We can use the regional enkf ctest in PR #750 to test the changes in this PR, right?

@RussTreadon-NOAA Yes, we can use the regional EnKF case from PR #750. But need to turn on the case manually.

@RussTreadon-NOAA Yes, we can use the regional EnKF case from PR #750. But need to turn on the case manually.

Good. So we can test if initializing variables to zero has any impact ... at least in the context of the regional enkf ctest.

@RussTreadon-NOAA Thank you for reviewing the changes. I will start an offline test.

Those added initialization are in "dummy " subroutines ( as shown by those lines

write(6,*)'this is a dummy subroutine, running this means something wrong ,stop'
   call stop2(555)

So, in my opinion, no need to do extra test to see if the will have impacts on the results if we are sure there are no changes outside of those dummy subroutines ( while the whole piece (observer_fv3reg.f90) is for those dummy subroutine)

@ShunLiu-NOAA , what is the status of this PR?

@ShunLiu-NOAA , what is the status of this PR?

Test project /scratch1/NCEPDEV/da/Shun.Liu/gsi_regression/GSI/build
Start 1: global_4denvar
Start 2: rtma
Start 3: rrfs_3denvar_rdasens
Start 4: hafs_4denvar_glbens
Start 5: hafs_3denvar_hybens
Start 6: global_enkf
1/6 Test #3: rrfs_3denvar_rdasens ............. Passed 859.46 sec
2/6 Test #6: global_enkf ...................... Passed 1298.20 sec
3/6 Test #5: hafs_3denvar_hybens .............. Passed 1403.31 sec
4/6 Test #2: rtma ............................. Passed 2170.92 sec
5/6 Test #4: hafs_4denvar_glbens .............. Passed 2186.12 sec
6/6 Test #1: global_4denvar ................... Passed 2886.83 sec

100% tests passed, 0 tests failed out of 6

@hu5970 ctest on WCOSS and HERA passed. The PR is ready to merge into develop.

Thank you @ShunLiu-NOAA for running ctests. My previous approval of this PR remains. These changes address a WCOSS Implementation Standard

5. Code must compile without errors or warnings. Errors and warnings may not be suppressed, and the compiler warning level ("-W" options) must be at least the default one.

@hu5970 , is it possible for you to run the regional enkf ctest for the sake of completeness? Initializing variable to izero or zero should have no impact on regional enkf results. It would be nice to confirm this.

@RussTreadon-NOAA Thank you. I will coordinate with @hu5970 for a regional enkf test.

I run a regression test suite with the regional EnKF case on and there are the results:

Test project /scratch1/BMC/wrfruc/mhu/code/gsi/GSI/build
    Start 1: global_4denvar
    Start 2: rtma
    Start 3: rrfs_3denvar_rdasens
    Start 4: hafs_4denvar_glbens
    Start 5: hafs_3denvar_hybens
    Start 6: rrfs_enkf_conv

1/6 Test #6: rrfs_enkf_conv ...................   Passed  303.82 sec
2/6 Test #3: rrfs_3denvar_rdasens .............   Passed  492.80 sec
3/6 Test #2: rtma .............................   Passed  1093.14 sec
4/6 Test #5: hafs_3denvar_hybens ..............   Passed  1109.30 sec
5/6 Test #4: hafs_4denvar_glbens ..............   Passed  1349.62 sec
6/6 Test #1: global_4denvar ...................   Passed  2028.61 sec

100% tests passed, 0 tests failed out of 6

Total Test time (real) = 2028.63 sec

I also updated the ctest instruction on how to run ctest with regional enkf case.