Cellular Automata Fire Spread Modeling

Abstract

Large-scale land and structure losses each year due to both small and large wildfires, and the longer fire seasons due to anthropogenic climate change make the prediction of fire spread a matter of concern. There are multiple physical fire spread models, but a Bayesian categorical model can also be used to determine which factors are particularly important when predicting fire spread. To study this a multi-state cellular automata model was used to predict the spread of a controlled wildfire. Information like land usage, wind speed, and other local covariates along with information from neighboring cells was used in the prediction. The model takes into consideration how the rate of spread varies both over time and through space.

Nicholas Grieshop

Graduate Student

My research interests are spatio-temporal statistics.