中国东南地区复杂地形下降水概率预报的订正研究

Calibration of the probabilistic forecast of precipitation over complex terrain in Southeast China

使用TIGGE (the THORPEX interactive grand global ensemble)资料集下欧洲中期天气预报中心(the European Centre for Medium-Range Weather Forecasts, ECMWF)逐日起报的预报时效为24~168 h的日降水量集合预报资料,集合预报共包括51个成员,利用左删失的非齐次Logistic回归方法(left-Censored Non-homogeneous Logistic Regression, CNLR)和标准化的模式后处理方法(Standardized Anomaly Model Output Statistics, SAMOS)对具有复杂地形的中国东南部地区降水预报进行统计后处理。结果表明:采用CNLR方法能够有效改进原始集合预报的平均绝对误差(Mean Absolute Error, MAE)和连续分级概率评分(Continuous Ranked Probability Score, CRPS),提升了降水的定量预报和概率预报的预报技巧。而使用SAMOS方法对数据进行预处理,考虑地形等因素的影响,能在CNLR方法的基础上进一步订正由于地形影响造成的预报误差,并得到更加准确的全概率的降水概率预报。

This study is based on the daily 24-to 168-hour ensemble precipitation forecast datasets derived from the European Centre for Medium-Range Weather Forecasts and extracted from the TIGGE(The Interactive Grand Global Ensemble)dataset.The ensemble forecast comprises 51 ensemble members.The study applies the left-censored non-homogeneous logistic regression method(CNLR)and the standardized model post-processing method(SAMOS)to calibrate the precipitation forecasts in Southeast China.The results show that the CNLR method can effectively reduce the mean absolute error(MAE)and continuous ranked probability score(CRPS)of the raw ensemble forecast,and improve the forecasting skills of quantitative and probabilistic precipitation forecasts.Using the SAMOS method to preprocess the data and considering the impact of topography and other factors,the forecast error caused by the terrain influence can be further corrected on the basis of the CNLR method,thereby obtaining a more accurate probabilistic forecast of precipitation.