NOAA-EMC / global-workflow

Global Superstructure/Workflow supporting the Global Forecast System (GFS)
https://global-workflow.readthedocs.io/en/latest
GNU Lesser General Public License v3.0
70 stars 161 forks source link

Ocean_prod task not being triggered when REPLAY_ICS is set to YES #2725

Open EricSinsky-NOAA opened 3 days ago

EricSinsky-NOAA commented 3 days ago

What is wrong?

When REPLAY_ICSis set to YES, the ocean_prod tasks are not triggered.

What should have happened?

The ocean_prod tasks should run even when REPLAY_ICS is set to YES.

What machines are impacted?

All or N/A

Steps to reproduce

  1. Set REPLAY_ICS to YES in config.base.
  2. Make sure ocean is one of the components set to run in your experiment.
  3. Run global workflow using REPLAY ICs.

Additional information

This issue is caused by OFFSET_START_HOUR not being 0. This offset causes filename mismatches between the actual raw data being produced by MOM6 and the links produced by the global workflow. This mismatch results in ocean hist files not being saved in COMROOT, resulting in the ocean_prod task dependencies not being satisfied. MOM6_output_bug

Do you have a proposed solution?

A fix has been added in PR #2680. In this fix, a separate diag_table file has been created that is only used when OFFSET_START_HOUR is not 0. This diag_table allows the correct filenames to be produced by MOM6 (see image below). In this fix, when OFFSET_START_HOUR is 3 and FHOUT_OCN is 6, the first lead time is averaged from f003 to f009, the second lead time is averaged from f006 to f012, the third lead time is averaged from f012 to f018, and so on. image