Abstract:

The European Centre for Medium-Range Weather Forecasts (ECMWF), National Center for Environmental Predictions (NCEP), and Canadian Meteorological Centre (CMC) ensemble forecasts and ERA-Interim reanalysis data were analyzed to understand the skill of the circulation system in Eurasian middle-high latitude in summer based on the anomaly correlation coefficient, root mean square error, and relative operating characteristic curve method.The results show that it captures the observed features of the best applicability for the ECMWF ensemble forecast and the worst applicability for the CMC ensemble forecast, but underestimates applicability showing a decreasing trend with increasing forecast lead time.A skillful forecast can be obtained up to 192 h for the ECMWF ensemble forecast, 180 h for the NCEP ensemble forecast, and 168 h for the CMC ensemble forecast.The ensemble-averaged forecast is more reliable than the control forecast, and the control forecast is slightly higher than the other members.For available predictive timeliness, the ensemble-averaged forecast is 12 h longer than the control forecast and 36 h longer than other members.There are obvious differences in available predictive timeliness and accuracy of the same numerical forecast model at different starting times.Although the predictive effects of the blocking high and the West Pacific subtropical high are periodic in the study period, the rate of deterioration in forecast ability slows down after 240 h in the blocking high areas, there is even an increasing trend in the Ural blocking high and the prediction effect is the best near the 300 h.The available predictive timeliness of the West Pacific subtropical high is lower than that of the blocking high, but its accuracy is obviously higher than that of the blocking high.

Key words: Middle-high latitude, Circulation system, Ensemble forecasting, Prediction performance