I’m trying to refine a northwestern Pacific with NEMO4.2.x.
Below is AGRIF_FixedGrids.in
1
65 210 250 416 5 5 3
0
The parent domain (ORCA05) has no cyclic east-west boundaries (nx = 720 & ny = 510).
However, the model stopped with below error message.
STOP
AGRIF zoom is East-West cyclic, imin must = 1
This is strange, because the nest domain doesn’t cross the cyclic points (northwestern Pacific).
Do I need to set additional options in namelist_cfg or modify the number in AGRIF_FixedGrids.in?
Meanwhile, this model stably runs with cyclic east-west boundaries in parent domain (722x511).
although the parent domain “lost” its dublicated columns and rows along the boundaries with the transition from NEMO4.0.x to 4.2.x it still has cyclic east-west bundaries. jperio should be 6 in domain_cfg.nc.
Your nest obviously doesn’t have cyclic boundaries, consequently jperio must be 0 in 1_domain_cfg.nc.
How did you create your (1_)domain_cfg.nc files? Here are the guidlines for how to create a consistent set of input-files.
Thank you for your kind suggestions. Following your suggestion, I’ve tested the model with new domain cfg files (host’s jperio=6 and nest’s jperio=0). However, the model was stopped due to the volume (averaged bathymetry) mismatch between the host and nest. I’ve tried some more tests by changing some options, but it’s still in swamp.
Attached files are the (1_)namelist_cfgs, which are for the current model test.
Many thanks again for any hints or help!
I don’t know all the details of how the connection between the parent and child bathymetries is done in the current version of the DOMAINcfg tool, so the following is kind of “just a guess” (maybe @jchanut could give some better advice):
In your namelist_cfg I spotted bathy_updated.nc as topography file for the unnested ORCA05 configuration. From the file-name, I assume you provide the updated parent bathymetry from the old (NEMO3.6-based ?) NPAC10 configuration. This would include modifications according to the connection between parent and child, as implemented in earlier versions of the nesting tools and additionally according to the partial cells definition you had for that configuration. (Is the latter the same as in your current namelist_ref? If not, don’t forget to add the correct values to namelist_cfg.) I could imagine some interference here…
So my suggestion is to provide the original ORCA05 bathy_meter.nc (720x510) file to generate the new (1_)domain_cfg.nc files.
