Use stochastic restart patterns on rerun #3077
Conversation
There was a problem hiding this comment.
looks good to me. No comments.
Please invite review from @pjpegion @NeilBarton-NOAA
WalterKolczynski-NOAA
added
the
CI-Hera-Ready
emcbot
added
CI-Hera-Building
|
Experiment C96_atm3DVar FAILED on Hera in Build# 1 in |
|
Experiment C48mx500_3DVarAOWCDA FAILED on Hera in Build# 1 with error logs: Follow link here to view the contents of the above file(s): (link) |
|
Experiment C48mx500_3DVarAOWCDA FAILED on Hera in Build# 1 in |
|
Experiment C96C48_ufs_hybatmDA FAILED on Hera in Build# 1 with error logs: Follow link here to view the contents of the above file(s): (link) |
|
Experiment C96C48_hybatmDA FAILED on Hera in Build# 1 with error logs: Follow link here to view the contents of the above file(s): (link) |
|
Experiment C96C48_ufs_hybatmDA FAILED on Hera in Build# 1 in |
|
Experiment C96C48_hybatmaerosnowDA FAILED on Hera in Build# 1 with error logs: Follow link here to view the contents of the above file(s): (link) |
|
Experiment C96C48_hybatmDA FAILED on Hera in Build# 1 in |
|
Experiment C96C48_hybatmaerosnowDA FAILED on Hera in Build# 1 in |
|
Experiment C48_S2SWA_gefs FAILED on Hera in Build# 1 with error logs: Follow link here to view the contents of the above file(s): (link) |
|
Experiment C48_S2SWA_gefs FAILED on Hera in Build# 1 in |
emcbot
added
CI-Hera-Failed
|
CI Failed on Hera in Build# 1 |
The stochastic pattern restart files were not being copied into the input directory when restarting the model after a segment/failure. These files are now copied in. Additional changes appear to be needed for them to be used by UFS. The files are NOT copied in for non-rerun warm starts. Refs NOAA-EMC#2937
Adds the namelist flag so the stochastic pattern restart files get read when doing a RERUN. Refs NOAA-EMC#2937
The forecast job would attempt to copy the ca_data tiled restart files even when cellular automata was off. Now it only copies them if cellular automata is on.
WalterKolczynski-NOAA
force-pushed
the
feature/stoch_restart
branch
from
984eee5
to
3dfcb68
Compare
The stage and archive jobs were assuming that the ca_data restart files will always be present. However, these are only produced when cellular automata is run. So now they are conditionally handled based on `DO_CA`. To accomodate this, the `DO_CA` setting had to be moved out of `config.fcst`/`config.efcs` and into `config.base`.
Reverts cellular automata to be on for all situations to match current develop
WalterKolczynski-NOAA
added
CI-Hera-Ready
| @@ -448,6 +448,15 @@ export INCVARS_EFOLD="5" | |||
| export netcdf_diag=".true." | |||
| export binary_diag=".false." | |||
|
|
|||
| # Cellular automata | |||
| case ${RUN} in | |||
There was a problem hiding this comment.
I think this is leftover from mixed DO_CA settings, but is the case section necessary now?
WalterKolczynski-NOAA
removed
the
CI-Hera-Ready
Description
The stochastic pattern restart files were not being copied into the input directory when restarting the model after a segment/failure. These files are now copied in and the stochini flag set to
.true.in the namelist on a rerun.The files are NOT copied in for non-rerun warm starts.
Also removes the restriction that stochastic physics cannot be run on member 0, as this is desired down the line. Additional settings are added to the fcst and efcs configs to retain the current behavior.
A bug was also discovered and corrected during this work. The stage, forecast, and archive job all assumed that
ca_datatile files are always present, but these are only created when cellular automata is on. Nowca_datafiles are handled if CA is on. To implement this change, theDO_CAsetting had to be moved from the forecast configs to base so it is available to the stage_ic and archive jobs.Resolves #2937
Type of change
Change characteristics
How has this been tested?
Multi-segment test on Hercules
Checklist