province and sub-province don't conform to SWEET

"province" or "sub-province" terms are used in a way that doesn't conform to the SWEET definitions. The feature type for these provinces and sub-provinces may need to be changed to align with whatever generic spatial subdivision SWEET has in mind.

What is the SWEET definition of a Province and Sub-Province? I don’t think they really know, but you may correct me here.

Stephen last said re Provinces (and this was solely in relation to sedimentary basins) : “I don't think 'sedimentary basin' should be a kind of province. Brainstorming here (I take this to infer there is no clear definition of a Province). Distinction could be: a 'province' is a composite of various features related to a shared geologic/tectonic history in a connected (at the time of the defining history) region, not denoting any particular process. A sedimentary basin is an individual feature characterized by processes of subsidence and sedimentation. A sedimentary basin can be partOf a province, but a basin by itself would not be considered a province”.

The sedimentary basin system that comprises the Sydney, Gunnedah, Bowen, Galilee and Couper basins would fit his “brainstorming” definition. This system has a shared geologic/tectonic history, and, while not denoting any particular process, it comprises multiple processes, not just subsidence and sedimentation.

I was planning to propose to Stephen that we do not assign sedimentary basins to provinces, but that they have the same status as other provinces that we have identified.

“The feature type for these provinces and sub-provinces may need to be changed to align with whatever generic spatial subdivision SWEET has in mind”. <<< SWEET has not indicated what they have in mind re spatiality, but clearly we have to account for this.

I did some digging and it would seem the core issue here stems from the fact that the GA class of Province is an administrative feature (albeit with a very loose geological rationale for delineating provinces), while the attempts at a SWEET definition are clearly for a geological feature class.

We obviously can't persist the GA-imposed useage of province for our features AND conform to SWEET at the same time, nor do I think any proposed sweet definition of province is compatible with GA's definition of "The administratoriv unit 'province'" link

Most of the features in our database have a name that specifies genesis, such as the many features that have basin in the title (the geological feature type) but have a rank of province (GA's administrative feature type). For example, the unit number 254 - Bamaga Basin - Province should be of SWEET class:Basin. For the features that are named as provinces, we will need to propose a new term as you agree above. "Domain" might be a candidate as it has been used in the QLD Geology book for "provinces" in Mount Isa. Whatever we decide (as soon as possible), there are only a couple of hundred instances in CPF_STRUCTURAL_UNITS that would need a SWEET classification which is achievable in a couple of days at most.

For the ontology, we could perhaps link the appropriate SWEET geological feature classes, to [http://pid.geoscience.gov.au/def/voc/ga/featureofinteresttype/province] to show that these classes are considered to be an administrative FOI by GA. Then we could remove the administrative class of province from our geological feature dataset and implement a geological feature class of province with a true geological definition in SWEET.

I agree that the Early-Permian EARS (Province containing a few large basin features) and the Basin and Range Province (A province containing very many tiny graben "basins" and extensional structures) both are very similar features, but because our grabens are still buried beneath the basin, we call the whole thing a basin, whereas they can see and define each graben or half graben as a basin within the province, hence the push for Basin partOf Province. GA can uphold that a sweet basin is a subclass of their admin feature Province if they want to, while sweet can say that Basin/Graben isPartOf Province without conflating administrative and geological features.

I hope the lengthy text above makes sense because I feel like we just cracked the case and can reconcile GA against the world without any unhappy compromises.

@johnmckellar wrote: Thanks Derek. You certainly have gone to a lot of trouble with your digging and lengthy explanation (much appreciated ☺☺☺), and especially finding the link for the FOI type ‘Province’ as an admin unit. This surprised me and was not the opinion that I obtained from: https://www.ga.gov.au/scientific-topics/energy/province-sedimentary-basin-geology

Let’s get Nick’s input into this as well, seeing that he is the author of the admin unit province. My mind boggles a little as to what we may need to do with, e.g., the Mount Isa Province, which has been divided into 15 domains. We have the domains there, OK, but what do we then term the collection of them, if we don’t call them a province? Extrapolating all this this would require some major changes to our terminology, and this may not be an easy process, as provinces are well and truly embedded in the GoQ.

I don’t think it will be that difficult, for Mt Isa or for any of the other structural units we have in the DB. If we go sweet we will surely need to reclassify many of our featuresFrom a GA perspective, the Mt Isa Inlier is a province, with the various domains being sub-provinces. From a geological perspective, the Mt Isa Inlier is an inlier (surprise surprise), and the inlier has been broken up into a set of geological domains (where the definition of a domain is similar to the def of sub-province we were working towards in sweet.) For Mt Isa I would suggest that Inlier would indeed be a subclass of a geological province.
The part I don’t have absolute clarity on is the method for attributing each of the features with a geological AND administrative feature type. Possibly this is something GA, rather than GSQ would need to do. If we/GA can make the link between GA’s feature types and the SWEET ontology as properties of the classes, or if we implement a collection (either in sweet or GA) for what classes are a GA province I can see how our data can stay linked to GA and SWEET without making structural changes to the DB and/or feature data model. Anyway, despite the lack of a clear definition of Province from GA, I don't think adopting GA's peculiar usage of 'Province" should or even could be persisted, at least not for geological features. Even if we wanted to do this, I don't think we are able to make the GA usage of province match with SWEET or any other international standard because nobody else does it like that. If it is indeed too much work, maybe we need to have a business review/SWOT analysis of the decision to go SWEET over GA, and GA may need to be consulted as they are a major consumer/stakeholder here. I'm not sure if the benefits of conforming to SWEET over GA, unless the ultimate plan is to get GA on board with SWEET.

If GA uses Province as an Administrative thing, then there's nothing stopping you using Province, defined differently, as a geological thing. When you call X a geofeatures Province it just is that, it may or may not also be an admin Province.

Sure, it's not perfect having to similar things with the same names in different ontologies but you have the power of using one or both of the terms. The full IDs of each Province mean that they won't collide in any data (i.e. the URIs for geo & admin Province are different, even though the ID part of Province is the same).

Nick, in the geofeatures.ttl, where you have: The core of this ontology is a class hierarchy starting with \"geologic province\", as it is defined in SWEET, and extending down into multiple subclasses of it."""@en ;

Did you mean to write: The core of this ontology is a class hierarchy starting with \"geologic feature\", as it is defined in SWEET, and extending down into multiple subclasses of it."""@en ; ???

Because that is what the hierarchy starts with—and it is what lies at its core!?

And, I see, in SWEET: https://github.com/ESIPFed/sweet/search?q=geologic+features&unscoped_q=geologic+features

rdfs:label "SWEET Ontology Realm Geologic" ;

29 | dcterms:license https://creativecommons.org/publicdomain/zero/1.0/ ; … | 177 | skos:example "Individuals: Lake Eyre Basin, Sarmatian Craton, Victoria Point Sandbar, Mount Erebus Volcano. Subclasses: Basin, Craton, Shield, Province, Sub-Province."@en ; 178 | skos:scopeNote "Geologic Features include sedimentary basins, stratigraphic units, non-stratigraphic (lithodemic) units, stratigraphic event features, provinces, tectonic and structural features, georesource accumulations, and geologically significant sites, among others"@en ;

… that there is already much the same sort of set-up (in terms of features) as to what we are proposing/have proposed. But, nowhere (in SWEET) can I find a definition of soreag:GeologicProvince. This would help, if we can. But, as I said previously, with Stephen’s admission of ‘brainstorming’ provinces, I suspect that there may not be a definition. Nick, please advise if SWEET has a definition for ‘province’ that we can use.

had provisionally eliminated ‘province’ as a generic class/subclass, and this is what stands in the present version of the ontology. Thus, provinces (specified types of: mineral, igneous, etc.), basins, cratons, orogens, etc. are all regarded as tectonic entities and presently have equal subclass status (in our provisional ontology). This kind of got around regarding them all as provinces, including basins as ‘sedimentary basin provinces’.

Accordingly, we have not defined ‘provinces’ per se (although I did so in my initial spreadsheet, but this definition included basins). However, if we need to link to classes as they exist in MERLIN, I see that we need to keep the generic terms of province and subprovince. Derek, you have indicated that the heirarchy presented in GitHub (reproduced above) needs to be reformed to reflect the outcome of SWEET discussions. What exactly are you referring to here? The main stumbling block was the inclusion of basins in provinces, following not only GA, but also the USGS. Are you referring to anything else?

Also, you have indicated that the number of classes must be expanded (<<< please see my query below) to allow us to link all sites to features as they are linked in MERLIN. The list below is all of the classes in MERLIN: Feature types in MERLIN PROVIN		SUBPRO		SUPER		RIDGE		STRU		NOTQLD		MISCEL		OBSOLE		INLIER		IGNPRO		IGNSPR		BLOCK

Derek, in a nutshell, are you are asking me to give definitions for the above to Nick for him to add to the ontology? Also, you say that the list above represents ALL of the feature classes in MERLIN. As I do not have access to MERLIN, please advise if the current-state list (the preceding one above) also includes feature classes in MERLIN, and, thus, if you want features in both of the above lists to represent classes in GitHub?

I do not know what the following is: There is an attribute called Preferred_structural_unit_key that I’m not sure of the relevance. Perhaps @BlakePaul can advise. Yes, ask Paul Blake.

Cheers,

John

Did you mean to write: The core of this ontology is a class hierarchy starting with "geologic feature", ...

Yes! This is a missed word update. I'll fix it.

skos:scopeNote "Geologic Features include sedimentary basins, stratigraphic units,

That's right; I added that scope note to Geologic Feature to presage the work we're doing now.

...nowhere (in SWEET) can I find a definition of soreag:GeologicProvince

It's in the file realmGeol.ttl, line 182 but it currently has no definition, just:

soreag:GeologicProvince rdf:type owl:Class ;
                      rdfs:subClassOf soreag:GeologicFeature ;
                      rdfs:label "geologic province"@en .

so, our enhancement, as originally proposed is:

soreag:GeologicProvince rdf:type owl:Class ;
                        rdfs:subClassOf soreag:TectonicEntity ;
                        dcterms:contributor <http://linked.data.gov.au/def/gsq> ,
                                            <https://orcid.org/0000-0001-5489-9590> ,
                                            <https://orcid.org/0000-0002-8742-7730> ;
                        skos:definition "Geologic provinces are tectonic entities that include cratons/shields (in the continental realm), orogens, sedimentary basins, and tectonised/metamorphosed and/or mineralised regions, as well as large igneous provinces (LIPs). (Definition created here)"@en ;
                        skos:prefLabel "geologic province"@en .

So we've shifted it from being a subClassOf soreag:GeologicFeature to being a subClassOf soreag:TectonicEntity which is, in turn, a subClassOf soreag:GeologicFeature, supplied a definition, for the first time, and indicated that we created the definition here, for SWEET (as opposed to sourcing it from somewhere).

geological-survey-of-queensland / qldgeofeatures-dataset

province and sub-province don't conform to SWEET #1

rdfs:label "SWEET Ontology Realm Geologic" ;