Analyses on the Multimodel Wind Forecasts and Error Decompositions over North China

In this study, wind forecasts derived from the European Centre for Medium-Range Weather Forecasts (ECMWF), National Centers Environmental Prediction (NCEP), Japan Meteorological Agency (JMA) and United Kingdom Office (UKMO) are evaluated lead times of 1–7 days at 10 m multiple isobaric surfaces (500 hPa, 700 850 hPa 925 hPa) over North China 2020. The straightforward multimodel ensemble mean (MME) method is utilized to improve forecasting abilities. addition, forecast errors decomposed further diagnose error sources forecasts. Results indicated that there little difference in performances four models terms direction (DIR), but obvious differences occur meridional (U), zonal (V) speed (WS) Among them, ECMWF NCEP showed highest lowest abilities, respectively. MME effectively improved more evident superiorities higher levels longer times. Meanwhile, all manifested consistent trends increasing (decreasing) U, V WS (DIR) with rising height. On other hand, main source both was sequence component (SEQU), which rose rapidly deficiency less proficient model could mainly be attributed bias (BIAS) SEQU, Furthermore, tended produce lower SEQU than layers, However, a slight reducing BIAS distribution errors. results not only recognized detail, also provided important references use business departments associated scientific researches.