Environmental susceptibility model for predicting forest fire occurrence in the Western Ghats of India

Abstract

Forest fires are a recurrent management problem in the Western Ghats of India. Although most fires occur during the dry season, information on the spatial distribution of fires is needed to improve fire prevention. We used the MODIS Hotspots database and Maxent algorithm to provide a quantitative understanding of the environmental controls regulating the spatial distribution of forest fires over the period 2003–07 in the entire Western Ghats and in two nested subregions with contrasting characteristics. We used hierarchical partitioning to assess the independent contributions of climate, topography and vegetation to the goodness-of-fit of models and to build the most parsimonious fire susceptibility model in each study area. Results show that although areas predicted as highly prone to forest fires were mainly localised on the eastern slopes of the Ghats, spatial predictions and model accuracies differed significantly between study areas. We suggest accordingly a two-step approach to identify: first, large fire-prone areas by paying special attention to the climatic conditions of the monsoon season before the fire season, which determine the fuels moisture content during the fire season; second, the most vulnerable sites within the fire-prone areas using local models mainly based on the type of vegetation.