Prepare for UGRID versions greater than 1.0

[sadielbartholomew]

The code presently registers and stores any valid version ID (1.0, 2.10.4, 6.2.3, etc.) for the UGRID Conventions, but only processes any UGRID features with code supporting v1.0. This is probably harmless for now with v1.0 being the latest and greatest version (and that included in CF for the present CF, v1.11), but as-is the behaviour will be mis-leading for subsquent UGRID versions upon release. To be ready in case datasets are around ready/compliant for new UGRID releases, we should at least warn that the code is assuming and processing features according to UGRID v1.0.

As discussed offline. To be fixed in #296.