Add constituent tendency capability by peverwhee · Pull Request #584 · NCAR/ccpp-framework

peverwhee · 2024-08-22T22:40:48Z

Summary
Adds constituent tendency array and enables the use of individual constituent tendency variables.

Description
In order to facilitate time-splitting, this PR adds/enables the following:

ccpp_constituent_tendencies: standard name for constituent tendencies array
standard names of the form "tendency_of_CONSTITUENT" now handled by the framework
- must also include "constituent = True" in metadata for the tendency

User interface changes?: Yes, but optional
User can now use the following in scheme metadata:

ccpp_constituent_tendencies (standard name for the array of constituent tendencies)
tendency_of_CONSTITUENT (tendency array for a specific constituent)
constituent = True (specifies variable will be handled by constituents object)

Testing:
test removed: None
unit tests: Added doctests to new check_tendency_variables function in host_cap.F90
system tests: Modified advection_test to use tendency variable and tendency array
manual testing: Updates run in CAM-SIMA

… constituent-tend

mwaxmonsky

Looks great so far @peverwhee! Still working through the details but just had a few preliminary questions/comments on the python side of things.

climbfuji

Thanks for this @peverwhee ! One or two most likely dumb questions, other than that looks fine with @mwaxmonsky's comments applied.

climbfuji · 2024-08-29T14:40:52Z

+                    else:
+                        tend_dim = dim
+                    # end if
+                    if const_dimensions[index] == 'horizontal_loop_extent':


This adjustment for tendency/constituent vars is just for matching horizontal dimensions, correct?

yep! not a permanent change, just a way to make sure that the dimensions match between the tendency variable and the corresponding constituent "state" variable

mkavulich

Some initial comments while I try to wrap my head around things.

Side note: the word "tendency" doesn't look real anymore.

…tdname

climbfuji · 2024-09-16T13:06:19Z

+            const_kind = const_var.get_prop_value('kind')
+            if tend_stdname_split == const_stdname:
+                # We found a match! Check units
+                if tend_units.split('s-1')[0].strip() != const_units:


This assumes that s-1 is always at the end. But it is entirely valid (from a math standpoint) to write kg s-1 m-2 instead of kg m-2 s-1. And it's of course also valid to write m s-2 for a tendency of the constituent had m s-1 ...

Probably not an issue, but we should document this in the updated (!) CCPP technical documentation

dustinswales

@peverwhee I think this is fine.
I do have a question about extending some existing code rather than adding your own routine (See inline comments).

dustinswales · 2024-10-03T15:03:04Z

        # end for
    # end for
+    # Check that all tendency variables are valid
+    errmsg, errflg = check_tendency_variables(tend_vars, const_vars)


Most of this code/logic in check_tendency_variables already exists in the the VarCompatObj (and here).
Could the ability to check for "compatibility between any variable, not just a const variable, and its tendency" be folded into there? (This would make it more useable if someone wanted to do something similar with "state_variables" in the _pre/_post scheme calls.)
I'd be happy to help.

I took a stab at modifying compatible to validate tendencies. Let me know what you think!

… constituent-tend

climbfuji · 2024-10-16T13:45:22Z

@peverwhee Please re-request a review from all the reviewers once you've addressed the questions/concerns from @mwaxmonsky and @dustinswales (I keep coming back to re-review but then remind myself that there are still comments that need to be addressed). Thank you!

climbfuji

Thanks for addressing my comments and those of @mwaxmonsky . Looks good to me now!

dustinswales

@peverwhee These changes look awesome!
I'm glad to see that the compatibility object was able to handle the logic w/o many modifications.

dustinswales · 2024-10-23T15:24:02Z

-        if var1_stdname != var2_stdname:
+        # If it's a tendency variable, it's assumed the standard name is of the
+        #  form "tendency_of_var2_stdname"
+        if not is_tend and  var1_stdname != var2_stdname:


Should we add a check that starts with tendency_ when is_tend=True? Or is this handled somewhere else? If redundant, ignore.

good call! as it's currently used, it's only called when a variable starts with "tendency_of" but I added the check to prevent future (by me) implementation issues!

climbfuji · 2024-10-31T13:00:36Z

@peverwhee Can you please update the branch from develop? There are merge conflicts that need to be resolved. Thanks!

peverwhee · 2024-10-31T19:46:07Z

@DomHeinzeller done!

I think I'll give @gold2718 a bit more time if he'd still like to review this before merging. Say, by next wednesday (11/6)?

Tag name (required for main): Originator(s): peverwhee Summary (include the keyword ['closes', 'fixes', 'resolves'] and issue number): Brings in and enables new apply_constituent_tendencies scheme (from ESCOMP/atmospheric_physics#111) Describe any changes made to build system: Exclude constituent tendencies from required variables (now handled by the CCPP Framework in NCAR/ccpp-framework#584) Describe any changes made to the namelist: None List any changes to the defaults for the input datasets (e.g. boundary datasets): None List all files eliminated and why: None List all files added and what they do: None List all existing files that have been modified, and describe the changes: (Helpful git command: `git diff --name-status development...<your_branch_name>`) M .gitmodules - update CCPP-Framework and atmospheric_physics externals to point to updates for constituent tendencies M src/data/write_init_files.py - exclude constituent tendency variables (and array) from required variables list If there are new failures (compare to the existing-test-failures.txt file), have them OK'd by the gatekeeper, note them here, and add them to the file. If there are baseline differences, include the test and the reason for the diff. What is the nature of the change? Roundoff? derecho/intel/aux_sima: derecho/gnu/aux_sima: CAM-SIMA date used for the baseline comparison tests if different than latest: --------- Co-authored-by: Courtney Peverley <courtneyp@izumi.cgd.ucar.edu>

Courtney Peverley and others added 8 commits August 12, 2024 14:01

tendencies working

1ee7a16

tendencies working

922a4dd

better implementation; tendencies working

44179da

merge

a6f8dd3

update const tend stdname

0dbc881

better testing, add constituent metadata attribute

1387cf2

code cleanup

dc3f3f3

remove orphaned end if

fb08863

peverwhee added enhancement capgen bugs, requests, etc. that involve ccpp_capgen labels Aug 22, 2024

peverwhee requested review from climbfuji, dustinswales, gold2718, grantfirl, mkavulich and mwaxmonsky as code owners August 22, 2024 22:40

Merge branch 'develop' of https://github.com/NCAR/ccpp-framework into…

a7eb174

… constituent-tend

mwaxmonsky reviewed Aug 23, 2024

View reviewed changes

climbfuji reviewed Aug 29, 2024

View reviewed changes

mkavulich reviewed Aug 29, 2024

View reviewed changes

Comment thread scripts/host_cap.py Outdated

Comment thread scripts/host_cap.py Outdated

peverwhee added 2 commits September 13, 2024 23:32

review comments; handle if "tendency_of" is not at the start of the s…

1f7387d

…tdname

cleanup in check_tendency_variables

d49fcfa

peverwhee requested review from climbfuji, mkavulich and mwaxmonsky September 14, 2024 05:49

climbfuji reviewed Sep 16, 2024

View reviewed changes

dustinswales reviewed Oct 3, 2024

View reviewed changes

peverwhee added 2 commits October 7, 2024 11:12

Merge branch 'develop' of https://github.com/NCAR/ccpp-framework into…

07e6a1b

… constituent-tend

amend compatibility functionality/testing to validate tendencies

17db5c0

peverwhee requested review from climbfuji and dustinswales October 16, 2024 15:48

climbfuji approved these changes Oct 18, 2024

View reviewed changes

dustinswales approved these changes Oct 23, 2024

View reviewed changes

mwaxmonsky approved these changes Oct 25, 2024

View reviewed changes

peverwhee added 2 commits October 31, 2024 09:28

add check for "tendency_of" in VarCompatObj

69186d1

merge to head of develop

cd2452f

fix merge

e070589

mkavulich merged commit ce381cf into NCAR:develop Nov 7, 2024

peverwhee mentioned this pull request Nov 27, 2024

Bring in constituent tendency updater ESCOMP/CAM-SIMA#292

Merged

peverwhee deleted the constituent-tend branch November 18, 2025 23:43

Conversation

peverwhee commented Aug 22, 2024

Uh oh!

mwaxmonsky left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

climbfuji left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

climbfuji Aug 29, 2024

Choose a reason for hiding this comment

Uh oh!

peverwhee Sep 14, 2024

Choose a reason for hiding this comment

Uh oh!

mkavulich left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

climbfuji Sep 16, 2024

Choose a reason for hiding this comment

Uh oh!

dustinswales left a comment

Choose a reason for hiding this comment

Uh oh!

dustinswales Oct 3, 2024

Choose a reason for hiding this comment

Uh oh!

peverwhee Oct 16, 2024

Choose a reason for hiding this comment

Uh oh!

climbfuji commented Oct 16, 2024

Uh oh!

climbfuji left a comment

Choose a reason for hiding this comment

Uh oh!

dustinswales left a comment

Choose a reason for hiding this comment

Uh oh!

dustinswales Oct 23, 2024

Choose a reason for hiding this comment

Uh oh!

peverwhee Oct 31, 2024

Choose a reason for hiding this comment

Uh oh!

climbfuji commented Oct 31, 2024

Uh oh!

peverwhee commented Oct 31, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants