Final Technical Report MuSIKAL: Multiphysics Simulations and Knowledge Discovery through AI/ML Technologies

Summary

The MuSiKAL project, funded by DOE Award Number DE-SC0022211, developed a coastal digital twin platform integrating diverse data, multiscale models, and scientific machine learning (SciML) for predictions of storm surge and heavy rainfall impacts on the U.S. Gulf Coast. The project, led by the University of Texas at Austin, validated its algorithms against historical hurricanes including Ike (2008), Harvey (2017), Ida (2021), and Ian (2022). Key accomplishments include identifying a tropical cyclone decay threshold of approximately 33 m/s and demonstrating computational savings of about 21 seconds with 0.7 cm MAE for flood mapping using hybrid ML models. New real-time forecasting capabilities with up to a 7-day horizon for tidal forecasts, trained on 3-year hindcasts, are planned for NOAA's operations by 2026.

Key Facts

Source Document

https://example-government.gov/policy-document-link