Xylar Asay-Davis

Okay, thanks @darincomeau. Nothing new, then. That's presumably good news, and we need to find a solution in the physics and parameterizations, not the mesh.

Thank @jonbob. I'll want to see the comparison with the 3rd run before making any assessment.

@vanroekel, for context, I had requested the 3 runs just so we can be sure what the relative effects of this change are compared with turning on submesoscale. @jonbob agreed...

@jonbob, I'll try to find out why the hovmoller plots didn't work but I don't think that should hold up this merge.

@jonbob, just so you know, I found the cause of the MPAS-Analysis issue you had and have a fix here: https://github.com/MPAS-Dev/MPAS-Analysis/pull/900 It is unrelated to this PR (except that it...

@vanroekel, yep, looks good!

I don't feel qualified to review. @cbegeman may wish to.

I would be willing to do compass testing, though, if you would like.

License files have been recovered from: https://github.com/talkiq/gcloud-aio/tree/7793ee034ab335da85fabfdc5c3169587e553e87 They were deleted in: https://github.com/talkiq/gcloud-aio/commit/bee30d049e7d52c7d859184987e0ebe9f425104a as reported in: https://github.com/talkiq/gcloud-aio/issues/517

Several of these dependencies are needed by https://github.com/conda-forge/apache-airflow-providers-google-feedstock/pull/20