不同资料同化对登陆台风菲特(2013)短时预报的影响研究

Relative Impact from Surface,Radiosonde,Aircraft,Satellite,and Radar Observations on Short-Time Forecasting for Typhoon Fitow (2013) at Landfall

为定量评估地面、探空、飞机报、卫星辐射亮温、雷达反射率及径向风等不同观测资料同化对台风预报性能的影响,本文以2013年严重影响我国的登陆台风菲特为例,利用WRF模式与GSI-3DVAR同化系统开展观测系统试验(OSE)研究,探讨了不同类型资料同化对"菲特"(2013)路径、强度、形势场和降水短时预报的相对贡献及可能影响机理。结果表明:(1)不同类型资料同化对模拟结果贡献程度有明显差别,其中探空、雷达反射率和飞机报对模拟结果有较大影响,分别"拒绝"这三种资料后模式模拟的高空各要素均方根误差分别上升约54.8%~62.0%、9.2%~16.5%和6.1%~6.4%。(2)对于不同的台风预报效果评估参数,各类资料的贡献率大小排序不同。对高空场和台风路径模拟影响较大的是探空和飞机报,对台风强度模拟影响较大的依次是雷达径向风、反射率、探空和飞机报,而对降水模拟影响较大的依次是雷达反射率、探空和飞机报。(3)各类资料对降水模拟的贡献率随时间变化不同。雷达反射率资料对降水的贡献随着模式积分时间明显下降,而飞机报、探空资料等对降水的贡献在模式积分3 h之后开始出现。(4)资料同化对降水模拟的改进与其对台风路径、水物质及强度模拟改进有关,因此影响高空场、台风路径和强度较大的雷达反射率、探空和飞机报资料,也是对降水模拟贡献较大的资料。

An assessment is presented on the relative forecast impact on the short-time forecasting for ty- phoon from six different observation data types： surface, radiosonde, aircraft, satellite, radar reflectivity, and radar radial velocity. In this paper, the landfall Typhoon Fitow, which seriously affected China in 2013, is chosen to investigate the relative impact from different data types on the typhoon track, intensity, and precipitation short-time forecasting by using Observation System Experiment （OSE） method with the Weather Research and Forecasting model （WRF） and Community Gridpoint Statistical Interpolation （GSI- 3DVAR） system. The results show that.. （1） the contribution of different data types to the simulation re- suits is obviously different. On the whole, the radiosonde, radar reflectivity and aircraft report have a great influence on the simulation of temperature, humidity, and wind of the full troposphere. The root mean square error （RMSE） increases by about 54.8%-62.0%, 9.2%-16.5% and 6.1%-6.4% for the ＂denied＂ experiments （NRS, NRF, NAC, respectively） compared to the control experiments. （2） Differ- ent sequence of contribution rates can be found in different assessment parameters. The radiosonde and aircraft report have the greatest influence on the simulation of temperature, humidity, and wind of the full troposphere. The sequence of contributions on typhoon intensity simulation is radar radial velocity, radar reflectivity, radiosonde, and aircraft report. Radar reflectivity, radiosonde, and aircraft report have great influence on the precipitation simulation. （3） The time-variations of the contributions to precipitation sim- ulation are different among different data types. The contribution of radar reflectivity data to precipitation decreases with the time of model integration, while the contribution of aircraft report and radiosonde data to precipitation appear after 3 h of model integration. （4） The improvement of precipitation simulation is related to the improvement of typhoon track, intensity and hydrometeors simulation. Therefore, the radar reflectivity, radiosonde and aircraft report, which have high influences on the simulation of the full tropo- sphere, typhoon track, and intensity, also have a great contribution to the precipitation simulation.