The Met Office and University of Exeter have published a paper about an AI-based model they have developed, which has been used to forecast ocean currents in the Gulf of Mexico and which can be adapted to be trained and run on a laptop or desktop.
Machine learning for low-cost offshore modeling
According to the report, the AI-based model is already helping to redefine how new technologies can support marine operational decision-making, with the work recently recognized by the American Society of Civil Engineers (ASCE) Offshore Technology Conference (OTC) Best Paper Award 2025.
The latest paper extends the previous work of the team to show how the machine learning for low-cost offshore modeling (MaLCOM) framework – originally designed for the regional prediction of ocean waves in UK coastal waters – can be adapted and applied to forecasting the currents in the Gulf of Mexico.
The researchers state that what makes this framework particularly exciting is its flexibility and efficiency, having been developed to directly leverage hyper-sparse observational measurements for the basis of its predictions, and its ability to be trained and run using a laptop or desktop computer.
In addition, the architecture of the model also enables easy interrogation of its temporal and spatial behavior, the Met Office an University of Exeter says, which means its characteristics can be better unpicked and explained – building trust in its outputs and providing a path to inform future improvements.
Dr Edward Steele, Met Office IT Fellow for Data Science and the paper’s lead author, said, “AI-based forecasting could revolutionize ocean prediction in a number of ways. Our research shows the exciting potential of very low-cost, observations-driven, AI-based models in delivering promising results, even when constrained by the scope of the available data.”
“Although still at an early stage of refinement – with further development needed to fully realize the anticipated benefits – we show how AI could quickly and reliably support the forecasting of ocean processes, with the work poised to support a range of possible use cases with key offshore energy, marine search and rescue, and defense applications (among others).”
Dr Ajit Pillai, University of Exeter senior lecturer and Royal Academy of Engineering research fellow, commented, “This is an exciting application of the MaLCOM framework to new parameters and new geographical regions, demonstrating the versatility of AI-based approaches and providing new decision-making capability to help offshore safety and workability.”
Development timeline
Further upgrades of the model are planned to further tailor the MaLCOM framework for ocean current forecasting, complementing the major focus of the team’s research on regional wave prediction.
This initial concept of what would later evolve into the MaLCOM framework originally began almost five years ago as an experimental research project led by the University of Exeter, in which the Met Office were supporting participants.
Together, the team worked to develop the approach and benchmark the AI-based model results, achieving parity of performance with the operational Met Office physics-based wave model under typical (non-extreme) conditions at short-range forecast horizons out to 12 hours ahead, prior to testing new approaches, use cases, regional domains and ocean variables. Throughout, they worked directly with industry operators to make sure developments complement their decision-making and workflow.
Dr Steele continued, “This is an example of the benefits of academic, government and industry organizations working together to develop new approaches and capabilities that are useful, usable and used. As well as the collaboration being scientifically stimulating and enjoyable, in rapidly evolving fields such as machine learning, partnerships are particularly essential to realizing the ambition and vision for the use of AI in weather and climate science and services.”
Dr Pillai said, “It is always exciting for research to deliver real impact, and the recognition of the team’s work through the ASCE OTC Best Paper Award 2025 is reflection of some of our progress to date – these are exciting times ahead.”
For more on the Met Office’s work with AI, read the organization’s opinion piece “How will AI weather forecasts make maximum impacts for users?”