Develop and test new infrared cloud detection routines

[wx20jjung]

We plan to test two different techniques to identify clear/cloudy infrared channels within the GSI. The two targeted are CO2_slicing (Wylie and Menzel) and dual regression (McNally and Watts).

[wx20jjung]

A branch was created with initial code changes for the CO2_slicing technique, jung_cloud_tests.

[RussTreadon-NOAA]

WCOSS2 ctests
Build wx20jjung:CADS_for_Andrew at ac16db4 on Cactus. Run ctests with the following results

Test project /lfs/h2/emc/da/noscrub/russ.treadon/git/gsi/pr616_fail_ctests.17nov/build
    Start 1: global_4denvar
    Start 2: rtma
    Start 3: rrfs_3denvar_glbens
    Start 4: netcdf_fv3_regional
    Start 5: hafs_4denvar_glbens
    Start 6: hafs_3denvar_hybens
    Start 7: global_enkf
1/7 Test #4: netcdf_fv3_regional ..............***Failed  542.82 sec
2/7 Test #3: rrfs_3denvar_glbens ..............   Passed  605.84 sec
3/7 Test #7: global_enkf ......................   Passed  669.95 sec
4/7 Test #2: rtma .............................   Passed  1102.28 sec
5/7 Test #6: hafs_3denvar_hybens ..............***Failed  1217.38 sec
6/7 Test #5: hafs_4denvar_glbens ..............***Failed  1343.83 sec
7/7 Test #1: global_4denvar ...................***Failed  1442.46 sec

43% tests passed, 4 tests failed out of 7

Total Test time (real) = 1443.22 sec

The following tests FAILED:
          1 - global_4denvar (Failed)
          4 - netcdf_fv3_regional (Failed)
          5 - hafs_4denvar_glbens (Failed)
          6 - hafs_3denvar_hybens (Failed)
Errors while running CTest

All four tests failed due to non-reproducible results between the updat and contrl gsi.x.

Make the following changes in a working copy of wx20jjung:CADS_for_Andrew on Cactus:

• src/gsi/cads.f90
  • move source code for subroutinecads_imager_calc from setuprad.f90 to cads.f90
  • remove radmod from the subroutine argument list and, instead, make radmod a local structure within cads_imager_calc
• src/gsi/crtm_interface.f90
  • remove if cads logic around block which computes chan_level. This means all calls to call_crtm compute chan_level
  • alternatively, we could move the chan_level calculation outside the omp parallel section and into its own loop inside an if cads block.
• src/gsi/qcmod.f90
  • notice that subroutine qc_geocsr does not follow the cld & cloudp pattern found in other qc routines. Add cloudp and use it to load cld as is done in other routines.
  • this change is probably not related to lack of WCOSS2 reproducibility.
• src/gsi/setuprad.f90
  • remove source code for cads_imager_calc
  • initialize imager_* variables to zero for all observations. This change may not be necessary.
  • remove radmod from the call to cads_imager_calc
  • replace the constant integer indices for imager_* arrays in the calling list for cads_imager_calc with :.

Recompile working copy of wx20jjung:CADS_for_Andrew with the above changes and rerun ctests on Cactus. All tests pass.

Test project /lfs/h2/emc/da/noscrub/russ.treadon/git/gsi/pr616_pass_ctests.17nov/build
    Start 1: global_4denvar
    Start 2: rtma
    Start 3: rrfs_3denvar_glbens
    Start 4: netcdf_fv3_regional
    Start 5: hafs_4denvar_glbens
    Start 6: hafs_3denvar_hybens
    Start 7: global_enkf
1/1 Test #4: netcdf_fv3_regional ..............   Passed  483.14 sec
2/7 Test #3: rrfs_3denvar_glbens ..............   Passed  485.42 sec
3/7 Test #7: global_enkf ......................   Passed  609.92 sec
4/7 Test #2: rtma .............................   Passed  969.69 sec
5/7 Test #6: hafs_3denvar_hybens ..............   Passed  1211.04 sec
6/7 Test #1: global_4denvar ...................   Passed  1322.89 sec
7/7 Test #5: hafs_4denvar_glbens ..............   Passed  1331.60 sec

100% tests passed, 0 tests failed out of 7


Total Test time (real) = 1331.62 sec

The modified working copy of wx20jjung:CADS_for_Andrew has been copied to Hera. Look in directory /scratch2/NCEPDEV/stmp1/Russ.Treadon/pr616_pass_ctests.17nov

Hera(hfe11):/scratch2/NCEPDEV/stmp1/Russ.Treadon/pr616_pass_ctests.17nov$ git status
On branch CADS_for_Andrew
Your branch is up to date with 'origin/CADS_for_Andrew'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git checkout -- <file>..." to discard changes in working directory)

        modified:   src/gsi/cads.f90
        modified:   src/gsi/crtm_interface.f90
        modified:   src/gsi/qcmod.f90
        modified:   src/gsi/setuprad.f90

no changes added to commit (use "git add" and/or "git commit -a")

[wx20jjung]

My latest CADS test with CrIS also look correct. Please start the pull request again.

[RussTreadon-NOAA]

@wx20jjung , please go to PR #616 and request reviews from @ADCollard, @erinjones2, and @RussTreadon-NOAA. For Andrew and me, click on the Re-request review circle icon.