Put symlinks for ncclimo in unique directories

[xylar]

This prevents us from mixing data from the reference year with that for the "main" climatology.

Checklist

• Testing comment in the PR documents testing used to verify the changes

Fixes #1145

[xylar]

Testing

I am running:

ncclimo --no_stdin -4 --clm_md=mth -a sdd -P mpaso -p bck -j 12 -v timeMonthly_avg_activeTracers_salinity,timeMonthly_avg_activeTracers_temperature,... --seasons=JFM,JAS,ANN -s 0003 -e 0010 -o /pscratch/sd/x/xylar/analysis_testing/pm-cpu/fix-ncclimo-file-list/wc_defaults/clim/mpas/avg/unmasked_oQU240wLI mpaso.hist.am.timeSeriesStatsMonthly.0003-01-01.nc ...

where I have abbreviated the long list of variables and files. What I get out is:

Started climatology generation at Fri Apr 10 02:26:42 AM PDT 2026
Running climatology script ncclimo from directory /global/u2/x/xylar/MPAS-Analysis/fix-ncclimo-file-list/.pixi/envs/default/bin
NCO binaries version 5.3.7 from directory /global/u2/x/xylar/MPAS-Analysis/fix-ncclimo-file-list/.pixi/envs/default/bin
Parallelism mode for files = background
Background parallelism processing monthly climos in 12/job_nbr = 12/12 = 1 sequential batches each handling job_nbr = 12 months concurrently
Producing standard climatology from raw input files in directory /pscratch/sd/x/xylar/analysis_testing/pm-cpu/fix-ncclimo-file-list/wc_defaults/clim/mpas/avg/source_symlinks_mpaso_0003-0010
Output files to directory /pscratch/sd/x/xylar/analysis_testing/pm-cpu/fix-ncclimo-file-list/wc_defaults/clim/mpas/avg/unmasked_oQU240wLI
Climatology from 8 years of contiguous raw data touching 8 calendar years from YYYYMM = 000301 to 001012
Input filenames will be read (not constructed) from globbing, positional arguments, or stdin
Each input file assumed to contain statistic for one month
Winter statistics based on seasonally discontiguous December (jfd-mode): Means labeled as "DJF" are formed from JFD months from the same calendar year
Annotation for CF climatology attribute and climatology_bounds variable will be performed
This climatology will not be regridded
Generating climatology...
Climatological monthly mean for month 1 ...
Climatological monthly mean for month 2 ...
Climatological monthly mean for month 3 ...
Climatological monthly mean for month 4 ...
Climatological monthly mean for month 5 ...
Climatological monthly mean for month 6 ...
Climatological monthly mean for month 7 ...
Climatological monthly mean for month 8 ...
Climatological monthly mean for month 9 ...
Climatological monthly mean for month 10 ...
Climatological monthly mean for month 11 ...
Climatological monthly mean for month 12 ...
Climatological seasonal mean for JFM ...
Climatological seasonal mean for JAS ...
Climatological seasonal mean for ANN ...
Quick plots of last computed climatology:
ncvis /pscratch/sd/x/xylar/analysis_testing/pm-cpu/fix-ncclimo-file-list/wc_defaults/clim/mpas/avg/unmasked_oQU240wLI/mpaso_ANN_000301_000812_climo.nc &
Completed 8-year climatology operations for input data at Fri Apr 10 02:26:54 AM PDT 2026

The key issue is that the final year of the climatology is 0008, not 0010 as expected.

[xylar]

Will re-test with 5.3.8 and hopefully merge this fix, and make a new RC.

[xylar]

Testing

This fix together with nco 5.3.8 seems to be working well. Many runs are still underway or in the queue, so I'll update when all have finished:
https://portal.nersc.gov/cfs/e3sm/xylar/analysis_testing/pm-cpu/fix-ncclimo-file-list/

Update: Indeed, all tests passed.