Out of date docstring for concat_dim in open_mfdataset

In the open_mfdataset docstring:

    concat_dim : str, or list of str, DataArray, Index or None, optional
        Dimensions to concatenate files along.  You only
        need to provide this argument if any of the dimensions along which you
        want to concatenate is not a dimension in the original datasets, e.g.,
        if you want to stack a collection of 2D arrays along a third dimension.
        ...

This is true for the default combine='_old_auto', but when using combine='nested' it is required while it is not used by combine='by_coords'. It would be clearer to make that distinction here.

Update the open_mfdataset docstring in xarray/backends/api.py, around the concat_dim section linked in the issue. Read how the documentation currently describes combine='_old_auto', combine='nested', and combine='by_coords', then revise the text to distinguish when concat_dim is required or unused. Done means the docstring accurately reflects all three combine modes.