Impact of data assimilation on Eulerian versus Lagrangian estimates of upper ocean transport

Abstract

Using four-dimensional variational analysis, we produce an estimate of the state of a coastal region in Northern Norway during the late winter and spring in 1984. We use satellite sea surface temperature and in situ observations from a series of intensive field campaigns, and obtain a more realistic distribution of water masses both in the horizontal and the vertical than a pure downscaling approach can achieve. Although the distribution of Eulerian surface current speeds are similar, we find that they are more variable and less dependent on model bathymetry in our reanalysis compared to a hindcast produced using the same modeling system. Lagrangian drift currents on the other hand are significantly changed, with overall higher kinetic energy levels in the reanalysis than in the hindcast, particularly in the superinertial frequency band.