Use of ICs for SFS Beta Version#3839
Conversation
NeilBarton-NOAA
requested review from
DavidHuber-NOAA,
KateFriedman-NOAA,
TerrenceMcGuinness-NOAA and
aerorahul
as code owners
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| @@ -0,0 +1,19 @@ | |||
| experiment: | |||
There was a problem hiding this comment.
Neil, These new case files look strong (along with their overwriting defaults). Do you have these sfs cases tested in automated scripts yet?
There was a problem hiding this comment.
@TerrenceMcGuinness-NOAA we've been using these edits for testing SFS and will be automated for the reforecasting effort.
There was a problem hiding this comment.
What framework will you be using for automation? global-workflow is currently developing framework in VLab's GitLab that uses these case files.
There was a problem hiding this comment.
@TerrenceMcGuinness-NOAA I'm not sure if this answers your question, but to run the full set of SFS runs, which is currently November 1st and May 1st start dates, we add
${YEAR}${MONTH}010000 ${YEAR}${MONTH}010000 24:00:00
to the rocoto .xml file before the rocoto run as rocoto does not support (to my knowledge) starting once a month on the first.
| "ocean_model", "LwLatSens", "LwLatSens", "@[MOM6_OUTPUT_DIR]/ocn%4yr%2mo%2dy%2hr","all",.true.,"none",2 | ||
| "ocean_model", "Heat_PmE", "Heat_PmE", "@[MOM6_OUTPUT_DIR]/ocn%4yr%2mo%2dy%2hr","all",.true.,"none",2 | ||
|
|
||
| # scalar fields |
There was a problem hiding this comment.
@jiandewang - Can you look at this from a GFS perspective --- I don't know if we want to add these fields or not for GFS at this stage of development....
There was a problem hiding this comment.
- Adding this output is a trivial (almost nothing in CPU costs and disk usage) addition.
- For any simulation that is < 1 mon, this output isn't going to help with anything - particularly if you have never used a
metricthat factors in global means and their time evolution.
Hope this helps make your choices!
There was a problem hiding this comment.
GFS simulations are less than <1 month, so why are we adding any cost/disk for that case?
There was a problem hiding this comment.
Compute and disk space are things that we would want to not add to for GFS - even if it's small if it's not going to be needed.
I think we need to find a way to toggle this on/off.
There was a problem hiding this comment.
PmE can be always derived from precipitation and latent heat fluxes saved in the atmos output.
|
We can remove heat_PME. |
|
@sanAkel I'm tagging Santha as these additions are his suggestion for evaluation |
Co-authored-by: Rahul Mahajan <aerorahul@users.noreply.github.com>
Co-authored-by: Rahul Mahajan <aerorahul@users.noreply.github.com>
NeilBarton-NOAA
requested review from
DeniseWorthen and
junwang-noaa
as code owners
|
@NeilBarton-NOAA Is this test using the Replay ICs? I had been working on the SFS RT w/ the assumption that the ICs would be from CORe. I only ask because if the Replay ICs are used, the missing value flag in the sfc ICs needs to be changed from The reason it is an issue is because we're trying ensure consistency between the RTs and the G-W. I can't pass the debug test in the RTs w/ the Replay ICs unless the sfc ICs are fixed. |
|
@DeniseWorthen These changes have updates to run with replay and CPC ICs. Yes, we need to fix the ICs that we are using for these tests outside of g-w. I'll ask @Yangxing.Zheng when he is back from A/L |
There was a problem hiding this comment.
ShellCheck found more than 20 potential problems in the proposed changes. Check the Files changed tab for more details.
| # Do Monthly Average ocean output and move to output ocean history directory (valid up to 12 month forecasts only) | ||
| if [[ "${RUN}" == sfs ]]; then | ||
| local last_fh_output files f_name | ||
| last_fh_output=$(ls -f "${COMOUT_OCEAN_HISTORY}/${RUN}.ocean.t${cyc}z.${FHOUT_OCN}hr_avg.f${FHMAX_GFS}.nc" ) |
Check notice
Code scanning / shellcheck
Possible misspelling: FHMAX_GFS may not be assigned. Did you mean FHMAX_HF? Note
|
I'm closing this PR as it is now being merged into dev/sfs |
9 participants
Description
These updates allow g-w to run with ICs used for testing in the SFS forecast-only framework. It supplies sample yaml files and needed updated. In addition, it add the archiving tasks for SFS.
Type of change
Change characteristics
How has this been tested?
Testing performed on GAEAC6 in forecast-only model
Checklist