Repository navigation
CTSM Changes Needed with a new Ocean Mask t233 #3766
Description
CESM is considering some changes to some points in the t232 land/ocean mask for MOM6. So the new grid would presumably be called t233. Presumably a few grid cells that are currently land would become ocean (so not run in CTSM anymore), and a few points currently ocean would become land (so would be newly run by CTSM).
This issue is just to think through how this might impact CTSM and what we'll need to do to respond to it.
- Update to the new ccs_config with the new t233 grid (has mesh and mapping files for ROF)
- Change our testing to use the t233 mask (CTSM, RTM, MOSART, mizuRoute) (Move our testing from POP masks to MOM masks (mt233) #2540)
- Use init_interp for the current finidat files with the t232 mask (this should just happen)
- Create new finidat files where we spinup longer to get the new land points spunup
- Full spinup with the new t233 grid to make sure everything is fully spunup
The ccs_config update will contain both a mesh file for the t233 grid to be used for fully coupled cases and a mask file that CTSM will use for offline cases. So almost all of the work for this happens in ccs_config.
Since, these changes will likely just be for a few gridcells it seems likely that initially relying on our current finidat files should be fine. However, these new land points are likely to be over glacier areas so will need to be spunup to get a reasonable snowpack. So doing a simulation for say 50 years starting from a t232 finidat file seems like it should be OK to do. Eventually, we'll redo spinup so the final step will happen anyway.
Activity
wwieder commentedon Feb 18, 2026
Naive question here, but what happens if landfrac on the surface dataset = 0 and the ocean mask =0 (i.e. says a gridcell is a land point)? Is this something that happens already? Could it with the new t233 ocean mask?
ekluzek commentedon Feb 18, 2026
Yes, there's a separate land-mask on the surface dataset. But, the model uses the land/ocean mask determined by the ocean model. So there are discrepancies. What we used to do was to run points that the surface dataset identified as ocean, but the ocean mask as land -- as wetland. Now what we do is we run those points as the most common vegetation type in the cell (if there is any vegetation) or as bare-ground if there isn't any.
So we'll probably have a few new land points that will run as bare ground.
ekluzek commentedon Feb 18, 2026
I'm leaving next here, but I realized this doesn't need much discussion. It will be good to verify with everyone that we've thought of everything in the top list.
In talking with @briandobbins and @wwieder on this, we don't need this in cesm3_0_beta08 (although the ccs_config tag will need to go into cesm3_0_beta08). But, we will need this done by the time the spinup starts. So we'll want this in beta09.
slevis-lmwg commentedon Apr 15, 2026
Putting back "next" to update the date on this issue's milestone. This issue's priority may be increasing again.
3 remaining items
wwieder commentedon Apr 17, 2026
Looks like the ccs_config PR has been merged, so we can start working on this issue on the land side of things.
samsrabin commentedon Apr 23, 2026
From CTSM SE meeting:
slevis-lmwg commentedon Oct 12, 2026
I'm marking "done" for sprint accounting. Other issues and PRs will continue to track in the next Sprint.