Modeling coastal circulation in Norway using a high-resolution 4D-Var ocean assimilation system

The circulation along the Norwegian coast is characterized by many transient smallscale features such as eddies and meanders that are challenging to reproduce by means of numerical modeling. In this thesis I investigate the use of advanced data assimilation (DA) techniques in high-resolution coastal models to improve the circulation estimates. One particularly interesting observational platform for the coastal ocean is highfrequency (HF) radars, which measure surface currents in the coastal zone up to 200 km offshore. The suitability of such observations for use in high-resolution coastal DA systems is assessed by quantifying which components of the near-surface current field are observed by the HF radars. Our results show that there are no contributions from wave drift in the measurements, thus they are suitable for use in coastal DA. Assimilation of HF radar currents in a high-resolution model shows clear improvement in the circulation estimates. Further improvement is obtained when CTD profiles of temperature and salinity are included in the assimilated data set. A reanalysis of a period, during which in-situ observations were abundant in the study area, is used to assess how an observational network dense enough to constrain the water mass distribution affects the upper ocean circulation estimates. Our investigations of the results show a weakening of the topographically steered currents and, as the stratification increases the effective resolution of the model, more small-scale circulation features are developed. Such changes may have a significant effect on upper ocean transport. Finally, the potential of using high-resolution coastal reanalyses to study specific physical processes is demonstrated for the case of the mechanisms causing variability in the Norwegian Coastal Current