In this study, we investigate the impact of Multi-Center Grand Ensemble (MCGE) forecasts, consisting of three operational ensemble forecasts by the Japan Meteorological Agency (JMA), the National Centers for Environmental Prediction, and the Canadian Meteorological Center. We verified the skill of MCGE forecasts in comparison with that of JMA ensemble forecast using root mean square error, anomaly correlation, and Brier skill score for 500 hPa geopotential height and 850 hPa temperature in the Northern Hemisphere (20°N-90°N) in September 2005.
Our results show that MCGE forecasts are more skillful than single-center ensemble forecast without considering weight among ensemble members and bias corrections. This implies that considering weight or bias corrections may result in further improvement of MCGE forecasts, specifically in probabilistic forecasts.