摘要
采用CRA、邻域TS评分、FSS等多种空间检验方法,对多个不同尺度业务数值模式在“21·7”河南极端暴雨过程中的预报性能进行了综合检验评估,并从低空急流、水汽辐合和热力条件等方面对模式偏差原因进行诊断分析。结果表明:1)在24 h大暴雨降水位置预报偏差上,WARMS预报性能最优,GRAPES_3 km次优;大暴雨降水预报范围与实况相当时,RMAPS降水强度预报较实况明显偏强,且落区较实况出现持续偏西的特征;2)GRAPES_3 km和WARMS预报3 h累积降水的位置偏差更多表现在经向方向上,且离散度较大,但在更临近预报时效经向偏差明显减小,而纬向偏差则随预报时效变化较小;RMAPS的位置偏差主要表现在纬向方向上,19—21日分别有86.7%、91.3%、72.7%的降水个体预报偏西;3)WARMS对低空急流和水汽辐合预报偏弱导致其对19—20日降水强度估计不足,EC和RMAPS对19—20日低空急流预报明显偏西是导致降水落区位置存在偏差的主要原因;MESO对20日急流和水汽辐合发生时间及位置预报较好,但明显偏弱的热力条件导致其缺乏对极端强降水的预报能力;4)21日,MESO和RMAPS预报低空急流过多的偏东分量导致其在太行山陡峭地形处预报了偏强的地形增幅降水;EC和WARMS预报低空急流风向更接近实况,但对低层水汽辐合强度和时间的预报偏差导致预报降水个体出现了较明显的经向位置偏差。
During the“21·7”extreme rainstorm in Henan,the precipitation of mutil-model showed great differences and divergence.This paper uses multiple spatial verification methods(such as CRA,neighborhood TS score and FSS)to comprehensively evaluate the forecast performance of multiple numerical models during the“21·7”Henan extreme rainstorm process,and diagnoses the cause of model deviation from the aspects of low-level jet,water vapor convergence and thermal conditions.Results show that:(1)For the position forecast deviation of precipitation over 100 mm,the forecast performance of WARMS is the best,followed by GRAPES_3 km.When the heavy rain area predicted by RMAPS is comparable to the observation,the predicted precipitation intensity is obviously stronger,and the predicted rain area continues to appear westward compared with the observation.(2)The position deviation of 3 h cumulative precipitation predicted by GRAPES_3 km and WARMS is more manifested in the meridional direction and has large dispersion.The meridional deviation decreases obviously when the forecast time is nearer,while the zonal deviation changes slightly with the forecast time.The position forecast deviation of RMAPS is mainly manifested in the zonal direction,with 86.7%,91.3%and 72.7%of the individual forecasts of precipitation being westward from the 19th to the 21st,respectively.(3)The weak forecast of low-level jet and water vapor convergence by WARMS leads to the underestimation of precipitation intensity from 19th to 20th.The main reason for the location deviation of precipitation area is that the low-level jet predicted by EC and RMAPS is obviously westward from the 19th to 20th.MESO has a good forecast of the occurrence time and location of jet stream and water vapor convergence on the 20th,but it lacks the ability to predict extreme heavy precipitation due to the obviously weak thermal conditions.(4)On the 21st,too many easterly components of low-level jet predicted by MESO and RMAPS leads to the strong terrain increasing precipitation forecast at the steep terrain of Taihang Mountain.The wind direction of low-level jet stream forecasted by EC and WARMS is closer to the observation,but the forecast deviation of intensity and time of low-level water vapor convergence leads to the obvious meridional position deviation of individual precipitation forecast.
作者
栗晗
王新敏
朱枫
LI Han;WANG Xinmin;ZHU Feng(CMA Henan Key Laboratory of Agrometeorological Support and Applied Technique,Zhengzhou 450003,China;Henan Meteorological Observatory,Zhengzhou 450003,China;Xinxiang Meteorological Observatory,Xinxiang 453000,China)
出处
《大气科学学报》
CSCD
北大核心
2022年第4期573-590,共18页
Transactions of Atmospheric Sciences
基金
中国气象局创新发展专项(CXFZ2021Z008,CXFZ2022J014)
国家超级计算郑州中心创新生态系统建设科技专项(201400210800)
河南省气象局重点项目(KZ202101)
中国气象科学研究院灾害天气国家重点实验室开放课题(2021LASW-A09)。
关键词
极端暴雨
对象检验
邻域空间检验
模式诊断分析
extreme rainstorm
object-based verification
neighborhood spatial verification
model diagnostic analysis