Project: Longer Time Steps for Weather and Climate Prediction
Supervisors: Hilary Weller (UoR), Pete Sweby (UoR), Colin Cotter (Imperial), Jennifer Scott (UoR) & James Kent (UK Met Office)
Project Description:
There has been recent progress creating implicit time stepping methods for transport that enable longer time steps in idealised models of the atmosphere (eg Weller et al, 2023). However more work is needed to create methods that will enable more efficient weather and climate prediction. Work so far has used simple numerical methods to increase time steps, in simplified models. More advanced numerical methods are needed, drawing on the extensive literature on Runge-Kutta methods (eg Weller et al, 2013). The project will also develop improved methods to combine long time step transport with accurate representation of gravity and acoustic waves and the heating associated with phase changes of water in the atmosphere (associated with precipitation).
The work will start in a simplified setting, writing Python code to solve equations representing idealised versions of transport, waves and phase changes in one spatial dimension. The simplified setting will enable numerical analysis of the new methods. Advanced implicit Runge-Kutta methods will be applied to increase time steps and accuracy. Promising new methods can then be applied in a more complete model of the atmosphere running on a parallel computer, leading towards more efficient weather and climate prediction.
