(Assuming we get Dan's permission to use his data file in the test suite), all good.

The only thing I'd comment before saying merge away is that the sample data file Dan provided has all masked data (I guess from context it's a small subspace of a ocean field masked over land?) and for testing purposes it might be better to have it or a similar dataset with at least some unmasked elements (unless a fully-masked dataset is useful in itself for testing on, perhaps?). Otherwise we could assume something went wrong when using it for testing. But maybe I'm overthinking it...