search

firedrakeproject / gusto_case_studies

A repository for Gusto case studies
1 stars 0 forks source link

Is Coriolis force correct for baroclinic wave in a channel? #25

Open tommbendall opened 3 weeks ago

tommbendall commented 3 weeks ago

I'm not sure if the Coriolis force is correct for the baroclinic wave in a channel ...

The specified Omega vector was coriolis = omega*sin(phi0)*domain.k. But since the domain is meant to be a channel centred on a latitude of 45 degrees, I don't know if domain.k is correct? This implies a rotation that points vertically upwards from the rectangular surface, whereas I think we want to replicate the direction of the Earth's axis by having something like: coriolis ~ (0, omega*cos(phi0), omega*sin(phi0))?

Adding the contribution to the y-direction should give a contribution to the vertical velocity equation.

This would actually be harder to implement in Gusto now, as we have hard coded that the rotation vector is (0, 0, Omega) ...

Also, the paper allows f-plane and beta-plane investigations, so the case study could be enhanced to provide this option.

colinjcotter commented 3 weeks ago

Doesn’t it depend on whether you are doing the traditional approximation?

tommbendall commented 3 weeks ago

Ah yes, so in that case maybe we should allow the compressible Euler equations to have a traditional_approximation option to control this