The Effects of Fire-Atmosphere Coupling on Predicting Fire Front Behavior

The complex interactions between atmospheric and fire-induced winds are a persistent obstacle to accurately predicting wildfire front behavior. There are a multitude of wildfire spread models, with one primary distinction being the level of fire-atmosphere coupling in each. Coupling of fire-induced winds and ambient winds in numerical models is carried out through linking the heat and mass fluxes from the wildfire with the surface energy fluxes in the atmospheric model. A tradeoff in numerical wildfire models is that highly coupled models are computationally expensive, while computationally efficient codes lack important features of both fire-induced flows and fire-atmosphere interactions.

We examined the effects of various levels of coupling on rate of spread in an idealized homogeneous grassland fire over flat terrain. The levels of coupling included an uncoupled model, a model coupled through parameterizations in a fast-response framework, and a model coupled through the full prognostic momentum and energy equations using large eddy simulation. A simplified physics rate of spread model was used in each of the coupling regimes, where only the level of fire-atmosphere coupling varied. It was found there was a non-linear relationship in the error between models with forward ROS when the fire-induced winds were within 20% of the ambient winds. This non-linear relationship shows a clear need to capture the key details in the feedback between fire-induced and ambient winds.