Supercomputing the Seasonal Weather Prediction

Most of the WMO Global Producing Centres for Long-Range Forecasts use ensemble prediction systems together with an ensemble re-forecast dataset, also called hindcast. Both ensembles usually come from integrations of a global coupled atmosphere, ocean, land-surface and sea ice model. This article describes SLAV–INMIO–CICE coupled model developed at Marchuk Institute of Numerical Mathematics RAS, Shirshov Institute of Oceanology RAS and Hydrometeorological Centre of Russia. The model components are coupled using the own-developed Compact Modeling Framework. SLAV–INMIO–CICE model requires several thousands of cores of CRAY XC40 system. The performance of the model is comparable to the latest European Centre for Medium-Range Weather Forecasts SEAS5 operational model.