Analysis and Evaluation of Forecast Jumpiness Characteristics in EC-EPS and EC-DERF

Based on the European Centre for Medium-Range Weather Forecasts Ensemble Prediction System (EC-EPS) and the Dynamical Extended Range Forecast (EC-DERF), this study systematically evaluates the forecast stability of the 500 hPa geopotential height field over the Northeast Cold Vortex activity region (Region B: 118°E~150°E, 25°N~60°N) and the Southwest Vortex activity region (Region A: 70°E~130°E, 20°N~40°N). The analysis focuses on the performance of ensemble mean and control forecasts across different seasons and lead times, as well as the influence of vortex systems on forecast uncertainty. The main findings are as follows: (1) The stability of the EC-EPS ensemble mean forecast is generally better than that of the ensemble control forecast. Region B shows higher forecast stability in the annual mean and summer compared to Region A, whereas Region A exhibits better stability in winter. (2) In terms of jump frequency, single-point jumps occur most frequently, while triple-point jumps with alternating signs are the least common. The ensemble mean forecast shows lower jump frequency than the control forecast at longer lead times, and Region B has higher long-lead jump frequency than Region A. Beyond 120 hours, the differences in jump frequency between ensemble control and mean forecasts become more pronounced in both regions. (3) Within 120 hours, the Northeast Cold Vortex and Southwest Vortex have little influence on the forecast jumpiness index. Between 120–240 hours, the ensemble mean jumpiness index in Region B is slightly lower when the Northeast Cold Vortex is absent than when it is present, while the opposite is observed in Region A. After 240 hours, the ensemble control jumpiness index in both regions is slightly higher when a vortex is present, indicating that both vortex systems increase the instability of the geopotential height field in the medium to extended range. (4) For EC-DERF, the jumpiness indices from Monday and Thursday initializations are generally similar. The ensemble mean forecast is more stable than the control forecast, and summer forecasts are less stable than those in other seasons. Region B shows higher summer forecast stability than Region A. This study reveals the stability characteristics of ensemble forecasts across regions and seasons, providing a scientific basis for improving the accuracy and reliability of meteorological prediction systems.