Established patterns for modelling of species occurrence?

[ashleysommer]

@nicholascar and I have some ideas for proposing one or two new classes and properties in the TERN ontology to more cleanly model species occurrences, for use in the BDR.

Before we go ahead and write up the proposal, is there any example data or real production data that shows how TERN models individual incidental species occurrence, and/or systematic survey results of species occurrence (plants or animals)?

Partially related to #59 Partially related to #69

[smguru]

Hello @ashleysommer yes we have data which is completely animal occurrence in Ecoplots https://ecoplots.tern.org.au/search?dataset=wet-tropics-vertebrate#/

Occurrence is a feature type, and we manage it using a FoI and feature type. For example, animal occurrence and plant occurrence are all related to occurrence. We have added a survey class based on Nic's recommendation, which is another option. We call that class optional.

FYi @javiersan18 @nicholascar @junrongYu

[ashleysommer]

Thanks @smguru

Our conversations in the BDR have been around modelling an occurrence as something that happens at a specific time. So in SOSA an occurrence would be a subclass of an Observation or a Sampling. In current modelling design we have a Sampling of a geometry produces a Sample (the animal specimen), then an Observation of the specimen produces the scientific_name result. We are treating the Observation as the occurrence, but we are considering switching the perspective to consider the Sampling as the "occurrence" as it happens at a time, and produces a specimen.

I suppose it comes down to the question of; is an "occurrence" an event at a time and place, or a thing with properties?

Our intention with this conversation was to propose a subclass of tern:Sampling called, called something like:

• OccurrenceSampling
• OccurrenceEvent
• OccurrenceActivity

Then we could more cleanly model the species occurrence without forcing occurrence-specific properties and restrictions on a tern:Sampling.

[ashleysommer]

I have diagrammed the modelling of a single record of the wet-tropics-vertebrate dataset, showing the way TERN has done their modelling, with the Feature-of-interest centric view:

[ashleysommer]

This is how we are currently modelling it using Sampling as the Occurrence activity:

[smguru]

Hello @ashleysommer long time back we did a demo on occurrence data you can look at the example here https://github.com/ternaustralia/bdr-faealla-worked-example you can play with the data as well.

FYI @javiersan18 @junrongYu

[smguru]

@ashleysommer probably subclass is not required and I will keep this issue on freeze