boundary problem

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
shifei
Posts: 30
Joined: Thu Aug 01, 2013 3:11 pm
Location: Ocean University of China

boundary problem

#1 Unread post by shifei »

Hi,

I try to simulate the northwestern pacific region. Near the boundary, the sst appears discontinuous; the ssh appeares a gradient which results virtual velocity along the boundary. I take the boundary condition like these: Cha for free surface, Fla for 2D momentum, RadNud for 3D momentum and tracers. The data used for boundary file is from OFES.
Attachments
romsssh.png
romstemp.png

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: boundary problem

#2 Unread post by kate »

Welcome to regional ocean modeling. You've given it a tough problem with a strong inflow coming in at an angle to your boundary. Can you move the western boundary over so that the flow comes in perpendicular to the boundary? Not that that's a guaranteed fix.

So, how bad does it get? Does it affect your region of interest? If it really is a problem, I'd add a nudging to a monthly climatology from that OFES, just around the boundary. What is the spatial and temporal resolution of OFES, anyway? And your spatial resolution?

shifei
Posts: 30
Joined: Thu Aug 01, 2013 3:11 pm
Location: Ocean University of China

Re: boundary problem

#3 Unread post by shifei »

Dear kate,
I'm interested in the vortex coming from the east boundary. The model grid space is 1/10*1/12 and OFES data resolution is 0.1*0.1 and monthly interval.
The strong velocity from the west boundary turns into unstable and forms lots of vortex in the interior region. The simulated resoluts compared with OFES data is not good.
It really is a problem. I have added a nudging to a monthly climatology from that OFES over 10 points just around the boundary, but the results are not really improved. And I tryed the RADIATION_2D in the .h file, the results just improved a little. Are there others methods? Or I have to adjust my boundary position, like enlarging the region.

shifei
Posts: 30
Joined: Thu Aug 01, 2013 3:11 pm
Location: Ocean University of China

Re: boundary problem

#4 Unread post by shifei »

This is the results from OFES. We can see obvious difference near the west boundary.
Attachments
ofesssh.png

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: boundary problem

#5 Unread post by kate »

I have not been in the position of using boundary fields at more or less the same spatial resolution as my model. If at all possible, I would suggest trying higher temporal resolution in your boundary fields. Your OFES picture shows an eddy near the western boundary. That thing is not going to just sit there for a month. Next would be to expand your domain so that the strongly varying region is inside. Even so, in the Atlantic, getting the Gulf Stream separation to look realistic is not trivial - probably true in the Pacific too.

Post Reply