E3SM-Project/e3sm_diags

[Bug]: CDAT Migration branch failing integration test due to regrid not maintaining correct units

Closed this issue · 2 comments

What happened?

The following plots differ because the variable's actual units are "mb" when the expected is "K":

  • ERA-Interim-T-850-ANN-global
  • ERA-Interim-T-850-ANN-CONUS_RRM -- the diff of the third plot shows large differences in metrics too. Need further investigation with this one.

What did you expect to happen? Are there are possible answers you came across?

This should be resolved by in PR #752, which includes a fix to maintain the variables units after regridding from hybrid to pressure levels.

Fix: #752 (comment)

Minimal Complete Verifiable Example (MVCE)

No response

Relevant log output

No response

Anything else we need to know?

ERA-Interim-T-850-ANN-global

Actual:

ERA-Interim-T-850-ANN-global_actual

Expected:
ERA-Interim-T-850-ANN-global_expected

Diff:
ERA-Interim-T-850-ANN-global_diff

ERA-Interim-T-850-ANN-CONUS_RRM

Actual:
ERA-Interim-T-850-ANN-CONUS_RRM_actual

Expected:
ERA-Interim-T-850-ANN-CONUS_RRM_expected

Diff:

For some reason, the bottom plot has large diffs too.
ERA-Interim-T-850-ANN-CONUS_RRM_diff

Environment

cdat_migration_fy24 branch

For some reason, the integration tests pass in CI/CD but not locally on perlmutter. For example, the build in PR #774 passes (link).

I think this is fixed now from recent merged commits.