Generating Stochastic Ellipsoidal Forest and Wildland Fire Scar Scenarios for Strategic Forest Management Planning under Uncertainty

Abstract

Explicit consideration of the possibility of wildfires enhances strategic planning for landscape management; however, incorporating the stochastic processes that govern fire occurrence and spread in optimization models that are used to inform landscape management presents new challenges. In this article, we describe a method of generating spatially explicit fire scenarios in the form of elliptical fire scars with associated probabilities. Computational experiments demonstrate that good scenarios can be achieved using currently available computer technologies. We also describe how more complex fire ignition and growth process models can be used to generate more realistic fire pattern scenarios and provide insight into the quality and effort required for various approximations. We conclude by discussing how our methodology could be used to perform some of the tasks for which fire managers use traditional burn probability models.