Abstract

A relatively small number of forest fires are responsible for a very high proportion of the total damage. The proportion due to the fraction p of largest fires, when plotted against p, is a measure of variability of fire sizes that is especially sensitive to the important extreme events. We find the theoretical form of the plot for several commonly used distributions and show how the results can be used in the analysis of empirical plots from actual fire size distributions. Some simple graphical methods are suggested for the fitting of truncated Pareto and lognormal distributions to empirical data. We use the plots to compare the firesize heterogeneity of several fire climate regions of the western United States; the proportion of area burned by the 1% of largest fires ranges from 80%-96%. We also compare patterns of the larger fires for Southern California and Baja California. For. Sci 35(2):319-328.

This content is only available as a PDF.