| 摘要: |
| 为客观评价大范围降温天气中的气温预报效果，本文以传统检验方法为基础，借鉴面向对象的检验方法，构建了降温预报综合检验指标：降温范围偏差指标和降温中心强度偏差指标分别对降温过程的范围和强度进行描述，日气温大误差率指标描述气温预报性能，三者共同构成降温预报偏差综合指数。以2020年6月1日-2021年5月31日期间中国区域的大范围降温日为样本，对ECMWF、NCEP客观模式1~7d时效的气温预报开展了降温时间预报的传统检验，和降温预报综合检验。结果表明：临近时段各预报场预报的降温日都较可靠，但漏报较多；降温预报偏差综合指数揭示在降温空间分布上，各预报场有降温范围偏小或降温中心偏离的缺陷，ECMWF的降温预报效果优势明显，而NCEP预报效果欠佳，主要因NCEP气温预报偏差大，但其对降温形态的预报具有参考性。对降温预报检验方法开展讨论，发现引入气温变化后的降温预报偏差综合指数结果与传统检验有差异，因此针对降温天气开展检验是有必要的；通过各降温过程比较和典型个例讨论，降温预报综合检验各项指标具有检验意义，所构建的降温预报偏差综合指数能够客观反映大范围降温日的气温、降温预报性能，可用于开展各降温日、各预报的比较；降温预报时间检验和降温预报综合检验分别关注大范围降温天气下降温预报在时间和空间上偏差，可拓展应用到不同区域范围。 |
| 关键词: 降温；气温预报；检验 |
| DOI： |
| 投稿时间：2023-11-01修订日期：2024-04-28 |
| 基金项目:中国气象局创新发展专项(CXFZ2022J071)“低纬山区心脑血管疾病气象预报技术研究与应用”，贵州省气象局科技项目（黔气科登[2021]05-03号）“黔中山地小气候空气污染气象条件预报技术研究”。 |
|
| Temperature Forecast Verification for Large-scale Temperature-drop Days |
| 唐,CHEN Lei,DU Zheng-jing,SHANG Yuan-yuan,SONG Dan |
| (Guizhou Meteorological Service) |
| Abstract: |
| In order to objectively verify temperature forecast in large-scale temperature-drop days, this paper constructed a verification method focusing on temperature-drop. Based on traditional verification, and referring to object-oriented verification, the forecasting performance on temperature-drop range and intensity are described by the temperature-drop range deviation indices and temperature-drop intensity deviation indices respectively, together with the daily temperature large error rate describing temperature forecast performance, we construct the temperature-drop deviation composite index. Taking large-scale temperature-drop days from June 1, 2020 to May 31, 2021 for samples, both traditional verification on temporal scale, and spatial composite verification by temperature-drop deviation composite index, were taken out for ECMWF and NCEP temperature forecasts in China area. Result showed that the forecasted temperature-drop days were reliable, especially in 1~3d forecasting periods, but all forecasts tended to miss some temperature-drop days. The temperature-drop deviation composite index showed that, all forecast fields showed smaller range or offset centre in the spatial distribution of temperature-drop, and ECMWF outperforms NCEP, primarily because NCEP had larger temperature drop deviation, while it’s temperature-drop distribution was referable. The verification result of temperature-drop deviation composite index differed from traditional verification, according to the introduction of temperature change. Therefore, it is necessary to carry out temperature-drop verification. The temperature-drop deviation composite index and its related indices provide a comprehensive description of temperature and temperature-drop forecasting performance, so it could be used for the comparison of temperature-drop days and forecast fields. Finally, temporal traditional verification, and spatial composite verification can be adjust to fit different regional scales. |
| Key words: temperature-drop; temperature forecast; verification |
