Test Ufuk's feature/era5 refactor

[NickSzapiro-NOAA]

Description of changes

Specific notes

Contributors other than yourself, if any:

CDEPS Issues Fixed (include github issue #):

Are there dependencies on other component PRs (if so list):

Are changes expected to change answers (bfb, different to roundoff, more substantial):

Any User Interface Changes (namelist or namelist defaults changes):

Testing performed (e.g. aux_cdeps, CESM prealpha, etc):

Hashes used for testing:

[gemini-code-assist]

Code Review

This pull request introduces a skip_field_check flag to bypass field validation in the DATM component, adds support for several new ERA5 stream variables, and refactors unit conversion and stream reading logic. The review feedback highlights several critical issues in datm_datamode_era5_mod.F90, including a scaling bug in Faxa_swnet unit conversion, a missing error status argument in ESMF_MeshGet, a memory leak from an undeallocated pointer, and flawed logic for determining stream size. Additionally, the reviewer noted potential segmentation faults and undefined behavior due to missing association checks on wind components and uninitialized temperature/pressure variables, an omitted unit conversion for Faxa_snow, and an unused cosfactor calculation.

[gemini-code-assist]

Multiplying Faxa_swnet(n) by 3600.0_r8 on line 635 is incorrect when lwmax < 1.0e4_r8 (i.e., when flux variables are already in W/m^2). In that case, the division by 3600.0_r8 in the unit conversion section is skipped, leaving Faxa_swnet 3600 times too large. Since strm_Faxa_swdn already has the correct units (either W/m^2 or J/m^2), Faxa_swnet naturally inherits the correct units from it. Removing line 635 completely resolves this bug for both cases.

      if (associated(strm_Faxa_swdn)) then
         if (.not. associated(strm_Faxa_swnet) .and. associated(Faxa_swnet)) then
            avg_alb = ( 0.069 - 0.011*cos(2.0_R8*yc(n)*degtorad ) )
            Faxa_swnet(n) = strm_Faxa_swdn(n)*(1.0_R8 - avg_alb)
         end if
      end if

[gemini-code-assist]

The call to ESMF_MeshGet does not pass the rc argument, but the subsequent line checks rc using ChkErr. This means the error check will not detect a failure of this call and will instead check the status of the previous ESMF call. Please add rc=rc to the ESMF_MeshGet call.

    call ESMF_MeshGet(sdat%model_mesh, ownedElemCoords=ownedElemCoords, rc=rc)

[gemini-code-assist]

ownedElemCoords is declared as a local pointer and allocated on line 192, but it is never deallocated, leading to a memory leak. Since it is only used locally to populate yc, it is highly recommended to declare it as allocatable instead of pointer. This ensures it is automatically deallocated when the subroutine exits, even in case of early returns.

    real(r8), allocatable :: ownedElemCoords(:) ! mesh lat and lons

[gemini-code-assist]

The logic to determine lsize will always fail and exit if Sa_pslv is not associated, even if Sa_z is associated and has been used to set lsize. If the intention is that at least one of Sa_z or Sa_pslv must be associated, the logic should be refactored to only fail if both are unassociated.

    lsize = -1
    if (associated(Sa_z)) then
       lsize = size(Sa_z)
    else if (associated(Sa_pslv)) then
       lsize = size(Sa_pslv)
    end if

    if (lsize < 0) then
       if (mainproc) write(logunit,*) subname,' Neither Sa_z nor Sa_pslv is given. Exiting!'
       call shr_log_error(subname//'ERROR: Sa_z and/or Sa_pslv must be associated for era5 datamode', rc=rc)
       return
    end if

[gemini-code-assist]

The wind components Sa_u, Sa_v, Sa_u10m, and Sa_v10m are accessed and assigned to without checking if they are associated. Since these pointers are retrieved from the export state with allowNullReturn=.true., they can be unassociated, which would cause a segmentation fault. Please add association checks before accessing them.

      !--- calculate wind components if wind speed is provided ---
      if (associated(strm_Sa_wspd)) then
         if (associated(Sa_u)) Sa_u(n) = strm_Sa_wspd(n)/sqrt(2.0_r8)
         if (associated(Sa_v)) Sa_v(n) = strm_Sa_wspd(n)/sqrt(2.0_r8)
      end if
      if (associated(strm_Sa_wspd10m)) then
         if (associated(Sa_u10m)) Sa_u10m(n) = strm_Sa_wspd10m(n)/sqrt(2.0_r8)
         if (associated(Sa_v10m)) Sa_v10m(n) = strm_Sa_wspd10m(n)/sqrt(2.0_r8)
      end if

[gemini-code-assist]

If neither strm_Sa_t2m nor strm_Sa_tbot is associated, tbot remains uninitialized. Similarly, if neither strm_Sa_pslv nor strm_Sa_pbot is associated, pbot remains uninitialized. This can lead to undefined behavior, division by zero, or floating-point exceptions when calculating e and qsat. Please initialize them to safe default values or guard the calculation block to ensure they are associated.

      !--- calculate specific humidity from dew point temperature ---
      if (associated(strm_Sa_tdew)) then
         tbot = 0.0_r8
         if (associated(strm_Sa_t2m)) then
            tbot = strm_Sa_t2m(n)
         else if (associated(strm_Sa_tbot)) then
            tbot = strm_Sa_tbot(n)
         end if

         pbot = 0.0_r8
         if (associated(strm_Sa_pslv)) then
            pbot = strm_Sa_pslv(n)
         else if (associated(strm_Sa_pbot)) then
            pbot = strm_Sa_pbot(n)
         end if

         if (tbot > 0.0_r8 .and. pbot > 0.0_r8) then
            if (td2max < 50.0_r8) strm_Sa_tdew(n) = strm_Sa_tdew(n) + tkFrz
            e = datm_eSat(strm_Sa_tdew(n), tbot)
            qsat = (0.622_r8 * e)/(pbot - 0.378_r8 * e)
            if (associated(Sa_q2m)) Sa_q2m(n) = qsat
            if (associated(Sa_shum)) Sa_shum(n) = qsat
         end if
      end if

[gemini-code-assist]

Faxa_snow is missing from the unit conversion section. When precmax >= 0.01_r8, all other precipitation variables (Faxa_rain, Faxa_rainc, Faxa_rainl, Faxa_snowc, Faxa_snowl) are converted from m to kg/m^2/s, but Faxa_snow is left unconverted. Please add Faxa_snow to the conversion list.

       if (associated(Faxa_rain))  Faxa_rain(:)  = Faxa_rain(:)/3600.0_r8*rhofw
       if (associated(Faxa_rainc)) Faxa_rainc(:) = Faxa_rainc(:)/3600.0_r8*rhofw
       if (associated(Faxa_rainl)) Faxa_rainl(:) = Faxa_rainl(:)/3600.0_r8*rhofw
       if (associated(Faxa_snow))  Faxa_snow(:)  = Faxa_snow(:)/3600.0_r8*rhofw
       if (associated(Faxa_snowc)) Faxa_snowc(:) = Faxa_snowc(:)/3600.0_r8*rhofw
       if (associated(Faxa_snowl)) Faxa_snowl(:) = Faxa_snowl(:)/3600.0_r8*rhofw

[gemini-code-assist]

cosfactor (and the intermediate rday) is calculated here but never used anywhere else in the subroutine. If this calculation is not needed, these lines and the declarations of cosFactor and rday should be removed to avoid redundant computations.