摘要
WRF(Weather Research and Forecast)模式参数化方案众多,不同参数化方案对降雨模拟效果有较大的影响。为研究WRF模式中云微物理、陆面过程及积云对流参数化方案组合对长江上游地区降雨模拟效果的影响,以长江宜宾至宜昌干流区间为研究区域,选取5种云微物理参数化方案、2种陆面过程参数化方案和3种积云对流参数化方案共30种方案组合对该地区4场不同降雨中心落区的典型强降雨过程进行模拟,并以实测站点数据和GPM(Global Precipitation Measurement)降水数据产品为参考,采用TS评分、空报率、漏报率、准确率、平均绝对误差以及均方根误差作为评价指标,充分考虑各指标的不同影响,运用基于熵权的密切值法对降雨模拟结果进行综合评价。结果表明,WRF模式中云微物理WSM3(WRF Single-Moment 3-class)方案,陆面过程Noah(Noah Land Surface Model)方案和积云对流Grell-D(GrellDevenyi Ensemble)方案的参数化方案组合对长江上游地区降雨模拟效果较好,在该参数化方案组合下,WRF模拟的24 h日面雨量平均值的模拟结果与实测资料基本一致,模拟得到的四场典型降雨的24 h面降雨中心和降雨量级分布带均与实际降雨较为接近,大部分地区的模拟误差小于10 mm,且对小雨和中雨的模拟效果更好,此外WRF模拟降雨精度会受地形变化的影响,地形起伏加剧会降低降雨模拟精度,西南山区降雨模拟精度稍低于东北平原地区。研究结果可为流域水文预报提供重要参考。
There are many parameterization schemes for the WRF(Weather Research and Forecast)model,and different parameterization schemes have a great influence on the rainfall simulation effect.In order to study the influence of the combination of cloud microphysics,land surface process and cumulus convection parameterization scheme in the WRF model on the rainfall simulation effect in the upper reach⁃es of the Yangtze River,this paper takes the mainstream from Yibin to Yichang in the Yangtze River as the research area,selects five kinds of cloud microphysics parameterization schemes,two kinds of land surface process parameterization schemes and three kinds of cumulus con⁃vection parameterization schemes,a total of 30 kinds of scheme combinations are used to simulate the typical heavy rainfall processes in the four different rainfall centers in this area,and uses the measured site data and GPM(Global Precipitation Measurement)precipitation data product as a reference and uses TS score,empty alarm rate,misses alarm rate,accuracy rate,mean absolute error and root mean square er⁃ror as evaluation indicators,fully considers the different influences of each index and uses the close relationship based on entropy weight.The numerical method is used to comprehensively evaluate the rainfall simulation results.The results show that the WRF-Single-Moment3(WRF-Single-Moment3-class)scheme,the Noah(Noah Land Surface Model)scheme and the cumulus convection Grell-D(Grell-Deve⁃nyi Ensemble)scheme are parametric schemes in the WRF model.The combination has a good effect on the rainfall simulation in the upper reaches of the Yangtze River.Under this parameterization scheme combination,the simulation results of the average 24-h daily surface rain⁃fall simulated by WRF are basically consistent with the measured data,and the simulated 24-h surface rainfall of four typical rainfalls The center and the distribution belt of rainfall magnitude are relatively close to the actual rainfall.The simulation error in most areas is less than 10 mm,and the simulation effect of light rain and moderate rain is better.In addition,the precision of WRF simulation rainfall will be affect⁃ed by terrain changes,and the terrain is undulating.The aggravation will reduce the rainfall simulation accuracy,and the rainfall simulation accuracy in the southwestern mountainous areas is slightly lower than that in the northeastern plains.The research results serve as an important reference for the basin hydrological forecast.
作者
于伟
彭杨
姚礼双
张雪敏
YU Wei;PENG Yang;YAO Li-shuang;ZHANG Xue-min(School of Water Conservancy and Hydropower Engineering,North China Electric Power University,Beijing 102206,China)
出处
《中国农村水利水电》
北大核心
2023年第5期98-105,共8页
China Rural Water and Hydropower
基金
国家自然科学基金面上项目(51679088)。
关键词
WRF
长江上游
参数化方案
熵权法
密切值法
降雨模拟
WRF
upper Yangtze River Basin
parameterization scheme
entropy weight method
osculating value method
rainfall simulation