Do not use watershed groups for upstream analysis

[smnorris]

Watershed codes / watershed groups / assessment watersheds are not quite consistent enough to use for pre-aggregated upstream summaries.

For example, looking for area upstream of this fundamental watershed:
watershed_feature_id = 8924583
wscode=100.190442.244975
localcode=100.190442.244975.541514

If we look upstream using the fundamental watersheds, almost of the Nicola River / NICL watershed group is included.
If we look upstream using assessment watersheds or watershed groups, the NICL watershed group is excluded because the watershed group is coded with local code 100.190442.244975.539934 - there is a tiny bit of river polygon with that code present in the group.

Do not use watershed groups (or assessment watersheds, that are also based on group boundaries) for upstream/downstream queries.

FWA_WatershedAtMeasure and fwa_watersheds_upstream_area.sql will have to be fixed.

[smnorris]

Also - because they are not consistent enough to produce results equivalent to using fundamental watersheds, dropping the watershed codes from the watershed groups might be a good idea.

[smnorris]

fixed for fwa_watersheds_upstream_area with 6f03ce6

[smnorris]

When watershed codes are removed from fwa_watershed_groups_poly :

• how does fwa_basins_poly get generated? Just copy/paste watershed group codes/ids manually into the query?
• is there any other convenient aggregation that could replace watershed groups in fwa_watershedatmeasure()?