EAGER: Understanding Wildfire Smoke Transport though Satellite Observations, Machine Learning, and Numerical Modeling

The transport of wildfire smoke is dependent on atmospheric factors, topography and the characteristics of the fire itself. Current numerical forecast models struggle with smoke because the fire can flare up or subside on short time scales and the atmospheric conditions that guide the smoke are on small spatial scales. This project involves the development of a new, observation-based framework using machine learning techniques for smoke modeling that can provide rapid updates to the smoke plume location. The framework can be used for research and as a nowcasting tool for wildfire plumes, which can be a significant health hazard. Improved wildfire plume nowcasting also enables faster, more accurate predictions of smoke and fire behavior, helping more effective firefighting operations, air quality management, and emergency response preparedness. This project is for the development of the GeoCast system, which is an observation-driven, geostationary-satellite-based research and potential nowcasting framework. GeoCast integrates physically interpretable modules with neural network architectures to provide a continuous, high-fidelity view of smoke dynamics and leverages these observations to directly characterize plume behavior, complementing model-based smoke forecast systems and informing improvements in physical parameterizations. In this project, the GeoCast system will be used alongside a high-resolution 4-dimensional observational dataset created through fusion of geostationary and polar-orbiting satellite observations, and simulations from the Weather Research and Forecasting coupled with Chemistry (WRF-Chem) model to examine the relationships of fire intensity indicators and plume rise height with observed smoke transport properties. By linking fire energetics to transport behavior under diverse fire-weather scenarios, the research team will identify critical uncertainties (e.g., fire emission, plume rise height, and land-ocean-atmosphere interactions) in current plume injection calculations. 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.