Rework ecoregions hierarchy

[pbuttigieg]

Branching off #1556 and #1556 (comment)

With new efforts in the semantics of biogeography, we will revise the ecoregions class and associated instances.

[pbuttigieg]

See this comment for #1556 (comment), included below:

Comment from #1556

biogeographic realm =def. "An ecosystem which is surrounded by natural, macro- to megascale geographic barriers to species migration, and thus which comprises ecological communities with evolutionary histories largely delimited by those barriers."

This definition looks good to me. Would we then add Uvardy's realms (e.g. "Palearctic") as instances of biogeographic realm in the ontology?

Yes, likely as instances of "terrestrial biogeographic realm" (which we'll create alongside marine etc).

There is some minor disagreement between WWF & Uvardy, which could be captured in this way.

That's okay,- as long as the class definition holds true, we can have instances from different systems.

Linking directly to the WWF bioregion by realm description page.

Would we need terms for each of the levels? (see tree diagram):

• biogeographic realm

  • subrealm

    • bioregion

      • ecoregion

Yes, there already are ecoregions in ENVO (we should improve the definition), but bioregions and subrealms can be added. The subrealm notion bothers me a little - it again seems more of a convenience / bureaucratised grouping than anything ecological. Did you find any convincing differentia vs realm?

You can explore each of the links above to learn more details about bioregion groupings within the 52 habitable subrealms, the unique characteristics of all 185 bioregions, and the plant and animal assemblages contained within each of the 844 component ecoregions that make up One Earth’s Bioregions Framework.

It looks like they took the 844 component ecoregions from Dinerstein et al. 2017.

There's a full table in the supplement which may be useful for our purposes: SM 1 Table 1 14Mar17.xlsx

Interesting, worth analysing. Most are at the instance level, so should be easy to mint. We should cross-link those to any stable URLs that WWF provides for them.

The IUCN places its ecosystem functional groups as coming between bioregion and ecoregion in terms of scale. They reference Dinerstein:

Ecoregionalisations (e.g. Spalding et al., 2007; Abell et al., 2008; Dinerstein et al., 2017) serve as simple and accessible proxies for biotic composition based on biogeographic boundaries and have recently been shown to delineate biodiversity patterns effectively, at least on land (Smith et al., 2018).

Perhaps we should then use the Dinerstein et al material more directly.

I think the proposition of the IUCN's functional groups is that these functional arrangements can (and tend to) occur in various areas depending on micro-scale conditions (sensu landscape ecology).

Sure, but this is true of many things, so it doesn't make a good ontological class - we have to find the differntia that allow us to separate these classes from all others.

Is it possible that a way forward is defining classes for subrealm, bioregion, ecosystem functional group, and ecoregion, then creating instances of these classes that correspond to terms in the WWF and IUCN taxonomies?

Yes for all except the ecosystem functional group - we need a convincing definition of EFG that isn't also true for other classes.

[timalamenciak]

Thinking through some proposed definitions from #1556 (specifically), I am starting to think the approach might be to agree on a semantic structure for the classifications themselves, using their considerations as the differentia. In this way we can create an ontology of ecosystem classifications.

biogeographic realm - A division which considers geophysical barriers
bioregion - A division which considers geophysical barriers AND climatic conditions
ecoregion - A division which considers distinct species assemblage AND geophysical barriers AND climatic conditions

Some terms could move to subclasses of one of these: e.g. ecological corridor, ecological province, ecosystem fragment would all be subclasses of ecoregion as they all consider all three differentia. Similarly, ecosystem functional group could be minted as a subclass of ecoregion that considers the functional traits of species (we also need a term for functional trait).

There are a whole bunch of terms that are current subclasses of ecosystem that should perhaps become instances in this new conceptualization. Non-exhaustive list:

• agricultural ecosystem and subclasses
• alpine ecosystem
• Calyptra habitat