indexing multi-geometry objects

[ksonda]

This is philosophical.

Many conceptual features like regulatory ones are associated with multiple spatial locations and NHD segments that a data system does not really keep track of individually. For example, in this map I have taken the geometries associated with a single EPA project in the reach address database.. Note how it includes 11 nhdpv2 reachcodes which EPA judges to be relevant to 7 polygons where projects associated with a grant took place. Sometimes there are points which represent project locations that are not necessarily on a flowline, at least a medium res flowline.

I'm wondering if there can be a best practice for this kind of feature. Is this something that should be in the NLDI at all?

Maybe it should only be the reachcodes, where the id is just the reachcode, but the uri can be an identifier for the project (and thus duplicated) and then ideally the project has geoconnex landing content that includes the other geometries as different representations of the project?

[dblodgett-usgs]

My knee jerk here is to register each hydrologic location of the project individually but have each point to the same project landing page where you would see all the parts as a whole. The NLDI is just indexing hydrologic locations with associated information, so we have a lot of flexibility there.

The project would then need to hold information describing a list of hydrologic locations that it is associated with.

Another (maybe easier) way to deal with this would be to just use the most downstream hydrologic location associated with the collection as the representative one. This would only be valid for collections that are within a single drainage basin, but for many cases, it would work fine.