Simulation of variations in the speed of fire propagation, during a prescribed burn

Abstract

Fire behavior simulations are generally based on standardized models of forest fuels, which can hardly represent all possible variations in forest ecosystems, which can result in biased estimates of fire behavior. Due to this, in the present work a comparative analysis of simulations of fire behavior (propagation rate [m/min]) is made, during the course of a prescribed burning in favor of the slope, in a pine-oak forest stand from the state of Chihuahua. These simulations varied in relation to: a) fuel models; b) fuel loads; and c) wind speed. The models tested were: 1) the one that resulted from a multiple regression (considering temperature, relative humidity and slope); 2) the fuel model M-8 and its variant M-8R (includes the values of fuel charges obtained in the field); and 3) the fuel model M-9 and its variant M-9R (includes the values of fuel charges obtained in the field). The equations of the trends defined by the correlation between the relative humidity (%) and the speed of fire propagation (m/min) were generated. Although neither model was found to be statistically significant, the best fit was defined with the regression equation: RP= 21.689e^-0.037HR, which can be used in a practical way for a rough estimation of the rate of fire spread (RP). It is concluded that, in general, the selected models were not appropriate for simulating fire behavior during prescribed burning. This highlights the need to generate specific fuel models for the analyzed vegetation and its variations in density, structure and composition.