Re-factor initialize-from-z to work on individual tracers

[adcroft]

The initialize-from-z code works in two modes, one assuming the isopycnal layer coordinate and the second assume T and S are the only inputs.

The second mode needs to be re-factored so we can call it for individual tracers and thus provide an initialize-from-z function to COBALT.

[adcroft]

@MJHarrison-GFDL and I were discussing this earlier today.

• We think we need a self-contained (more kernelized) function that takes inputs of the lateral grid, vertical grid and a filename and returns a scalar field using the same algorithm used to T/S.
• We can test the new routine by making sure it gives the same answers when used for T/S.
• Some options need to be passed to the remapping routines (e.g. order of accuracy, extrapolation choices) but few enough that I think we can pass them as arguments and avoid using types.
• The routine MOM_temp_salt_initialize_from_Z() is currently close to 1k lines and has several "contained" subroutines. Should this function be in a self-contained module?

[MJHarrison-GFDL]

...working on finalizing a new interface for generic tracers.

current version based on discussions with @nikizadehgfdl and Jasmin John:

MOM_initialize_tracer_from_z(h,tr,G,PF,src_file,src_var,unit_conversion,recnum, homogenize,useALE,remappingScheme,src_gridspec)

italics==optional

Defaults are based on USE_REGRIDDING and REMAPPING_SCHEME parameters.

[nikizadehgfdl]

@MJHarrison-GFDL , do you mean to create a new routine to generalize MOM_temp_salt_initialize_from_Z() to step through all the known tracers (COBALT tracers, ...) and initialize the ones that are requested to be initialized by the generic_tracer package?
So the call sequence in MOM_initialize.F90 would be something like:

    if (from_Z_file) then
!     Initialize thickness and T/S from z-coordinate data in a file.
      if (.NOT.use_temperature) call MOM_error(FATAL,"MOM_initialize : "//&
         "use_temperature must be true if INIT_LAYERS_FROM_Z_FILE is true")

      call MOM_temp_salt_initialize_from_Z(h, tv, G, PF, dirs)

>     call  MOM_initialize_tracer_from_z(h,tr,G, ...)   !Loops through tracers in tr and asks generics which should be initialized and mimic the above temp_salt call for them 

      call pass_var(h, G%Domain)
    else

[nikizadehgfdl]

MOM6 does not know about the generic tracers until after the call to tracer_flow_control_init() in MOM.F90
So a call like MOM_initialize_tracer_from_z above must appear after the call to tracer_flow_control_init() in MOM.F90
and cannot happen along with T&S.

So why not replace the existing function call to tracer_Z_init() in initialize_MOM_generic_tracer() with MOM_initialize_tracer_from_z() ?

[adcroft]

Status on this issue?

[Hallberg-NOAA]

This capability has existed via calls to MOM_initialize_tracer_form_Z for 5 years now (since September 1, 2015). This issue can be closed. If there are specific issues that remain, please open a new issue ticket describing them.