TerraHydro: A Community Digital Twin for AI-Driven Hydrologic Prediction and Decision Support

Presenter: Craig Pelissier
Organization: NASA GSFC
Co-Authors: Carlos Cruz, Deepthi, Raghunandan, Mahmoud Saeedimoghaddam, Arun Saranathan, Scott Sinno, Brandon Smith

Abstract

Recent advances in cloud computing, artificial intelligence, Earth observation, and web technologies have enabled a new generation of Earth Information Systems (EISs) that are near real-time, adaptable, accessible, and interactive. These capabilities have accelerated the development of Earth System Digital Twins (ESDTs)—persistent, service-oriented information systems that are continuously updated, configurable, and designed to provide far more than static public data products. By delivering modeling, analysis, and decision-support capabilities through open, cloud-based platforms, ESDTs advance Open Science while supporting NASA's Earth Science to Action initiative.

TerraHydro is a NASA Earth Science Technology Office (ESTO)-funded project developing a community AI-enabled ESDT for terrestrial hydrology that predicts the evolution of Earth's hydrologic state, quantifies its societal impacts, and delivers actionable information to researchers, stakeholders, and decision makers. TerraHydro combines fast, flexible AI-based models and surrogates with scalable cloud computing hosted on NASA's Science Cloud and an intuitive web-based interface. Ongoing integration with NASA's Hydrology Copilot will provide a natural language interface for configuring simulations, executing workflows, and visualizing results.

TerraHydro currently generates daily global hindcasts, nowcasts, and forecasts of key hydrologic variables, along with selected future climate scenario projections. The platform is also being extended through collaborations with industry and government partners to assess the impacts of Amazon deforestation on Brazil's agriculture, energy, and economic sectors and to support NASA's contributions to the Famine Early Warning Systems Network (FEWS NET).