Abstract

As computer resources and power increase, more climate modelling centres around the world are testing the value of increasing resolution in climate models in the hope that their simulations will become more realistic and provide more trustworthy climate predictions and projections. However, these models require very substantial resources to be operated and analysed, while, for many aspects of the climate system, we still do not fully understand the effect of resolution on climate simulations. As our community is already planning a move towards high-resolution Earth System modelling, it is imperative that we focus on the fundamental uncertainties in the physical aspects of climate models, particularly the atmosphere-land surface-ocean-sea ice components. In the JWCRP High-Resolution Global Climate Modelling group, a long-term collaboration between the Met Office and the National Centre for Atmospheric Science in Reading, we use high-resolution global models to better understand the mean and variability of climate features, their teleconnections and emerging processes. The first part of this seminar will introduce where we are in terms of high-resolution global climate modelling in the UK, as well as the opportunities such models offer. By highlighting several examples from atmosphere-only and coupled atmosphere-ocean global simulations, focus will be put on the features that do depend on resolution such as tropical cyclone variability, ocean-to-land atmospheric moisture transport, organisation of convective events, while counter examples will also be shown, such as blocking events and atmospheric rivers for which the role of resolution is not clear yet. These results highlight that the main challenge that remains in high-resolution global climate modelling (more important than supercomputer resources and data storage) is the need for a community approach for data analysis to assess confidence in high-resolution simulations. This will be discussed in the last part of the seminar.