Various wave updates: Reformat WW3 inputs; enable PIO for wave restarts; enable WW3 cycling; synchronize UFS restart writes; add uglo_15km WW3 grid #3190
Conversation
Conflicts: .gitmodules
…reading if desired
…combineforww3testing
|
@JessicaMeixner-NOAA Thank you for brining in these new cycling capabilities for WW3! In addition to my suggestions, could you also update the description to include the relevant UFS PR? Also, could you review the Wave section of the documentation and see if any new updates need to be added? |
DavidHuber-NOAA
changed the title
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
Co-authored-by: David Huber <[email protected]>
JessicaMeixner-NOAA
changed the title
| @@ -25,6 +25,7 @@ local SHOUR=${model_start_date:8:2} | |||
| local FHROT=${IAU_FHROT:-0} | |||
| local DT_ATMOS=${DELTIM} | |||
| local RESTART_INTERVAL="${FV3_RESTART_FH[*]}" | |||
| local RESTART_FH="${CMEPS_RESTART_FH:-" "}" | |||
There was a problem hiding this comment.
I am not sure introducing a CMEPS variable dependency in FV3 would be clean, since that means there is a dependency on a coupled model variable for an atm-only configuration. Need to think a bit on this.
There was a problem hiding this comment.
The RESART_FH variable is not used by FV3, however it is in the model_configure file.
There was a problem hiding this comment.
Also - please note this is what is being done in this PR: ufs-community/ufs-weather-model#2419 so if you want changes we'll potentially need changes to ufs-weather-model
There was a problem hiding this comment.
If the FV3 does not use this variable, can we just set this as:
local RESTART_FH=" "
There was a problem hiding this comment.
This assumes the restarts are written according to local RESTART_INTERVAL="${FV3_RESTART_FH[*]}", correct?
There was a problem hiding this comment.
The RESTART_FH variable is needed for the non-FV3 components to write restarts at the FV3 restart times when we have IAU, so no we cannot have it as RESTART_FH=" " to my understanding.
There was a problem hiding this comment.
I misunderstood this comment that said the FV3 does not use this variable. Does the model_configure also now use this variable to specify restart write times for all model components? (In the past, model_configure was fv3atm specific)
How does this relate to RESTART_INTERVAL that is currently specifying the FV3 restart times? Does RESTART_FH override RESTART_INTERVAL?
There was a problem hiding this comment.
From @NickSzapiro-NOAA in an email:
3 restart triggers:
restart_n+restart_option: frequency-based ESMF restart trigger (via ufs.configure) *Not currently in FV3
restart_interval: just in FV3 (via model_configure) and can mean different things based on "encoding"
restart_fh: UFS wraps ESMF trigger (via model_configure). *Not currently in FV3
There was a problem hiding this comment.
@aerorahul - let me know if you'd like me to run a test case and save the output showing that we're writing all restarts at the same time when using IAU for all components. I can also run an fv3 standalone ci test to ensure that we are not negatively impacting atm-only runs too.
There was a problem hiding this comment.
@aerorahul - is there any change you'd like to see with this? Or any specific tests you'd like to see output with?
ush/forecast_predet.sh
Outdated
| @@ -134,6 +134,29 @@ common_predet(){ | |||
| # Several model components share DATA/INPUT for input data | |||
| if [[ ! -d "${DATA}/INPUT" ]]; then mkdir -p "${DATA}/INPUT"; fi | |||
|
|
|||
| # For CMEPS, CICE, MOM6 and WW3 determine restart writes | |||
| # Note FV3 has its own restart intervals | |||
| cmeps_restart_interval=${restart_interval:-${FHMAX}} | |||
There was a problem hiding this comment.
This isn't a common variable if it starts w/ cmeps.
There was a problem hiding this comment.
What would you prefer it be called? This is for CMEPS, WW3, ICE, and MOM6
There was a problem hiding this comment.
It is not the name, but the place these are introduced.
The function CMEPS_predet IMO, would be appropriate to introduce these variables related to cmeps_ (and more general the coupled model since they share these concepts)
There was a problem hiding this comment.
Thanks for the additional explanation. I will move it to CMEPS!
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| local restart_interval_start=${cmeps_restart_interval} | ||
| local restart_interval_end=${FHMAX} | ||
| fi | ||
| CMEPS_RESTART_FH="$(seq -s ' ' "${restart_interval_start}" "${cmeps_restart_interval}" "${restart_interval_end}")" |
Check warning
Code scanning / shellcheck
CMEPS_RESTART_FH appears unused. Verify use (or export if used externally). Warning
There was a problem hiding this comment.
Any recommendations on how to fix this @DavidHuber-NOAA or @aerorahul ?
Description
This PR adds the following:
Note this PR requires the following:
Co-author: @sbanihash
Related Issues:
Type of change
Change characteristics
How has this been tested?
Tests were done by adding waves to C48mx500 3DVar tests for cycling.
High resolution tests cases were also added.
This likely needs a new round of testing before CI and after staged ICs & new fix files are added.
Checklist