Abstract:

Abstract The aim of this study was to examine the major factors influencing the occurrence of fires by analyzing the frequency ratio (FR) of areas with the greatest occurrence of managed fires and wildfire and to generate the susceptibility map of these occurrences for the Indigenous Territory of Kraholândia (Tocantins state), Brazil. A supervised classification method was apply using the Mahalanobis algorithm, and the fire scars were delimited from 2003 to 2014 based on images obtained from the Landsat-5, CBERS-2, ResourceSat-1, and Landsat-8 satellites. The higher recurrence fire class was using as a reference to assess the following variables: topography; type of land use; soil classification; and distances from human settlement, roads, and waterways. The FR of managed fire or wildfire occurrences was determined for each variable and a subsequent map for all the variables was create wiht the FR value of the sum of each pixel was denoted as the forest fire susceptibility index (FFSI). The factors showing the greatest correlation with the highest frequencies of managed fire and wildfire occurrences were vegetation types of farmland and rock field, areas used for agriculture, areas with slopes higher than 30%, altitudes above 350 m, areas in ridgelines, and proximity to roads. The FFSI map can be an effective tool in planning and controlling forest fires.

Keywords: