Abstract: |
We describe a method to produce probabilistic fire spread predictions which can be used to allocate firefighting resources optimally according to a specified decision rule. The fire spread prediction system utilizes a high resolution weather model, the FARSITE fire behavior modeling system, and a probability model for the errors of the predicted fire spread. The output includes the predicted fire perimeter and associated error bounds that delimit the location of the actual fire perimeter with specified probability. With data that describes the economic value of the resources in the area of interest, it is therefore possible to compute the expected cost plus net value change, and therefore to allocate firefighting resources to minimize that quantity, if that is the desired optimization strategy. On the other hand, trade-off analyses can be performed to evaluate the consequences of different allocation strategies in terms of the area protected resource losses and fire fighting resource allocation efficiency. |
