2024‐09‐26

Sept 26, 2024

Agenda:

1. Updates
2. ADF climatology presentation
3. x4c climatology presentation
4. Climatology in CUPiD discussion

Notes

• Updates:
  • manage_externals -> git submodules [update README with advice for updating existing sandboxes]
• Discussion:
  • ADF
    • monthly climatology calls scripts/averaging/create_climo_files.py
      • inputs are time series files
      • parallelizes by variable
    • process_variable() does the heavy lifting
      • currently hard-coded for time bounds as CAM, but CUPiD should have code for other components
    • may want to support daily climo
  • x4c
    • also does time series -> climatology
      • create a Timeseries Case object
      • call gen_climo()
      • can regrid on the fly with this
      • can specify specific variables, or just a component
    • For seasonal / annual, currently does straight mean rather than weighting by day
    • Currently seasons are hard-coded (DJF, etc) -- sea ice might prefer flexibility to look at JFM, ...
  • geoCAT also offers general climatology generation tools, but might not handle CESM's time bounds correctly
  • xarray has resampling, but Dave B had trouble: wanted to reshape from (1200 months, x, y) to (100 years, 12 months, x, y) but xarray couldn't get there
  • Questions / next steps:
    • Should we compute all time series and then move on to climatologies, or compute time series and climatology one variable at a time?
    • ADF team will put together a key metric notebook that looks at a couple plots (including climatology)

Slides