NTR: site established for making environmental observations

[dr-shorthair]

I've been fossicking around ENVO trying to see if there is a class that could be used for geospatial locations where observations are made or samples taken. Nothing is really standing out. LTER sites, flux-tower locations, traps, etc are all related concepts. Is this a gap or am I missing something obvious?

[dr-shorthair]

locality description http://purl.obolibrary.org/obo/BCO_0000025 appears to be close. Is it expected that this could be established for repeated visits, with multiple processes occurring at discrete short times over a long total time interval?

[dr-shorthair]

However, I see that http://purl.obolibrary.org/obo/BCO_0000025 is the description since it is under http://purl.obolibrary.org/obo/IAO_0000030 (information content entity)

The thing in the world would probably be under http://purl.obolibrary.org/obo/BFO_0000029 (site)

[wdduncan]

Sites in BFO are immaterial entities. LTER sites sound like physical entities to me. But maybe I am misunderstanding?

[dr-shorthair]

I think the Site (which is a region of space) may be immaterial, while the equipment installed or located there is material. These are not the same thing. Is that how it works?

[wdduncan]

Yes. The site is the region of space. There has been some debate about how to best represent these kinds of situations. Consider, for instance, cities and states. Are these material entities, sites, both? Some of this, I think, depends on the context. When I say, "My home is in Buffalo, NY", I am obviously referring to Buffalo in the sense of a spatial region, and not the physical bits of earth that constitute Buffalo.

So, for the example of LTER sites, I think site makes the most sense. But, I'm not familiar with all the use cases you may have in mind.

[smrgeoinfo]

As I understand the 'site' concept, it is a location where one or more observations are made. The intention/utility is that another observer could re-occupy the same site and make a subsequent (for time-varying phenomenon) or duplicate (for static phenomonon) observation at the same site.
The problem is that 'location' can be specified relative to various reference systems. In a simple case, that might be a coordinate relative to a spatial reference system for the Earth (e.g. lat, long, elevation, WGS84, MSL2000). Location of a site might also be referenced to other spatial things: relative to 'current high tide level', 'center of river channel', '200 mb elevation in atmosphere'.. I agree that the 'site' is a 'region of space', with the caveat that it is distinguished from a generic 'region of space' by its relationship to a sampling or observation strategy to enable 'resampling/reobservation' at the same site.

[wdduncan]

@smrgeoinfo A BFO site is necessarily bounded by a material entity. So, an observational site would/could make sense as a subclass. We just need to work out what the right differentia would be. Perhaps something like:

observational site is a
  - site
  - and participates in some observational process

The participates in some observational process differentia is quite loose, but it may work for our purposes. The class observational process would also need to be added to envo, and its usefulness would depend on how we define it.

We could also delve into assigning observational sites particular kinds of roles. This may help provide more "bite" to the notion that observations are performed within observational sites (i.e., observational sites are not mere "participants").

Envo also has an observing system, and this might be further used to provide more differentia for some kinds of observation sites.

[smrgeoinfo]

I don't think an observation site is necessarily bounded by a material entity in the sense of the examples given for http://purl.obolibrary.org/obo/BFO_0000029, so the subclass is not appropriate. I'd suggest subclass from BFO spatial region

observational site is a
  - spatial region
  - and plays role in some observational process

(according to editor note: BFO 2 Reference: Spatial regions do not participate in processes.)

[wdduncan]

@smrgeoinfo Nice catch on the editor note regarding spatial regions and process!

In that case, I think the correct definition would be:

- spatial region
- contains process some observational process

contains process [http://purl.obolibrary.org/obo/BFO_0000067] is not showing up on Ontobee for some reason, but you can find in the Relation Ontology.

I think the two axioms (above) would only be necessary conditions at this point. It seem overly strong to make them necessary and sufficient (i.e., equivalent).

Also, observation process still needs to be defined.

I don't think an observation site is necessarily bounded by a material entity in the sense of the examples given for http://purl.obolibrary.org/obo/BFO_0000029

Can you give some examples?

[wdduncan]

Here is a definition for observing process to consider:

A process in which a person or machine sees or detects a material entity and selects it as worthy of observation, and which has as output an information content entity about the selected material entity.

It may need refinement, though, to fit this use case.

[dr-shorthair]

http://purl.obolibrary.org/obo/BCO_0000003 was essentially what I had in mind when I kicked off this issue.

This definition makes observing process essentially a subclass of sosa:Observation (for which the 'feature of interest' is not limited to material entities). Also perhaps a match to oboe:Measurement.

contains process can be seen at http://www.ontobee.org/ontology/ENVO?iri=http://purl.obolibrary.org/obo/BFO_0000067 but is not showing up in Ontobee for BFO or RO :-(

[smrgeoinfo]

@wdduncan -- the examples for site in BFO_0000029 are : "a hole in the interior of a portion of cheese; an air traffic control region defined in the airspace above an airport; a rabbit hole; your left nostril (a fiat part – the opening – of your left nasal cavity); the interior of a kangaroo pouch; the hold of a ship; the lumen of your gut; the interior of your bedroom; Manhattan Canyon); the interior of the trunk of your car; the interior of your office; the interior of your refrigerator; the Piazza San Marco; the Grand Canyon; the cockpit of an aircraft"

Many observation sites are simply point locations specified by a geospatial coordinate. For instance when I measure the orientation of bedding in some rock unit, I pick a location based on where I see the bedding clearly enough to measure-- there's not particular material boundary. When one is doing a gravity survey, observation sites are located along a traverse or grid. Each observation site is a lat-long point location with no material boundary that distinguishes it from other sites. An observation of the solar wind flux made by a satellite in space is made at some site located in an astronomical reference system but there is no material boundary for the observation site.

BFO sites are specifically dependent on some material entity that defines a boundary of the site. Observation sites have a location that is based (dependent) on some reference system, which you could argue must be anchored in some material entity, but it is not the boundary of the site.

[dr-shorthair]

Very clear explanation Steve. (You could have been a teacher ;-) )

[wdduncan]

@smrgeoinfo It depends on how you are understanding what it means for a region to be bounded by a site. The airport example, for instance, is not how many usually think of boundaries:

an air traffic control region defined in the airspace above an airport

As you note:

you could argue must be anchored in some material entity, but it is not the boundary of the site

But, in the airport example, it is airport (the physical entity) that "bounds" (or as you say "anchors") the air traffic control region. This is similar to things such as no-fly zones.