Allow a UGRID grid to be used for FMRC

[drf5n]

There are a few workarounds to build FMRC-like functionality towards serving UGRID data:

1. create subsets externally and Aggregate them per http://www.unidata.ucar.edu/mailing_lists/archives/thredds/2013/msg00502.html

2. Add a degenerate coordinate dimension to make the datatype recognizable as dataType="grid" per UGRID with TDS and edal-java/ncwms #526 (comment)

3. Use the <logicalSection section="(0:24,:,:,:)"> in aggregations per Clarify logicalSection in docs / Use logicalSection to hyperslice subsets? #778 (comment)

So per #526 (comment) it looks like a real FMRC aggregation should be feasible.

[cofinoa]

can we have a use-case dataset? may be it would better start with just a "horizontal" ugrid?

[drf5n]

I've two sample files at http://web.vims.edu/~drf/NETCDF/ if that would help. @rsignell-usgs might be able to point at more.

[rsignell-usgs]

Here's another UGRID dataset:
http://www.smast.umassd.edu:8080/thredds/forecasts.html?dataset=gom3_nocache