Allow complex number data type

[shunt16]

From Zhav Loizeau in #1:

In interferometric SAR, I think the data should be stored as complex numbers, no? Possibly relevant

[shunt16]

NetCDF does not natively support complex numbers as a data type. However, you can store complex numbers in a NetCDF file by separating the real and imaginary parts into two variables or as a dimension in a single variable. My understanding is that this approach is taken for InSAR (correct me if I'm wrong though!). For simplicity, I would be inclined to recommend users take this approach to maintain adherence to NetCDF (and therefore close this issue).

Any thoughts otherwise?

[EmmaWoolliams]

I agree with Zhav that we need to describe how to do that, or how it is generally done. We'd definitely want to describe the covariance between two separate variables though as the real and imaginary parts almost certainly have error correlation!