Assimilating temperature and salinity profile observations using an anisotropic recursive filter in a coastal ocean model

Abstract

In order to improve the ocean forecasting in the North Sea and Baltic Sea, an assimilation scheme based on a bottom-topography-dependent anisotropic recursive filter has been used in this study. This scheme can stretch or flatten the shape of a local representative contour surface of the background error covariance function into the form of an ellipse. Furthermore, the computing efficiency has been largely improved due to implicit computation of the background error covariance. A two-month experiment has been used for verifying the impact of assimilating ocean profile observations on ocean forecasting. The results indicate that the use of temperature and salinity profiles can largely improve the oceanic forecasting. The root mean square differences between the forecasts and observations for temperature and salinity have been reduced by 36% and 18% in the experiment period, respectively. Moreover, it is found that the anisotropic recursive filter approach is especially efficient in areas with complex coastlines and sharp fronts, e.g., inner Danish waters. The results also show that the propagation of observation information from an observation position to its neighboring grid points is closely related to currents.