SOCA updates

[guillaumevernieres]

We need to take care of the following when we decide to update to more recent version of develop:

• [x] the Rossby radius of deformation file was converted to netcdf
• [x] latlon from soca::Geometry is no more. It's used in the hybrid weight app
• [x] Might not be able to make a subset copy of an increment anymore (changes in ensemble stuff needed)
• [x] diffusion will be moved to oops, which will require some updates to diagb (probably not going to be merged soon ...)
• [x] diag B failing: Exception: Could not find edge 0 for Quadrilateral elem 0 with nodes ( 1 73 74 2 )

[RussTreadon-NOAA]

Update the following GDASApp submodules to the head of their respective develop as of 13:30 EST on 5/17/2024

        modified:   fv3-jedi (new commits)
        modified:   ioda (new commits)
        modified:   iodaconv (new commits)
        modified:   oops (new commits)
        modified:   saber (new commits)
        modified:   soca (new commits)
        modified:   ufo (new commits)
        modified:   vader (new commits)

build.sh -v aborted with

In file included from /work2/noaa/da/rtreadon/git/global-workflow/update/sorc/gdas.cd/bundle/gdas-utils/soca/gdas_socahybridweights.cc(1):
/work2/noaa/da/rtreadon/git/global-workflow/update/sorc/gdas.cd/bundle/gdas-utils/soca/gdas_socahybridweights.h(34): error: class "soca::Geometry" has no\
 member "latlon"
        geom.latlon(lats, lons, halo);
             ^

compilation aborted for /work2/noaa/da/rtreadon/git/global-workflow/update/sorc/gdas.cd/bundle/gdas-utils/soca/gdas_socahybridweights.cc (code 2)
make[2]: *** [gdas-utils/soca/CMakeFiles/gdas_socahybridweights.x.dir/gdas_socahybridweights.cc.o] Error 2
make[2]: Leaving directory `/work2/noaa/da/rtreadon/git/global-workflow/update/sorc/gdas.cd/build'
make[1]: *** [gdas-utils/soca/CMakeFiles/gdas_socahybridweights.x.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....

This is checklist item 2.

[guillaumevernieres]

Thanks for testing @RussTreadon-NOAA . I'll start working on this in a few mn ...

[RussTreadon-NOAA]

@guillaumevernieres , no need to rush on a Friday afternoon.

[guillaumevernieres]

@RussTreadon-NOAA , do you have a branch I could submit a PR fix to? There's only 2 issues that I know need to be fixed, the latlon issue that you mention above and a ascii file that became netcdf and changed name ...

[RussTreadon-NOAA]

@guillaumevernieres , I do not have a branch. I'm working from a local copy of GDASApp develop.

[RussTreadon-NOAA]

@guillaumevernieres , I just created GDASApp feature/soca_update. Feel free to push to this branch.

[RussTreadon-NOAA]

As a test, revert soca back to current GDASApp hash, soca @ 18f7813. After this ./build.sh completes. Run test_gdasapp including g-w tests on Orion. test_gdasapp_atm_jjob_var_run and test_gdasapp_atm_jjob_ens_run failed with

1: Exception: lapse_rate does not exist in ufo::PredictorFactory
1: Exception: oops::Variational<FV3JEDI, UFO and IODA observations> terminating...
1: Exception: level: 0
1: lapse_rate does not exist in ufo::PredictorFactory

Given the var_run and ens_run failures, the downstream var_inc, var_final, and ens_final also failed. A partial answer to the failure is found in fv-3jedi PR #1201. The var and ens failures will be addressed via a new GDASApp issue.

[CoryMartin-NOAA]

I think the VarBC predictors now are named like lapseRate rather than lapse_rate.

EDIT: guessing since you linked to the fv3-jedi PR you already knew that :-)

[guillaumevernieres]

@RussTreadon-NOAA , @CoryMartin-NOAA , I fixed the build issues, but I'm stuck with the diag B application failing. I'm going to postpone work on this until I talk to @travissluka later today.

[guillaumevernieres]

Fix is here: https://github.com/NOAA-EMC/GDASApp/tree/feature/submodules_soca