AI-Based Surrogate Modelling of Marine Environment in the Baltic Sea

AI-Based Surrogate Modelling of Marine Environment in the Baltic Sea

The PhD topic is developing AI-based emulators for Baltic Sea ocean dynamics, including circulation, tracer transport, heat evolution, biogeochemistry, and ecology. The work involves harmonising marine datasets, identifying key process drivers, developing deep learning models for spatiotemporal prediction, and applying them in Baltic Sea case studies. Applicants should have a relevant Master’s degree, programming skills, and interest in scientific machine learning.

Research description:

The Baltic Sea is a highly dynamic marine environment shaped by strong atmosphere–ocean interactions, river inflows, stratification, coastal processes, and pronounced seasonal variability. Numerical ocean models are essential for resolving circulation, transport, and thermal dynamics in this region, but they are often computationally expensive and operationally demanding at high spatial and temporal resolution. This PhD research will develop advanced AI-based emulators for key physical oceanographic variables and processes in the eastern Baltic Sea, with emphasis on circulation dynamics, tracer transport, and heat evolution.

The candidate will investigate how convolutional, recurrent, graph-based, transformer, and neural-operator architectures can be developed and combined to reproduce the spatiotemporal dynamics of circulation, transport pathways, and thermal variability with high fidelity. Particular emphasis will be placed on integrating data-driven modelling with physics-informed deep learning to improve computational efficiency while maintaining physical realism and predictive robustness.

The research is designed to attract both Earth science and computing-oriented candidates. It offers an opportunity to work on a scientifically meaningful environmental system while developing expertise in neural surrogates, scientific machine learning, uncertainty-aware prediction, hybrid AI-physics workflows, and reproducible research pipelines for large environmental datasets. The position is embedded within the newly established AIMES project, funded by the Estonian Research Council.

Responsibilities and (foreseen) tasks:

• harmonising multi-source datasets, including model forcings, state variables, and validation data, onto consistent spatial and temporal grids;
• identifying the dominant drivers of selected physical targets, such as circulation patterns, tracer transport, pollution dispersion, and heat dynamics;
• designing, implementing, and benchmarking deep learning architectures for spatiotemporal prediction, including advanced hyperparameter optimisation and rigorous evaluation of uncertainty, generalisation, and computational efficiency;
• conducting targeted case studies to address key knowledge gaps in the physical oceanography of the eastern Baltic Sea, with scope for extension to coupled biogeochemical applications;
• contributing to scientific publications, conference presentations, software development, documentation, and collaborative research activities within the team.

Applicants should fulfil the following requirements:

• a Master’s degree in a relevant discipline within computer engineering or computational science, such as data science, artificial intelligence, or applied mathematics, or within Earth sciences, such as oceanography or atmospheric science;
• strong academic performance, with high graduation marks; prior research experience and topic-relevant publications will be considered an advantage;
• strong knowledge of, and clear interest in, scientific machine learning, data-driven modelling, and applied computational methods, together with the creativity to work across disciplinary boundaries;
• good programming skills, preferably in Python, together with familiarity with Linux-based scripting, modern machine learning frameworks such as PyTorch, and high-performance computing environments;
• ability to work independently and as part of an interdisciplinary research team;
• motivation to publish research results in international journals and present them at scientific conferences;
• good command of spoken and written English

We offer:

• supervision by experienced researchers in marine dynamics, coastal processes, and applied AI;
• a fully-funded fixed-term 4-year PhD Early-stage researcher position with a gross monthly salary of 2300 euros; an internationally active and interdisciplinary research environment at the interface of marine science and artificial intelligence;
• access to high-value environmental datasets, computational infrastructure, and collaborative research networks;
• support for conference participation, research visits, and international collaboration;
• training in advanced modelling, reproducible computational workflows, scientific writing, and science communication.

Supervisors:

Main supervisor: Senior Researcher Ilja Maljutenko, School of Science: Department of Marine Systems:

Co-Supervisor: Researcher Mariliis Kõts, School of Science: Department of Marine Systems:

Tallinn University of Technology (TalTech) is an international scientific community with approximately 9,000 students and 2,000 employees; it is one of the largest universities in Estonia, the leading EU country in digitalisation. The university's strengths are broad multidisciplinary study/research interests, a modern research environment, and strong collaboration with international educational and research institutions. TalTech is aiming to be an organisation leading the way to a sustainable digital future.

The Modelling and Remote Sensing of Marine Dynamics team at the Department of Marine Systems, TalTech, is an internationally active research group working in marine science, remote sensing, operational forecasting, and artificial intelligence. The team investigates atmosphere–ocean interactions, marine environmental change, and physical ocean processes using high-performance computing, numerical models, satellite observations, and large-scale environmental datasets.

In recent years, the group has expanded its work on artificial intelligence and machine learning, applying data-driven methods to satellite image processing, marine forecasting, and the analysis of complex ocean simulations. The team provides an interdisciplinary environment linking Earth system science with computational innovation. It also has strong expertise in operational oceanography, supporting public authorities and society with information on water levels, ice conditions, and other marine parameters.

For information about the admission process, please visit the PhD Admission homepage

Applications can be submitted from 14.08.2026 to 13.09.2026