Supervisory Team • Profs. Samar Khatiwala (Earth Sciences, University of Oxford) and Coralia Cartis (Maths Institute, University of Oxford)
Numerical models are one of the key tools for understanding the climate system and predicting its future evolution under anthropogenic forcing. However, as these models get more complex, they also become far more computationally expensive, limiting scientists’ ability to perform simulations and make more effective use of them. It also increases the need to better calibrate the models against data. To addresses these challenges, one or more projects is available to develop novel mathematical and numerical techniques, and apply them to a range of problems in ocean and climate science. The ultimate goal is efficient and scalable software implementations that can be used in state-of-theart climate models of the kind participating in the IPCC assessments.
The student(s) will be co-supervised by Profs. Samar Khatiwala in Earth Sciences and Cora Cartis in the Mathematical Institute. While the specific project will depend on the interests of the student, topics include: Newton-Krylov and vector sequence acceleration methods; fast PDE solvers; derivative-free parameter optimization; surrogate modelling and low rank approximations of complex systems via linear sketching; and machine learning. Of particular interest are applications to oceanography, marine biogeochemistry, and climate modelling, for example, ocean heat and carbon uptake, the ocean biological pump, and climate sensitivity.
Students with a strong first degree in natural sciences, mathematics, computer science or other engineering fields are encouraged to apply. See the full advert below for more information.