Ana C. L. Sá

Spectral characterisation and discrimination of burnt areas

Spectral properties of recent burns are characterised, in the visible, near infrared, mid-infrared, thermal infrared, and microwave spectral domains. Fire-induced reflectance changes are also compared for various ecosystems and biomes, and discussed in terms of the ecological effects of phytomass combustion. The spectral signatures of combustion products and of burnt areas are compared with those of various plant material and land cover types, in order to graphically represent relevant aspects of burnt area spectral discrimination. A series of colour composite images, based on Landsat Thematic Mapper data is used to illustrate the appearance of burnt surfaces in various tri-spectral spaces, and in contrast with healthy forests, agricultural fields, and urban areas. The temporal evolution of the spectral properties of burns is also demonstrated, with a five-year time series of Thematic Mapper images of two conifer forest burns in central Portugal. Finally, a series of conclusions is proposed, concerning the distinctive spectral properties of burnt surfaces, and implications for discrimination and mapping of such areas.

Assessing the feasibility of sub-pixel burned area mapping in miombo woodlands of northern Mozambique using MODIS imagery

The goal of this study was to evaluate the feasibility of sub-pixel burned area detection in the miombo woodlands of northern Mozambique, using imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS). Multitemporal Landsat-7 ETM+ data were acquired to produce a high spatial resolution map of areas burned between mid-August and late September 2000, and a field campaign was conducted in early November 2000 to gather ground truth data. Mapping of burned areas was performed with an ensemble of classification trees and yielded a kappa value of 0.896. This map was subsequently degraded to a spatial resolution of 500 m, to produce an estimate of burned area fraction, at the MODIS pixel size. Correlation analysis between the sub-pixel burned area fraction map and the MODIS reflective channels 1-7 yielded low but statistically significant correlations for all channels. The better correlations were obtained for MODIS channels 2 (0.86 µm), 5 (1.24 µm) and 6 (1.64 µm). A regression tree was constructed to predict sub-pixel burned area fraction as a function of those MODIS channels. The resulting tree has nine terminal nodes and an overall root mean square error of 0.252. The regression tree analysis confirmed that MODIS channels 2, 5, and 6 are the best predictors of burned area fraction. It may be possible to improve these results considering, as an alternative to individual channels, some appropriate spectral indices used to enhance the burnt scar signal, and by including MODIS thermal data in the analysis. It may also be possible to improve the accuracy of sub-pixel burned area fraction using MODIS imagery by allowing the regression tree to automatically create linear combinations of individual channels, and by using ensembles of trees.

Relationships between Human Population Density and Burned Area at Continental and Global Scales

We explore the large spatial variation in the relationship between population density and burned area, using continental-scale Geographically Weighted Regression (GWR) based on 13 years of satellite-derived burned area maps from the global fire emissions database (GFED) and the human population density from the gridded population of the world (GPW 2005). Significant relationships are observed over 51.5% of the global land area, and the area affected varies from continent to continent: population density has a significant impact on fire over most of Asia and Africa but is important in explaining fire over < 22% of Europe and Australia. Increasing population density is associated with both increased and decreased in fire. The nature of the relationship depends on land-use: increasing population density is associated with increased burned are in rangelands but with decreased burned area in croplands. Overall, the relationship between population density and burned area is non-monotonic: burned area initially increases with population density and then decreases when population density exceeds a threshold. These thresholds vary regionally. Our study contributes to improved understanding of how human activities relate to burned area, and should contribute to a better estimate of atmospheric emissions from biomass burning.

Regional-scale burnt area mapping in Southern Europe using NOAA-AVHRR 1 km data

A brief review of studies dealing with burnt area mapping using coarse spatial resolution satellite imagery is presented, followed by an analysis of areas burnt in Iberia during the 1991 and 1994 fire seasons, two of the worst on record in Portugal and Spain, respectively. In order to detect and map burnt areas, new multitemporal image compositing algorithms were developed. Burnt area mapping for the 1991 fire season relied on the Global Environment Monitoring Index (GEMI), albedo, and surface temperature. A rule-based classifier was induced from training data, using the classification and regression trees (CART) algorithm. Fire size estimates compared well to those derived with Landsat TM imagery, but appear unreliable for burns smaller than about 2000 ha. The 1994 fire season data were analysed with a two-phase procedure. First, a new spectral index was specifically designed for burnt area detection. Images of this index were thresholded to detect clearly burnt pixels and to avoid commission errors. Secondly, a distance-based multicriteria analysis technique was applied, combining spectral similarity and spatial contiguity criteria, to map the burns. The method detected over 80% of all large fires, and proved especially effective at mapping burns larger than 1000 ha.

Analysis of the relationship between spatial pattern and spectral detectability of areas burned in southern Africa using satellite data

Fires in Africa affect atmospheric emissions and carbon sequestration, landscape patterns, and regional and global climatic conditions. Studies of these effects require accurate estimation of the extent of measurable fire events. The goal of this study was to assess the influence of burned area spatial patterns on the spectral detectability of burned areas. Six Landsat‐7 ETM+ images from the southern Africa were used for burned area mapping and spatial pattern analysis, while contemporaneous MODIS 500 m spatial resolution images were used to measure the spectral detectability of burned areas. Using a 15 by 15 km sample quadrats analysis, we showed that above a burned area proportion threshold of approximately 0.5 the spectral detectability of burned areas increase due to the decrease in the number of mixed pixels. This was spatially related to the coalescence of burned patches and the decrease in the total burned area perimeter. Simple burned area shapes were found at the Botswana site, where the absence of tree cover and the presence of bright surfaces (soil and dry grass) enhanced the spectral contrast of the burned surfaces, thus enabling better estimates of burned area extent. At the Zambia and Congo sites, landscape fragmentation due to human activity and the presence of a tree vegetation layer, respectively, contribute to the presence of small burned area patches, which may remain undetectable using moderate spatial resolution satellite imagery, leading to less accurate burned area extent estimates.

Estimation of combustion completeness based on fire¿induced spectral reflectance changes in a dambo grassland (Western Province, Zambia)

An experimental burn was performed in a dambo grassland, in the Western Province of Zambia, during the SAFARI 2000 Third Intensive Field Campaign. The main goal of this study was to analyse the possibility of estimating combustion completeness based on fire‐induced spectral reflectance changes in surface. Inverse, nonlinear relationships were obtained between combustion completeness and pre‐fire to post‐fire spectral reflectance changes, in the green, red, and near‐infrared spectral domains (equivalent to Landsat 7 ETM+ channels 2, 3, and 4). The coefficient of determination (R 2) varied from 0.50 for channel 4, to 0.57 for channel 3, and all the regressions were significant at the 95% confidence level. Thus, it may be feasible to treat combustion completeness as a variable whose values can be remotely estimated. However, its relationship with fire‐induced spectral reflectance changes is expected to exhibit some dependence on vegetation structure. The experimental burn was performed simultaneously with overpasses from the Terra satellite, and from the NASA ER‐2 research airplane carrying the 50‐channel MODIS Airborne Simulator (MAS) image spectrometer. Our results may be used in conjunction with imagery from these sensors, to support the development of operational approaches for combustion completeness estimation from remotely sensed data.

Probabilistic fire spread forecast as a management tool in an operational setting

BackgroundAn approach to predict fire growth in an operational setting, with the potential to be used as a decision-support tool for fire management, is described and evaluated. The operational use of fire behaviour models has mostly followed a deterministic approach, however, the uncertainty associated with model predictions needs to be quantified and included in wildfire planning and decision-making process during fire suppression activities. We use FARSITE to simulate the growth of a large wildfire. Probabilistic simulations of fire spread are performed, accounting for the uncertainty of some model inputs and parameters. Deterministic simulations were performed for comparison. We also assess the degree to which fire spread modelling and satellite active fire data can be combined, to forecast fire spread during large wildfires events.ResultsUncertainty was propagated through the FARSITE fire spread modelling system by randomly defining 100 different combinations of the independent input variables and parameters, and running the correspondent fire spread simulations in order to produce fire spread probability maps. Simulations were initialized with the reported ignition location and with satellite active fires. The probabilistic fire spread predictions show great potential to be used as a fire management tool in an operational setting, providing valuable information regarding the spatial–temporal distribution of burn probabilities. The advantage of probabilistic over deterministic simulations is clear when both are compared. Re-initializing simulations with satellite active fires did not improve simulations as expected.ConclusionThis information can be useful to anticipate the growth of wildfires through the landscape with an associated probability of occurrence. The additional information regarding when, where and with what probability the fire might be in the next few hours can ultimately help minimize the negative environmental, social and economic impacts of these fires.