Skip to content

Conversation

@wlin7
Copy link
Contributor

@wlin7 wlin7 commented Nov 6, 2025

Previously it was set to true for all configurations that use ne120 grid for atm.
This would affect EAM and EAMxx F-cases that use ne120. This change limits its
use to cases that use v3HR grid (ne120pg2 for atm and RRSwISC6to18E3r5 for ocean/ice).
It applies to both coupled and uncoupled configurations.

[BFB] No impact except for cases using ne120 atm grid with other ocean masks


Verified no impact for v3.HR.piControl-CMIP7 spin-up that uses ne120pg2_r025_RRSwISCE3r5,
previously also tested that nlmaps_atm2srf_conserve=.false. for F1850-CMIP7 that uses the same grid
when the use was restricted to coupled cases. This requirement is now dropped (removing the compset attribute)

@wlin7 wlin7 requested a review from rljacob November 6, 2025 20:16
@wlin7 wlin7 added the BFB PR leaves answers BFB label Nov 6, 2025
@rljacob rljacob added Coupler Related to code in driver-mct or driver-moab or component connections to the coupler. Coupled Model labels Nov 6, 2025
@rljacob
Copy link
Member

rljacob commented Nov 6, 2025

I wonder if we do want this on for F-cases that use the same grid? EAMxx F-cases are using ne120pg2_ne120pg2 so would not pick this up just from the new grid spec.

@mt5555
Copy link
Contributor

mt5555 commented Nov 6, 2025

I second Rob's suggestion - it should be based on the grid, but not compset. If using the a%ne120np4.+oi%RRSwISC6to18E3r5 grid in an F case, I think one would still benefit from nonlinear maps.

@ndkeen
Copy link
Contributor

ndkeen commented Nov 6, 2025

Note we had several fails (with ERROR: seq_nlmap_init_a2l_cons ERROR: Nonlinear map not set) reported on cdash that used ne120pg2_ne120pg2.F2010-SCREAMv1 -- this PR allows at least one of those tests to pass and remains BFB.

@ndkeen ndkeen self-requested a review November 6, 2025 21:27
@ndkeen ndkeen added the EAMxx C++ based E3SM atmosphere model (aka SCREAM) label Nov 6, 2025
@wlin7
Copy link
Contributor Author

wlin7 commented Nov 6, 2025

@rljacob @mt5555 . The grid in theory could be used, for example, to spin-up land that uses the same ocean mask.

@mt5555
Copy link
Contributor

mt5555 commented Nov 6, 2025

agreed. I'm suggesting that instead of:

<value grid="a%ne120np4.+oi%RRSwISC6to18E3r5" compset="_EAM.+MPASO">.true.</value>

we use:

<value grid="a%ne120np4.+oi%RRSwISC6to18E3r5" >.true.</value>

@wlin7
Copy link
Contributor Author

wlin7 commented Nov 6, 2025

agreed. I'm suggesting that instead of:

<value grid="a%ne120np4.+oi%RRSwISC6to18E3r5" compset="_EAM.+MPASO">.true.</value>

we use:

<value grid="a%ne120np4.+oi%RRSwISC6to18E3r5" >.true.</value>

I think this is good. The problem with the EAMxx F2010 test is mainly due to not having such a map, not that it couldn't use one. Changes submitted and the PR message to be updated.

@wlin7 wlin7 changed the title Restrict nlmaps_atm2srf_conserve=.true. to coupled v3HR Restrict nlmaps_atm2srf_conserve=.true. to v3HR grid Nov 6, 2025
wlin7 added a commit that referenced this pull request Nov 6, 2025
Restrict nlmaps_atm2srf_conserve=.true. to v3HR grid

Previously it was set to true for all configurations that use ne120 grid for atm.
This would affect EAM and EAMxx F-cases that use ne120. This change limits its
use to cases that use v3HR grid (ne120pg2 for atm and RRSwISC6to18E3r5 for ocean/ice).
It applies to both coupled and uncoupled configurations.

[BFB] No impact except for cases using ne120 atm grid with other ocean masks
@wlin7
Copy link
Contributor Author

wlin7 commented Nov 6, 2025

Merged to next.

@wlin7 wlin7 merged commit 951485c into master Nov 7, 2025
6 checks passed
@wlin7 wlin7 deleted the wlin/reset_ne120_nlmaps_atm2srf branch November 7, 2025 16:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

BFB PR leaves answers BFB Coupled Model Coupler Related to code in driver-mct or driver-moab or component connections to the coupler. EAMxx C++ based E3SM atmosphere model (aka SCREAM)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants