基于WRF-3DVAR同化多源融合数据对近海风模拟的改进试验

Improving experiment on wind simulation based on asimilation multi-source wind data of WRF-3DVAR

本文利用WRF模式及其3DVAR同化系统,以2008年4月20日00时—23日00时的江苏近海10 m风场为研究个例,对Quik SCAT、Wind SAT、多源测风融合数据进行同化试验,通过比较WRF-3DVAR同化系统对模拟风场初始场和预报场的改进,检验了同化不同类型资料后WRF模式对研究区域内单点及区域近地层风速的预报效果。结果表明:同化试验对初始场有改进,且对预报场的改进较FNL资料明显;不同资料对风场模拟的影响不同,同化星星、星地多源融合资料效果最佳,Quik SCAT次之,Wind SAT最差。此外,在模式分辨率一定的情况下,提高观测资料的分辨率并不一定能够改善模拟效果,资料的稀疏分辨率存在最佳选择。

By using the Weather Research and Forecasting （WRF） model and its 3DVAR assimila- tion system and selecting a wind field at 10 m over offshore region of Jiangsu province from 00：00 on A- pril 20 to 00：00 on April 23, 2008 as research case, assimilation experiments on various schemes of QuikSCAT, WindSAT and multi-source wind data were conducted. Moreover, through comparing the im- provement effects of the initial and forecast fields of wind field with WRF-3DVAR system, the forecast effects of the research region for near-surface wind was also verified. Results show that compared to the FNL forecast directly, assimilation of wind data improves initial fields obviously. Different data has differ- ent effects on wind field simulation, among which, the effects of assimilating the multi-source data of star and star-ground are the best, QuikSCAT is the second and WindSAT is the last. In addition, under the certain model high resolution data, increasing the resolution of the observation data is not always possibleto improve the result of assimilation, and best thinning resolution is available.