Sort mesh indices before saving culled global ocean meshes

[xylar]

We need to create the graph file only after sorting has happened.

Checklist

• Document (in a comment titled Testing in this PR) any testing that was used to verify the changes

[xylar]

Tesitng

So far, the ecwisc30to60 test suite has run successfully (except for the failing validation in #331) through the dynamic_adjustment phase using this branch!

I'm still running files_for_e3sm but don't expect any surprises there.

[xylar]

Further testing

After updating to the released MPAS-Tools package in #662, I'm retesting through files_for_e3sm with the following meshes:

• IcoswISC480E3r1
• IcoswISC240E3r1
• IcoswISC120E3r1
• ECwISC30to60E3r1 - dynamic adjustment is failing due to Still seeing temperature extrema (likely from Redi) in mesh spin-up #331
• SOwISC12to60E3r1
• WCwISC14to60E3r1