Case Study On Observation Sensitive Region Of Heavy Rainfall In Beijing Area

The adaptive observation is an effective method to improve the predictability of HighImpact Weather (HIW), is one of the hot issues of the numerical prediction all around theworld, and becomes the core objectives of the international atmospheric research in the nextdecade. The determining sensitive areas are the most important issues of adaptive observation.With more and more attention focused on adaptive observation field, the method ofdetermining sensitive area gained great improvement. On one hand, several approaches suchas ETKF, CNOP, and SV, have been proposed to determine the sensitive areas for adaptiveobservation in recent years. On the other hand, ensemble sensitivity analysis can also used toevaluate the relationship between forecast errors and initial condition errors; this method isalso known as based on ensemble forecast statistical correlation method. At first,theobservation sensitive region of a heavy rainfall took place around Beijing in23th July2009isanalyzed. Then by using WRFDA, an ensemble of30initial samples is constructed, and then12hours forecast is carried based on the initial samples. In order to determine the sensitivevariables and their regions, the sensitivity of the accumulated precipitation in the verificationregion (Beijing area) to the basic variables of initial condition is studied by using the methodmentioned above. Moreover, the data assimilation of observing system simulation experiment(OSSE) is used to verify the sensitive region. Results indicate that:1. The heavy rainfall was induced by a squall line, the maximum precipitation falls area insoutheastern Beijing and in the border between Beijing and Tianjin, the6hour maximumaccumulated precipitation located in ChaoYang District is74mm. Beijing area is in front of thecenter of lower lever vortex, accompanied by strong rising movement. Water vapors channel islocated south west of Beijing and the unstable energy accumulated at14o'clock, so the liftingeffect of cold front lead to the result of energy release.2. Utilizing WRF3.1model, both the synoptic and radar echo are simulated well afterchoosing the best physics parameterization schemes. Except that the simulated radar echoesevolution was1-2hour earlier and bit of stronger than observations. Thus lay the foundationfor forming a reasonable amount of Ensemble Prediction System.3. By using WRFDA, an ensemble of30initial samples is constructed, and then12hoursforecast is carried based on the initial samples. In order to determine the sensitive variables and their regions, the sensitivity of the accumulated precipitation in the verification region(Beijing area) to the basic variables of initial condition is studied by using the methodmentioned above. Results show that both water vapor and the temperature are the sensitivevariables, the corresponding sensitive regions are at south-west side and north-east side ofBeijing, and have clear physical meaning.4. Moreover, the data assimilation of observing system simulation experiment (OSSE) isused to verify the sensitive region. It is shown show that assimilating sensitive variable watervapor in its sensitive region can improve the precipitation forecast accuracy; assimilating thesensitive variable temperature in its sensitive region also can improve the forecast skill, itindicate the correctness of the sensitive variables and sensitive areas. Assimilating sensitivevariable water vapor in its sensitive region while assimilating sensitive variable temperature inits sensitive region can exert a more positive improvement on forecast skill,and it illustratesthat the effect of temperature and water vapor contribution to improve rainfall predictionaccuracy maximum.5. Further analysis showed that, by both assimilating sensitive variable water vapor in itssensitive region and assimilating sensitive variable temperature in its sensitive region canimprove the initial status of temperature and moisture conditions, then through the modelforecast, the result shown that the CAPE have largely improved in14h,thus improve the within6hour accumulated precipitation forecast results for care region ultimately.