Research on data assimilation and features analysis of storm surge in the Shanghai offshore areas

Storm surge is a complex ocean phenomenon which is sensitive to many factors and has attracted much attention. In this paper, based on covariance localization method of ensemble Kalman filter(EnKF), the storm surge data of different sources, different error and different spatial and temporal resolution calculated by tide gauge stations and FVCOM model were assimilated and fused for the first time. Taking the storm surge process of typhoon 201810 landing in Shanghai for example, the optimal solution of 72-hourly storm surge in the Shanghai offshore areas was obtained and verified, the setting range of set sample number and Schur radius were given. The results show that the root mean square error of storm surge calculated by the observed stations and the model is 0.20 m, while calculated by the observed stations and assimilation is 0.07 m, which is improved by 65%, the root mean square error calculated by independent observation and assimilation is 0.09 m, the ratio of set dispersion to root mean square error is 0.90, the assimilation effect is better and credible. The assimilated storm water increment field can clearly and accurately depict the characteristics of double peaks storm surge, typhone eye surge and frontal surge which can be better used for the research of storm surge, correction of numerical simulation and marine disaster prevention.