Fire occurrence zones: kernel density estimation of historical wildfire ignitions at the national level, Greece

Abstract

The focus of our study was to create a Main Map of fire occurrence zones from historical wildland fire ignition observations at national level in Greece using a Kernel Density estimation procedure. Kernel density estimation, a non parametric statistical method for estimating probability densities, has been widely used for home range estimation in wildlife ecology. It has the advantage of directly producing density estimates that are not influenced by grid size and localization effects. Furthermore, it produces densities of any shape and analyzes any data distributed multi-modally or non-normally. Under this perspective, kernel density surfaces have been created to construct fire occurrence zones. Their observed distribution was statistically significantly different than the expected one that arises under complete spatial randomness. A smoothing effect is certainly observed when increasing the bandwidth size of the kernel density interpolation. Excluding the kernel size of 1000 meters, then the results do not prove any influence of kernel size or control points on the kernel density surfaces.