Three-Dimensional Reconstruction of Ocean Circulation from Coastal Marine Observations: Challenges and Methods

Monitoring and investigating the dynamics of coastal currents is crucial for the development of environmentally sustainable coastal activities, in order to preserve marine ecosystems as well as to support marine and navigation safety. This need is driving the set-up of a growing number of multiplatform operational observing systems, aiming to the continuous monitoring of the coastal ocean. A significant percent of the existing observatories is today equipped with land-based High Frequency Radars (HFR), which provide real-time currents with unprecedent coverage and resolution, limited however, to the surface layer. The combination of data from HFR with complementary data from in-situ platforms providing information of the currents at subsurface layers (ADCP moorings) is investigated here to reconstruct the 3D current velocity field from in-situ observations. For this purpose, two methods based on different approaches are used. On the one hand, the Reduced Order Optimal Interpolation which is fed, in this case, with a spatial covariance matrix extracted from a realistic numerical oceanic simulation; and on the other hand, the Discrete Cosine Transform Penalized Least Square, which is a data gap-filling method based on penalized least squares regression that balances fidelity to the data and smoothness of the solution. As a proof of concept, we test the methods’ skills by using emulated observations of currents, extracted from a numerical simulation (3D reference field). The test set-up emulates the real observatory scenario in the study area (south-eastern Bay of Biscay), which includes a long-range HFR and two ADCP moorings inside the HFR footprint area. Then, the reconstructed fields (outputs of the methods) are compared with the 3D reference fields. In general, the results show satisfactory 3D reconstructions with mean spatial (for each depth level) errors between 0.55–10.94 cm s−1 for the first 150 m depth. The methods perform better in well sampled areas, and although different performances between the methods are observed, both show promising skills for the computation of new operational products integrating complementary observations, broadening the applications of in-situ observational data.