Kopera wins fellowship to advance ocean modeling with artificial intelligence

Michal Kopera, associate professor of mathematics and director of the NUMO Lab at Boise State University

Michal Kopera, associate professor of mathematics at Boise State University, has been awarded a National Science Foundation (NSF) EPSCoR Research Fellowship to support research at the intersection of scientific computing and machine learning.

MIT collaboration advances ocean forecasting

As part of this award, Kopera will spend a portion of his upcoming sabbatical, beginning in summer 2026, at the Massachusetts Institute of Technology in the Multi-Scale Estimation and Assimilation (MSEAS) Laboratory. There, he will collaborate with Professor Pierre Lermusiaux, a leading expert in ocean modeling and data-driven prediction.

The project focuses on developing machine learning algorithms for adaptive mesh refinement - advanced computational methods that dynamically adjust the resolution of numerical simulations based on need. Similar to how a contemporary camera might automatically center on the most important part of a scene, these smart simulations would focus computational resources on the most critical areas of complex models. By training artificial intelligence to focus on the more important parts of vast models like those used for ocean studies, researchers can yield more relevant information more efficiently.

"Ocean modeling is important because the ocean covers two-thirds of the Earth's surface," Kopera said. "That's a massive system that's just there, and it's not sitting idly. It's really controlling, like an AC system for the Earth. So if it's changing how it's working, what we feel as the weather is going to change too."

Implications for weather forecasting

Conceptual sketch of Adaptive Mesh Refinement (AMR) driven by a neural network deployed on an ocean simulation

By integrating data-driven decision-making, this research aims to improve accuracy and cost-effectiveness of complex system modeling simulations used in ocean circulation and environmental forecasting. This work has important implications including improved understanding of how oceans influence weather patterns and more accurate and timely weather predictions.

There is great potential for predicting ocean currents, climate patterns like El Niño, and coastal hazards like tsunami waves and hurricane-induced storm surges. Because these events threaten lives and property, timely forecasts are critical for coastal community safety. The techniques developed in this project could also be applied in other fields from engineering to astrophysics.

Students join cutting-edge research

From left to right: Antone Chacartegui, Michal Kopera and Hailey Stubbers.

The fellowship builds on Kopera's work as director of Boise State's Numerical Modeling (NUMO) Laboratory, which uses data-driven modeling to solve complex problems across disciplines.

NUMO Lab students have tackled problems ranging from predicting March Madness results to modeling the expansion and contraction of the universe.

"This is really like the space for students to show up with their ideas and not be afraid," Kopera said. "We have students from different disciplines come in and say, 'Hey, I have a science problem which needs some math'. And I love doing that. It's so much fun because they and I learn so much."

Students involved in the NSF fellowship include Antone Chacartegui, from the computing doctoral program, and Hailey Stubbers, from the mathematics master's program. Chacartegui will join Kopera at MIT for part of the visit, contributing directly to the collaborative research.

This fellowship will advance fundamental research in computational mathematics and establish collaboration between Boise State and MIT. This serves to strengthen ties between the NUMO Lab and the MSEAS group and create opportunities for student involvement in cutting-edge research.

This publication was made possible by the NSF Idaho EPSCoR Program and by the National Science Foundation under award number OIA-2531614.

Share This

• Share on Facebook
• Share on X
• Share on LinkedIn
• Share through Email