Parameterize impact of tides

[dougiesquire]

Once explicit tidal forcing is disabled in all our configs, we need include parameterisations for the impact of tides.

I haven't looked into this much at all, but see:

• MOM_tidal_mixing parameters
• Related parameters in MOM_surface_forcing_nuopc
• Possibly also the INTERNAL_TIDES parameter in MOM_diabatic_driver
• ...

[minghangli-uni]

OM2 used Simmons et al (2004) method to parameterise tidal mixing via CVMix, which we may not use for the future configuration updates?

In OM2, LEE_WAVE_DISSIPATION is used to parameterise vertical mixing and viscosity due to the drag of barotropic tides on the bottom , but it is not compatible with CVMix in OM3, so USE_CVMix_TIDAL and LEE_WAVE_DISSIPATION cannot be enabled together. In this case, I think OM2 might not be a suitable reference for our configuration choices? I can review the parameters and provide a list of required parameters, then @aekiss can take a look.

[minghangli-uni]

Does anyone remember how to generate the three tide-related files in /g/data/vk83/experiments/inputs/access-om2/ocean/tides/global.025deg/2020.05.30?

roughness_amp.nc roughness_cdbot.nc tideamp.nc

Since our grid resolution has changed, we can’t use these files directly for the 0.25° configuration.

[dougiesquire]

Unfortunately I think these were inherited and their origin is uncertain. The is what the OM2 tech report has to say:

The tidal speed data in tideamp.nc probably originates from Jayne and St. Laurent (2001) if it came
from GFDL.TODO: try to confirm this? The bottom roughness files seem basically the same, except that
roughness_amp.nc is on T points and roughness_cdbot.nc is on U points. They probably originate from
Jayne and St. Laurent (2001) or Jayne (2009), which in turn are derived from Smith and Sandwell
(1997).

[minghangli-uni]

MOM6 supports four different internal tidal dissipation schemes,

• Laurent et al 2002
• Polzin et al 2009
• Simmons et al 2004 (CVMix)
• Schmittner et al 2014 (CVMix)

The CVMix schemes require an external tidal energy flux to parameterise vertical diffusivity. The input is different from what MOM5 used roughness_amp.nc roughness_cdbot.nc tideamp.nc.

Below is one basic implementation using cvmix.

USE_CVMix_TIDAL = True          !   [Boolean] default = False
                                ! If true, turns on tidal mixing via CVMix
INT_TIDE_DISSIPATION = True     !   [Boolean] default = False
                                ! If true, use an internal tidal dissipation scheme to drive diapycnal mixing,
                                ! along the lines of St. Laurent et al. (2002) and Simmons et al. (2004).
CVMIX_TIDAL_SCHEME = SIMMONS    ! Use the Simmons et al (2004) tidal
                                ! mixing scheme.
TIDAL_ENERGY_TYPE = "JAYN"      !
                                ! The type of input tidal energy flux dataset. Valid values are   Jayne
                                !    ER03

---

The Laurent and Polzin schemes rely on current speed and bottom roughness to estimate tidal mixing, which is similar to the approach used in MOM5.

However, we currently do not have both files. I've reached out to Callum Shakespeare to kindly ask for suggestions.