Output 3 - highlight results
Global climate models are best suited to provide information on the scale of broad geographical regions, while users of climate predictions need geographically specific detail to make best use of them. In the CSRP consultation the need for higher spatial detail in forecasts was given high priority.
To provide higher spatial detail, predictions from global climate models need to be post-processed using a technique known as 'downscaling'. Regional Climate Models (RCMs), high resolution limited area climate models, are widely used for downscaling predictions from global models. They use information from the global model to provide the large-scale context and then add skilful high resolution details. By limiting their focus to specific regions, RCMs are able to provide much finer geographical detail without prohibitive increases in computing costs. CSRP is developing the science of RCM downscaling, with a focus on downscaling seasonal forecasts.
Highlight 1: Improved Regional Climate Modelling for Africa
Improvements to the HadGEM3 global climate model achieved in Output 1 have been implemented into the Met Office's Regional Climate Model. System tests with the RCM, in which climate information was generated on scales of 50km, 25km and 12km, (compared to the 135km scale of the global model) have been conducted as well as tests with schemes to represent important factors for Africa such as atmospheric dust loading. Finalised simulations are being contributed to the International Coordinated Downscaling Experiment (CORDEX) of the World Climate Research Programme. The new RCM will form the core of version 3 of the Met Office's PRECIS (Providing Regional Climates for Impact Studies) downscaling system. Work to implement the new RCM in an African centre to downscale seasonal forecasts is in progress.
More information on PRECIS
Significant improvements to the simulation of African rainfall achieved in the global version of HadGEM3 are also found in the regional model. In addition, the higher resolution of the regional model has brought additional performance benefits.
The work has ensured the best available configuration of the RCM is available for implementation in an African centre for 'proof of concept' downscaling of seasonal forecasts. The improved performance of the RCM for Africa will also enhance the quality of our contribution to the CORDEX dataset and our service to developing countries through the PRECIS system.