GRAPES沙尘数值预报系统对一次沙尘暴沙尘浓度同化模拟试验分析  

Model-Experimental Analysis of Variational Assimilation of Dust Concentration with GRAPES Sand-Dust Numerical Forecasting System in a Sand-Dust Storm

作　　者：赵建华[1] 李耀辉[1] 蒲朝霞[2] 刘世祥[3] 张良[1] 段海霞[1] 

机构地区：[1]甘肃省(中国气象局)干旱气候变化与减灾(开放)重点实验室/中国气象局兰州干旱气象研究所,甘肃兰州730020 [2]美国犹他大学大气科学系,美国盐湖城84112 [3]兰州中心气象台,甘肃兰州730000

出　　处：《高原气象》2012年第3期697-705,共9页Plateau Meteorology

基　　金：民勤生态变化对(风蚀)起沙影响的特征观测与数值基础研究项目(2008CB417213);国家自然科学基金项目(40975060);"十一五"国家科技支撑计划重点项目"沙尘暴遥感监测与预报集成技术研究-省级沙尘暴遥感监测预警评估与服务应用示范(2008BAC40B04)";科技部公益性行业(气象)科研专项(GYHY201006023)共同资助

摘　　要：为了提高沙尘模式的预报准确率,通过在区域天气数值模式GRAPES(Global/Regional Assimilation and PrEdiction System)的三维变分同化系统中增加沙尘浓度这一控制变量的方法建立了GRAPES_3DVAR_DUST沙尘同化系统。利用中国北方8个观测站提供的沙尘PM10数据和沙尘模式(GRAPES_CUACE/Dust系统)提供的背景场,对2008年2月29日-3月1日发生在中国北方的一次沙尘暴天气进行了控制试验和一次同化与间断同化的敏感性试验,结果表明:(1)引入该同化系统后,一次同化和间断同化试验模拟的地面沙尘浓度分布较未同化的控制试验结果更接近卫星监测,而间断同化的结果又好于一次同化;(2)一次同化试验与控制试验对单站PM10浓度的演变预报较差;(3)间断同化试验较准确地再现了单站PM10浓度的连续演变;(4)间断同化试验效果整体上优于一次同化试验。总体而言,引入沙尘同化系统在一定程度上可以提高沙尘模式对沙尘天气的预报准确率。In Order to improve the accuracy rate of dust model, a dust assimilation system named GRAPES_3DVAR DUST was developed through adding a new control variable （namely, dust concentra- tion） in the three-dimensional variational data assimilation system of regional GRAPES （Global/Regional Assimilation and Prediction System） model which is developed by Chinese scientists by themselves. Three schemes of control experiment, one-time assimilation experiment and intermittent assimilation experiment with observed dust PM10 concentration from 8 sand-dust-monitored stations in North China and background fields provided by a sand-dust model （GRAPES_CUACE/Dust system） are performed to test the assimila- tion influence on dust-concentration forecast through modeling a sand-dust storm occurring from February 29 to March 1, 2008. The sensitivity test results show： （1） comparing with the satellite-monitoring im- age, the modeled spatial distributions of surface dust concentration of one-time assimilation experiment and intermittent assimilation experiment are more precise than those from control experiment which has no data assimilation, while the result of intermittent assimilation experiment is the most precise; （2） one-time as- similation experiment fails to grasp the continual evolutions of PM10 concentration of the majority of single station as well as control experiment; （3） the results of intermittent assimilation experiment capture the main features of continual evolutions of PM10 concentration at the single station; （4） to some degree, the effect of intermittent assimilation experiment is better than that of one-time assimilation experiment. In a total, GRAPES_3DVAR_DUST system is able to improve the dust prediction of dust model.

关 键 词：GRAPES_3DAVR_DUST系统 沙尘模式 PM10浓度 

分 类 号：P425.55[天文地球—大气科学及气象学]