Spatial Autocorrelation Analysis and Weighting Functions in the Distribution of Wildland Fires

Abstract

Spatial autocorrelation and spatial weighting functions are examined in relation to the distribution of wildland fires. Records of fire activity and vegetation for the Idyllwild 15-minute quadrangle of the San Jacinto Mountains, San Bernardino National Forest, California, are processed in a geographic information system to generate six spatial weighting functions: contiguity, area, boundary length, distance, area-distance combined, and boundary-distance combined. The weighted coefficients and significance tests show that contiguity best represents the spatial relationship in the distribution of wildand fires. Distance introduces a scaling effect which enhances the magnitude of Moran's I coefficient of spatial autocorrelation.