Recent Developments in Global Ensemble Prediction at the Met Office

The Met Office has recently upgraded their global ensemble system (MOGREPS-G) to an ensemble of data assimilations (EDA). This new scheme uses a 4d ensemble variational scheme on an ensemble of 44 members which includes inflation schemes and a costefficient mean-pert minimization method. Static and flow-dependent background error covariances are combined in a hybrid approach. Model error is handled by SKEB and SPT stochastic physics schemes, and we introduce a new approach called Additive Inflation. This paper describes the new ensemble configuration and shows a comparison with the previous ETKF system.