Regional capabilities for precip variability

[acordonez]

• Subset grid by latitude and longitude - CropLatLon function and parser updates
• Add "standard" calendar checks
• Update filenames for arbitrary grid size
• Update RegridHoriz function for arbitrary lat/lon resolution
• Update Avg_PS_DomFrq to add custom region option
• Add option to crop dataset via shapefile
• Check for any remaining cdat reference to remove
• Clean up new code
• Resolve conflicts
• Document new features

[acordonez]

Ready to review but running into CI issues due to same issue documented here: #1103

[lee1043]

@acordonez could you remind me an example commend to test this capability? Do we have a shape file example as a part of the demo download file?

[acordonez]

@acordonez could you remind me an example commend to test this capability? Do we have a shape file example as a part of the demo download file?

@lee1043 I am not aware of a shapefile in the demo file. I've been running this with an NCA5 regions shapefile that I can point you to on perlmutter for testing.

[lee1043]

@acordonez I am running Demo 7 notebook with this branch, and getting below error. Could this be related to the change in this PR?

Commend:

variability_across_timescales_PS_driver.py -p basic_precip_variability_param.py

Error message:

demo_data_tmp/CMIP5_demo_timeseries/historical/atmos/day/pr/
pr_day_GISS-E2-H_historical_r6i1p1_*.nc
[2000, 2005]
730 365
2
demo_output_tmp/precip_variability/GISS-E2-H/
demo_output_tmp/precip_variability/GISS-E2-H/
demo_output_tmp/precip_variability/GISS-E2-H/
['demo_data_tmp/CMIP5_demo_timeseries/historical/atmos/day/pr/pr_day_GISS-E2-H_historical_r6i1p1_20000101-20051231.nc']
GISS-E2-H.r6i1p1
['demo_data_tmp/CMIP5_demo_timeseries/historical/atmos/day/pr/pr_day_GISS-E2-H_historical_r6i1p1_20000101-20051231.nc']
GISS-E2-H.r6i1p1 365_day
syr, eyr: 2000 2005
Complete regridding from (2190, 90, 144) to (2190, 90, 180)
/home/lee1043/.conda/envs/pcmdi_metrics_dev_20240220/lib/python3.10/site-packages/pcmdi_metrics/precip_variability/lib/lib_variability_across_timescales.py:315: FutureWarning: 'H' is deprecated and will be removed in a future version, please use 'h' instead.
  climtime = pd.period_range(
Complete calculating climatology and anomaly for calendar of 365_day
Complete power spectra (segment:  730  nps: 5.0 )
Traceback (most recent call last):
  File "/home/lee1043/.conda/envs/pcmdi_metrics_dev_20240220/bin/variability_across_timescales_PS_driver.py", line 64, in <module>
    precip_variability_across_timescale(
  File "/home/lee1043/.conda/envs/pcmdi_metrics_dev_20240220/lib/python3.10/site-packages/pcmdi_metrics/precip_variability/lib/lib_variability_across_timescales.py", line 87, in precip_variability_across_timescale
    psdmfm_forced = Avg_PS_DomFrq(ps, freqs, ntd, dat, mip, "forced", regions_specs)
  File "/home/lee1043/.conda/envs/pcmdi_metrics_dev_20240220/lib/python3.10/site-packages/pcmdi_metrics/precip_variability/lib/lib_variability_across_timescales.py", line 496, in Avg_PS_DomFrq
    if "Ocean" in dom or regions_specs[dom].get("value", -1) == 0:
TypeError: 'NoneType' object is not subscriptable

[acordonez]

@lee1043 I've pushed a fix that I think should resolve this issue (it worked on the command line on perlmutter for me).

[lee1043]

@lee1043 I've pushed a fix that I think should resolve this issue (it worked on the command line on perlmutter for me).

Thanks, it helped resolving the aforementioned error. I am continuing to test the rest part of demo 7, and once they all run without error, I will try the shape file, and keep you updated.

[lee1043]

In demo 7, it looks like there was a naming change, did I get that right?:
PS_pr.day_regrid.180x90_area.freq.mean_GISS-E2-H.r6i1p1.json to PS_pr.day_regrid.3x3_area.freq.mean_GISS-E2-H.r6i1p1.json (180x90 to 3x3)

@acordonez Could you please double check running demo 7 all cells on your end and make any necessary updates?

[acordonez]

@lee1043 The file should not be renamed, that was a bug. I've pushed the fix and the notebook should run without error now.

[lee1043]

Confirmed that demo notebook 7 works fine with the PR.

[acordonez]

@lee1043 Thank you for your review!