Use Separate geostrophic thresholds for h and norm. vel.

[xylar]

We want to have a higher threshold on velocity convergence since it is better behaved than water-column thickness.

Checklist

• User's Guide has been updated
• Testing comment in the PR documents testing used to verify the changes

[xylar]

Testing

I ran the ocean/spherical/qu/geostrophic test case on Chrysalis, once with a normal velocity tolerance that was too strict (1.8) and once with the current one that worked (1.3). In the latter case, the analysis test passed, whereas it failed as it should in the former case with the expected error message:

Order of convergence for water-column thickness: 0.402
Order of convergence for normal velocity: 1.352
Error: order of convergence for normal velocity
  1.352 < min tolerance 1.8
          execution:        ERROR
Exception raised while running the steps of the task
...

[xylar]

@cbegeman, I saw how you handled different tolerances for different fields in sphere_transport and wanted to fix that here as well. A quick review by inspection is all that is needed here.

[cbegeman]

@xylar Looks great!

[xylar]

Thanks @cbegeman!