Tidy up more compiler warnings

[NickSzapiro-NOAA]

Address remaining CCPP physics warnings in ufs-community/ufs-weather-model#2703 of

.../ufs-weather-model/FV3/ccpp/physics/physics/Interstitials/UFS_SCM_NEPTUNE/sfcsub.F:3097:72:
 3097 |               data(imax-i+1,jj) = work(i,j)
Warning: 'jj' may be used uninitialized [-Wmaybe-uninitialized]

.../ufs-weather-model/FV3/ccpp/physics/physics/GWD/cires_ugwpv1_triggers.F90(301): warning #6717: This name has not been given an explicit type.   [TLIM_FGF]
          if (dtot >= tlim_fgf ) kex = kex+1

.../ufs-weather-model/FV3/ccpp/physics/physics/GWD/cires_ugwpv1_oro.F90(14): warning #6843: A dummy argument with an explicit INTENT(OUT) declaration is not given an explicit value.   [ZLWB]
                      zobl, zlwb, zogw, tau_ogw, dudt_ogw, dvdt_ogw,      &

ufs-weather-model regression tests pass bit-for-bit on hera. As for the changes:

1. I'm mostly guessing at the "fix" for jj based on neighboring logic. I don't think any RTs exercise dlat > 0.0, dlon <= 0.0 for sfcsub.F
2. tlim_fgf is defined above this in get_spectra_tau_nstgw but not in get_spectra_tau_okw
3. Setting zlwb=0 is as in code and Issue 1984 fix: reduce warnings for ccpp-physics #149

[grantfirl]

@NickSzapiro-NOAA To speed up the UFS merge queue, I'm testing this combined with #295 on Hera. If successful, I'm going to combine this into that PR.

[grantfirl]

I'm not an expert in the sfcsub.F code, but I would have done the same fix, so perhaps two people guessing are correct?

[grantfirl]

It's possible that @GeorgeGayno-NOAA might have a better idea if the changes to sfcsub.F are correct.

[grantfirl]

OK, the combined Hera RTs were successful except for a timeout for hafs_regional_atm_wav_intel, so I'm going to combine into #295 .

[GeorgeGayno-NOAA]

It's possible that @GeorgeGayno-NOAA might have a better idea if the changes to sfcsub.F are correct.

I am not too familiar with that routine, but your changes look ok. That appears to be a big error. What is the impact of the code fix on your tests?

[NickSzapiro-NOAA]

There's no change from this in ufs-weather-model RTs (presumably as we don't get into this dlat > 0.0, dlon <= 0.0 logic). I don't know if UFS_UTILS has tests/datasets that do that

[GeorgeGayno-NOAA]

The sfcsub.F module is used by the UFS_UTILS global_cycle program. I made the code fix, recompiled global_cycle and reran the regression tests. One of the tests mimics how global_cycle is used in GFS v16 OPS. All tests passed. And that particular 'if' branch was never accessed. That routine appears to remap the input data so it has a consistent N/S E/W orientation. Apparently, we don't have any input data that requires that type of remapping. Otherwise, we would have discovered this bug sooner.

[grantfirl]

The sfcsub.F module is used by the UFS_UTILS global_cycle program. I made the code fix, recompiled global_cycle and reran the regression tests. One of the tests mimics how global_cycle is used in GFS v16 OPS. All tests passed. And that particular 'if' branch was never accessed. That routine appears to remap the input data so it has a consistent N/S E/W orientation. Apparently, we don't have any input data that requires that type of remapping. Otherwise, we would have discovered this bug sooner.

@GeorgeGayno-NOAA Thanks for your input/help!