Combining observations and machine learning to infer ocean turbulence
Supervisors: Laura Cimoli and Robert Rouse
Ocean turbulence plays an important role in sustaining the global ocean circulation and in the mixing of heat, carbon, nutrients, and other tracers within the ocean interior. Turbulent mixing is technically challenging to measure and is often inferred from measurable quantities (such as temperature (T), salinity (S) and depth (Z)) using parameterisations that are based on numerous simplifying assumptions about the physics of turbulence.
Previous work (Mashayek et al., 2024) has shown that machine learning models can be trained on the existing turbulence data and be used to infer turbulent mixing from observations of T, S, Z measured routinely by global observational programs.
In this project, we aim to expand that work, and apply the trained algorithm to existing datasets. In particular, the intern will employ the trained model to infer the mixing produced in the wake of a large iceberg by using observations recently collected with autonomous vehicles in West Antarctica.
Recent studies showed that iceberg calving can lead to large changes in ocean stratification and biogeochemical cycles (Lucas et al., 2025), but such impacts are largely unrepresented in climate models. Hence, the estimate calculated here will be useful to have a first assessment of the net contribution of iceberg-driven mixing to the Southern Ocean stratification and water mass properties.
The intern will work with Dr Cimoli and Dr Rouse, who will provide complementary expertise in physical oceanography/ocean turbulence and machine learning modelling. The project will use publicly available code and datasets.
Working environment
The student will work on their own and with both supervisors. We expect the student to have weekly meetings with either one of the supervisors or both, depending on the phase of the project. We generally find it good practice for students to work in the Centre for Mathematical Sciences during normal office hours, though some remote working is fine, so long as the student is still available for a weekly virtual meeting.
The student will have the option of attending the ICCS Summer School.
References
Mashayek et al., 2024; Deep Ocean Learning of Small Scale Turbulence, published in Geophysical Research Letters
Lucas et al., 2025; Giant iceberg meltwater increases upper-ocean stratification and vertical mixing, published in nature geoscience.
Essential knowledge, skills and attributes
- Background in physical sciences or computing sciences (eg. mathematics, physics, earth sciences, computing)
- Proficiency in Python
- Prior knowledge in fluid dynamics/physical oceanography is desirable but not essential
- Ability to work independently as well as in a team.
Contact us
If you have questions about this project, email Laura Cimoli