[MITgcm-support] cartesian grid

[Martin Losch]

Himansu,

the model really only knows two horizontal direction: the i-direction and the j-direction. (for curvilinear grids you can give it the actual coordinates etc, but that's even more complictated than what you need). So you have to give the model the wind in the i-direction ("taux"=zonalWindFile, ) and the j-direction ("tauy"=meridWindFile,), the naming is admittedly a little unfortunate.
I mean that you have to rotate your wind vector (I asume with east-west and north-south components) according the orientation of the your grid.

vectorOnGrid = RotationMatrix*vectorOnLatLon, which rotationMatrix = (cos(phi) sin(phi); -sin(phi) cos(phi)). where phi the angle between north and your grid north. Don't nail me down on indices and direction of rotation (you can look that up in any linear algebra book) ...

Martin


On Aug 6, 2012, at 1:44 PM, Himansu Pradhan wrote:

> Hello MITgcm users,
> 
> I am using "Cartesian Grid" in the model. The Wind stress field  in my model domain looks like figure feb_tech.jpg (monthly climatological wind stress)
> 
> Plotting the same Wind Stress field in matlab (before making Input binary files) the plot looks like Input.jpg. (attached below both the figures)
> 
> In Comparing both the plots it is seen , that the DIRECTION OF THE VECTORS of figure "Input.jpg" has been rotated (clearly seen in SOUTHERN region) . this means that i am providing wrong WIND STRESS (Input) to the MODEL.
> 
> I want to know how other modelers (those using Cartesian coordinates in the MODEL) deals with it  ???? / Should i rotate my model domain (how to do this in the MODEL)????? / Or any kind of suggestion/help
> 
> Thanking all in advance,
> 
> -himansu 
> <feb_tech.jpg><Input.jpg>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support