site vs. geographical entity

[ddooley]

I was trying to compose a class of animal related outbreak locations:

And was reminded that a farm is a site whereas the other items are not, leading to unsatisfiability. I'm wondering if all items in this list belong in site, or to anthropogenic geographic feature instead?

d.

[smrgeoinfo]

Looks to me like 'exposure event location' should be an objectProperty with range 'Location', and domain something like 'infection outbreak', which would be an Event.

[ddooley]

The above issue is about how entities for 'agricultural fairground, farm, and petting zoo would seem to be of the same class, but at moment cannot be made so without unsatisfiability under BFO. A location class/domain can't have those three items in it as one is a "site" whereas the other two are material entities. Unless "location" class sits up above site and material entity, under "independent continuant". Re. objectProperties, my own modelling preference is to have as few of those as possible, and instead attach semantic relations to say 'outbreak animal exposure event location' itself. I would model "outbreak event" with 'has location' some 'animal outbreak exposure event location' which is admittedly a bit longwinded. (along with has start, has end, has pathogen etc.)

[pbuttigieg]

I tried to deconvolute the area where farming occurs from the actual physical entity (mostly due to influence from geospatial semantics). Subsequent conversations have convinced me this was overkill - why would we talk about the site separately from its physical instantiation? There are some exceptions (e.g. where an area is "declared" something, despite not really being that thing), but those should be handled separately.

I'll move farm into a material entity branch to prevent this.

[pbuttigieg]

The textual definition (sourced from USGS way back) cites an area of land; however, I think this was not thought about with ontologies in mind.

I won't obsolete the term when I move it, as the definition hasn't been through OBOfication. It will be in the agricultural feature hierarchy

[ddooley]

Sounds good!

[smrgeoinfo]

farm subclass of 'agricultural feature' makes more sense and is consistent with other subclasses there. Seems like pasture should be there as well. I figure there's probably been plenty of discussion on this I don't know about, but shouldn't 'cropland biome' be partOf farm, not vice versa? The farm my father grew up on was not all cropland--there was a big house, a yard, big barns and granaries, a big machine shop/garage for maintaining equipment, chicken houses, as well as pasture land, and orchards. What is a farm is defined by functional intention (produce food from primary sources)-- it might include crops, pasture, rangeland, as well as living space for people (houses) or animals (barns...), storage facilities....

[cmungall]

I think this makes a lot of sense. The move seems justified to me.

On Wed, Apr 29, 2020, 10:46 Pier Luigi Buttigieg notifications@github.com wrote:

I tried to deconvolute the area where farming occurs from the actual physical entity (mostly due to influence from geospatial semantics). Subsequent conversations have convinced me this was overkill - why would we talk about the site separately from its physical instantiation? There are some exceptions (e.g. where an area is "declared" something, despite not really being that thing), but those should be handled separately.

I'll move farm into a material entity branch to prevent this.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/EnvironmentOntology/envo/issues/931#issuecomment-621363058, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMOKVUZDIRJKDK25NOJTRPBRWLANCNFSM4MQOR7KA .

[pbuttigieg]

farm subclass of 'agricultural feature' makes more sense and is consistent with other subclasses there. Seems like pasture should be there as well.

To fit more with its definition, I'm moving this under "grassland ecosystem" and adding a 'participates in' some 'agricultural process' axiom to it. I've added the same axiom to agricultural feature so they're linkable.

Keep in mind however all "features" will be relabelled and may be moved around a little as the semantics of "feature" are fraught. The IRIs and the practical meaning will stay the same though, as this is just coping with lingo.

I figure there's probably been plenty of discussion on this I don't know about, but shouldn't 'cropland biome' be partOf farm, not vice versa? The farm my father grew up on was not all cropland--there was a big house, a yard, big barns and granaries, a big machine shop/garage for maintaining equipment, chicken houses, as well as pasture land, and orchards. What is a farm is defined by functional intention (produce food from primary sources)-- it might include crops, pasture, rangeland, as well as living space for people (houses) or animals (barns...), storage facilities....

The scale of the Ellis et al anthropogenic biomes is more synoptic, so it would include all of those systems at a finer grain as long as the larger ecosystem they nest in is primarily used for crop cultivation.

[pbuttigieg]

Frankly, these features are all ecosystems, so perhaps that would be the best way forward: Move the ag feature hierarchy under an ecosystem type and use the agricultural process axiom as an equivalence axiom to autopopulate it with strays like pasture

[pbuttigieg]

agricultural feature has been converted into agricultural ecosystem with an equivalence axiom:

ecosystem and 
('participates in' some 'agricultural process') or 
('formed as result of' some 'agricultural land conversion process')

Which neatly ties it in to some sustainable development semantics.

pasture is now reasoned into that hierarchy.

Thoughts?

[dr-shorthair]

(I'd like to see 'paddock' somewhere ;-) ) https://www.macquariedictionary.com.au/blog/article/472/