Verification of Seasonal Ensemble Forecasts Based on the INM-CM5 Earth System Model

The issues related to the verification of ensemble seasonal forecasts obtained using a new technology implemented on the basis of the INM-CM5 Earth system model are considered. The forecast objects are global gridded ( ) fields of ensemble mean anomalies and probabilities of three gradations of the anomalies (below normal, normal, above normal) for each of the five characteristics: 500 hPa geopotential height ( ), mean sea level pressure (SLP), 850 hPa air temperature ( ), surface air temperature ( ), and precipitation (Prec) for six periods of time averaging (at monthly intervals for the 1st, 2nd, 3rd, 4th months and at two seasonal intervals: season 1 (the 1st–3rd months) and season 2 (the 2nd–4th months)). It is noted that forecast skill scores vary greatly depending on a region, a season, and a meteorological parameter. The best results for all parameters were obtained for the tropics, where the main sources of long-term predictability of the atmosphere are located. In extratropical latitudes, the quality of forecasts, especially at the monthly intervals, decreases and approaches the level of climate forecasts. The skill of precipitation forecasts is significantly inferior to the quality of forecasts of other meteorological parameters. It is noted that the success in ensemble seasonal probabilistic and deterministic forecasting of the main meteorological elements with the INM-CM5 model, both on a global scale and in individual regions, is comparable with the skill scores of similar forecasts of the foreign meteorological centers participating in the LC MME-WMO project. The introduction of the new climate model into the scientific and operational practice of the Hydrometcenter of Russia/North Eurasia Climate Centre will contribute to improving the quality of monthly and seasonal forecasts and developing specialized climate services.