Epscor research fellows: NSF: developing a practical physics-informed AI benchmark for estuarine physical-biogeochemical prediction and data assimilation

The National Science Foundation Office of Integrative Activities awarded the University of South Alabama $247,885 on February 1, 2027, under the Integrative Activities program (CFDA 47.083) to support an EPSCOR Research Fellows project developing physics-informed artificial intelligence methods for estuarine prediction and data assimilation.

The recipient will provide a fellowship to a research assistant professor and training for a graduate student to investigate physics-informed AI and machine learning techniques that integrate physical laws with data-driven learning to predict coastal water movement and water quality. The research develops and validates physics-informed neural operator frameworks that emulate computationally intensive partial differential equation solvers for time-dependent advection-diffusion-reaction processes in coastal environments. The work integrates a 50-year high-resolution hydrographic hindcast with a decade of water-quality observations to evaluate the reliability and generalizability of these frameworks for representing coupled physical-biogeochemical processes, improving computational efficiency while preserving physical fidelity and predictive accuracy. The project supports STEM workforce development in AI-enabled marine and environmental science and strengthens Alabama's research capacity in coastal ocean modeling.

Dr. Ruoying He at North Carolina State University collaborates on the research.

Period of performance runs from February 1, 2027 through January 31, 2029, with work performed in Mobile, Alabama.

Abstract

This research infrastructure improvement (RII) epscor research fellows project provides a fellowship to a research assistant professor, and training for a graduate student, at the university of south alabama. This work is conducted in collaboration with DR. Ruoying he at north carolina state university. Through the fellowship, the PI will investigate physics-informed artificial intelligence and machine learning (AI/ML) methods that integrate physical laws with data-driven learning to predict coastal water movement and water quality in real-world environments. The research will bridge advances in AI methodology with practical oceanographic applications by developing rapid, reliable prediction tools for coastal environmental monitoring and resource management. The fellowship will strengthen alabama's research capacity, foster interdisciplinary collaboration at the interface of marine science and artificial intelligence, and support stem workforce development in AI-enabled marine and environmental science. This project will evaluate the reliability and generalizability of physics-informed neural operators (pinos) for representing coupled physical?biogeochemical processes in coastal oceans, advancing trustworthy and adaptive AI methodologies for coastal ocean prediction. The research will integrate a 50-year high-resolution hydrographic hindcast with a decade of water-quality observations to develop and validate pino frameworks that emulate computationally intensive partial differential equation (PDE) solvers for time-dependent advection?diffusion?reaction processes. By embedding physical principles into data-driven learning, the framework will improve computational efficiency while preserving physical fidelity and predictive accuracy. The fellowship will expand the PI's expertise from traditional coastal ocean modeling to physics-informed AI, establishing interdisciplinary capabilities at the interface of oceanography and artificial intelligence. Collaboration with the host institution will transfer knowledge through graduate student training and collaborative research, strengthening the university of south alabama's research capacity in AI-enabled coastal ocean science and contributing to stem workforce development in alabama. This project is supported by the epscor research infrastructure improvement program: epscor research fellows (ERF). The ERF program supports early- and mid-career investigators in eligible jurisdictions to develop collaborations at the nation?s private, government or academic research institutions this award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.