Adding surface flux analysis notebook for OM4_05 (heat is not closing)

[gustavo-marques]

This PR has the following purposes:

1. Add all necessary fields in the diag_table.MOM6, using double precision, so that the budget analysis employed in the notebook provided can be closed.

2. Add surface flux analysis notebook for the OM4_05 test. This notebook has been tested using the latest MOM6 dev/master. Mass and salt budgets close within round-off values. However, heat budget is not closing (see below) -- perhaps I am missing something. Notebook figures have been omitted to reduce the amount of data. A complete example of this notebook can be found here.

Before running this notebook the following must be done:

1. cp diag_table.MOM6 diag_table
2. Set days = 3 in input.nml

Since ice_ocean_SIS/GOLD_SIS is no longer tested, I suggest we change the link for the surface flux budgets tutorial in ice-ocean mode to this newly added notebook.

[StephenGriffies]

The figures for the surface heat fluxes all look fine; self-consistency across all terms. Even so, the global heat content change does not agree with the net heat entering across the boundaries; that is clear. I do not have a good idea.

But one question: does the heat budget close when changing the number of time steps from one time step (which is a new setting) to the original setting of three days? Perhaps there is an issue with lagging? Could you try that?

[gustavo-marques]

@StephenGriffies, I don't understand your question.

The new setting, which is used in ice_ocean_SIS2/OM4_05, has the following:

• DT = 1800.0 s
• DT_THERM = 4 x DT s
• dt_cpld = 3600 s
• THERMO_SPANS_COUPLING = True (MOM will take thermodynamic and tracer time steps that can be longer than the coupling time step).

The original setting, which is used in ice_ocean_SIS/GOLD_SIS, has the following:

• DT = 3600.0 s
• DT_THERM = 2 x DT s
• dt_cpld = 7200 s
• THERMO_SPANS_COUPLING = False

The budgets are performed over one day, as in the original setting.

Are you suggesting that we should try these original values and see if the heat budget closes?

[StephenGriffies]

@adcroft and I spoke. We think it prudent to test the THERMO_SPANS_COUPLING = False option, as the code with THERMO_SPANS_COUPLING = True was developed after the development of the heat conservation tests. We are unsure why salt should be fine but heat is problematic. But as a point of reference, it would be useful to return to these original settings.

Also, my suggestion to return to 3 days was based on what I thought was now a move to a single time step. I do not recall where I read that you switched to a single time step; perhaps that is not the case, in which case ignore my suggestion concerning length of the test.

So bottomline is that it would be useful to start with the same settings as for the original case, in particular to set THERMO_SPANS_COUPLING = False.

[gustavo-marques]

I've tested the following options:

• DT = 3600.0 s
• DT_THERM = 7200 s
• dt_cpld = 7200 s
• THERMO_SPANS_COUPLING = False

Heat budget still does not close:

I also tried changing just 1) THERMO_SPANS_COUPLING = False, 2) USE_MEKE = False, 3) DIABATIC_FIRST = True and DT_THERM = 3600.0, and 4) KHTH_USE_EBT_STRUCT = False, but heat budget does not close.