Feature request: add opendap/netcdf authorization to backend_kwargs in open_[mf]dataset

This is a paraphrase of a SO Q: https://stackoverflow.com/questions/66178846/read-in-authorized-opendap-url-using-xarray

Is your feature request related to a problem? Please describe.
I would like to use xarray to read data stored at UCAR's Research Data Archive. To access data you have to create an account with a username and password.

Describe the solution you'd like

url = ""https://rda.ucar.edu/thredds/dodsC/files/g/ds084.1/2020/20200101/gfs.0p25.2020010100.f000.grib2"
ds = xr.open_dataset(url, backend_kwargs={"auth": ("username", "password")})

Describe alternatives you've considered
You can read this file by create dot files in your home directory (https://stackoverflow.com/a/66179413/6046019)

Additional context
The idea of "auth" comes from siphon (https://unidata.github.io/siphon/latest/examples/Basic_Usage.html#sphx-glr-examples-basic-usage-py).

Start by tracing how open_dataset and open_mfdataset pass backend_kwargs to the backend, then compare the requested auth tuple with Siphon's authorization approach and the linked Stack Overflow context. Done means an authorized OPeNDAP/NetCDF URL can be opened with backend_kwargs containing credentials, including through the requested dataset-opening interfaces.