Bug fix for HRRR IFI: Use ifi_apcp if it is 0

[SamuelTrahanNOAA]

This is the bug fix described in detail here:

• fixes IFI.F can use junk values for hourly accumulated precipitation #1126

It corrects a bug in how accumulated hourly precipitation is sent to IFI from within UPP.

I've tested it with the regression tests in this branch:

• Add IFI to regression tests (and put regression tests in repository) #1119

With these changes, the HRRR UPP IFI and standalone IFI match, which they don't in develop now.

Also, in testing, @WenMeng-NOAA discovered another problem which this branch fixes:

• run_post_hrrr_ifi_HERA.sh copies CRTM fix files it doesn't need #1156

EDIT: The original title of the Issue was wrong. The troubled script is "run_post_hrrr_ifi_HERA.sh"

[SamuelTrahanNOAA]

This PR is expected to change IFI results due to the bugfix in IFI.F

It is not expected to change results of other tests. Also, it shouldn't change results of the "ifi missing" tests.

[SamuelTrahanNOAA]

My tests produce a change I don't expect:

.../upp-HERA/rap_2020072316_pe_test/WRFPRS.GrbF16.diff:1:708:195559853:CDCON:convective cloud layer:rpn_corr=-nan:rpn_rms=undefined

There are no commits in this PR which should change the result of the rap_2020072316_pe_test

[WenMeng-NOAA]

My tests produce a change I don't expect:

.../upp-HERA/rap_2020072316_pe_test/WRFPRS.GrbF16.diff:1:708:195559853:CDCON:convective cloud layer:rpn_corr=-nan:rpn_rms=undefined

There are no commits in this PR which should change the result of the rap_2020072316_pe_test

@SamuelTrahanNOAA This is a glitch in the UPP. So far, there is no smoking gun for this metadata change. The workaround is updating WRFPRS.GrbF16.diff in the baseline.

[WenMeng-NOAA]

@SamuelTrahanNOAA My UPP RTs including IFI are /scratch1/NCEPDEV/stmp2/Wen.Meng/upp-HERA on Hera. Somehow, the test 'hrrr_ifi_2020060118' loaded unused crtm fix files, e.g. imgr_*. I don't see that loading in ci/jobs-dev/run_post_hrrr_ifi_HERA.sh. Do you any ideas why these crtm fix files were copied into the run directory?

[SamuelTrahanNOAA]

Is this doing it?

#copy fix data                                                                                                                                                                                                                                                                    
cp $homedir/fix/fix_2.3.0/*bin .

[WenMeng-NOAA]

Is this doing it?

#copy fix data                                                                                                                                                                                                                                                                    
cp $homedir/fix/fix_2.3.0/*bin .

@SamuelTrahanNOAA That's it. Please remove this line. Thanks!

[SamuelTrahanNOAA]

I've removed the unneeded cp command. Can someone please retest this?

[SamuelTrahanNOAA]

I opened an issue for the unneeded cp:

• run_post_hrrr_ifi_HERA.sh copies CRTM fix files it doesn't need #1156

Merging this PR will automatically close the issue.

EDIT: The original title of the Issue was wrong. The troubled script is "run_post_hrrr_ifi_HERA.sh"

[WenMeng-NOAA]

I've removed the unneeded cp command. Can someone please retest this?

@SamuelTrahanNOAA The issue of crtm fix files was solved in my latest run at /home/Wen.Meng/stmp2/upp-HERA/hrrr_ifi_2020060118. I would think the changes in 'IFIFIP.GrbF04.diff' are expected with this PR.

[WenMeng-NOAA]

@gspetro-NOAA This PR is ready for the UPP RTs on R&D machines. From the EPIC end, there should be no baseline changes since IFI tests are not included. You may copy the new baseline for hrrr_ifi from my testing at /home/Wen.Meng/stmp2/upp-HERA/hrrr_ifi_2020060118 on Hera.

[SamuelTrahanNOAA]

Any IFI GRIB file with libIFI enabled is expected to change results due to the bug fix.

No other files should change, including the "IFI missing" tests.

[gspetro-NOAA]

RTs pass on Hera, Orion, and Hercules with no baseline changes for non-IFI RTs. I updated the baseline for hrrr_ifi.
@WenMeng-NOAA Should there be an update for fv3r_ifi?

[SamuelTrahanNOAA]

The fv3r_ifi probably won't change. If it does, that doesn't indicate anything is wrong.

[WenMeng-NOAA]

The fv3r_ifi probably won't change. If it does, that doesn't indicate anything is wrong.

From my testing, that's right.

[WenMeng-NOAA]

This PR is ready for merging.