Fesom2.7 recom3.1

[ogurses]

Update REcoM in FESOM2.7

[ogurses]

I am not allowed to add the following yml file.

.github/workflows/fesom2_recom.yml

name: "FESOM2: REcoM Biogeochemistry"

Controls when the action will run. Triggers the workflow on pull request only.

on:
workflow_dispatch: {}
pull_request:
types:
- opened
- synchronize
- reopened
branches:
- main

A workflow run is made up of one or more jobs that can run sequentially or in parallel

jobs:
general_test:
# Containers must run in Linux based operating systems
runs-on: ubuntu-latest
# Docker Hub image that container-job executes in
container: ghcr.io/fesom/fesom2_docker:fesom2_test_refactoring-master

# Service containers to run with `gfortran_ubuntu`
steps:
  # NK: this changes working directory to fesom2
  - uses: actions/checkout@v2

  - name: Git safe directory
    run: |
      git config --global --add safe.directory ${PWD}
  - name: Compile model with REcoM
    run: |
      ./configure.sh ubuntu -DRECOM_COUPLED=ON
  - name: Create REcoM test run
    run: |
      mkrun recom test_pi_recom -m docker

  - name: Add REcoM namelists
    run: |
      # mkrun doesn't create namelist.recom, so copy from config and patch
      # Using 2p3z2d (30 tracers) - more complete/tested than 2p1z1d
      cp config/bin_2p3z2d/namelist.recom work_recom/
      cp config/bin_2p3z2d/namelist.tra work_recom/

      # Patch file paths in namelists to use local files
      cd work_recom

      # Remove nam_rsbc section (not used by code, causes parsing errors)
      sed -i '/^&nam_rsbc/,/^\//d' namelist.recom

      # Set REcoM_restart to false for initialization from climatology
      sed -i "s|REcoM_restart.*=.*\.true\.|REcoM_restart         =.false.|" namelist.recom

      # Patch tracer file list in namelist.tra  
      sed -i "s|fe_pisces_opa_eq_init_3D_changed_name\.nc|fe5deg.nc|" namelist.tra
      sed -i "s|woa18_all_o00_01_mmol_fesom2\.nc|oxy5deg.nc|" namelist.tra
      sed -i "s|woa13_all_i00_01_fesom2\.nc|si5deg.nc|" namelist.tra
      sed -i "s|GLODAPv2\.2016b\.TAlk_fesom2_mmol_fix_z_Fillvalue\.nc|talk5deg.nc|" namelist.tra
      sed -i "s|GLODAPv2\.2016b\.TCO2_fesom2_mmol_fix_z_Fillvalue\.nc|tco2_5deg.nc|" namelist.tra
      sed -i "s|woa13_all_n00_01_fesom2\.nc|din5deg.nc|" namelist.tra
      sed -i "s|phc3\.0_winter\.nc|woa18_netcdf_5deg.nc|g" namelist.tra

      # Enable atmospheric pressure forcing for REcoM CO2 flux calculation
      sed -i "s|l_mslp *= *\.false\.|l_mslp=.true.|" namelist.forcing
     
      echo "--- Verifying patches ---"
      echo "REcoM restart: $(grep 'REcoM_restart' namelist.recom)"
      echo "Atmospheric pressure: $(grep 'l_mslp' namelist.forcing)"
     cd ..

      # Copy REcoM input data files
      cp test/input/recom/*.nc work_recom/

      # Create dust file with expected name in the source directory
      # (code reads from ClimateDataPath, not work directory)
      cd test/input/recom
      ln -s DustClimMonthlyAlbani_pimesh.nc DustClimMonthlyAlbani.nc
      cd ../../..
      
      # Copy forcing data files (from 1948)
      cp test/input/global/*.1948.nc work_recom/
      cp test/input/global/runoff.nc work_recom/
      cp test/input/global/PHC2_salx.nc work_recom/
        
  - name: FESOM2 REcoM test run
    run: |
      cd work_recom
      echo "--- Running REcoM test (32 timesteps, 1 day) ---"
      chmod +x job_docker_new
      ./job_docker_new

  - name: Check REcoM results
    run: |-
      cd work_recom
      fcheck .

[JanStreffing]

Hey @ogurses, & @patrickscholz. I suggest we first merge #863, then #802. Then we update this branch and merge here.

[suvarchal]

@copilot resolve the merge conflicts in this pull request

[Copilot]

I am not allowed to add the following yml file.

...

Added the workflow file at .github/workflows/fesom2_recom.yml in commit 9363d33. I also added explicit contents: read permissions to the GITHUB_TOKEN as required by security scanning (commit 4f84f2b).

@copilot resolve the merge conflicts in this pull request

Merge conflicts with origin/main have been resolved in commit 6cffefa. All REcoM 3.1 changes are preserved, and the new changes from main have been integrated. No conflict markers remain.

[ogurses]

@copilot resolve the merge conflicts in this pull request

I have the following problem now

/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4368): error #6457: This derived type name has not been declared. [MEANDATA0D]
type(Meandata0D), pointer :: entry0D
---------^
/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4462): error #6457: This derived type name has not been declared. [MEANDATA0D]
type(Meandata0D), intent(inout) :: entry0D
---------^
/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4378): error #6404: This name does not have a type, and must have an explicit type. [IO_XIOS_IS_ON]
if (io_xios_is_on()) return
--------^
/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4378): error #6341: A logical data type is required in this context. [IO_XIOS_IS_ON]
if (io_xios_is_on()) return
--------^
/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4380): error #6404: This name does not have a type, and must have an explicit type. [IO_NSTREAMS0D]
do n = 1, io_NSTREAMS0D
--------------^
/albedo/home/ogurses/test_case/trunk/src/io_meandata.F90(4381): error #6404: This name does not have a type, and must have an explicit type. [ENTRY0D]
entry0D => io_stream0D(n)
--------^

Some module is missing @suvarchal @JanStreffing

[JanStreffing]

788d0a3 — restored the 0D scalar-stream infrastructure that the Copilot merge resolution had dropped from src/io_meandata.F90. The merge kept the writer subroutines (output_0D_streams, create_0D_file) but removed (i) the use io_xios_module, only: io_xios_is_on, ... import block, (ii) def_stream0D in the public :: list, (iii) type Meandata0D ... end type Meandata0D, (iv) the module-level storage type(Meandata0D), save, target :: io_stream0D(50) and integer, save :: io_NSTREAMS0D = 0, and (v) the registrar subroutine def_stream0D(...). All five were restored verbatim from main, so the six errors you saw on MEANDATA0D / IO_XIOS_IS_ON / IO_NSTREAMS0D / ENTRY0D are resolved.

c2b4213 — renames enable_AWICM to recom_coupled_to_atmosphere and moves the declaration from recom_config into g_config. recom_config picks it up via use g_config, only: recom_coupled_to_atmosphere, so the &parecomsetup namelist binding in namelist.recom still works in recom builds, and gen_surface_forcing.F90 no longer needs preprocessor guards around its two reference sites. This fixes the 'Symbol enable_AWICM has no IMPLICIT type' errors on the non-recom CI jobs (fesom2_main, fesom2_icepack, fesom2_openmp, ...).

Rename motivation: in our group we typically use "AWICM" for the AOGCM-only coupled setup (FESOM + ECHAM, without biogeochemistry). The switch in &parecomsetup is only read when REcoM is active and controls SBC behaviour when REcoM is forced by a coupled atmosphere model, which is the ESM regime (AWESM3 / AWI-ESM), not the AOGCM regime, so the old name was a bit misleading. Semantics are unchanged and the default is .false., so existing setups continue to work; update_daily_flag is now initialised to .false. so the runoff_data_source=='AWICM' consumer is safe when the switch is off. Namelist defaults in config/bin_2p1z1d/, bin_2p3z2d/, bin_4p3z2d/ updated to the new key.