variable resolution u,v,theta surface data

[marqh]

new fields file test data, with u, v and theta fields on an Arakawa C grid

[pelson]

This adds 17Mb to the test data. Ideally I'd like to strip this down to <1Mb unless absolutely necessary.

[esc24]

I agree, but I think @marqh has done what he can in this regard, given the tools available. Do you have any suggestions on how to reduce the filesize?

[pelson]

Do you have any suggestions on how to reduce the filesize?

That is the tricky part. It'd be nice if we could write fieldsfiles wouldn't it 😄.
Perhaps we can just unit test the appropriate bits instead of having a real file at all. I'll take a look at that option.