[Mitgcm-support] RE: grid scale oscillations

[mitgcm-support at dev.mitgcm.org]

Hi Dan,

 I haven't looked at the figure - sorry!
 The CD scheme does reduce numerical noise,
however at these resolutions it is not the
most obvious numerical issue. 
 Things to check are
 o biharmonic friction value v. lower limit
   for frictional boundary and upper limit
   from numerical stability.
 o dissipation for temperature
 o is it a standing pattern i.e. does it appear
   in time averages.

 We have some configs. at both slightly higher and
 slightly lower resolutions to this that look fine.
 We'll try and track down the parameters they are using.

Chris
 

-----Original Message-----
From: Daniel Lea [mailto:daniel.lea at jhuapl.edu]
Sent: Friday, September 06, 2002 4:30 PM
To: support at mitgcm.org
Subject: grid scale oscillations



Hi,

I'm having a problem with grid-scale "waves" in the temperature and
salinity fields. I am running a 30 level model at 1/12 degree with open
boundaries. I attach my data file and some example output showing the
temperature of the topmost layer. You can see the "waves" I describe in
the centre of the domain. Notice they all run east to west. The model
doesn't crash, but they are obviously unrealistic and ugly.

I was wondering if you had seen anything like this before and whether you
had any suggestions for getting rid of these waves?

I should tell you also that as consequence of having OBCs I had to #undef
INCLUDE_CD_CODE so maybe that's the problem?

Cheers

Dan