Moi-sture of extinction. — Rate of spread is directly proportional to the moisture 

 damping coefficient {\^) , which enters the mathematical model as a multiplier of reac- 

 tion intensity. 



CMf) CM.)2 (M )3 



% = 1 - 2.59 ^ ^ 5.11 ^ - 3.52 - 



XX X 



where : 



= moisture content of fuel, fraction of dry weight 



= moisture of extinction, fraction of dry weight 



The moisture of extinction is the moisture content of fuel at which fire cannot 

 sustain itself. Criteria for choosing M^^ are not well established; thus use of this 

 function in the model is rather subjective. We used an of 0.24. Working backwards 

 through the model using fuel inputs and measured rates of spread for each plot, we 

 found that an average }-\^ of 0.11 for the ponderosa pine and 0.13 for Douglas-fir would 

 have provided perfect agreement between observed and predicted spread rates. values 

 that permitted perfect agreement ranged from 0.065 to 0.45 for all plots. There is good 

 reason to suspect that an of 0.24 is too high for the fuels tested. If so, this 

 would account for much of the deviations between observed and predicted spread rates. 



A comparison of with packing ratio and loading values suggests that a relation- 

 ship exists between and these fuel properties. In this study, plots with higher 

 loadings developed more intense fires. The would be expected to increase at higher 

 intensities because more energy is available for preheating fuel to ignition. Plots 

 having lower packing ratios (more porous) had greater percent deviations between 

 observed and predicted spread rates. Apparently this was in part caused by discontin- 

 uities in fuel and heat flux. Discontinuities should tend to lower U-^. 



Surface area -in/Zwence. --Weighting the input variables, especially a, by the amount 

 of surface area in each particle size class possibly leads to overprediction of spread 

 rate. The thinnest fuel particles in a fuel complex receive heavy weighting for their 

 influence on fire spread rate. This seems appropriate, but perhaps too much weighting 

 is received. Further study should refine the accuracy of this function in the model. 



Regression Analysis for Fire Spread 



A step wise multiple regression analysis using the combined data for both species 

 of slash resulted in the following equation: 



Y = 0.109(Xi) + 2.01 (X2) + 0.00827(X3) - 1.96(Xi+) (5) 



where : 



Y = rate of spread, ft./min. 



Xi= surface area of fuel to 1 cm. in diameter per square foot of ground, 

 dimension less 



X2= total fuel loading, Ib./ft.^ 



X3= windspeed ft./min. 



X,+= bulk density of fuel, Ib./ft.^ ^ 

 r2= 0.71 



Standard error of estimate (0.05 confidence level) = 1.04 



17 



