Enable MALI Greenland mesh overlap with ocean mesh, through a buffer defined by a geojson file#876
Conversation
andrewdnolan
added
land ice
in progress
andrewdnolan
force-pushed
the
moderate_refinement
branch
from
a632621 to
589e70e
Compare
| gridded_dataset=source_gridded_dataset_1km, | ||
| projection=src_proj, geojson_file=None) | ||
| gridded_dataset=source_gridded_dataset_1km, projection=src_proj, | ||
| geojson_file="greenland_only_outline_45km_buffer_latlon_singlepart.geojson", # noqa: E501 |
There was a problem hiding this comment.
@trhille , will this line make this new geojson buffer the new behavior? Is that what we want?
There was a problem hiding this comment.
Yes, that will make it cull to the geojson file if you set cull_distance < 0 in the .cfg file. If cull_distance > 0, it will cull both to the geojson file and to the uniform distance from the ice margin.
We could set geojson_file in the .cfg file instead of hard-coding it here, for some extra control.
There was a problem hiding this comment.
To answer your second question, I think we want this to be the new default behavior, but that could still be introduced through the default .cfg file instead of hardcoded in the script.
There was a problem hiding this comment.
A config option seem like the right call. We can have it on by default if, but also the option turn it off and/or use a different buffer geojson file.
TestingI successfully created a 4–40km mesh with these changes: However, I notice that the geojson file seems to extend beyond our bed topography data set in a few places and we get artifacts with bedTopography = 0 at the edge of the mesh: I'm not sure what's the best way of dealing with this. We could manually edit the geojson to stay within the bounds of the gridded dataset. Or we could set a mask that is the intersection of the two. It's unlikely to ever actually interact with ice, but it's definitely not a desirable feature. |
andrewdnolan
changed the title
trhille
left a comment
trhille
left a comment
There was a problem hiding this comment.
@andrewdnolan , we need to figure out how to avoid those bed topography interpolation artifacts in the screenshot posted above. I think that's within the scope of this PR.
|
Using a cull distance of -100, the latest commit removes the areas of zero bed topography: |
Make Greenland geojson filename user configurable instead of hardcoded. Also update docs to reflect changes to landice.mesh.
|
My testing confirms that those interpolation artifacts: |
trhille
left a comment
trhille
left a comment
There was a problem hiding this comment.
Thanks @andrewdnolan! Approved based on testing above.
3 participants
This pull request adds a geojson file for 45 km buffer off of the coast of Greenland. This is then used by the Greenland mesh generation testcase.
api.rst) has any new or modified class, method and/or functions listedTestingin this PR) any testing that was used to verify the changes