Adding IcoswISC30E3r5 ice initial condition with DIB #6221
Conversation
|
Tested
|
|
@darincomeau, would you be willing to cherry-pick the commit from E3SM-Ocean-Discussion#77 onto this branch? I think it would be good to add a test that would have caught this issue so we make sure this configuration stays functional. |
|
@xylar I agree on adding the test, thanks for adding that to Ocean-Discussion, I missed that. I can add it here, but do you think it should be it's own PR? |
|
@darincomeau, I think we should add those tests here because a big part of the motivation in adding them is to show that the configurations work that this PR is meant to fix. I can split them into a separate PR if there are objections to including them here. |
There was a problem hiding this comment.
I ran both of the following tests successfully, whereas they failed without the changes on this branch:
SMS_D_Ld1.ne30pg2_r05_IcoswISC30E3r5.CRYO1850.chrysalis_intel
SMS_D_Ld1.ne30pg2_r05_IcoswISC30E3r5.CRYO1850-DISMF.chrysalis_intel
Ideally, I'd like to have them added to this branch but I won't let that hold up the merge.
|
Ok I understand. I've cherry-picked that test commit over here and updated the description above. @jonbob let us know if there's concern on the content here. |
darincomeau
requested review from
jonbob
and removed request for
proteanplanet
|
@rljacob -- this PR adds two new tests to e3sm_integration. Do you want to move one to e3sm_prod, since they really represent a cryo production configuration? I can run tests using the testing pelayout and document how long they take to run |
|
Sure move one of the cryo cases to production. |
|
Also, I'm not sure where is a good place to document this, but cases without DIB can use the sea ice IC with DIB, so for future meshes we can just make one IC with DIB to avoid the extra file. |
Adding IcoswISC30E3r5 ice initial condition with DIB
This adds an initial condition for MPAS-Seaice for cases where the
IcoswISC30E3r5 mesh is paired with a CRYO* compset, which utilize data
iceberg (DIB) forcing. The new initial condition is from a restart from
a one month GMPAS-JRA1p5-DIB-PISMF case.
The current sea ice IC for this mesh in B-cases ('spun-up' ICs) was not
run with DIB forcing, and as such MPAS-Seaice cannot use it as a restart
for a case with DIB.
This also adds CRYO1850 and CRYO1850-DISMF tests with this mesh to the
e3sm_integration test suite, which prior to this would not pass.
[BFB] Only impacts configurations with IcoswISC30E3r5 and a CRYO*
compset, which prior to this would seg fault in ice during
initialization
|
Passes:
Also verified that both new tests:
merged to next |
|
merged to master and new baselines requested for compy_prod, chrysalis_prod, and chrysalis_integration. Will need to do the same for pm-cpu_integration when it reports next, as well as anvil_prod which is failing all tests with the new Icos mesh. I'm currently testing to see if those fails are related to the size of the test pelayout on anvil |
This adds an initial condition for MPAS-Seaice for cases where the
IcoswISC30E3r5mesh is paired with aCRYO*compset, which utilize data iceberg (DIB) forcing. The new initial condition is from a restart from a one monthGMPAS-JRA1p5-DIB-PISMFcase.The current sea ice IC for this mesh in B-cases ('spun-up' ICs) was not run with DIB forcing, and as such MPAS-Seaice cannot use it as a restart for a case with DIB.
This also adds
CRYO1850andCRYO1850-DISMFtests with this mesh to the e3sm_integration test suite, which prior to this would not pass.[BFB] Only impacts configurations with
IcoswISC30E3r5and aCRYO*compset, which prior to this would seg fault in ice during initialization.Credit to @xylar for rooting out the problem.