Thomas Wilder

I am a research scientist at the University of Reading within the National Centre for Atmospheric Science interested in machine learning and climate modelling. In my current role within the EU Horizon project (AI4PEX), I am applying various machine learning techniques to improve current mesoscale eddy closures, with the aim of producing more realistic ocean dynamics.

My primary research focuses on the development of mesoscale eddy parameterisations. In this area I have explored how the role of different viscosity closures, such as 2D and QG Leith, can modulate the ocean circulation (see this paper). I am currently working on a machine learning closure for the GEOMETRIC eddy parameterisation at the eddy-permitting resolution (see ML code here) by applying a U-Net model. Much of this work directly feeds into the enhancement of the NEMO ocean model, and for future versions of the UK Earth System Model (UKESM).

Beyond the parameterisation world, I am keen to explore the future development and application of climate emulators.