A Modified Level Set Framework for Physics and Uncertainty Informed Wildfire Spread

A wildfire refers to an uncontrollable fire that spreads through vegetation, such as forests and grasslands, and poses a serious risk to ecosystems, infrastructure and human lives. Because wildfires are common throughout the world, it is very important to understand and model their spread so that we can better prepare our communities for their potential consequences. Most existing fire models, which were developed primarily, focus on only a few variables (i.e., wind velocity, slope, or Rothermel spread rate), making them less reliable in replicating how a wildfire propagates in reality. Therefore, this research presents a modified model for estimating the spatial extent of wildfires based on the use of the Level Set method and uncertainty. The factors considered in the modified model include wind speed and direction, slope, fuel type, moisture content, slope corrections, and uncertainty; and the slope corrector also considers slope steepness and tightness of the fuel pack in its definition, while the uncertainty considers the actual distance travelled from the ignition point. All computations were performed in MATLAB; and the shape of the fire was studied over time. Future studies will include detailed research on complicated ignition geometries, the effect of heterogeneity in the environment, and the impact of uncertainty on fire evolution, thus improving the dependability and utility of our model to predict and controlling wildfires. The study results demonstrate that the Level Set model can correctly depict the evolution of a fire's boundary.