CONCLUSIONS 



The mathematical model's predictions correlated reasonably close to observed rates 

 of spread, although the model tended to overestimate. The study showed that M^, dis- 

 continuities, and possibly the method of weighting by surface area in the mathematical 

 model need additional study. The Mx appears to be a function of particle size, loading, 

 and packing ratio. Its functional relationships in living fuels probably involve other 

 properties as well. Weighting the input parameters by surface area in each size class 

 seems appropriate for heterogeneous fuels. However, the contribution of the very fine 

 fuels to spread rate may be overemphasized. 



Although imprecise field methods prevented extensive analysis of 1^ and related 

 flame length, refinement of Rothermel's model should improve their prediction. Con- 

 tinued development of mathematical models for fire spread and intensity should be pur- 

 sued because they will provide a sounder basis for quantitative fuel appraisal. 



23 



