Chunky downscaled SMB over Antarctica in NorESM test case

[mpetrini-norce]

We have run the first NorESM/CISM test cases with multiple Greenland and Antarctic ice sheets. The SMB over Antarctica looks a bit strange to me. After an initial discussion with Heiko we thought this might be to the coarse precipitation pattern from CLM (which is not downscaled with ECs like melt), however after looking at the same SMB field in a coupled test ran by Kate I am wondering if there could be something strange happening in the remapping.

This is how the SMB looks like in Kate's B-run (here on Derecho: /glade/derecho/scratch/katec/archive/b.e23_alpha17d.BLT1850MOM-AISGRIS.ne30_t232_ais8.test4/cpl/hist)

This is how the same field looks like in our I-runs (DATM, LND, CISM) on Betzy.

What is making me think there might be some issue with remapping/downscaling at play is that in our case (1) the SMB field is not smooth at all, (2) there is no imprint of high res CISM topography in the Transantarctic Mountains or near the coast, except for the tip of the Antarctic peninsula, (3) there are some slightly elongated features that I’ve seen before in some cases of bad remapping (not with CESM/CISM); these are even more visible in the screenshot below using different min/max values.

It is also possible that our SMB pattern is related to our specific compset and there’s nothing wrong with the remapping/downscaling (I really hope so). This is the compset that we have used for this test run:

COMPSET=1850_DATM%GSWP3v1_CLM50%SP_SICE_SOCN_MOSART_CISM2%AIS-EVOLVE%GRIS-EVOLVE_SWAV
RESOLUTION=f09_tn14_ais8gris4

This is the CLM and CISM version that we are using:

[clm]
protocol = git
tag = ctsm5.1.dev167-noresm_1
repo_url = https://github.com/NorESMhub/CTSM
local_path = components/clm
externals = Externals_CLM.cfg
required = True

[cism]
protocol = git
branch = feature/update_to_cismwrap_2_1_99
repo_url = https://github.com/mvertens/CISM-wrapper
local_path = components/cism
externals = Externals_CISM.cfg
required = True

Tagging those who could potentially be interested, feel free to include others! @Katetc @mvertens @hgoelzer @whlipscomb @gunterl

[mpetrini-norce]

Discussion resolved: the chunky pattern was caused by specific atmospheric dataset used in this compset.