search

ESCOMP / CTSM

Community Terrestrial Systems Model (includes the Community Land Model of CESM)
http://www.cesm.ucar.edu/models/cesm2.0/land/
Other
296 stars 301 forks source link

Install dual polar VR grid into CTSM #2386

Open adamrher opened 4 months ago

adamrher commented 4 months ago

Install the dual polar grid into CTSM. The main CAM issue is here: https://github.com/ESCOMP/CAM/issues/979. We are aiming to complete this in time for the CESM3 release. I can do the PR, with some help from the SEs.

I do have surface datasets for this grid, but they are not using the new ctsm5.2 dataset creation tool. I presume we want CESM3 running with the 5.2 datasets, and so I will plan to create those.

I would probably just copy the ARCTICGRIS grid for setting up the finidat file. We have spun-up inic for that grid, and so I think we should utilize them.

grid dualpoles ce

FYI, we've decided to label this the 'POLARCAP' grid.

ekluzek commented 4 months ago

OK, so this adds a new grid: ne0np4.POLARCAP.ne30x4. We will want the surface dataset for this created with the new ctsm5.2 mksurfdata_esmf tool. Since, that tag is really close at this point (see #2372), I propose we wait to bring this in until after ctsm5.2.0 is created. This will also need an update to ccs_config. In #2385 we update the externals to close to the latest in ccs_config, so after that point it would be easy to bring in the ccs_config update needed for this.

@adamrher is waiting roughly a few weeks on this to do ctsm5.2.0 and have updated externals OK for this work? We can definitely do it before CESM3. This is also something that could/should go on the b4b-dev to make it easier to bring in.

adamrher commented 4 months ago

@adamrher is waiting roughly a few weeks on this to do ctsm5.2.0 and have updated externals OK for this work?

No problem. I just wanted to get it on the docket for CESM3. If u recall our grids spreadsheet, we did plan for this grid. We also planned to bring in 2 MPAS VR grids, but that's looking increasingly unlikely (https://github.com/ESCOMP/CAM/issues/507#issuecomment-1960429035).

ekluzek commented 4 months ago

Do I recall the grids spreadsheet? Do I recall it? Ohhhh, yes. That spreadsheet has been the bible for putting together the ctsm5.2 branch. It's something that @slevis-lmwg and I have been constantly consulting. And it's really been great to know we are referencing a document that we got shareholders across CESM to agree to. That allowed us to drop a ton of non-important grids, but make sure we get all the important ones in place.

I just took a look and updated that document about this one. Thanks for the update that the 2 MPAS VR grids are unlikely.

The two other grids in red outside of the above three are: C5, and ne3.np4 do you know anything about their status @adamrher ?

Oh, and one other thing there seems to be a mpasa15-3conus grid in ccs_config, but there are also reference to mpasa15-3. What is going on there? I assume it should just be mpasa15-3conus -- because mpasa15-3 doesn't give a hint on where the high resolution 3km part is.

adamrher commented 4 months ago

Good! Sorry I didn't mean to come off as patronizing.

I've made some adjustments to the document. You'll notice there are now only two grids in the "install in CESM3" columns: the polarcap grid and MPAS 15 km. I'm not sure about the status of the MPAS 15 km grid so Ill look into that.

The ne3np4 and C5 grids are not installed anywhere, and since there's no X in the "install in CESM3" column, I think you can just ignore them [note - the ne3pg3 grid is our primary testing grid]. I'd like to hold another meeting before I start actually deleting lines from this spreadsheet.

Re: mpasa15-3conus. That is installed in ccs_config, along with some other hi-res mpas grids to support the EarthWorks branch. I'm not seeing references to just "mpasa15-3" though? I agree that is an inadequate grid name.

slevis-lmwg commented 4 months ago

When I type git grep mpasa15-3 in /ccs_config, I get:

component_grids_mct.xml:  <domain name="mpasa15-3conus">
component_grids_nuopc.xml:  <domain name="mpasa15-3conus">
component_grids_nuopc.xml:    <mesh driver="nuopc">$DIN_LOC_ROOT/share/meshes/mpasa15-3.conus_ESMF_desc.210816.nc</mesh>
modelgrid_aliases_mct.xml:  <model_grid alias="mpasa15-3_mpasa15-3" not_compset="_POP">
modelgrid_aliases_mct.xml:    <grid name="atm">mpasa15-3</grid>
modelgrid_aliases_mct.xml:    <grid name="lnd">mpasa15-3</grid>
modelgrid_aliases_mct.xml:    <grid name="ocnice">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:  <model_grid alias="mpasa15-3_mpasa15-3" not_compset="_POP">
modelgrid_aliases_nuopc.xml:    <grid name="atm">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:    <grid name="lnd">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:    <grid name="ocnice">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:  <model_grid alias="mpasa15-3-conus_mpasa15-3-conus" not_compset="_POP">
modelgrid_aliases_nuopc.xml:    <grid name="atm">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:    <grid name="lnd">mpasa15-3</grid>
modelgrid_aliases_nuopc.xml:    <grid name="ocnice">mpasa15-3</grid>

In my copy of component_grids_nuopc.xml I removed "conus" from the name to get things to work. If we prefer a different fix, I'm open to that.

adamrher commented 4 months ago

Ah, I see. So the problem is that the grid aliases are referencing "mpasa15-3" but it doesn't actually exist in component_grids_nuopc.xml. So we need to clean up moodelgrid_aliases_xxx to only reference "mpasa15-3conus". And it looks like a duplicate grid alias in nuopc.

Pinging @brianpm and @jtruesdal so they aware of this issue and can maybe comment on the origins of this 15-3 business. Brian and John we are not supporting this grid in CESM3.0, so I think this is more relevant to the EarthWorks branch.

ekluzek commented 4 months ago

@adamrher you did NOT come off as patronizing. I should have added the \<snarky>\<\/snarky> markers on my beginning comment, or at least a ;-) smiley face. I was actually trying to show my appreciation for that document through sarcasm. If you could've seen my face while I was writing you would've seen my smile. The impression I got was that you were under-appreciating that file and how pivotal it's been for us, so you thought that maybe we had forgotten about it. But, I also appreciate you talking about this so we could clear it up!

I'd love to have another confirmation meeting on what final CESM3 grids are supported.

And thanks for all the details you gave about the other grids, that's really helpful.

On the mpasa15-3conus grid, we might fix this after ctsm5.2.0 just because it sounds like it might take some time to figure out how to handle it. We also should have someone open an issue in ccs_config to talk about this. Should we have @slevis-lmwg start one, or would someone from CAM volunteer?

brianpm commented 4 months ago

I just reached out to @sherimickelson about this... we'll see if there is any EarthWorks input.

sherimickelson commented 4 months ago

From my understanding, EarthWorks is more interested in supporting the global uniform grids (at 120, 60, 30, 15, 7, and 3.75 km) vs. refined grids. This may change in the future and there are individual scientists tied to the project looking at refined grids, but the the current focus for support is the uniform grids.

adamrher commented 2 weeks ago

@ekluzek now that ctsm5.2 is on main I was going to go ahead and make a PR to resolve this issue. Is that OK?

As with the other SE-VR grids, Ill provide surface dataset support for 1850,1979,2000,1979-2026, which is what is listed in the CESM3 grids document.