Wildfires are increasing in both frequency and intensity, posing escalating risks to lives, property, air quality, infrastructure, and public health. This project investigates how the integration of physics-based and AI-driven models can enhance forecasting capabilities within a Wildfire Digital Twin (DT) framework for predicting wildfire progression and associated air quality impacts. The proposed DT will integrate ground-based observations, satellite data, and physics-based models within a unified spatiotemporal system. AI-enabled models will be tightly coupled with physics-based atmospheric and chemical transport models, including WRF-Chem, to improve forecast accuracy and to assess wildfire smoke dispersion and air quality impacts. The system will generate probabilistic forecasts with explicit uncertainty quantification that dynamically update as new observations become available. These uncertainty-aware forecasts will support emergency response operations, evacuation planning, and public health decision-making during active wildfire events.
