Evaluation of precipitation forecasts for Dabie Mountain area and application countermeasures for the reservoirs regulation

对降水预报进行性能评估及应用对策研究可以更好地发挥降水预报在水库调度中的决策支持作用.基于大别山库区近10 a汛期（2007-2016年5月1日-9月30日）24~168 h共7个预见期降水预报和地面降水观测资料，采用正确率、TS评分、概率统计、ROC曲线以及CTS等方法评估大别山库区降水预报性能，并以响洪甸水库为重点研究区域分析降水预报在水库调度中的应用对策.结果表明：1）大别山库区各量级的降水预报都有正预报技巧；24~72 h预见期降水预报的TS评分较高且空报率、漏报率也较低，具有较高的预报性能；但96 h及以上预见期降水预报性能明显下降，中雨以上量级空报率、漏报率较大，特别是对大暴雨及其以上量级的降水预报性能显著下降.2）大别山库区预报降水量级与实况降水量级基本符合，预报降水量级大于等于实况降水量级的概率超过75%；虽然降水预报量级上呈现出过度预报的现象，但降水过程预报对水库调度仍有较好的应用价值，应用时要考虑到降水预报量级可能存在偏差.3）转折性天气预报96 h及以上预见期CTS评分较低，但72 h以内预见期的性能明显改进，尤其是24 h预见期CTS评分也提高到了38.2%；水库调度可从长预见期的降水预报获取降水过程及其可能发生转折的信息，根据短预见期的降水预报进行调度方案调整.;The evaluation and application research of precipitation forecasts for reservoirs area can help to optimize the reservoir regulation. Based on the precipitation forecasts with 7 lead time(24-168 h) and the observation precipitation data in the flood season during 2007-2016, evaluation methods (e.g. Percentage Correct, Threat Score, Probability Statistics, ROC curve, CTS) were introduced to estimate the precipitation prediction skill, and the application countermeasures for the reservoir regulation in Xianghongdian Reservoir of Dabie Mountain area were analyzed. The results indicate that:1) The precipitation forecasting skills for all precipitation magnitudes are all positive in the Dabie Mountains reservoir area. The precipitation forecast performance of 24-72 h lead time in the Dabie Mountains reservoir area is the best, and the TS score is relatively high and false alarm rate and missing forecast rate are also relatively low. However, the forecasting performance of 96 h lead time and above is obviously decreased, the false alarm rate and missing forecast rate are getting higher for moderate rain, especially for heavy rain. 2) Overall, even though the rainfall scale predictions are generally consistent with the observations, the probability of precipitation forecast greater than or equal to the observation is more than 75%. Although it is inclined to the over forecast the precipitation scale, the precipitation forecast products still havea good application value. However, precipitation scale forecast deviation must be considered during application. 3) Compared to the forecasts with long lead time which have great uncertainty, forecasts with short lead time are more reliable and have better performance in heavy precipitation processes and transition weather. The CTS scores of 96 h lead time and above are relatively low,but the forecasting performance is improved significantly within 72 h lead time. Especially, the CTS scores of 24 h lead time increases to 38.2%. In practical application, the combination of precipitation forecasts with different forecast periods, including the latest forecasts, are strongly suggested to get the information of transition weather. Such a forecast combination can be utilized to adjust the reservoirs regulation timely when the transition weather is approaching. Our statistic results have justified the usability of precipitation forecast, and the research results could be a valuable reference for the reservoir regulation decision making.