AI-Embedded Digital Twin for Wildfire

Wildfires emerge as a critical environmental concern, causing significant ecological and economic harm while directly impacting human health through the release of hazardous airborne chemicals. As the frequency and intensity of wildfires continue to rise due to various environmental factors, the importance of early warnings or the ability to forecast these events becomes a crucial tool in our efforts to minimize their devastating consequences. Early warning systems play a vital role in saving lives by facilitating timely evacuations and also serve to safeguard livelihoods, properties, and essential natural resources.In this proposal, we will develop an advanced AI-embedded Digital Twin model with High-Resolution Predictive Models that integrate and use satellite and ground-based data to predict wildfire spread and progression in the hours-to-days ahead as well as the air quality impacts. The Digital Twin model provides real-time probabilistic guidance on fire spread and smoke dynamics to stakeholders and the public. The predictive model is based on advanced deep learning algorithms that integrates NASA satellite-based data with ground-based information, historical wildfire records, meteorological data, and land information to discover fire spatiotemporal patterns and predict fire progression in future. The preliminary results have demonstrated the effectiveness and high performance of the proposed system in predicting wildfire progression with high-resolution. The proposed predictive models and the integrated digital twin system will greatly support firefighters, first responders, and other authorities and personnel in resource allocation, precise wildfire containment, and ensuring the safe relocation of individuals to protect lives. Additionally, the proposed system will inform communities and the public about the most recent and future status of the fire, smoke dynamics and air quality impacts.Dr. Mohammad Pourhomayoun is the principal investigator (PI) of this proposal. PI Pourhomayoun is an associate professor of computer science and the Founding Director of the Artificial Intelligence and Data Science Research Lab at California State University Los Angeles (CSULA). The California State University Los Angeles (CSULA) is a Minority-Serving Institution (MSI) and a Hispanic-Serving Institution (HSI). Dr. Pourhomayoun’s expertise lies in AI, machine learning, and predictive analytics for Earth science applications including wildfires, air quality prediction/management, and climate change. Dr. Pourhomayoun has served as the PI and Co-I on numerous significant research projects in the field AI and data science for wildfire and air quality. He is a Co-I on the NASA funded project “Fire Alarm: Science Data Platform for Wildfire and Air Quality” in collaboration with NASA Jet Propulsion Laboratory (JPL). He is the Co-I on the NASA funded project “Predicting What We Breathe (PWWB)” in collaborations with policymakers in the City of Los Angeles. Dr. Pourhomayoun has published more than 120 peer-reviewed journal and conference papers in his field, and received five best paper awards