8.7 The Colorado Fire Prediction System: system description and evaluation

Jimenez, Pedro A., Branko Kosovic, Domingo Munoz-Esparza, Amanda Anderson, Bill Petzke, Jennifer Boehnert, Jim Cowie, John Exby, Kevin Sampson, Robert Sharman, Barbara Brown and William Mahoney, National Center for Atmospheric Research

Wildland fires are responsible for large socio-economic impacts. Fires affect the environment, damage structures, threaten lives, cause health issues, and involve large suppression costs. These impacts can be mitigated via accurate fire spread forecasts to inform the incident management team. To this end, the state of Colorado is funding the development of the Colorado Wildland Fire Prediction and Decision Support System. The system is based on the Weather Research and Forecasting (WRF) model with a fire behavior module based on the Coupled Atmosphere Wildland Fire Environment (CAWFE) model (i.e., WRF-Fire).

This presentation will describe the Colorado Fire Prediction system (CO-FPS) and its evaluation. The CO-FPS has a web interface that allows the user to configure and launch fire forecasts. The simulations are based on WRF-Fire configured at fine horizontal resolution (111m) to explicitly resolve atmospheric turbulence via large eddy simulations. The fire module has been upgraded to include improved parallelization of MPI tasks, a higher order scheme to advect the level set method that tracks the fire, re-initialization of the level-set to improve the stability of the fire propagation, initialization of fires from observed fire perimeters, as well as prediction of flame length and smoke. The forecasts are displayed via the web interface using a geographical information system. The forecasting system is evaluated against fires during the 2016 season. Results indicate the value of the CO-FPS to assist fire analysts to minimize the adverse impacts of wildland fires.