Snow hole in N Greenland snow from CRU-JRA I cases

[wwieder]

@gunterl noticed this blue wedge of 'missing snow' in N. Greenland in his land ice initializations.

I think you can see the same feature pretty clearly without remapping this from an I case at the f09 resolution below), but you also see it on the ne30 grid.

Temperatures over this region seem cold (not shown), but likely reflect low precipitation in the region (here an annual mean of monthly climatologies created with the LDF and ncra)

The obvious fix here would be to increase precipitation in this region. Not sure how best to do this and retain the integrity of the CRU-JRA dataset? Another option could be to fill these data with data from a GSWP3 - forced run, which is how we're filling in data Antartica for CRU-JRA. Either option seems like a slightly laborious task.

@gunterl, what's the end here here? What applications are you wanting to meet with this and what's the timeline this may need to be fixed on? Please bring others into this discussion too.

[wwieder]

One more note, you can see this clearly looking at SNOW (here a flux from the atmosphere) from a decade of a GSWP3 forced run (left) vs. the CRU-JRA run (right)

[ekluzek]

@wwieder I assume this must be a problem with the CRUJRA data itself, is this something that they would fix if it was pointed out? They also update their data yearly right? Is there another version available now or soon enough that it could be used?

[wwieder]

Great minds! I just looked at the data @adrifoster used for TRENDY2025 (new data this year, see /glade/campaign/cgd/tss/projects/TRENDY2025/history_outputs/S3/month_1), but it seems like it shares the same features over this region with a snow free widow's peak over the N of Greenland. 🫤

[wwieder]

The other thing I'll note is that CRUJRA is provided for the TRENDY project, which has a terrestrial carbon cycle focus. Motivation to improve precipitation at the North of Greenland may not be high. Remember, this is the same dataset that masks out Antarctica completely! Regardless, I'm not sure @gunterl wants to wait until next summer to initialize his ice sheet model.

[gunterl]

Looking at CRUJRA product, it looks like areas in mountainous areas and ice sheets are having some problems (e.g. High Mountain Asia and Greenland). Would it be possible to mix GSWP3 and CRU-JRA? (I assume if you used CRU-JRA, you liked it better for what you were doing?). Could we use GSWP3 for glacier regions (ice sheets and glaciers) and CRU-JRA otherwise? And @wwieder you are right, I would like to have a good Greenland product for CESM3. Unless we plan on postponing CESM3 release for another year :)

…

On Wed, Jan 7, 2026 at 2:38 PM will wieder ***@***.***> wrote: *wwieder* left a comment (ESCOMP/CTSM#3680) <#3680 (comment)> The other thing I'll note is that CRUJRA is provided for the TRENDY project, which has a terrestrial carbon cycle focus. Motivation to improve precipitation at the North of Greenland may not be high. Remember, this is the same dataset that masks out Antarctica completely! Regardless, I'm not sure @gunterl <https://github.com/gunterl> wants to wait until next summer to initialize his ice sheet model. — Reply to this email directly, view it on GitHub <#3680 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB6QV6G2SJBAZIGEV7CMHOL4FV4ENAVCNFSM6AAAAACQ7ULAD6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMRQHEYTOOJSG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Gunter Leguy, Ph.D Scientist NSF National Center for Atmospheric Research cell: (575) 418 1021 desk: (303) 497 1790

[wwieder]

@gunterl can you clarify what you mean by "I would like to have a good Greenland product for CESM3"?

In B cases we'll eventually be spinning up the land with model atmosphere data from CPL_HIST output. Since you say that Cecile's cases eventually fill in this hole, it seems like we'll have land initial conditions from CPL_HIST that would generate reasonable ice sheets in Greenland that are consistent with CESM3 climate. Is this adequate for your purposes?

Alternatively (or additionally) in I cases we'd need discuss potential modifications to DATM products if you want to initialize the Greenland ice sheet with initial conditions from an I case?

• Is this something you all regularly do?
• If so, can we just initialize it with a GSWP3 forced run, which seems to have more reasonable precipitation over Greenland? or
• Do we need to modify the CRU-JRA DATM precipitation inputs regionally for some reason?

A clearer picture of what initial conditions the land ice group needs would be helpful.

[gunterl]

Here are some caveats on not addressing this issue right now: 1. There is a possibility for us to forget about it and this error will live on. And if anyone at some would like to use this input file for an IG run, it will be bogus. A similar problem actually occurred with F compsets when Adam was trying to diagnose excessive melt over Greenland. His data atmosphere ended up being the problem and we did not know until now unless it had been forgotten. 2. It takes about 80 years to remove the memory of the pattern of the snow depth described in the CRU-JRA dataset in a B run (I am using run271 output for this diagnosis). In order to generate a climatology for the ice sheet and create an initial condition for CESM3 and a BG run, we typically use between 30 and 40 years. This means that a control run will need to run for at least 120 years before we can start and diagnose a spin-up before giving CESM3 an initial Greenland condition for the fully coupled run. If that is satisfactory, then fine. "Filling the whole" for snow is important. Until CLM reaches a snow depth of 10m, it does not compute a Surface Mass Balance for CISM. The 2 possibilities here would be to either: a. Be comfortable with option 2 above. b. Initialize land ice cells in CLM with a GSWP3 forced run. I am unclear on the details of this initialization process and how combursom it would be to "mix and match".

…

On Wed, Jan 7, 2026 at 2:56 PM will wieder ***@***.***> wrote: *wwieder* left a comment (ESCOMP/CTSM#3680) <#3680 (comment)> @gunterl <https://github.com/gunterl> can you clarify what you mean by *"I would like to have a good Greenland product for CESM3"*? In B cases we'll eventually be spinning up the land with model atmosphere data from CPL_HIST output. Since you say that Cecile's cases eventually fill in this hole, it seems like we'll have land initial conditions from CPL_HIST that would generate reasonable ice sheets in Greenland that are consistent with CESM3 climate. Is this adequate for your purposes? Alternatively (or additionally) in I cases we'd need discuss potential modifications to DATM products if you want to initialize the Greenland ice sheet with initial conditions from an I case? - Is this something you all regularly do? - If so, can we just initialize it with a GSWP3 forced run, which seems to have more reasonable precipitation over Greenland? or - Do we need to modify the CRU-JRA DATM precipitation inputs regionally for some reason? A clearer picture of what initial conditions the land ice group needs would be helpful. — Reply to this email directly, view it on GitHub <#3680 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB6QV6FUAHGYEPUW7HCJQLT4FV6KHAVCNFSM6AAAAACQ7ULAD6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMRQHE3TGNBSHA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Gunter Leguy, Ph.D Scientist NSF National Center for Atmospheric Research cell: (575) 418 1021 desk: (303) 497 1790

[samsrabin]

CTSM SE meeting: Replace all of Greenland with GSWP3 data. All variables, not just precip. Bill Sacks might be able to help with the masking for this process.

Do we still call it CruJra? Yes. We didn't change it when we filled Antarctica. Just make sure to clearly document in publications and website.

[samsrabin]

@wwieder will talk with @gunterl about timing for this work.