v3.NARRM L80 runs failed due to temporary vertical grid override

[tangq]

Due to the temporary vertical grid override for RRM cases #6112, the ongoing v3.NARRM (F20TR.northamericax4v1pg2_r025_IcoswISC30E3r5) tests failed when trying to use the L80 atmospheric initial condition (IC) interpolated from v3.LR production runs.

I am not sure about the best solution for now. Suggestions are welcome that can let us test v3.NARRM.L80 without needing to update all the L72 atm ICs used in RRM tests to L80. (Would it be better updating the L72 ICs until we have final v3.NARRM.L80 runs?)

I am temporarily developing v3.NARRM with this patch.

diff --git a/components/eam/cime_config/config_component.xml b/components/eam/cime_config/config_component.xml
index 33eaacb36b..21fc342d03 100755
--- a/components/eam/cime_config/config_component.xml
+++ b/components/eam/cime_config/config_component.xml
@@ -59,7 +59,7 @@
       <value compset="_EAM.*_BGC%*"            >-co2_cycle</value>
 
       <!-- Temporarily override the default v3 vertical grid (L80) and use L72 to maintain RRM test coverage -->
-      <value grid="a%ne0">-nlev 72</value>
+      <!-- value grid="a%ne0">-nlev 72</value -->
 
       <!-- Single column model (SCM) -->
       <value compset="_EAM%SCM_"       >&eamv3_phys_defaults; &eamv3_chem_defaults; -scam</value>

[zhangshixuan1987]

@tangq: I think one solution is to interpolate all initial condition files for the atmosphere model from 72 levels to 80 levels and use them for the initialization of the v3 NARRM with 80 levels.

[whannah1]

@zhangshixuan1987 interpolation to L80 will allow these to run - but the resulting initial conditions require some adjustment to be viable for science. I've found the prognostic stratospheric ozone in particular needs a couple years to adjust to a change in the vertical grid.

[tangq]

This PR (#6698) tries to fix the issue. We might continue the discussions on the PR page.