Simulating seasonal weather influences on wildfire behavior in Glacier National Park, Montana: A comparative analysis and geospatial visualization toolkit

DiBiase, Anthony

Wildfires create a critical role in ecosystem functionality througout Glacier National Park (GNP), but require accurate modeling to mitigate risks to human lives and property. The process of modeling fire behavior is a computationally intensive, multi-scalar effort involving approximation of interactions between wind, climate, fuel sources, and the fire itself; the degree of sophistication in how models approach these phenomena largely determine the projected impacts of a burn. Despite its importance to understanding fire behavior, the most commonly used fire model (FARSITE) does not integrate fire weather feedback. My analysis provides a deeper understanding of the seasonal behavior of fire in GNP by comparing the spread of numerous simulated fires during the height of summer against the end of the fire season in October. To explore the variance caused by each model's treatment of local weather feedbacks, I compare the commonly used FARSITE model-which is easy to use, but relies on steady state temperature and wind inputs-to the performance of the experimental WRF-FIRE model-which requires supercomputing capabilities, but provides the ability to model advanced weather dynamics and feedback loops at multiple spatial and time scales. As an intermediate approach, I added diurnal and orographic wind influences to FARSITE with WindNinja extension. I ran all models for a 24-hour period for two time periods, on 1 July and 20 October 2013, to determine the relative difference in burned area over the fire season. Across all time intervals, the July runs demonstrate a greater area burned than in October, but the magnitude of this variability immensely decreased with models that added complex wind-fire interactions. In addition to reducing seasonal variability, the addition of feedback mechanisms cause WRF-FIRE to predict overall more area burned and a faster rate of spread than with FARSITE. This pattern continues with the addition of diurnal and orographic wind dynamics with WindNinja, generating nearly twice of the total area burned compared to the standard FARSITE model. These results demonstrate that the fire-wind relationship (both via orographic and local-weather feedbacks) is critical for accurately modeling the impact of wildfires, and that fire-weather feedbacks largely override the impacts of seasonal climatic factors in terms of driving the amount of area burned. The results of these simulations provide powerful information to fire managers and ecologists in Glacier National Park, suggesting that models using wind dynamics are essential for understanding the impact of fire in the Northern Rocky Mountains.

Simulating seasonal weather influences on wildfire behavior in Glacier National Park, Montana: A comparative analysis and geospatial visualization toolkit

Collections