Improving weather forecast skill through reduced precision data assimilation
I studied physics at the University of Bristol. Through my final year project, which involved modelling the dynamics of knotted polymers, I became interested in programming. I further developed this interest with a year in software development. I have also worked in Earth Sciences, University of Bristol and at Surrey Space Centre on geophysical gravity surveys and space weather, respectively.
My research concerns data assimilation, the statistical process by which observations are incorporated into weather forecasting models. I will be investigating how to make this process more efficient by reducing the numerical precision of the underlying code. It is hoped that such a reduction will not impact the quality of the assimilation product, while making available additional computational resources that can be used elsewhere.
Sam Hatfield, Aneesh Subramanian, Peter Düben, Tim Palmer, Improving weather forecast skill through reduced precision data assimilation, Monthly Weather Review, accepted for publication