Ocean Surface Reconstruction From Direct Assimilation of Radar Measurements Using the Adjoint Method

Kurzfassung

We present an algorithm for phase-resolved reconstruction of ocean surfaces from marine radar measurements based on the adjoint method. No assumption is made about the spatio-temporal location of the radar measurements and they are directly processed in the adjoint model through the inclusion of the two main X-band radar modulation mechanisms, namely the tilt and shadowing modulations. This makes the proposed algorithm able to assimilate raw radar data of complex helix-shaped spatio-temporal structure. Using numerically generated radar measurements, we highlight the importance of the spatial discretization and show that the reconstruction accuracy converges with the measurement duration by mitigating the impacts of the strong measurements' spatio-temporal non-uniformity and of the tilt modulation. A detailed comparison between the reconstructed and reference surfaces reveals the local discrepancies, which can be explained by the amount of assimilated wave information and the different modulation mechanisms. The obtained accuracy is found to be high enough for detailed wave field analysis and phase-resolved predictions. Although linear waves are used in the presented applications, the chosen numerical framework allows a direct extension to nonlinear waves.