NTR: environmental structure: new grouping above astronomical body part

[cmungall]

I would like to propose a new grouping in ENVO, to collect together discrete entities (material entities that are not systems or materials). I will discuss definitions and nomenclature below, but first to give you an idea here is the proposed hierarchy:

• material entity
  • astronomical object
  • environmental material
    • water, sediment, sand, magma
  • environmental region [NEW]
    • rhizosphere, cryopeg, atmosphere, epilimnion, etc
  • environmental structure [NEW]
    • astronomical body part (ABP)
      • oceans, seafloor, deserts, volcanoes, etc
    • atmospheric structure [NEW]
      • meteor [sensu meteorology]
        • cloud
        • cloud condensation nucleus

Motivation: the current material/ABP distinction reflects what I think is a really useful intuitive distinction between discrete structures (rivers, forests, islands, mountains, glaciers, etc) and and materials/substances (water, sand, leaf litter, ice). However, ABP is an unusual label, and it doesn't group all discrete structures, the top level of ENVO is a little ragged.

The proposed structure gives useful split between materials/substances and discrete entities. It reflects a division that has worked well in anatomy ontologies.

Proposed definitions and annotations

id: NEW name: environmental structure is_a: material entity synonyms:

• environmental object
• connected environmental object
• discrete environmental object
• discrete
• environmental feature [RELATED] def: A material entity that is a single connected structure with inherent 3D shape that is a discrete part of an astronomical body, or part of the atmosphere of an atmospheric body, excluding environmental systems, layers, and organisms examples: mountain, cloud, coral reef, city, archipelago, iceberg, grain of sand, pebble, marine sanctuary, air mass disjoint_from: environmental material disjoint_union_of: astronomical body part, atmospheric structure

id: NEW name: atmospheric structure is_a: environmental structure synonyms:

• atmospheric object
• atmospheric feature [RELATED] def: An environmental structure that is part of the atmosphere of an astronomical body examples: cloud, cloud nucleus

id: NEW name: environmental region is_a: material entity synonyms:

• zone or layer [RELATED] def: TBD

Note: driven by discussion in #963

[smrgeoinfo]

The distinction of regions and structures will be tricky.
the rhizosphere a cryopeg the atmosphere an epilimnion (this lake's epilimnion) an ocean (the marine hydrosphere) the seafloor a desert a volcano

are the 'the ...' things regions, and the 'a/an ...' things structures?

[pbuttigieg]

On first evaluation, I don't really see this as much of an improvement on feature: when does something become structured?

[cmungall]

@smrgeoinfo good point. There are always tricky subjective decisions involved in these upper ontology groupings, no matter what ontologists tell you!

I am not wedded to having this, but it was there to accommodate some things already in envo such as 'layer' as well as things under zones (currently categorized as immaterial entities which I think is not useful).

But to advance the idea further: I would say ocean, desert, volcano are discrete, even though like most things they may have fuzzy boundaries. regions and layers would be subdivisions of structures with relative continuity on two axes with transition (sharp or gradual) on another.

But if this isn't a useful way of organizing ENVO for domain scientists, I'm happy to lump this into 'environmental structure'

I do like having a structure/material distinction.

I think having some level beneath the often unintuitive BFO classes is useful

This is the current hierarchy (ENVO classes in bold):

[pbuttigieg]

New groupings are fine, but I can see no way in which "environmental structure" is semantically useful, rather than useful UX. If a group of curators need a different visual interface / browsing, that should be provided as an export (as is being attempted in #960) or they should create a filter on their side.

[smrgeoinfo]

@cmungall -- wading into DEEP water here! I like many aspects of the BFO foundation, but as you allude, the devil is in the details. There are many entities in the hierarchy you show that I'd be interested in clarifying. One example : is a cave entrance a site? There's a physical hole in the ground (speaking loosely) that most observers would agree on. Is that really the same kind of thing as an administrative boundary? (I don't think so...its invisible and a completely human construct).

My focus is on representing geology-- the structure and composition of the solid earth. For that purpose, my take is that the key distinctions are 'objects' (things with identity, unity, location, composition), 'material' (substance, has identity, no unity or location), and configurations (structures defined by arrangement of objects, has identity, parts). I'm wrestling with how to harmonize this view with schemes like BFO/ENVO or SWEET.

[pbuttigieg]

One example : is a cave entrance a site? There's a physical hole in the ground (speaking loosely) that most observers would agree on. Is that really the same kind of thing as an administrative boundary? (I don't think so...its invisible and a completely human construct).

In terms of BFO:site, it could be any spatial region anchored by (relativised by) one or more material entities. Administrative boundaries are indeed made by fiat, but there are at least some material things (rivers, roads, buildings) that relativise them. Are there counterexamples we can look at?

My focus is on representing geology-- the structure and composition of the solid earth. For that purpose, my take is that the key distinctions are 'objects' (things with identity, unity, location, composition), 'material' (substance, has identity, no unity or location), and configurations (structures defined by arrangement of objects, has identity, parts). I'm wrestling with how to harmonize this view with schemes like BFO/ENVO or SWEET.

Objects as you describe them can fit into BFO:object reasonably well, materials into the ENVO's environmental material. The configuration /structure thing is iffy to me: Is there some sort of generalisable criteria we can use to determine whether something is 'structured' or 'unstructured'?

[pbuttigieg]

I think this top-level class structuring should be the theme of the release after the next one

[cmungall]

@smrgeoinfo:

My focus is on representing geology-- the structure and composition of the solid earth. For that purpose, my take is that the key distinctions are 'objects' (things with identity, unity, location, composition), 'material' (substance, has identity, no unity or location), and configurations (structures defined by arrangement of objects, has identity, parts). I'm wrestling with how to harmonize this view with schemes like BFO/ENVO or SWEET.

I think proposal gets at what you are after with the first two distinctions, but it's always good to ground in examples. And also use cases for making these distinctions. We ontologists like to spend a lot of time with upper level distinctions, but frequently these have no bearing on any practical usage of the ontology, and don't contribute to ease of maintenance.

For your 3rd, I think this would come under environmental quality which is unaffected by this proposal.

is a cave entrance a site?

I have much to say about BFO:sites, most of it not favorable. But my proposal leaves this untouched, to keep things simple(r) and incremental. I would love to discuss this on a separate ticket (or maybe over a beer on zoom)

@pbuttigieg:

I think this top-level class structuring should be the theme of the release after the next one

+1000

[wdduncan]

Interesting discussion.
@cmungall You propose environmental structure as a type of material entity. Would it also have sites as parts?