FW: [MITgcm-support] obcs for lock-exchange problem (fwd)

[samar khatiwala]

Chris, Michael: I don't know anything about this problem, but this does
not seem right:

> it seems strange to set h(1)=0. (i think i/m missing something here).

Won't this give rise to a solid wall? So I suspect you don't want h(1)=0

Non zero OB_Ieast/west indices will then give you open BCs. For example:

 OB_Jnorth=500*0,
 OB_Jsouth=500*0,
 OB_Ieast=3*-1,
 OB_Iwest=3*1,
 useOrlanskiEast = .TRUE.,
 useOrlanskiWest = .TRUE.,
 useOrlanskiNorth = .FALSE.,
 useOrlanskiSouth = .FALSE.,

Samar

---------- Forwarded message ----------
Date: Sat, 17 Apr 2004 21:37:19 -0400
From: Chris Hill <cnh at mit.edu>
To: 'samar khatiwala' <spk at ldeo.columbia.edu>
Subject: FW: [MITgcm-support] obcs for lock-exchange problem

Samar,

 Do you have any setups or advice that might help Michael?

Thanks,

Chris

> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of
> Michael Schaferkotter (Contractor)
> Sent: Saturday, April 17, 2004 12:44 PM
> To: mitgcm-support at mitgcm.org
> Subject: [MITgcm-support] obcs for lock-exchange problem
>
> i/ve setup the lock-exchange problem on 320x60 x-z grid with
> 1.0m resolution in each extent.
>
> the temperature difference is 1.25 deg with cool on left.
>
> the OBCS are set as:
>
> # Open-boundaries
>  &OBCS_PARM01
>  OB_Ieast=-1,
>  OB_Iwest=1,
>  useOrlanskiEast=.TRUE.,
>  useOrlanskiWest=.TRUE.,
>  &
>
> it took me awhile to figure out how to turn off period
> boundary condtions.
> it seems strange to set h(1)=0. (i think i/m missing something here).
>
> the documentation in the /doc directory said that one should
> follow the docs on the web site, which i couldn/t locate.
>
> the plumes start off as they should warmer over cooler and
> velocity < 0 in upper domain and velocity (u) > 0 in lower domain.
> the rotors develop quite nicely.
>
> when the front hits the west edge (first) the front bends
> down the boundary as though the boundary is closed and
> eventually bends around.
>
> any pointers here?
>
> until i put h(1)=0 the periodicity was getting in the way.
>
> michael
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>