Towards global km-scale climate simulations at the Met Office Hadley Centre

Convection-permitting global climate simulations are seen as a promising route to reducing long-standing biases in conventional global climate models, and thereby may offer more reliable climate projections. 

The Met Office and academic partners have developed a global-regional model hierarchy, comprising several global models that drive multiple nested limited-area models (LAMs), utilising a range of different resolutions and model physics configurations. This framework enables assessment of the upscale impacts of explicit convection and has been run for a year following the DYAMOND3 protocol. 

A novel member of the hierarchy is a convection-permitting global model (grid length of ~5km), which uses a physics configuration employed in regional km-scale NWP and climate modelling at the Met Office. We demonstrate that through tuning of model cloud properties, a realistic top-of-atmosphere (TOA) energy balance is obtained, establishing a suitable configuration for climate change experiments. Initial results show realistic large-scale conditions and improved intensity of mesoscale phenomena relative to the other models in the hierarchy. Finally, we discuss plans for multi-year idealised climate change experiments (Cess-Potter +4K SST) with km-scale global models, aiming to begin to understand how cloud feedbacks differ from conventional global climate models.