廊坊地区5种数值模式降水预报性能检验与评估

Verification and assessment of precipitation forecast by five numerical models in Langfang

为了提高廊坊地区降水预报数值产品的释用能力,利用廊坊市9个气象观测站24 h的实况降水资料和日本(JPN)、德国(GER)、GRAPES、T639及MM 5模式的降水预报产品资料,对目前降水预报业务中广泛使用的5种数值模式的预报效果进行检验分析。结果表明:日本(JPN)和德国(GER)模式对≥10.0 mm及≥25.0 mm量级降水预报的TS评分比其他模式高10.0%—40.0%,夏季MM 5模式对≥50.0 mm量级降水的预报表现出一定的优越性;T639、GRAPES模式分别对大雨及以上量级的降雨和小雪预报效果较好,日本、德国模式对中雪和大雪的预报表现出一定的优势,T639模式对暴雪预报的TS评分达33.3%,高于其他模式。

In order to improve the application of precipitation forecast products from numerical models,forecast results were tested using observed precipitation from 9 weather stations of Langfang and forecast products from five numerical models,namely,Japan,Germany,T639,GRAPES and MM5 models.The results indicate that TS scores of Japan and Germany models for more than or equal to 10 mm and 25 mm precipitation are 10.0%-40.0%higher than those from the other models.The MM5 model for more than or equal to 50 mm precipitation has better forecast results in summer.T639 and GRAPES models showbetter prediction ability for heavy rain or more grade and light snow,while Japan and Germany models are better for moderate and heavy snow.TS score of T639 for rainstorm reaches 33.3% and higher than those of the other models.