Changes to improve vegetation health at high latitudes #2348
Conversation
There was a problem hiding this comment.
@olyson this is really nice to see become a PR. It'll be a nice update.
I have a couple things that we need to do here. You won't necessarily be the one that does them, but I marked them as something that needs to happen.
It also looks to me like this will also change answers for clm45 and clm50 physics, and I wasn't sure if that's acceptable or not? So there's some questions about that...
|
Just FYI on the latest I heard about the cesm2_3_beta17 tag. The next alpha tag is close to being done. And there are two tags after it. The first one has several CAM changes that need to come in that are likely a few weeks away. It would be really great to get this into cesm2_3_beta17, so hopefully that will give time for this to be figured out before then. So the beta tag is roughly a few weeks to a month away... |
|
@olyson the scientific tests you ran look good. Is it worth discussing this? |
|
Maybe we can discuss at our meeting tomorrow. I think it is probably as good as we can get it for now. A question that came up in the LMWG meeting was whether to do some kind of elevation threshold for the Sturm approach, but that would require more research, etc. And maybe that would be more important for high-resolution simulations. We'll also have to think about what initial file we'll want to use in conjunction with these changes in the next coupled model set of simulations. |
|
I agree, the alpine vs. tundra distinction is something that needs more scientific evaluation as we as conversations about SE support. I suggest we table this option for now. Moreover, for regional or point runs users can still switch between schemes using the namelist options. We should discuss the initial condition question, however. Is a longer spinup needed to provide better initial conditions? Maybe this would be a good idea anyway if the 5.2 PR is close to coming to main? We can discuss tomorrow. |
|
@slevis-lmwg cool! I moved it up in upcoming tags. @rgknox has the next tag for FATES history dimension control. And this should then be next. |
There was a problem hiding this comment.
I went through this again since it will be coming in right away. @slevis-lmwg already knows this, but I just pointed out about moving files to CESM inputdata.
There's some discussion about namelist settings and specifying the default for all clm physics versions. Technically if you do...
<namelist_thing>blah</namelist_thing>
<namelist_thing phys="clm5_1">bigger blah</namelist_thing>That's equivalent to
<namelist_thing phys="clm4_5">blah</namelist_thing>
<namelist_thing phys="clm5_0">blah</namelist_thing>
<namelist_thing phys="clm5_1">bigger blah</namelist_thing>BUT, the advantage of the later is that if you don't set ALL the clm physics versions -- it'll complain if that namelist setting is required. So it'll flag an error when we bring in a new physics version for the first time. I think it's also clearer.
Set initial t_soisno=272K for soils and 274K for urban road slevis resolved conflicts: bld/namelist_files/namelist_defaults_ctsm.xml cime_config/testdefs/testmods_dirs/clm/ciso_cwd_hr/user_nl_clm src/main/atm2lndType.F90
|
I will hold off running the test-suites until I update to dev172. |
|
@olyson could you look over my ChangeLog entry and let me know what you think? |
slevis resolved conflicts: doc/ChangeLog doc/ChangeSum
|
@ekluzek helped me resolve the ./manage_externals error with the following (I will post this in the CTSM_SE chat, too): Now I have submitted the test-suites. |
|
I was just looking at the test results on derecho for this which I assume is: /glade/derecho/scratch/slevis/tests_0314-145540de I noticed that there are differences with baseline for some Clm50 tests (all fates tests). Based on my original testing, I wouldn't expect this. My results showed no differences for all Clm50 tests. Do we know why these baseline comparisons fail? |
|
Sorry for overlooking that. I do not know why... |
|
Given the discussion I saw regarding fates externals, I wonder if it has something to do with that? |
|
I was about to say the same. |
|
I am open to suggestions on how to proceed @ekluzek |
|
Just noting that there are a couple of Clm45Fates DIFF failures as well. |
|
Here's the list of changed tests that I'm seeing (for non Clm51): ERP_Ld9.f45_f45_mg37.I2000Clm50FatesRs.derecho_intel.clm-FatesColdAllVars |
|
I'm looking through this more closely, but I think that the change to "dr" should propogate into FATES results as well, which should explain the diffs. |
@olyson didn't think so according to this conversation, but - just in case - I'm trying I changed line 1659 to what's in dev174 (same as Ryan's link in the last post). |
|
The last test returned: Going up the above list of tests:
BUT... I restored the code to dev173 and reran the very last test and got exactly the same result: Which tells me that the dev173 baseline is wrong? We will meet to discuss in 4 minutes. |
|
@rgknox will rerun the dev173 baselines. |
Changes to improve vegetation health at high latitudes Details in PR ESCOMP#2348 Conflicts: .git-blame-ignore-revs bld/unit_testers/build-namelist_test.pl cime_config/testdefs/testlist_clm.xml python/ctsm/test/test_sys_lilac_build_ctsm.py
Changes to improve vegetation health at high latitudes Details in PR ESCOMP#2348 Conflicts: bld/unit_testers/build-namelist_test.pl
|
I reran the following tests and confirmed that Clm45 and Clm50 are bit-for-bit against dev173 now: |
Changes to improve vegetation health at high latitudes Details in PR ESCOMP#2348 slevis resolved conflicts: bld/unit_testers/build-namelist_test.pl
and add Clm45 default for urban_explicit_ac Changes to improve vegetation health at high latitudes Details in PR #2348
Description of changes
Goal is to improve vegetation health/survivability at high latitudes (address LMWG_dev discussion #3)
Specific notes
These changes include:
Remove snicar_snobc_intmix from EXPERIMENTAL endrun (allow it to be true)
Remove flg_snoage_scl in SNICAR such that xdrdt can have an effect (fixes #2298 )
New parameter file and namelist values for clm5_1:
Contributors other than yourself, if any: @wwieder , @dlawrenncar , @slevis-lmwg , LMWG members
CTSM Issues Fixed (include github issue #): LMWG_dev discussion #3 and #2298
Are answers expected to change (and if so in what way)? Yes, new climate
Any User Interface Changes (namelist or namelist defaults changes)? Yes, new namelist defaults including new parameter files:
ctsm51_params.c240208.nc
clm50_params.c240208.nc
clm45_params.c240208.nc
ctsm51_ciso_cwd_hr_params.c240208.nc
Testing performed, if any:
Replicate simulation in LMWG_dev issue #51 as bfb when snow5d_thresh_for_onset on ctsm51_ciso_cwd_hr_params.c240208.nc was set to original value of 0.1.
aux_clm on Derecho:
No unexpected failures.
NLCOMP failures as expected.
clm51 compsets fail baseline comparison as expected.
clm50 and clm45 pass baseline comparisons as expected.
These EXPECTED FAILURE tests passed:
PASS SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Bgc.derecho_gnu.clm-default--clm-NEON-NIWO RUN time=62 (UNEXPECTED: expected FAIL)
PASS SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Bgc.derecho_gnu.clm-NEON-MOAB--clm-PRISM RUN time=79 (UNEXPECTED: expected FAIL)
PASS SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Fates.derecho_gnu.clm-FatesFireLightningPopDens--clm-NEON-FATES-NIWO RUN time=52 (UNEXPECTED: expected FAIL)
PASS SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Fates.derecho_gnu.clm-FatesPRISM--clm-NEON-FATES-YELL RUN time=35 (UNEXPECTED: expected FAIL)
PASS SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51SpRs.derecho_gnu.clm-default--clm-NEON-TOOL RUN time=33 (UNEXPECTED: expected FAIL)