FES2022 global tide atlas: a step towards SWOT-compliant tidal corrections

Abstract. The FES2022 Global Tide Atlas represents a significant advancement in global tidal modelling, specifically designed to meet the high-resolution tidal correction requirements of modern altimetry missions like the Surface Water and Ocean Topography (SWOT) mission, which demand enhanced resolution and accuracy in the coastal seas. Building on the success of its predecessor FES2014, which performed well in most ocean regions but exhibited residuals issues in coastal, shelf, and polar areas, FES2022 introduces a tenfold increase in mesh resolution, prioritizing coastal and shelf regions. To some extent, this resolution aligns with SWOT’s KaRIn instrument capabilities, ensuring compatibility with its high-resolution ocean elevation measurements in shallow water regions. The bathymetry dataset is a composite of GEBCO_2020 and regional patches from other datasets, with a 2-meter depth threshold to better capture shallow coastal dynamics.

The underlying hydrodynamic model (T-UGOm) includes both frequency-domain and time-stepping solvers, with the latter providing more accurate schemes for non-linearities in tides. Bottom friction and wave drag parameters were regionally calibrated. The model also incorporates atmospheric forcing from ERA5 to account for the radiational component of tides when relevant. The data assimilation process uses an ensemble Kalman filter with 358 members, generated by perturbing bathymetry, bottom friction, and wave drag. 16 tidal constituents were assimilated, along with 18 additional waves derived from time-stepping simulations. The assimilation was performed in two iterations (the second one resulting in the final FES2022b atlas) to incorporate the FES2022a Loading and Self-Attraction solution, enhancing the consistency between the model and the observations.

The validation of FES2022b demonstrates substantial improvements over FES2014b and other state-of-the-art models. In the frequency domain, the Root Sum Square (RSS) errors for tidal elevations were reduced to 7.6 mm at altimetry crossovers in the deep ocean, 11.3 mm on shelves, and 42.3 mm at tide gauges compared to FES2014b. Temporal validation showed a variance reduction in Sea Surface Height (SSH) of -2.26 cm² for Jason-3 and -5.75 cm² for AltiKa when compared to FES2014b, with even greater reductions (-16.6 cm² for AltiKa) when compared to GOT4.10c. In polar regions, FES2022b significantly reduced the Sentinel-3A SSH variance in the Arctic (Baffin Bay, Siberian Shelf) and Antarctic (Ross and Weddell Seas). Coastal zones benefited from more than 40 cm² of variance reduction within 20 km of the shore. Tidal current validation using 119 current-meter stations showed a 56 % improvement compared to FES2014b, with a global RMSD of 2 cm/s for the major axis of tidal ellipses. GRACE gravity field residuals were also reduced, particularly around the Vancouver Island, from 10 cm to 3 cm EWE (Equivalent Water Elevation). Despite these advancements, some local residual errors persist due to bathymetry inaccuracies, shoreline misalignments, and limited high-latitude data coverage. These issues will be addressed in the future FES release.

FES2022 is operational for altimetry de-aliasing corrections and available for download on the AVISO+ website, along with the prediction code on the CNES GitHub. Future improvements will focus on integrating latest bathymetry databases, refining shoreline databases, and enhancing high-latitude coverage of data assimilation through dedicated missions. The model’s unprecedented accuracy in coastal and polar regions makes it a recommended replacement for FES2014 in operational and scientific applications.