Open tommbendall opened 3 weeks ago
Doesn’t it depend on whether you are doing the traditional approximation?
Ah yes, so in that case maybe we should allow the compressible Euler equations to have a traditional_approximation
option to control this
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 ifdomain.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.