[Mitgcm-support] Re: basic equations and grid

[mitgcm-support at dev.mitgcm.org]

Jianwu Chen wrote:
> 1) What kinds of governing equations in MITGCM? Are they Reynolds Averaged
> Navier-Stokes(continuity and momentum) and transport equatins for tracer?

MITgcm solves either the Hydrostatic Primitive Equations (HPEs) or
non-hydrostatic
equations which are both Boussinesq and so are not the Navier-Stokes equations.

> 2) In both curvilinear and Cartesian grid, how to dealt with land area inside
> domain? For example, in coastal area, there are islands inside the domain. How
> t o do this?

Land points are masked. We use the pressure method which does not have
difficulty
with islands, unlike the stream function method.

> 3) Do you use a set of governing equations for Cartesian, Curvilinear and
> spherical grid? I mean, such set of governing equations are written in
> orthogonal coordinates. What  do you think about
> such idea: set a submodel and use a set of  Cartesian basic equations for
> Cartesian grid,  set a submodel and use a set of equations transformed in
> curvilinear coodinates? This may make model much
> longer, but maybe easier to understand for users prefering Cartesian
> coordinates.

The model is written in orthogonal curvilinear coordinates of which geographic
and
Cartesian coordinates are special cases. There is no fundamental difference
between the equations in each coordinate system and so the same code can be
used in each. We do patch together different coordinate systems in a gridding
of the sphere generated by a conformal projection of a cube onto the sphere.
Each
face of the cube is a separate coordinate system.

> One suggestion: If possible, put papers of model developer on website, in ps
> or pdf type.

Model documentation and references are at

 http://mitgcm.org/sealion/

We'll be adding downloadable re-prints some time soon.

Thanks for the interest,

Alistair.