Tides influence the ocean temperature and salinity in many ways, in particular in Antarctic seas. In their review of tidal effects on ice sheets, Padman et al. 2018 distinguish tidal processes occurring seaward of the ice shelves, such as tidal vertical mixing and residual currents, from those directly affecting heat exchanges at the ocean/ice-shelf interface.

Tidal vertical mixing is caused by (i) the vertical shear as tidal currents rub upon the seafloor (in particular in shallow areas where these currents tend to be relatively swift); and by (ii) the breaking of internal tidal waves generated by the interaction of barotropic tidal currents with steep topography. In cold regions such as the Ross and Weddell Seas, tidal vertical mixing can mix the relatively cold ice-shelf melt water with the underlying and relatively warmer waters (High Salinity Shelf Water). Tides not only induce mixing but also generate a mean residual circulation through the Stokes drift and non-linear dynamics. These residual currents can affect heat exchanges across continental shelves and ice-shelf edges.

Within ice shelf cavities, tides primarily affect ice/ocean interactions by increasing velocities and therefore turbulent heat exchanges along the ice base. This effect is relatively more important for large and cold cavities, such as Ross and Filchner-Ronne, where tidal currents can be significantly stronger than buoyancy-driven currents. The extra melting caused by tides induces an additional buoyancy-driven residual circulation, which in turn can increase ice-shelf melting. Estimating the relative importance of each of these tidal processes is a prerequisite for better prescribing or parameterizing the effect of tides on ice shelf cavities.

Understanding the interplay between all these processes is important because it tells how tidal effects should be precribed in the climate models that do not explicitly represent tides. In our new paper, we consider the example of the Amundsen Sea, and run numerous NEMO regional simulations to analyse the multiple effects of tides on ice-shelf melting.

Simulated area including 7 ice shelves (from Getz to Abbot), the Dotson-Getz Trough (DGT) and the Western and Eastern Pine Island Troughs (PITW and PITE respectively). The blue contour indicates the ice sheet edge, while the gray to black contours indicate isobaths). Shadings indicate the barotropic stram function (the difference between 2 points on the map gives the ocean transport in millions of m3/s through the corresponding vertical section).

In this paper, we show that diurnal tidal constituents generate strong tidal velocities over the continental shelf break. Steep topography indeed generates tidal waves that cannot propagate away from their generation site that are poleward of the critical latitude (where inertial frequency equals tidal frequency and beyond which tidal waves do not propagate freely). These strong currents enhance vertical mixing over the continental shelf break. We also bring evidence for a significant residual circulation that flows westward along the continental shelf break and southward in Dotson-Getz Trough (see figure below).

Barotropic stream function of the tidal residual circulation in the absence of ocean stratification and ice shelf melting (the difference between 2 points on the map gives the ocean transport in millions of m3/s through the corresponding vertical section).

While these processes can be identified seaward of the ice shelves, they do not affect ocean temperatures to a significant extent and have therefore little effects on ice-shelf melting. By contrast, tidal velocities in ice shelf cavities strengthen ocean turbulence near the ice base, and therefore enhance melting underneath all the ice shelves of the Amundsen Sea. The cold meltwater produced by tides slightly cools the ocean waters at the ice-sehlf base, which cancels approximately a third of the melt increase that would occur in response to enhanced turbulence.

Overall, the representation of tides in our regional simulations enhances ice-shelf melting, with weakest effects for Pine Island (+1%) and Thwaites (+8%) and strongest effects for Dotson (+30%), Cosgrove (+34%) and Abbot (+39%). The relatively weak tidal effect on Pine Island and Thwaites is likely due to the thick water column that makes the resonnant quarter wavelength much larger than the typical horizontal cavity size. By contrast, the Amundsen cavities with shallower water columns tend to be resonnant for semi-diurnal tides. The strong sensitivity of the tidal effect to the water column thickness shows that the aforementioned numbers have to be considered carefully because the high uncertainty on bathymetry and ice-shelf drafts.

These results indicate that adding a tidal velocity into the equation of the turbulent heat flux is a good approach to account for tide-induced melting in ocean models that do not explicitly represent tides. It is nonetheless important to keep the horizontal patterns of tidal velocities, and prescribing uniform tidal velocities leads to large errors. Overall, we find that prescribibg 66% of the tidal velocity from a barotropic tidal model is find reproduces remarkably well tide-induced melting.

The work leading to this publication was supported by the TROIS AS ANR project.

Reference

Jourdain, N. C., Molines, J.-M., Le Sommer, J., Mathiot, P., Chanut, J., de Lavergne, C. and Madec, G. (2018). Simulating or prescribing the influence of tides on the Amundsen Sea ice shelves. Ocean Modelling, in press. doi/10.1016/j.ocemod.2018.11.001