-
Notifications
You must be signed in to change notification settings - Fork 371
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add script support for ARRM10to60E2r1 ocn/ice grid #5178
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good @darincomeau.
Let me add that I also tested this on cori: two simulations are running successfully with this MPAS mesh and with ne30 and the ArcticRRM for the atmosphere/land mesh.
…in sea ice timestep" This reverts commit 2e4100b.
I've added the Do not merge tag until I can generate land ICs for these configurations. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It seems like at least 2 changes are needed here.
In addition, I would request that the diagnostics
files get copied into place from:
/lcrc/group/e3sm/ac.xylar/compass_1.2/anvil/test_20220730/ec_arrm_for_e3sm/ocean/global_ocean/ARRM10to60/PHC/files_for_e3sm/assembled_files/diagnostics
to
/lcrc/group/e3sm/public_html/diagnostics
Please make sure they are all in the E3SM group and have group read/write and world read permissions.
These files being available will significantly speed up MPAS-Analysis runs. I will sync these files over to other E3SM supported machines as part of the E3SM-Unified emergency release (1.7.1) that I'm working on.
<config_Redi_min_layers_diag_terms>6</config_Redi_min_layers_diag_terms> | ||
<config_Redi_min_layers_diag_terms>15</config_Redi_min_layers_diag_terms> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If I undrestand correctly, this is changing Redi for all configurations that have GM on. This needs to only be for ARRM10to60E2r1
. I believe this is correct:
<config_Redi_min_layers_diag_terms>6</config_Redi_min_layers_diag_terms>
<config_Redi_min_layers_diag_terms ocn_grid="ARRM10to60E2r1">15</config_Redi_min_layers_diag_terms>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oops, that was definitely my mistake commit, I certainly should have included the grid specification. Thanks for the catch!
@milenaveneziani I just want to confirm that we do indeed want this separate setting for this mesh, before I make the change (or just revert the commit)?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yes, we want it at 15 for this mesh.
decomp_date = '220730' | ||
decomp_prefix = 'mpas-o.graph.info.' | ||
restoring_file = 'sss.PHC3.0_monthlyClimatology.ARRM10to60E2r1.220802.nc' | ||
analysis_mask_file = 'ARRM10to60E2r1_mocBasinsAndTransects220730.nc' |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think this file should have been given the date it was created. The date stamp on the mocBasinsAndTransects
files is a date stamp for the regions used in the masks, rather than the masks themselves. This won't matter in E3SM but it will be a problem if the file to be added todiagnostics
was modified in this way. In general, I think it would be better to leave the file with the date stamp created in compass
:
ARRM10to60E2r1_mocBasinsAndTransects20210623.nc
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For a little more background, this date is controlled and updated in geometric_features
:
https://github.com/MPAS-Dev/geometric_features/blob/master/geometric_features/aggregation/__init__.py#L56-L58
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for pointing this out. I just diffed the file with the 220730
datestamp and the 20210623
one and they are the same, so I must have mv
ed this to match the other datestamps? I shouldn't have done that - I'll fix it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we did go through a couple of mesh iterations between June and July though..
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we did go through a couple of mesh iterations between June and July though..
@milenaveneziani, that date stamp is from June 2021, not 2022. It is the date that we added the separate Atlantic and Atlantic+Mediterranean MOC basins to geometric features and has nothing to do with when the file or the mesh itself was created.
@xylar Done -thanks for pointing out the missed step! |
@darincomeau, I just synced the files over from Anvil to Cori. Thanks for putting them in place! I'm trying to run MPAS-Analysis (with
? |
Opened - thanks for testing @xylar ! |
…aving 6 as general default
…insAndTransects20210623.nc
@xylar the last two commits addressed your changes. As part of |
Thanks @darincomeau. As soon as I have permission, I'll rerun the analysis and then (hopefully) approve. |
@xylar that Cori run should be opened now. Note that the |
@darincomeau, ah, thanks. I was expecting you to change group rather than make world readable so didn't notice. No, the |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@darincomeau and @milenaveneziani, I think this mesh is working okay and the necessary files are in place. With @darincomeau's recent changes, I think the code changes look reasonable.
I reran MPAS-Analysis here:
https://portal.nersc.gov/project/e3sm/xylar/analysis/arrm/20220803.WCYCL1950.ne30pg2_ARRM10to60E2r1.baseline.cori-knl/ts_0011-0020_climo_0011-0020/
The mapping and mask files were taken from the diagnostics
cache, as expected.
There are some fairly serious temperature and especially salinity biases in the WOA18 plots. I'm merely pointing those out as something that HiLat will presumably need to investigate. It isn't clear to me whether changes in the mesh (e.g. resolution or passages) would have a mitigating effect.
Thanks @xylar !
Yeah that's been a pretty persistant problem, actually looks a little bit better with the new mesh. Here's an run with an older mesh for comparison: |
Okay, thanks @darincomeau. Nothing new, then. That's presumably good news, and we need to find a solution in the physics and parameterizations, not the mesh. |
@xylar: this is great. I believe the new mesh is so much better than the old one, and the new bathymetry is so much smoother. The B-cases with the low-res atmosphere look quite awful at the moment, but that has nothing to do with the mesh (imho). The coupled cases with the atmRRM mesh should look much better. |
I've added a default land |
Removed the Do not merge flag - @jonbob I think this is ready whenever's a good time. |
Add script support for ARRM10to60E2r1 ocn/ice grid Brings in support for the ARRM10to60E2r1 ocn/ice mesh. MPAS-Dev/compass#414 has details on the mesh generation, which is an updated version of the oARRM60to10 mesh. This PR includes all necessary mapping, runoff, and domain files for the following configurations: * ne30pg2_ARRM10to60E2r1 * arcticx4v1pg2_ARRM10to60E2r1 * TL319_ARRM10to60E2r1 (JRA-forced) * T62_ARRM10to60E2r1 (CORE-forced) [BFB] for all currently tested configurations
test merge passes:
verified that all new files are available on the inputdata repo merged to next |
merged to master |
Brings in support for the
ARRM10to60E2r1
ocn/ice mesh. MPAS-Dev/compass#414 has details on the mesh generation, which is an updated version of theoARRM60to10
mesh. This PR includes all necessary mapping, runoff, and domain files for the following configurations:ne30pg2_ARRM10to60E2r1
arcticx4v1pg2_ARRM10to60E2r1
TL319_ARRM10to60E2r1
(JRA-forced)T62_ARRM10to60E2r1
(CORE-forced)Testing
The following 5 day smoke tests have been run on Chrysalis:
WCYCL1950.ne30pg2_ARRM10to60E2r1
WCYCL1850.arcticx4v1pg2_ARRM10to60E2r1
GMPAS-JRA1p4.TL319_ARRM10to60E2r1
GMPAS-IAF.T62_ARRM10to60E2r1
A
WCYCL1950.ne30pg2_ARRM10to60E2r1
B-case is currently running on Cori as the final check for this mesh, and is in year 26. Here is the MPAS-Analysis for years 11-20[BFB] for all currently tested configurations.