The Inter-Comparison of LANL—FIRETEC and NIST—FDS Idealized Grass Fire Simulations

Advances in computational technology and in the understanding of fire phenomenon since the 1970s have allowed fire scientists to develop a new generation of fire behavior models. These new models employ combustion and geophysical fluid mechanics simulations to explicitly resolve many of the processes that impact fire behavior in realistic conditions of variable fuels, weather, and topography. Two examples of this new generation of models are: 1) FIRETEC developed at Los Alamos National Laboratory; 2) WFDS developed at the National Institute of Standards and Technology.

Both FIRETEC and WFDS employ physical equations of combustion and fluid mechanics to predict fire spread in three-dimensional domains. However, the two models approach the numerical simulation of fire spread and fire behavior using significantly different approaches. The WFDS was developed to simulate wildland-urban interface and structure fires, while FIRETEC was designed to simulate wildland fires in exclusively organic fuel beds. Both models attempt to capture the fire-fuel and fire-atmosphere interactions that drive the overall fire behavior.

The enormous complexity of the physical modeling approaches used in FIRETEC and WFDS and the numerous obstacles to repeatable field-scale fire experiments make evaluation of results from these two models difficult. The research presented in this talk will involve a comparison between the models for the relatively simple scenario of grassland fires on flat terrain. By performing comparisons that address the consequences of the different model approaches and model results, and we will advance our overall understanding of the science upon which these models are based.

Initial simulations will be of fire spread in a homogeneous grass fuel bed and idealized atmospheric conditions. Results from both models will be compared to identify the physical modeling approaches responsible for the differences in the simulations, and to better understand the sensitivities of both models to variations in fuel conditions and combustion parameters. Spread rates, heat release rates, the evolution of the fire front and burned area will be presented from the two models and, when possible, from experiments and operational models. In the event that these comparisons yield insight into the underlying physics of the two models and the processes they simulate, additional grassland fire simulations, motivated by a series Australian experiments in the 1980s, will be conducted.