GSWP3V1 atmospheric forcing for 2011-2014 is "bad" over Antarctica

[olyson]

Brief summary of bug

When GSWP3V1 data was extended for 2011-2014, bad atmospheric forcing data (e.g., too warm temperatures) over Antarctica was introduced.

General bug information

CTSM version you are using: N/A

Does this bug cause significantly incorrect results in the model's science? Yes (over Antarctica)

Configurations affected: All configurations forced with 2011-2014 GSWP3V1 data including anomaly forcing experiments.

Details of bug

This is a known problem, but an issue hasn't been filed for it before.
Present in /glade/p/cgd/tss/CTSM_datm_forcing_data/atm_forcing.datm7.GSWP3.0.5d.v1.c170516.
Here is the latest information I have from Dave L. about the data (Feb 2018) and an anticipated fix :
"I informed Hyungjun about Antarctica. He says he will fix it next update (which won't be soon and which I think also will include fix to humidity and will actually be GSWP3v2). "

Important details of your setup / configuration so we can reproduce the bug

/glade/work/oleson/release-clm5.0.30/cime/scripts/clm50_r-clm5.0.30_1deg_GSWP3V1_iso_SSP245_AnomalyF2

Important output or errors that show the problem

https://webext.cgd.ucar.edu/IRCP/clm50_r-clm5.0.30_1deg_GSWP3V1_iso_SSP245_AnomalyF2/lnd/clm50_r-clm5.0.30_1deg_GSWP3V1_iso_SSP245_AnomalyF2.2025_2034-clm50_r-clm5.0.30_1deg_GSWP3V1_iso_SSP3-7_AnomalyF2.2025_2034/set6/set6_landf_Antarctica.png

[ekluzek]

OK. Since this is an issue in forcing data, is technically inside of cime. So I'll file an issue there as well. Since it's most visible in CTSM I think it still makes sense to have an issue here as well. And actually even when it's fixed in cime it's not fixed in CTSM until CTSM brings a cime update in, so you need two issues anyway.

[mvertens]

This should get added to CDEPS as well - and since we are splitting the source code out of cime - it will have to also be added to the new CESM specific data model repository for mct/cpl7.

…

On Tue, Feb 2, 2021 at 11:21 AM Erik Kluzek ***@***.***> wrote: OK. Since this is an issue in forcing data, is technically inside of cime. So I'll file an issue there as well. Since it's most visible in CTSM I think it still makes sense to have an issue here as well. And actually even when it's fixed in cime it's not fixed in CTSM until CTSM brings a cime update in, so you need two issues anyway. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1269 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4XCE3LZGNZ6QXESZMJZ3LS5A7CNANCNFSM4W67FDVA> .

-- Mariana Vertenstein CESM Software Engineering Group Head National Center for Atmospheric Research Boulder, Colorado Office 303-497-1349 Email: mvertens@ucar.edu

[ekluzek]

As I understand this problem the only "fix" we could do until "v2" is out is to limit forcing to end in 2010. But, we likely don't want to do that. So I'm thinking we are just going to warn people about this problem. Is that what others think as well? @dlawrenncar @wwieder @danicalombardozzi?

[wwieder]

@dlawrenncar suggested that we should just warn users about this, but not stop the run in 2010. Data across Antartica are likely not worth trusting in offline simulations (as data for the reanalysis are basically absent). Instead, this region can be masked out in global diagnostics to avoid generating spurious results.