Address entity-level converage inconsistencies about lter-core-metabase HOT 3 CLOSED

FWIW, I don't think it's that common to use entity level coverage either. But I do think it is useful to have a set of geographic coverages per dataset (otherwise with distributed datasets the bounding box gets huge). I generally don't bother splitting up the temporal coverage into subsets (i.e. for a dataset that was collected 1997 and 2001 I would just make the temporal coverage 1997-2001). I think(?) perhaps the entity sort order not actually meant to be datasetentities, but rather index 1 or more coverages within a dataset?

from lter-core-metabase.

I think(?) perhaps the entity sort order not actually meant to be datasetentities, but rather index 1 or more coverages within a dataset?

in earlier ports of metabase, we have used entity-level coverage, although it is somewhat uncommon (across the network, but still used by MCR). This table allows that, and this is what how entitysortorder is intended to be used.

from lter-core-metabase.

closing since we decided to keep the convention (0 in EntitySortOrder means dataset level coverage)
only remaining task is to add it to taxa. see #59

from lter-core-metabase.