Multi-Fidelity AI Foundation Model for Coupled Surface-Groundwater Predictions of Water Availability and Flood Hazards Across CONUS

A new AI foundation model blending multi-scale hydrologic simulations with real-time data will deliver fast, high-resolution predictions of water availability and flood hazards to support energy, infrastructure, and community resilience nationwide. Water availability and flood risks are major concerns for energy production and infrastructure resilience across the United States. Predicting these challenges accurately is difficult because current large-scale hydrologic models are often too computationally expensive to provide local detail at a continental scale and tend to oversimplify the complex interactions between surface water and groundwater. Many also do not fully represent human influence such as groundwater extraction, degrading the predictability of water availability in high-demand regions. This makes it hard for communities, industries, and energy providers to plan for droughts, floods, and changing water needs.

Argonne National Laboratory