On the interplay between horizontal resolution and wave drag and their effect on tidal baroclinic mode waves in realistic global ocean simulations.

The effects of horizontal resolution and wave drag damping on the semidiurnal M 2 tidal energetics are studied for two realistically-forced global HYbrid Coordinate Ocean Model (HYCOM) simulations with 41 layers and horizontal resolutions of 8 km (1 ∕ 12. 5 ∘ ; H12) and 4 km (1 ∕ 25 ∘ ; H25). In both simulations, the surface tidal error is minimized by tuning the strength of the linear wave drag, which is a parameterization of the surface-tide energy conversion to the unresolved baroclinic wave modes. In both simulations the M 2 surface tide error with TPXO8-atlas, an altimetry constrained model, is 2.6 cm. Compared to H12, the surface tide energy conversion to the resolved vertical modes is increased by 50% in H25. This coincides with an equivalent reduction in the tuned loss of energy from the surface tide to the wave drag. For the configurations studied here, the horizontal and not the vertical resolution is the factor limiting the number of vertical modes that are resolved in most of the global ocean: modes 1–2 in H12 and modes 1–5 in H25. The wave drag also dampens the resolved internal tides. The 40% reduction in wave-drag strength does not result in a proportional increase in the mode-1 energy density in H25. In the higher-resolution simulations, topographic mode-scattering and wave–wave interactions are better resolved. This allows for an energy flux out of mode 1 to the higher modes, mitigating the need for an internal tide damping term. The HYCOM simulations are validated with analytical conversion models and altimetry-inferred sea-surface height, fluxes, and surface tide dissipation. H25 agrees best with these data sets to within ∼ 10%. To facilitate the comparison of stationary tide signals extracted from time series with different durations, we successfully apply a spatially-varying correction factor. • The doubling of the horizontal resolution doubles the number of resolved vertical modes. • The increased resolution both affects surface and internal tide damping. • Higher-resolution simulations need less parameterized internal-tide damping. • The higher-resolution simulation agrees with validation data sets to within ∼ 10%. • A correction factor is applied to allow for the comparison of stationary tide signals. [ABSTRACT FROM AUTHOR]