摘要
本文通过在区域暴雨预报模式(AREM)的后向映射四维变分同化系统(AREM-B4DVar)中引入GPS掩星折射率局地和非局地两种观测算子,使得该系统具备了同化全球定位系统(GPS)掩星折射率资料的能力,并采用台湾地区与美国联合执行的气象、电离层和气候星座观测系统计划(COSMIC计划)探测得到的GPS掩星折射率资料和常规探空资料,对2007年7月4日至5日发生在我国江淮流域的暴雨个例进行了同化预报试验.结果表明,在同化系统中采用局地和非局地两种观测算子,加入GPS掩星折射率资料后,均可以提高观测资料附近初值的分析质量,从而在改进24小时的降水预报中起到正效果;基于非局地观测算子的掩星折射率资料同化可以通过大气非局地的约束,进一步改进基于局地观测算子掩星折射率资料同化的不足;在常规资料的基础上加入掩星折射率资料,可以使同化系统进一步改进初值分析质量和24小时预报效果,尤其能更好地发挥非局地观测算子的作用.
In this paper,the observational operators of local and non-local GPS radio occultation refractivities are added to the Backward-4DVar data assimilation system of the regional heavy rainfall forecast model AREM（AREM-B4DVar）,and thereby the assimilation system has the capability to assimilate GPS radio occultation refractivity data.Experiments to predict the heavy rainfall,which occurs on July 4,2007 to July 5,2007 in Yangtze-Huaihe river basin of China,with assimilation of the COSMIC GPS radio occultation refractivity data and the radiosonde data,are designed.The results indicate that the quality of initial analysis near the observation and thus the corresponding 24-h rainfall forecast are all improved after the GPS radio occultation data are incorporated into the initial condition using the observational operators of local and non-local in the AREM-B4DVar system.Through the non-local constraints of atmospheric variables,the performance of the assimilation with the non-local operator can further improve the insufficient existing in the assimilation with the local operator.Joining the assimilation based on radiosonde data in the AREM-B4DVar system,the radio occultation refractivity data can further improve the analysis quality of initial condition and the skill of 24-h rainfall forecast,in which the non-local operator plays a better role.
出处
《中国科学：数学》
CSCD
北大核心
2012年第5期377-387,共11页
Scientia Sinica：Mathematica
基金
国家重点基础研究发展计划(批准号:2010CB951604)
国家高技术研究发展计划(批准号:2010AA012304)资助项目
关键词
掩星折射率
后向映射四维变分同化
局地和非局地观测算子
数值模拟
暴雨
GPS radio occultation refractivity data
Backward-4DVar assimilation
local and non-local observational operator
numerical modeling
heavy rainfall
