Splitting h0 files to h0a and h0i will break some post-processing tools #3317
Description
Brief summary of bug
This issue becomes relevant once I merge #2445 this week and we start simulations with tags ctsm5.3.062 or newer. From today's conversation with @olyson and @wwieder:
- @slevis-lmwg started a 20-yr I1850 f09 simulation with ctsm5.3.062 to generate h0a output for post processing and a control with ctsm5.3.061: Test post-processing of new h0a/h0i files (ctsm53062_f09_1850) NCAR/LMWG_dev#109
- We expect ILAMB and CUPID (including LDF) to work with both h0 and h0a files, because history extensions get specified by the user before running these tools
- ILAMB: see caveats posted in Test post-processing of new h0a/h0i files (ctsm53062_f09_1850) NCAR/LMWG_dev#109
- CUPID: example notebooks should be checked for backwards compatibility
- LDF: really just need to modify the config file, but need to check for interoperability between funs that have
clm2.h0.andclm2.h0a.files (marking CUPID/LDF done here, as Will opened this issue CLM history variables name change,*clm2.h0a.*NCAR/CUPiD#252)
- Update the land diagnostic package: details in Test post-processing of new h0a/h0i files (ctsm53062_f09_1850) NCAR/LMWG_dev#109
- Consider revisiting tools that I updated in ctsm5.3.062: Put instantaneous and non-inst. fields on separate history files #2445 that will benefit from backwards compatibility:
- contrib/SpinupStability_*.ncl
- contrib/run_clm_historical
- site_and_regional/neon_gcs_upload (may revisit this one later if necessary)
- any others? (none have come up so far)
General bug information
CTSM version you are using: preparing to merge #2445 as tag ctsm5.3.062
Does this bug cause significantly incorrect results in the model's science? No
Configurations affected: All