提高月预报业务水平的动力相似集合方法

针对基于大气环流模式的月预报问题,提出了一种能有效减小预报误差并提高预报技巧的动力相似集合预报新方法。该方法着眼于动力模式与统计经验的内在结合,在模式积分过程中通过提取大气环流历史相似性信息,对模式误差进行参数化处理,形成多个时变的相似强迫量来扰动生成预报的集合成员。将这一集合新方法应用到中国国家气候中心业务大气环流模式(BCC AGCM1.0),一组10 a准业务环境下回报试验结果显示,相比于业务集合预报,动力相似集合预报方法能有效改进模式对于大气环流的纬向平均、超长波和长波预报,从而有效提高了月平均环流预报技巧(几乎达到业务可用标准)和逐日环流预报技巧,并显著降低了预报误差,合理增加集合离散度,使二者配置关系得以改善,有望在业务预报中应用。

The dynamical-analogue ensemble method for improving operational monthly forecasting

Focusing on the monthly forecasting problem based on the Atmospheric General Circulation Model (AGCM), a method of the dynamical-analogue ensemble forecasting (DAEF) is proposed to effectively reduce prediction errors and increase prediction skills. This method aims to the intrinsic combination of the dynamical model and statistical-empirical methods, which can establish perturbation members for ensemble forecasting by extracting the historical analogue information of the atmospheric general circulation, parameterizing empirically model errors and generating the multi-time-independent analogue forcing. Applying this new ensemble method to the operational AGCM in Beijing Climate Center (BCC AGCM1), a 10-yr monthly forecasting experiment under a quasi-operational condition shows encouraging results. Compared with the operational ensemble forecasts by the BCC AGCM1, the DAEF method is capable to improve effectively prediction skills of the monthly-mean and daily atmospheric circulation forecasts in which the former almost reaches the standard, available in the BCC operation, through effectively improving predictions of the zonal mean, ultra-long waves and long waves of the circulation. The results also show that prediction errors for the DAEF are significantly reduced and its spread of the ensemble members is reasonably increased, indicating an improvement in the relationship between the prediction errors and the spread. This study suggests a big potential application of the DAEF method in the BCC monthly forecasting operation.