The rate of spread distribution lies in a narrow band near 3 ft/min. See 

 figure 7. All values presented are less than 4 ft/min. All three averages-- 

 uniform, nonuniform, and weighted-- lie within the extremes of uniform sagebrush 

 and uniform grass. 



The fireline intensity distribution extends from to 30 BTU/s/ft and in- 

 cludes the uniform model averages. In contrast to the rate of spread, both the 

 uniform and nonuniform model averages lie outside the extremes of uniform sage- 

 brush and uniform grass--exceeding both values. 



Fires were not simulated in the grass-sagebrush mixture at 2 mi/h windspeeds 

 because the transition zone depth exceeded the cell size, 2.5 ft (76 cm). See 

 appendix I. The grouping of cells into larger cells is possible but in this case 

 the fundamental cell is the size of a sagebrush plant. Averaging such dissimilar 

 fuel cells at this time does not serve the purpose of this paper--to illustrate 

 the nonuniform modeling approach and the form of the result. 



DISCUSSION 



Fuel nonuniformities , and resulting nonuniform fire behavior, although recog- 

 nized as a problem, have been ignored because of a general inability to handle non- 

 uniformity both in its assessment and simulation and in the form of the result. His- 

 torically we have attributed a single rate of spread to a unit of land. This concept 

 existed before the formulation of the uniform fire spread model (Rothermel 1972) . 

 The assumption of uniformity in the first attempts to describe firespread in wild- 

 land fuels was a logical first step. The assumption may have enforced our view of 

 single valued results but is not the source of that viewpoint. The recognition of 

 fuel nonuniformity carries with it the realization that fire exhibits nonuniform 

 behavior. Consequently the result must be in the form of distributions, excursions 

 about the average. 



Two fuel arrays, slash and a grass -sagebrush mixture, are presented as examples 

 for describing the nonuniform fire behavior model. Slash is a single fuel type with 

 varying depth and load. The grass-sagebrush mixture is a mixture of fuel types. 

 Unlike slash, most any location in the mixture has a dichotomous description as 

 either grass or sagebrush. 



All possible combinations of obtaining average fire behavior results using the 

 uniform fire model were employed for comparison with the distributions. For slash, 

 the average depth and fuel load were used. For the grass-sagebrush mixture, there 

 were the additional results obtained from uniform grass, uniform sagebrush, 

 weighted average of grass and sagebrush fuels, and the weighted average of the 

 two results for uniform grass and uniform sagebrush. 



19 



