WRF数值预报模式气象资料的同化处理与对比分析  

Assimilation and Comparative Analysis of WRF Numerical Prediction Model Data

作　　者：潘江勇[1,2] 李发明 荆大为 

机构地区：[1]北京大学物理学院大气科学系,北京100871 [2]民航青岛空中交通管理站气象台,山东青岛266108

出　　处：《安徽农业科学》2010年第17期9098-9102,共5页Journal of Anhui Agricultural Sciences

摘　　要：应用WRF中尺度数值预报模式及其同化模块(WRFDA)对常规观测资料和青岛市气象台多普勒气象雷达资料进行同化试验。以山东半岛2009年7月8日夜间一次暴雨天气过程为例。首先,以7月8日00:00(文中所有时间均为国际时间)NCEP再分析资料为初始场,对00:00的常规观测资料进行同化,积分36 h;其次,将上述试验积分12 h后对7月8日12:00的常规观测资料和多普勒雷达资料进行循环同化试验,并积分24 h。将以上同化后得到的初始场结果和数值预报场结果分别进行对比分析,结果表明:常规观测数据的同化对初始场各层的风、温度、水汽等要素均有影响,但对雨水混合比影响较小;多普勒雷达资料的同化,使初始场的对流层中低层水汽增加,高层水汽减少,并且使中低层雨水混合比增大;多普勒雷达资料对初始场各层的风均有明显影响。多普勒雷达资料对初始场的调整比常规观测资料大一个量级以上,同时同化上述2种数据后显示,风、水汽、雨水混合比结果明显趋近于单独同化雷达数据的结果。以00:00为初始场同时同化常规观测资料并积分36 h的结果,对较大范围的强降水落区的修正效果最好:同化前降水中心位置和降水量与实际观测结果比较相差较大;同化后降水中心和降水量都与实况比较接近,但模式没有模拟出青岛地区小范围的暴雨天气。同化12:00雷达资料后,模式模拟的青岛地区的降水落区与降水量均与实况接近。试验结果说明,同化常规观测资料对中尺度环流形势有较好的改善,能较好地模拟出大范围的降水落区,但对小尺度的系统模拟效果较差;同化多普勒雷达资料对中小尺度强对流天气的模拟效果较好,但仅适用于短时临近预报;而循环同化可以将两者的优点相结合,综合利用2种资料能改善模拟效果,增加预报结论的可靠性。WRF numerical prediction model and its assimilation module（WRFDA） were used in this dissertation,to do assimilation experiment between the conventional observation data and Doppler radar data.The test mode was applied to numerically simulate a rainstorm in Shandong Peninsula on the night of Jul 8th,2009.Firstly,used NCEP reanalysis data on 00：00am（all the time mentioned in this paper was UTC time） as initial field,assimilated the conventional observation data at that time,integrated for 36 hours;secondly,used the above test result after integrating 12 hours to do cycle assimilation experiment between the conventional observation data and Doppler radar data at 12：00,integrated for 24 hours;finally,compared the outcome of the initial field after above assimilation with the numerical prediction field.The results showed that：the assimilation of the routine observation data has great affect on wind,temperature,water vapor and other factors on each layer of initial field,but has little affect on rain water mixing ratio;the assimilation of Doppler radar data increased water vapor of the low-level′s troposphere in initial field but reduced water vapor in high level troposphere,it also increased the rainwater mixing ratio in and low-level;Doppler radar data has a significant effect on the wind layer of initial field.The adjustment to initial field from Doppler radar data is one magnitude greater than that from observation data.After assimilated both kinds of data synchronously,it was found that,the test result for wind,water vapor and rain water mixing ratio is obviously higher than singly radar data assimilation.The result of using 00：00 data as initial field,assimilating the two kinds of data mentioned above synchronously and integrating for 36 hours,would mostly amend a broader range of heavy rain fall zone;refer to the precipitation heartland and rainfall,the result from WRF simulation after assimilation is much closer to the facts than before assimilation.But the model couldn＇t simulate the small-

关 键 词：WRF 数值预报 循环同化 RADAR 

分 类 号：P468[天文地球—大气科学及气象学]