Validation of Models on "Forest Fire Emergency Measure Simulation":

A Case Study of Kagamihara City.

Abstract

The appropriateness of Forest Fire Emergency Measure Simulation, which was developed for planning fire prevention measures, was examined based on a sensitivity analysis conducted on a forest fire which occurred in Kagamihara City. For the sensitivity analysis, one-way analysis of variance, Bonferroni multiple comparison and L16 (2¹⁵) orthogonal array variance analysis were applied to each factor. An examination of the area consumed by the fire, the rate and direction in which the fire spread and spotting numbers revealed that the most effective factors were as follows : the rate of fuel fire spread, wind direction, wind velocity, and mean stand height. This suggests that the data of spotting position was quite important for calculations related to the forest fire. An examination of appropriateness revealed that the utilization of average observed values for the calculations were the cause of error increment. Utilizing observed maximum values for the calculations was appropriate to a certain degree, but it was plagued with delays in spotting occurrence and undervaluation of the fire spread rate. To improve its compatibility, the following measures were suggested : re-examination of spotting parameters, incorporating the effect of intense heat on changing wind patterns, and examining of the rate of fuel fire spread. As it currently stands, this simulation can only be utilized for suggesting fire brigade allocations and locations for forest roads and fire breaks.