Ensemble-based assimilation of high-frequency radar surface currents in regional ocean models

The results of coastal ocean models depend critically on the accuracy of boundary and initial conditions and atmospheric forcing. The precision of coastal ocean models is limited among others by uncertainty in those forcing fields. Since high-frequency (HF) radar installations provide measurements over a relatively large area, the assimilation of these data has a high potential to reduce the errors in ocean models and to provide a dynamically consistent estimation of the ocean circulation. The assimilation of HF radar data is not without its own challenges: the spatial variation of the surface currents uncertainty, the high temporal resolution of HF radar data, the simultaneous presence of a wide range of processes with distinct spatial and temporal scales (tides and other surface gravity waves, mesoscale and wind-driven circulation), and the generally strong sensitivity of regional models to errors in the boundary conditions and atmospheric forcings. These processess are important aspects to consider in the application of data assimilation methods to HF radar measurements. The results of two data assimilation experiments on the West Florida Shelf (WFS) and the German Bight are presented. HF radar currents are assimilated in a nested West Florida Shelf based on an ensemble of model realizations with different wind forcings. The model is sequentially updated and a filter is implemented to reduce spurious surface-gravity waves. Results of the WFS model assimilating surface currents show an improvement of the model currents not only at the surface but also at depth compared to independent ADCP observations. This West Florida Shelf assimilation experiment does not include tides. Tides are not generated within the domain, but are rather propagated inside the domain through the boundary conditions. The potential of using HF radar data to reduce errors in tidal boundary conditions is shown in a model setup of the German Bight. For improving the modeled tidal variability it is not sufficient to update the model state without updating the boundary conditions. An ensemble smoother to improve the tidal boundary values is presented and validated with independent HF radar measurements and tide-gage data. The ensemble-scheme is also applied to improve the wind forcing by assimilation of surface currents. The improvement of the analyzed wind forcing is assessed by using in-situ wind measurements.