Model blow-up with refined grid

Discussion of how to use ROMS on different regional and basin scale applications.

Moderators: arango, robertson

Post Reply
Message
Author
txu68
Posts: 27
Joined: Wed Nov 11, 2015 4:49 pm
Location: Georgia Institute of Technology

Model blow-up with refined grid

#1 Post by txu68 » Wed Mar 29, 2017 5:31 am

Hello,

I have tried to incorporate a nested grid in a complicated estuary region. The model works fine without a nested grid. And I tried to generate the nested grid file by using coarse2fine.m and generate the contact point information by contact.m.

However, with the nested grid setting, the model blows up within 1 minute. By looking at the restart file, I see that the temperature at a corner of the wet points near the boundary becomes nan. So as the salinity. That is the only abnormal result I found from the restart file.

I suspect that the contact point information is not so accurate. So I re-generate another nested grid in a region with every point wet and also generate the corresponding contact point information. It seems that the model works fine.

Can anyone tell me what is wrong with my model setting?

Best,
Amy

txu68
Posts: 27
Joined: Wed Nov 11, 2015 4:49 pm
Location: Georgia Institute of Technology

Re: Model blow-up with refined grid

#2 Post by txu68 » Wed Mar 29, 2017 7:07 pm

An update to this problem.

I changed the model setting back to an old version of COAWST, the version does not require a contact file. And the cpp option for nesting is #define REFINED_GRID. The code I used to generate a nest grid is create_nested_grid.m instead of coarse2fine.m. With the child grid inside a complex esturary, the model seems to work so far.

Has the robust version of contact.m released?

Best,
Amy

Post Reply