四维变分同化技术在风暴潮数值模拟中的应用  

Application of four-dimensional variational data assimilation technique in storm surge simulations

作　　者：王宗辰 于福江[1,2] 原野[1,2] 

机构地区：[1]国家海洋环境预报中心,北京100081 [2]国家海洋局海洋灾害预报技术研究重点实验室,北京100081

出　　处：《海洋预报》2015年第1期1-9,共9页Marine Forecasts

基　　金：海洋公益性行业专项(200905013)

摘　　要：利用区域海洋模式ROMS(Regional Ocean Modelling System)及其四维变分同化模块,建立了具有资料同化能力的东中国海风暴潮数值模式,通过将海洋站水位观测资料同化到风暴潮模式中,提高了模式对风暴潮的模拟精度。四维变分同化技术能够在整个同化时间窗口保持动力协调,使模拟结果在该时间窗口内最大程度的靠近观测,同时,得到了最优预报初始场。利用该模式,对两次温带风暴潮过程进行了数值模拟,结果表明:在同化窗口内,同化对模拟精度有明显的提高;结束同化之后,得到的最优预报初始场对临近预报精度也有一定提高。Base on Regional Ocean Modeling System（ROMS） with four-dimensional variational（4D-Var） data assimilation modules, a model is established for storm surge simulation in the East China Sea, and the deterministic model output is corrected by assimilating the available tidal gauge station data. The 4D-Var technique is able to compensate for the errors between modeled outputs and observations by containing dynamically consistent and persistent during a period of time（also called assimilation window）, and the optimal initial condition for forecasting is obtained. Two applications on extra tropical storm surges that occurred in Bohai are performed by the model. The results show that the procedure is capable of strongly improving simulation accuracy, and the optimal initial condition effectively improves the forecasting accuracy in a short period.

关 键 词：区域海洋模式ROMS 四维变分同化 风暴潮数值模拟 最优预报初始场 

分 类 号：P731.23[天文地球—海洋科学]