Data-driven physical models for weather prediction

Home > Data-driven physical models for weather prediction

Data-driven physical models for weather prediction

Dr Lisa Kreusser, Department of Mathematical Sciences, University of Bath

Partner: Met Office

Accurate weather forecasts are of great importance for predicting the occurrence of extreme weather events such as floods, droughts, heat waves, cold snaps, and hurricanes, as these events can have a massive socio-economic impact on agriculture, transport, and energy use and production. While models based on physical laws have been used for decades for weather prediction, the use of machine learning has exploded in recent years. This led to a significant interest in the effectiveness of data-driven in addition to model-based methods for weather forecasts. My research aims to combine model-driven forecasts with weather data for accurate weather forecasts.