This project will use high-resolution models in combination with machine learning algorithms to improve the predictions of the behavior and spread of wildfires. Variables such as weather forecasts, the type of fuel burned, the topography of the landscape, and the extent of firefighting efforts will be included in the analyses. This effort is designed to improve the understanding of fire behavior by taking advantage of data from recent field campaigns and from higher-resolution satellites with enhanced capabilities.
Air quality simulations using the Weather Research and Forecasting with Chemistry (WRF-Chem) model driven by emission predictions will be performed for multiple fire events. Emissions will be predicted using the WRF model together with a fire spread model based on the Coupled Atmosphere-Wildland Fire Environment model, together known as WRF-Fire. Data from the two most recent summer fire seasons (July-September 2018 and 2019) over western North America will be included in the analyses.
This research will address the following science questions: (1) What methodologies can be used to effectively forecast biomass burning emissions in the context of air quality forecasts? Can these methodologies outperform persistence forecasts? (2) How skillful are these methodologies when evaluating air quality predictions driven by these emissions against ambient observations of smoke? (3) What factors (e.g., fire size, type of fuel, weather conditions, topography) control the skill of the methodologies? (4) At what time scales can the emissions be accurately predicted by these methodologies? Hourly? Daily? Up to how many days/hours in advance? (5) Can these methodologies provide information on the injection height of smoke and how do they compare to approaches previously developed?
This award reflects NSF’s statutory mission and has been deemed worthy of support through evaluation using the Foundation’s intellectual merit and broader impacts review criteria.