Neon datm mode#3850
Conversation
ekluzek
left a comment
ekluzek
left a comment
There was a problem hiding this comment.
@jedwards4b glad to see this work is getting started. I request some changes that come about because of previous work I've done where I found problems with a similar approach that I took as well. It likely might be good to chat about this and maybe include Will on the discussion. NEON is a cool project and it will be really awesome to get their data integrated into the model.
| <grid name="rof">null</grid> | ||
| </model_grid> | ||
|
|
||
| <model_grid alias="1x1_NEON" compset="DATM.+CLM|DATM.+SLND"> |
There was a problem hiding this comment.
It appears that 1x1_NEON is doing the same thing as CLM_USRDAT_NAME, I'm not sure we need an additional name here that's specific for NEON.
There was a problem hiding this comment.
I'd also suggest when using 1x1_ to use 1x1pt_ as used in other places since the default units for grids are in degrees, and the "pt" says that in this case it's "points". So 1x1pt_ is a single point regional grid, while something like 5x5pt_ is a grid that's has 5 points in latitude and also 5 in longitude.
There was a problem hiding this comment.
If you look at the line below you will see the reason for this name, it supports all of the NEON tower sites:
$DIN_LOC_ROOT/share/domains/single_point/datmdata_NEON/$NEONSITE/domain.lnd.fv
0.9x1.25_gx1v7_$NEONSITE.nc
How about we change the file name to: domain.lnd.1x1pt_$NEONSITE_20210210.nc
There was a problem hiding this comment.
Yes, since these files shouldn't change (except maybe adding new sites?) I think that solution makes perfect sense.
There was a problem hiding this comment.
The one other question I had about this though is I though @mvertens was doing something where single point sites didn't require domain files anymore for NUOPC? I'd have to look at it again to see how that worked though...
There was a problem hiding this comment.
OK, I looked at that again, and what she was doing was to eliminate the need for both mesh files AND domain files. But, it's still configured to need domain files.
| </domain> | ||
|
|
||
| <domain name="1x1_NEON"> | ||
| <nx>1</nx> <ny>1</ny> |
There was a problem hiding this comment.
The problem here that I ran into with CLM_USRDAT_NAME is that this domain file when updated will change ALL existing simulations being run with it. This is problematic. As such I came to the conclusion a better way to handle it is to have scripts that setup single-point create user-mod directories where the DOMAIN file is set to a particular path that includes a time-stamp in the filename. This way changes to files don't change the results of someone else's existing simulations.
|
A relevant issue in CTSM is this one ESCOMP/CTSM#1041 where I decided that having specific filenames hardcoded into CTSM wasn't useful (having CLM_USRDAT_NAME still is useful), and using user-mod directories was a better approach. An example user-mod directory put together for a single-point case is here: PTCLM is being deprecated, but I think the above kind of user-interface for what it does still makes sense. With the creation of these types of directories you can create a case with a single added option added to create_newcase. And that case won't suddenly have answer changes because someone else updated files that have generic names without time-stamps. |
|
superceeded |
### Description of changes Adds support for NEON tower data in datm. ### Specific notes Contributors other than yourself, if any: CMEPS Issues Fixed (include github issue #): Are there dependencies on other component PRs - [X] CIME (list) Requires branch neon_data_mode (PR ESMCI/cime#3850) - [ ] CMEPS (list) Are changes expected to change answers? - [X] bit for bit - [ ] different at roundoff level - [ ] more substantial Any User Interface Changes (namelist or namelist defaults changes)? - [ ] Yes - [X] No Testing performed: - [X] (required) aux_cdeps - machines and compilers: cheyenne intel - details (e.g. failed tests): SMS_Vnuopc_Ld5.TL319_t061.2000_DATM%JRA-1p4-2018_SLND_SICE_SOCN_SROF_SGLC_SWAV_SESP.cheyenne_intel fails in build with Invalid values ['CORE_IAF_JRA_1p4_2018'] - [ ] (optional) CESM prealpha test - machines and compilers - details (e.g. failed tests): Hashes used for testing: - [ ] CIME - repository to check out: https://github.com/ESCOMP/CESM.git - branch: master - hash:8b48a2e - [ ] CMEPS - repository to check out: https://github.com/ESCOMP/CESM.git - branch: master - hash: v0.9.0 - [ ] CESM - repository to check out: https://github.com/ESCOMP/CESM.git - branch: master - hash:
2 participants
Adds a 1x1 grid for NEON data sites. This is used with ctsm and cdeps to generate single column clm cases with this data.
Test suite: hand testing, scripts_regression_tests.py
Test baseline:
Test namelist changes:
Test status: bit for bit
Fixes
User interface changes?:
Update gh-pages html (Y/N)?:
Code review: