Version 27 (modified by frrh, 6 years ago) (diff) |
---|
VALID-12_CLEVY_CoupledInterface
The PI is responsible to closely follow the progress of the action, and especially to contact NEMO project manager if the delay on preview (or review) are longer than the 2 weeks expected.
Summary
Action | VALID-12_CLEVY_CoupledInterface |
---|---|
PI(S) | C. Levy |
Digest | Validation of NEMO coupled interface, using OASIS in the trunk (as future 4.0 release) |
Dependencies | |
Expected for | |
Ticket | #2147 |
Branch | NEMO/branches/$YEAR/dev_r{REV}_{ACTION_NAME} |
Previewer(s) | S. Masson, R. Hill, E. Maisonnave |
Reviewer(s) | S. Masson, R. Hill, E. Maisonnave |
Link | ExtractUrl(.)? |
Abstract
The NEMO OASIS interface is important for all coupled configurations. Testing that this section (mainly the coupling activated) is functional must be done before the 4.0 announcement.
Implementation
To check if the coupling interfaces are functional, three steps/tests will be set up:
STEP1:
Validation of the SAS reference configuration
STEP2:
Set up a test using NEMO (starting from the global reference configuration) with key_coupled, OASIS active and a toy atmospheric model
STEP3:
Using this configuration, check that the coupled interface is functional for ocean dynamics, sea ice and biogeochemistry
Reference manual and web pages updates
Not done yet
Preview
For this taks, no development is scheduled: the goal is only to check that the coupling interface of NEMO 4.0 is working.
The advice of previewer on the list of proposed tests is of course welcome.
Tests
TEST1: Validation of the SAS reference configuration
The reference configuration ORCA2_SAS_LIM has run sucessfully for one year using the set up defined in the NEMO shared reference (NEMO trunk rev 10272).
It has been compared with a one year run of the forced reference configuration ORCA2_ICE_PISCES (NEMO trunk rev 10272).
- Outputs of ORCA2_SAS_LIM after one year, in particular the ORCA2_SAS_5d*.nc output files, and the ORCA2_SAS_00005475_restart_ice.nc and ORCA2_SAS_00005475_restart.nc restart files
- Outputs of ORCA2_ICE_PISCES after one year
Validation of the results:
- Some questions raised on the forced configuration itself, see below
- After one year experiment, the outputs and ice restart files created in the SAS run make sense
- Having a complete an exact validation of this configuration will need some more work: here the ice is forced with an ocean input file containing the surface fields. These fields are not varying with time for now in the input file. To have a complete validation, it would need:
- to run a one year of ORCA2_ICE_PISCES full reference configuration, with outputs of these daily surface fields
- to use this output as input to the SAS configuration
- to make some subtle changes in SAS in order to have the exact same use of these fields in time (not the case for now, work done once by Sébastien and Eric, remaining to be taken in account again
- Once all this is done, the ice restart file should be strictly identical between the ORCA2_ICE_PISCES run and the ORCA2_SAS_LIM one
This work has not been done at this stage, so that for now the conclusion is:
The ORCA2_SAS_LIM configuration is running in NEMO 4.0, e.g. the SAS interface is functionnal in this case.
Comments on forced reference configuration ORCA2_ICE_PISCES :
The reference configuration is set up to start from rest (e.g. the very initial start, no ocean restart at all). In this case, using a call frequency of the ice =3 , nn_fsbc=3, is unstable: the run crashes after a few tenth of timesteps. Using nn_fsbc=5 (e.g. ice called every 5 timesteps, inducing less high frequencies in the forcing since the mean value used is calculated on a longer time scale), the model runs fine. This instabilities was not expected in the ORCA2 configuration, usually considered as very stable. Still this nn_fsbc=5 values has always been used up to now...
Once the ocean has started, after a few years, it seems that the nn_fsbc=3 option is indeed stable enough.
TEST2: Set up a test using NEMO (starting from the global reference configuration) with key_coupled, OASIS active and a toy atmospheric model
Using this configuration, check that the coupled interface is functional for ocean dynamics, sea ice and biogeochemistry
- OASIS3-MCT version is installed revision 2499
- a CPL_TOYATM configuration is created, based on ORCA2_ICE_PISCES, and with oasis cpp keys key_oasis3 key_oa3mct_v3 active
- First compilation problem, requests a fix in oasis for now (see attached file in the ticket #2147 )
- The toyatm developped by Eric Maisonnave is compiled as a simple toy for the atmosphere
- iodef.xml modified for OASIS use
- the number of (fake) fields is adjusted in the namcouple and namelist_cfg files
- Coupling frqequence is set to 1 (every time steps)
- Test is short (5 timesteps, since the atmospheric fields are constant in space and time and not realistic)
- The run is completed
Validation: Outputs available here. The NETCDF file of SST seen by the toy atmosphere ATSSTSST_toyatm_01.nc looks reasonable, see pot of Sea surface temperature seen by the atmospheric toy :
The global 2° configuration in a very simple coupled mode using OASIS is running fine with NEMO 4.0, e.g. the OASIS interface between ocean dynamics, sea ice and atmosphere is functionnal.
[[Image
Summary
As a result of the tests, a few things needs to be updated/fixed:
- sbccpl.F90
- namelist_ref
A new test case has been build with a toy for the atmosphere (proposed name for this new test case: YAZD, for Yardstick Atmosphere to Zero a Demonstrator, since yazd is an oasis in Iran...). It is proposed to add this configuration in the list of available test cases.
Review
Eric Maisonnave
Assessments:
- Is the proposed methodology now implemented? -yes-
- Are the code changes in agreement with the flowchart defined at preview step? -yes-
- Are the code changes in agreement with list of routines and variables as proposed at preview step? -yes-
- Is the in-line documentation accurate and sufficient? -yes-
- Do the code changes comply with NEMO coding standards? -don't know-
- Is the development documented with sufficient details for others to understand the impact of the change? -yes-
- Is the project literature (manual, guide, web, …) now updated or completed following the proposed summary in preview section? -don't know-
Finding:
- Is the review fully successful? If not, please indicate what is still missing
TEST1: the intermediate step is completed. the review is fully successful
TEST2: It should be mentioned that coupling validation is also done for sea-ice (but not biogeochemistry) related coupling fields.
The atmosphere toy model sources should be distributed within the NEMO directory (tools ?) the review is fully successful
R Hill review comments:
- I agree with the need to be explicit about which revision(s) of OASIS3-MCT have been used in testing and therefore which revisions of the coupler are officially supported by NEMO.
- Items discussed by email with Eric include the simple changes necessary to sbccpl.F90 referred to above to allow NEMO to be compatible with both OASIS3-MCT vn3.0 and vn4.0.
- 2nd order conservative remapping using coupling field gradient terms is not explicitly catered for or tested (at least not in the ocean->atmosphere direction). This is fine, but worth explicitly mentioning in documentation.
- Also worth mentioning in documentation that when using OASIS3-MCT vn4.0, 1st order conservative regridding cannot use weights files containing 2nd order terms (and of course 2nd order regridding cannot use 1st order weights).
Attachments (1)
- ATSSTSST_ts5.jpeg (391.0 KB) - added by clevy 6 years ago.
Download all attachments as: .zip