###
气象:2020,46(10):1272-1285
本文二维码信息
码上扫一扫!
2016—2018年全国智能网格降水及温度预报检验评估
韦青,代刊,林建,赵瑞霞
(国家气象中心,北京 100081)
Evaluation on the 2016-2018 Fine Gridded Precipitation and Temperature Forecasting
WEI Qing,DAI Kan,LIN Jian,ZHAO Ruixia
(National Meteorological Centre, Beijing 100081)
摘要
图/表
参考文献
相似文献
本文已被:浏览 633次   下载 3085
投稿时间:2019-05-05    修订日期:2020-08-29
中文摘要: 为综合评估智能网格预报业务的产品质量,基于站点观测和格点实况分析资料,开展了对2016—2018年国家级网格指导预报、省市级订正反馈网格预报的降水和温度要素的检验,并与GRAPES和ECMWF模式输出进行对比。分析结果表明:智能网格预报业务整体上实现了从数值模式系统、国家级指导产品到省市级订正反馈产品的逐级增加价值的业务链条;国家级指导产品的降水、温度在数值模式的基础上都有明显提升,相比于ECMWF模式,小雨ETS评分提升9%~37%,暴雨ETS评分提升最大达41%,日最高温的预报误差降幅可达16%;省市级订正反馈产品的日最高/低温在国家级指导产品基础上进一步降低误差,能给出最优预报结果;国家级指导产品需进一步改进,重点要控制对于强降水预报的空报率,改善午后降水峰值预报显著提前的偏差,提高对空间精细化降水特征的表现能力,改进对于最低温的预报统计模型;省市级订正反馈产品在降水预报、整点温度预报方面未能对国家级指导预报进行很好订正,需要进一步加强;智能网格预报产品已能替代城镇天气预报,有利于流程的集约高效。
Abstract:To comprehensively evaluate the fine gridded forecasts, based on site observation and grid point analysis data, the verification of precipitation and temperature forecasts of the 2016-2018 national gridded guidance forecast, provincial and municipal revised feedback gridded forecasts were carried out, and compared with the GRAPES and ECMWF model outputs. The results show that the fine gridded forecast operation can realize the chain of increasing value step by step from numerical model system, national guidance forecast to provincial and municipal revised feedback product. The national guidance precipitation, daily high and low temperature forecasts have a significant improvement on the basis of the model forecast products. Compared to ECMWF model, light rain ETS score of the precipitation is increased by 9%-37%, the maximum increase of rain storm ETS score reaches 41% and the maximum decrease of daily high temperature RMSE reaches 16%. The daily high/low temperature of the provincial and municipal revised feedback products can further have the reduced error on the basis of the national guidance forecast, and the optimal forecast results can be obtained. However, the national guidance products need to be further improved, with emphasis on controlling the false alarm rate of the heavy rain forecast, improving the significant advance deviation of the afternoon precipitation peak, improving the ability to characterize spatially refined precipitation, and improving the forecast statistical model for the lowest temperature. Besides, as the provincial and municipal revised feedback products have failed to correct the national guidance forecast well in the aspects of precipitation and hour temperature forecast, they need to be further strengthened. The fine gridded forecast products can replace the urban weather forecast, which is beneficial to the intensive and efficient forecasting process.
文章编号:     中图分类号:    文献标志码:
基金项目:国家重点研发计划(2018YFC1507205)、中国气象局预报关键项目[YBGJXM(2019)05-1]和国家气象中心预报员专项(Y201927)共同资助
引用文本:
韦青,代刊,林建,赵瑞霞,2020.2016—2018年全国智能网格降水及温度预报检验评估[J].气象,46(10):1272-1285.
WEI Qing,DAI Kan,LIN Jian,ZHAO Ruixia,2020.Evaluation on the 2016-2018 Fine Gridded Precipitation and Temperature Forecasting[J].Meteor Mon,46(10):1272-1285.