Javier Salas

Assessment of different topographic corrections in Landsat-TM data for mapping vegetation types (2003)

Different methods for topographic correction of Landsat Thematic Mapper images have been assessed in the context of mapping vegetation types. The best results were obtained with a variation of the C method, which takes into account the overcorrection of low illuminated slopes by the original C method. The performance of this method was tested using two criteria: the changes in the spectral characteristics of the image and the reduction in standard deviation of each vegetation type after the correction.

Mapping the spatial distribution of forest fire danger using GIS

ABSTRACT A geographical information system (GIS) is proposed as a suitable tool for mapping the spatial distribution of forest fire danger. Using a region severely affected by forest fires in Central Spain as the study area, topography, meteorological data, fuel models and human-caused risk were mapped and incorporated within a GIS. Three danger maps were generated: probability of ignition, fuel hazard and human risk, and all of them were overlaid in an integrated fire danger map, based upon the criteria established by the Spanish Forest Service. GIS make it possible to improve our knowledge of the geographical distribution of fire danger, which is crucial for suppression planning (particularly when hotshot crews are involved) and for elaborating regional fire defence plans.

Assessment of vegetation regeneration after fire through multitemporal analysis of AVIRIS images in the Santa Monica Mountains

Abstract Spectral mixture analysis (SMA) from Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) was used to understand regeneration patterns after fire in two semiarid shrub communities of the Santa Monica Mountains, California: northern mixed chaparral and coastal sage scrub. Two fires were analyzed: the Malibu Topanga fire (3 November 1993) and the Calabasas fire (21 October 1996). SMA was compared to the results of the Normalized Difference Vegetation Index (NDVI) to assess vegetation recovery. An unburned control plot (within the past 20 years), having similar environmental features, was used to generate two relative fire regeneration indices, Regeneration Index (RI) and Normalized Regeneration Index (NRI). Indices were calculated using the Green Vegetation (GV) endmember and the NDVI. These indices were determined to be largely independent of AVIRIS radiometric calibration uncertainty, minor errors in the atmospheric correction, topographic distortions, and differences in the phenological state of the vegetation because of interannual or seasonal differences. The temporal evolution of the two fires were combined to produce a longer observation period and used to fit a logarithmic regression model for each Mediterranean shrub community. The NRI developed from the GV endmember (NRIGV) produced the closest estimate for the time of recovery in both communities based on recovery times in the literature. The use of NDVI worked very well for recovery in the northern mixed chaparral, but was less successful in the coastal sage scrub, mainly because of extensive herbaceous cover during the first years of the regeneration process. Endmembers generated from hyperspectral images were more accurate because they are tuned to capture the greenness of the shrub type of vegetation. Use of matching plots having similar environmental features, but which were burned in different years were demonstrated to improve estimates of the recovery within each community.

Estimation of shrub height for fuel-type mapping combining airborne lidar and simultaneous color infrared ortho imaging

A fuel-type map of a predominantly shrub-land area in central Portugal was generated for a fire research experimental site, by combining airborne light detection and ranging (LiDAR), and simultaneous color infrared ortho imaging. Since the vegetation canopy and the ground are too close together to be easily discerned by LiDAR pulses, standard methods of processing LiDAR data did not provide an accurate estimate of shrub height. It was demonstrated that the standard process to generate the digital ground model (DGM) sometimes contained height values for the top of the shrub canopy rather than from the ground. Improvement of the DGM was based on separating canopy from ground hits using color infrared ortho imaging to detect shrub cover, which was measured simultaneously with the LiDAR data. Potentially erroneous data in the DGM was identified using two criteria: low vegetation height and high Normalized Difference Vegetation Index (NDVI), a commonly used spectral index to identify vegetated areas. Based on the height of surrounding pixels, a second interpolation of the DGM was performed to extract those erroneously identified as ground in the standard method. The estimation of the shrub height improved significantly after this correction, and increased determination coefficients from R2 = 0.48 to 0.65. However, the estimated shrub heights were still less than those observed in the field.

Integrating geospatial information into fire risk assessment

Fire risk assessment should take into account the most relevant components associated to fire occurrence. To estimate when and where the fire will produce undesired effects, we need to model both (a) fire ignition and propagation potential and (b) fire vulnerability. Following these ideas, a comprehensive fire risk assessment system is proposed in this paper,whichmakesextensiveuseofgeographicinformationtechnologiestoofferaspatiallyexplicitevaluationoffirerisk conditions. The paper first describes the conceptual model, then the methods to generate the different input variables, the approachestomergethosevariablesintosyntheticriskindicesandfinallythevalidationoftheoutputs.Themodelhasbeen applied at a national level for the whole Spanish Iberian territory at 1-km 2 spatial resolution. Fire danger included human factors, lightning probability, fuel moisture content of both dead and live fuels and propagation potential. Fire vulnerability was assessed by analysing values-at-risk and landscape resilience. Each input variable included a particular accuracy assessment, whereas the synthetic indices were validated using the most recent fire statistics available. Significant relations (P,0.001) with fire occurrence were found for the main synthetic danger indices, particularly for those associated to fuel moisture content conditions.

Assessment of forest fire danger conditions in southern Spain from NOAA images and meteorological indices

Traditionally, the estimation of fire danger is performed from meteorological danger indices that are computed for single locations, where the weather stations are located. Frequently, these locations are far from forested areas, and there is a need to spatially interpolate danger variables. Methods for spatial interpolation are always prone to error, especially for those variables that show a greater spatial variability (wind, mainly). Satellite images may be considered a good alternative for interpolation of danger values, since they perform a spatially exhaustive observation of the territory. This paper analyses the spatial distribution of the Canadian Drought Code (DC), part of the Canadian Forest Fire Weather Index System (CFFWIS), in the region of Andaluc{@a (south Spain) following two procedures. First, maps of DC values were obtained from spatial interpolation of a network of 30 weather stations using the squared inverse distance algorithm. These results were compared with interpolation based on linear regression analysis, using National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) derived bands as independent variables. The most significant variables found for these empirical fittings were relative greenness, the ratio of Normalized Difference Vegetation Index (NDVI) and surface temperature, and a temporal variable, which accounts for the variations in day length throughout the fire season. After several empirical fittings were obtained, the most precise estimation was found after adjusting the coefficients to the time period considered.

A multivariate analysis of biophysical factors and forest fires in Spain, 1991–2005

The main goal of this study was to explain the relationship between forest fires and different climatic, topographic and vegetation factors, establishing explanatory models from multivariate analysis. The study area comprised peninsular Spain. Two dependent variables were considered: probability of burning and fire size class, from a forest-fire map derived from visual analysis of satellite images from 1991 to 2005 (3337 fires greater than 25 ha). Logistic regression, discriminant analysis and regression trees were used to analyse the probability of burning. The models showed a significant relationship with land cover and slope, where the classification achieved an agreement of ~66%, and this was very similar for the three statistical methods used. Discriminant analysis and regression trees were used to model fire size class. These models appeared more related to ecozones and climatic variables (winter precipitation and mean summer temperature). In this case, the best classification results were obtained in the category of very large fires (>5000 ha), with an agreement above 80%. Regression trees achieved better results for fire size class models.