cam - errors in severe/radar verification

[PerryShafran-NOAA]

In testing PR #682 to change KEEPDATA to NO, it was noticed in testing of severe and radar prep and stats jobs that there were some errors and warnings that need to be addressed. Examples here:

In jevs_cam_severe_prep, the following warning occurred:

jevs_cam_severe_prep.o19257940:WARNING: 
jevs_cam_severe_prep.o19257940:WARNING: Zero observations retained!
jevs_cam_severe_prep.o19257940:WARNING: 

In several of the severe stats jobs, the following warnings existed:

echo 'WARNING: Missing file is /lfs/h2/emc/vpppg/noscrub/emc.vpppg/evs/v2.0/prep/cam/nam.20250310/namnest.t12z.MXUPHL25_A24.SSPF.2025031112-2025031212.f48.nc\n'

WARNING: 
WARNING: process_scores() -> Forecast and observation valid times do not match (20250312_120000 != 19700101_000000) for Prob_MXUPHL25_A24_geHWT(*,*) versus Fscale_mask(*,*).
WARNING: 

The above warnings were also seen in the href severe stats job, but also this error:

WARNING: 
WARNING: read_data_plane() -> Prob_MXUPHL25_A24_geHWT(*,*) not found in forecast file: /lfs/h2/emc/vpppg/noscrub/emc.vpppg/evs/v2.0/prep/cam/href.20250311/href.t00z.MXUPHL25_A24.SSPF.2025031112-2025031212.f36.nc
WARNING: 
ERROR  : 
ERROR  : process_scores() -> no requested data found!  Exiting...
ERROR  : 

A FATAL ERROR followed after that.

See the log files for prep here: /lfs/h2/emc/vpppg/noscrub/alicia.bentley/pr682test/EVS/dev/drivers/scripts/prep/cam
See the log files for stats here: /lfs/h2/emc/vpppg/noscrub/alicia.bentley/pr682test/EVS/dev/drivers/scripts/stats/cam

All these issues should be resolved for both the RRFS branch (in case these errors are found there as well) as well as the develop branch.

[AliciaBentley-NOAA]

@MarcelCaron-NOAA Did these errors in severe/radar verification get addressed in EVS v2.0, or are they things we need to fix in EVS v2.1?

CC @AndrewBenjamin-NOAA

[MarcelCaron-NOAA]

@AliciaBentley-NOAA The warning:

WARNING:
WARNING: process_scores() -> Forecast and observation valid times do not match (20250312_120000 != 19700101_000000) for Prob_MXUPHL25_A24_geHWT(,) versus Fscale_mask(,).
WARNING:

... is unresolved but we agreed with Russell to address it in a future release, which involves changing the default behavior of setting date metadata to 1/1/1970 for the practically perfect grid on days with no SPC reports. From Russell's Jan 14 2026 email:

"I agree that addressing the warning messages for evs_stats_cam_*_severe jobs can wait until the next upgrade. I will leave the bugzilla open and communicate to my team lead why it was left open when we review the IT documents."

The other issues were resolved in the following PR:
#721 ("Zero observations retained...")

Since then, all of the other warnings and errors have been resolved and haven't appeared in EVS logs or subsequent test PRs (including the recent #895, which tested rrfs severe stats). HREF severe stats have been clean in this way as far as I know (@BinbinZhou-NOAA can correct me if I'm wrong).