Include Omega fields needed for p* coordinate in initial conditions

[xylar]

This PR maps vertCoordMovementWeights to VertCoordMovementWeights in Omega and adds RefLayerThickness as the reference pseudo-thickness computed from restingThickness in Omega initial conditions.

We also add support for multiple iterations for computing pseudo-thickness from geometric thickness if the EOS is TEOS-10. The default is 6 iterations, which has been found to be sufficient to achieve machine precision in the horiz_press_grad test.

Checklist

• Testing comment in the PR documents testing used to verify the changes

[xylar]

While this should work now, it won't actually be useful until we bring in the p* vertical coordinate from Omega.

[xylar]

Seems to work in that I'm seeing bit-for-bit results with the omega_pr suite and I see the expected variables:

short_forward]$ ncdump -h init.nc 
...

	double PseudoThickness(time, NCells, NVertLayers) ;
		PseudoThickness:long_name = "pseudo-thickness" ;
		PseudoThickness:units = "m" ;
		PseudoThickness:note = "h_tilde = -dp / (RhoSw * g)" ;
	double RefLayerThickness(time, NCells, NVertLayers) ;
		RefLayerThickness:long_name = "pseudo-thickness" ;
		RefLayerThickness:units = "m" ;
		RefLayerThickness:note = "h_tilde = -dp / (RhoSw * g)" ;
...

	double SpecVol(time, NCells, NVertLayers) ;
		SpecVol:long_name = "specific volume" ;
		SpecVol:units = "m3 kg-1" ;
...

	double VertCoordMovementWeights(NVertLayers) ;
...

[cbegeman]

Are there plans to rename the Polaris variable RefPseudoThickness?

[cbegeman]

Seems that you have that in the pseudo-thickness rename PR. I'm cool with you merging this now and we can make that change after the pseudo-thickness PR is merged.

[xylar]

Yes, that's the plan.