Description: A German national project coordinates research on improving a global decadal climate prediction system for future operational use. MiKlip, an eight-year German national research project on decadal climate prediction, is organized around a global prediction system comprising the climate model MPI-ESM together with an initialization procedure and a model evaluation system. This paper summarizes the lessons learned from MiKlip so far; some are purely scientific, others concern strategies and structures of research that targets future operational use. Three prediction-system generations have been constructed, characterized by alternative initialization strategies; the later generations show a marked improvement in hindcast skill for surface temperature. Hindcast skill is also identified for multi-year-mean European summer surface temperatures, extra-tropical cyclone tracks, the Quasi-Biennial Oscillation, and ocean carbon uptake, among others. Regionalization maintains or slightly enhances the skill in European surface temperature inherited from the global model and also displays hindcast skill for wind-energy output. A new volcano code package permits rapid modification of the predictions in response to a future eruption. MiKlip has demonstrated the efficacy of subjecting a single global prediction system to a major research effort. The benefits of this strategy include the rapid cycling through the prediction-system generations, the development of a sophisticated evaluation package usable by all MiKlip researchers, and regional applications of the global predictions. Open research questions include the optimal balance between model resolution and ensemble size, the appropriate method for constructing a prediction ensemble, and the decision between full-field and anomaly initialization. Operational use of the MiKlip system is targeted for the end of the current decade, with a recommended generational cycle of two to three years.

Description: Low prediction skill in the tropical Pacific is a common problem in decadal prediction systems, especially for lead years 2–5 which, in many systems, is lower than in uninitialized experiments. On the other hand, the tropical Pacific is of almost worldwide climate relevance through its teleconnections with other tropical and extratropical regions and also of importance for global mean temperature. Understanding the causes of the reduced prediction skill is thus of major interest for decadal climate predictions. We look into the problem of reduced prediction skill by analyzing the Max Planck Institute Earth System Model (MPI-ESM) decadal hindcasts for the fifth phase of the Climate Model Intercomparison Project and performing a sensitivity experiment in which hindcasts are initialized from a model run forced only by surface wind stress. In both systems, sea surface temperature variability in the tropical Pacific is successfully initialized, but most skill is lost at lead years 2–5. Utilizing the sensitivity experiment enables us to pin down the reason for the reduced prediction skill in MPI-ESM to errors in wind stress used for the initialization. A spurious trend in the wind stress forcing displaces the equatorial thermocline in MPI-ESM unrealistically. When the climate model is then switched into its forecast mode, the recovery process triggers artificial El Niño and La Niña events at the surface. Our results demonstrate the importance of realistic wind stress products for the initialization of decadal predictions