The risk of wildfires has increased significantly in recent years and touched communities not previously at high risk. Effective mitigation of wildfire risk is essential to reduce the potential for catastrophic losses. Accurate assessment of the risk of wildfire on a parcel-by-parcel basis will enable fire departments and homeowners to effectively triage and plan to reduce the risk. We will develop a Wildfire Integrated Modeling, Prediction, and Learning Environment (WIMPLE), a hybrid AI tool for wildfire risk assessment. WIMPLE is based on our Scruff AI framework, which provides integration of different kinds of AI models, sharing and composition of models, with spatiotemporal flexibility in model composition. We will demonstrate WIMPLE by developing a new wildfire risk assessment method that integrates multiple model components such as fire propagation and climate models at different spatial and temporal scales, as well as learning from historical data. We provide a decision-support UI using explainable AI techniques to ensure that predictions and recommendations of WIMPLE can be understood and trusted by users.
WIMPLE will link work being done at NASA with the end-user community to support decision making about wildfire risk triage and mitigation. Using sources such as NASA Earth Observatory and NASA Visible Earth, and climate models such as the GISS GCM, WIMPLE will provide an avenue for these sources to directly support critical environmental decisions.
WIMPLE will support wildfire risk assessment at low cost for homeowners working with fire departments, for example through programs such as Marin County’s Community Wildfire Protection Plan (CWPP). WIMPLE will enable more rapid and proactive triage and mitigation of wildfire risk than current approaches.