Repository navigation
alpha-ctsm5.4.CMIP7.19.ctsm5.3.082: Fix LII2 COMPARE_base_no_interp failures - #3613
Conversation
|
derecho and izumi So far I have only looked at aux_clm on derecho:
Get these to point to the other f09 file OR (easier?) rerun LII2 as a Bgc test with c13/c14. |
|
@olyson turns out that you were correct that we still have a problem with the two fields VEGWP[LN,PD], though I only confirmed that when I reran a test against the newly generated baseline (see 4 above) and still had DIFFs in these fields. This means that these fields do need to go on restart files. This is my plan:
There is another set of failures (see 5 above) that I need to address. I have a clearer plan for that, so I may start with that. (After I finish "export compliance" training...) |
This comment was marked as outdated.
This comment was marked as outdated.
|
Thanks @slevis-lmwg . As we discussed, it isn't clear why none of the local noon solar variables have to be on the restart file. In looking at the code further though, it might be because they are set to spval at every non-local noon time step in SurfaceRadiationMod.F90, whereas vegwp_ln seems to be only set to spval at every non-local noon time step if it is daytime. Maybe it could be set at the end of PhotosynthesisHydraulicStress so that it is set at every time step. As you suggest, maybe the two fields could be removed for now and possible solutions could be experimented with in a new issue. |
|
Oh yeah, I still have problem (5) from the list above, so I need to deal with that: Rebuilt and reran the LII2 test, but with use_c13 and use_c14 = .true., so as to generate a finidat that I hope will work with the failing tests: On derecho and izumi ctsm_sci waiting for two RX* tests to finish, else OK |
Description of changes
LII2 tests require some steps in order to work when we update fsurdat files and/or when we update finidat files (details in this post). This PR handles those steps.
Unlike in previous updates (e.g. ctsm5.2 and ctsm5.3), when I updated for ctsm5.4, COMPARE_base_no_interp continued to fail in the LII2 tests due to two diagnostic fields appearing unnecessarily in restart files (details in #3252). This PR removes these two fields from restarts.
UPDATE
Found we needed the two fields in restarts, so I removed from history and filed an issue.
Specific notes
CTSM Issues Fixed (include github issue #):
Fixes #3252
Are answers expected to change (and if so in what way)?
Yes, because the LII2 tests point to new finidat files.
Any User Interface Changes (namelist or namelist defaults changes)?
Yes, to enable LII2 tests to pick up the correct finidat files.
Testing performed, if any:
The following now passes:
./create_test LII2FINIDATAREAS_D_P256x2_Ld1.f09_t232.I1850Clm60BgcCrop.derecho_intel.clm-default -c /glade/campaign/cgd/tss/ctsm_baselines/alpha-ctsm5.4.CMIP7.17.ctsm5.3.082The other LII2 test looks the same with the suffix
--clm-matrixcnOn_ignore_warningsadded to it, so I expect it to pass.