Biography
Thomas is a Schmidt AI in Science fellow and Senior Research Associated in the Department of Engineering Science. He lectures on information engineering, fluid mechanics, and natural hazards. He joined New College as the W.W Spooner Junior Research Fellow in 2025. He graduated from Bates College with a degree in Physics and Mathematics in 2022. He joined the Environmental Fluid Mechanics and Machine Learning research groups at the University of Oxford in 2022, funded by the Department of Engineering, and completed his DPhil in 2025. For his DPhil work, he won the Osborne Reynolds prize for the best UK doctorate in fluid mechanics and went on to win the Da Vinci prize for the best European doctorate in fluid mechanics.
He leads the Agile Sprint research project on improving flood resilience in the Thames Estuary. This work looks to operationalize his scientific machine learning approach to flood forecasting, RTide, for use by the UK government.
Research Interests
Thomas' primary research is on oceanic forecasting, specifically in the prediction of coastal flooding events. His work develops scientific machine learning methods for studying and predicting sea-levels, with an emphasis on understanding how ML can actually enable richer physical insights to be discovered. You can read more about this work here.
He also is interested in Bayesian machine learning, particularly using variational Bayes and neural methods such as neural processes to characterize the complex uncertainties in real-world systems. Thomas works closely with the UK and Dutch governments to help produce their operational surge forecasts and is also a member of the several international scientific teams studying sea-levels.
Current Projects
Variational Bayesian Harmonic Analysis: Developing a framework for tidal and mean sea surface corrections from and for the Surface Water Ocean Topography mission using a spatially coherent variational Bayesian harmonic analysis (under review JGR: Oceans).
Response Framework: Tidal analysis and prediction through physics-informed ML: A new non-parametric framework for analysis of complex tidal phenomena under external forcing such as storm surge, tidal rivers, and interactions with mean-sea-level.
(under review: ProcRSoc A) Tidal analysis from shortened records: Theoretical analysis of "super-resolution" using newly developed harmonic and Response methods.
Spatial characteristics of nonlinear coastal and estuarine tides from the Surface Water Ocean Topography mission: Leveraging new wide-swath satellite altimetry along with new empirical analysis techniques to study nonlinear characteristics of ocean tides in coastal regions.
(In progress) AutoSSA:A fully non-parametric singular spectrum analysis tool for intelligent signal decomposition and denoising (In progress)
Open Source Code: rtide: Python implementation of the "Response Framework" vtide: Python implementation of the variational Bayesian harmonic analysis
Research Groups
Related Academics
Publications
Tidal corrections from and for SWOT using a spatially coherent variational Bayesian harmonic analysis
Monahan T, Tang T, Roberts S & Adcock T (2025), Journal of Geophysical Research: Oceans, 130(3)
BibTeX
@article{tidalcorrection-2025/3,
title={Tidal corrections from and for SWOT using a spatially coherent variational Bayesian harmonic analysis},
author={Monahan T, Tang T, Roberts S & Adcock T},
journal={Journal of Geophysical Research: Oceans},
volume={130},
number={e2024JC021533},
publisher={American Geophysical Union},
year = "2025"
}
A hybrid model for online short-term tidal energy forecasting
Monahan T, Tang T & Adcock TAA (2023), Applied Ocean Research, 137
BibTeX
@article{ahybridmodelfor-2023/5,
title={A hybrid model for online short-term tidal energy forecasting},
author={Monahan T, Tang T & Adcock TAA},
journal={Applied Ocean Research},
volume={137},
number={103596},
publisher={Elsevier},
year = "2023"
}
