Welcome to the new DiscussCESM forum!
We are still working on the website migration, so you may experience downtime during this process.

Existing users, please reset your password before logging in here: https://xenforo.cgd.ucar.edu/cesm/index.php?lost-password/

Trouble with u,v sponges

I tried the sponges in my Arctic domain and it's running with just T,S sponges, failing with u,v sponges:
FATAL from PE 229: external field u does not have an associated record dimension (REQUIRED)
Odd because I have 12 records of all four fields.
 
Now it wants the u,v fields on the tracer points. I'd been told that, but hadn't thought that someone would give me u,v on u,v points when they knew that. The tracer sponges use
horiz_interp_and_extrap_tracer_record and now the u,v sponges do too. Maybe it's time for changes to the sponge code if the sponges are provided on the native grid.
 

MattHarrison

Matthew Harrison
New Member
Now it wants the u,v fields on the tracer points. I'd been told that, but hadn't thought that someone would give me u,v on u,v points when they knew that. The tracer sponges use
horiz_interp_and_extrap_tracer_record and now the u,v sponges do too. Maybe it's time for changes to the sponge code if the sponges are provided on the native grid.

u,v sponges are currently assumed to be provided on the tracer grid and shifted to the staggered locations in the sponge routine. This is causing a problem in regional cases since the values on the tracer grid external to the model domain are not given , hence the values at the edges are wrong. Values should instead be coming into the sponge routine on the C-grid locations. I think we should make this change , but how much of the original code should we support?
 
There is a PR on github for the u,v sponges, fixing the tile boundary issue I was having. Whether we should redo the horiz_interp_and_extrap_tracer_record code for u,v points is a separate discussion.
 
Top