):12005-12036.
Results from the implementation of the advanced algorithms
Burned area mapping
Gitas et al. (2010) BAS2, An In-House Developed Stand Alone GIS Application For The Estimation Of Burned Area Statistics. 3rdWSEAS International Conference on VISUALIZATION, IMAGING and SIMULATION (VIS '10), 3-5 November 2010, University of Algarve,Faro, Portugal.
Open-source GIS software for damage assessment – Burned Area Statistics (BAS2)
Fire Type mapping
Mitri G.H. and Gitas I.Z. (2006) Fire type mapping using object-based classification of Ikonos imagery. International Journal ofWildland Fire 15, 457-462.
Canopy burn
Surface burn
Forest canopy:
heavily burned?
Living trees in the plot:
scorched?
Vegetationno
no yes
yes
Mapping the type of fire using IKONOS imagery
Burn severity
Burn severity is a descriptive term that integrates the physical, chemicaland biological changes on a site as a result of fire disturbance.
(Jain,2004) FIRE IMPACT ON SOIL AND VEGETATION
Burn severity-Definition
Burn severity
LANDSAT TM images used
before and after the fire.
Estimation of NBR
(Normalized Burned ratio)
index and production of burn
severity map
Veraverbeke et al. (2010) Evaluating Landsat Thematic Mapper spectral indices for estimating burn severity of the 2007 Peloponnesewildfires in Greece. International Journal of Wildland Fire 19, 558–569
Estimation using LANDSAT imagery
Burn severity
LegendVery High
High
Moderate
Low
Estimation using WorldView-2 imagery, GEOBIA and Composite Burn Index (CBI)
(on-going research)
Burn severity field data were collected annually from 100 selected 60X60m plots (November of 2011, August of 2012, November of 2014).
Long-term monitoring
1995 2006
1989 Fire
1985 Fire 1985 Fire
1989 Fire
Legend
Coniferous Forest
Bare Land
Broadleaved Forest
Low Vegetation
Mixed Forest
Shrubs
Polychronaki, A.; Gitas, I.Z.; Veraverbeke, S.; Debien, A. Evaluation of ALOS PALSAR Imagery for Burned Area Mapping in Greece Using
Object-Based Classification. Remote Sens. 2013, 5, 5680-5701.
Monitoring of burned areas –Synergy of SPOT and RADAR ERS satellite data
Generation of fractional vegetation
cover maps with the use of Spectral
Mixture Analysis (SMA) method.
Veraverbeke et al. (2011) Spectral mixture analysis to assess post-fire vegetation regeneration using Landsat Thematic Mapper imagery:
Accounting for soil brightness variation, International Journal of Applied Earth Observation and Geoinformation, Volume 14, Issue 1, 1-11.
Long-term monitoring
Post-fire vegetation monitoring using LANDSAT data and SMA
Long-term monitoring
Breaks For Additive Season and Trend (BFAST)
Katagis Thomas, Gitas Ioannis Z., Toukiloglou Pericles, Veraverbeke Sander, Goossens Rudi (2014) Trend analysis of medium- andcoarse-resolution time series image data for burned area mapping in a Mediterranean ecosystem. International Journal of Wildland Fire23, 668–677.
Post-fire vegetation monitoring using trend analysis of satellite time series
Long-term monitoring
Mitri G.H. and Gitas I.Z. (2009) Mapping Postfire Vegetation Recovery Using EO-1 Hyperion Imagery, IEEE Transactions on Geoscience andRemote Sensing
Regeneration monitoring- Thasos island
The forest ecosystem succession
assessment is based on the comparison
of the pre-fire and the post-fire land
cover maps within a GIS environment.
Long-term monitoring
Vegetation succession monitoring –Thasos island
The range of RS applications related to forest fires is increasing as a result of the following:
A
• an increase in the number of sensors with different characteristics suitable for studying aspects of fire
B
• an increase in the number of sensors with different characteristics suitable for studying aspects of fire
C
• the development of new advanced digital image analysis techniques
D
• improved access to and availability of satellite data and derived products
Thank you very much for your attention