cm1 datasets not getting fields correctly

[matthewturk]

Something is up with how cm1 datasets are working. I'm using an example dataset from @leighorf, but the behavior I'm seeing is that when I define a derived field:

def vort_mag(field, data):
    return (data["cm1", "xvort"]**2 + data["cm1", "yvort"]**2)**0.5

The second field, yvort, is all zeros. This is not the case elsewhere, if I read yvort directly from disk. More detail:

https://gist.github.com/matthewturk/127c1ec0d7b9e09f7a228a96cf9f98d4

I think this is specific to the cm1 reader, but I am going to check.