卫星云图反演资料质量控制方案在暴雨模拟中的对比研究  

STUDY OF QUALITY CONTROL SCHEMES ON RETRIEVAL DATA FROM SATELLITE CLOUD IMAGES IN RAINSTORM SIMULATION

作　　者：赖绍钧[1] 沈桐立[1] 何芬[1] 孙桂平[1] 

机构地区：[1]南京信息工程大学气象灾害和环境变化重点开放实验室,南京210044

出　　处：《气象科学》2005年第6期551-559,共9页Journal of the Meteorological Sciences

基　　金：国家自然科学基金项目(编号:40075023)资助

摘　　要：本文以2003年7月9～10日发生在江淮流域的暴雨过程为例,应用FY-2B卫星红外云图灰度资料,利用统计回归的反演方法求出高分辨率的温度和湿度的反演场用于改进预报的初始场.设计和对比了几种质量控制方案,再将控制后的反演场与常规资料进行同化,最后用MM5模式进行了模拟对比试验.结果表明:在模式的初始场中引入卫星云图反演场能反映出更为细致的中尺度结构,有效地增强了对降水有重要影响的高湿区;锋生函数质量控制方案改善模式的初始场,从而改善了MM5模式降水预报的落区和强度,提高了降水预报准确性及Ts评分.In this paper, taken the rainstorm process of July 9 - 10, 2003 in YangtzeHuaihe river basin as an example, Grayness data of FY-2B satellite infrared cloud images are used for high resolution retrieved temperature and humidity fields through regression retrieved method to improve model initial fields for forecast. Several schemes are contrived and to contrasted to correct the errors from the image and to control the quality of retrieved fields. The variational assimilation is then performed of the retrieved fields with conventional fields by MM5 model. Results indicate that importing retrieved fields from satellite cloud images to model initial fields can well reflect more delicate mesoscale convection construction, efficiently enlarge high humidity region; moreover through quality control scheme of frontogensis function-it is possible to amend model initial fields and obviously enhance the intensity and the region of model precipitation forecast, so that the accuracy of MM5 model rainfall prediction and its Ts grades are greatly improved.

关 键 词：卫星资料反演 资料变分同化 质量控制 

分 类 号：P435[天文地球—大气科学及气象学]