Recommended terms for a DwC freshwater extension

[CecSve]

feedback for: https://docs.gbif-uat.org/freshwater-data-publishing-guide/en/index.en.html#freshwater-dwc-extension-terms

Most, if not all, of these terms are needed by a broader community than just the freshwater community and I would propose that the title instead should be Recommended terms to be included in DwC and that the freshwater* part is excluded from the term name:

• Biome (as mentioned in other comments, this is already a proposed term in DwC)
  • LakeZone and RiverMesohabitat could just be subcategories/children of the Biome concept
• EcosystemFunctionalGroup
• TypeOfContribution
• TypeoOfObservation
• OrganismGroup (this could be aggregated by the end user based on the provided taxonomy in the record and may be redundant)
• Season (this could easily be calculated by the user based on eventDate and may be redundant)
• SamplingEquipment
• SampleProcessing

The bold and italicized concepts are especially ones where a broader community could benefit from the addition of the terms to DwC, as I see it.

@ManonGros may have further comments

[ManonGros]

I agree. I think that recommending the use of a Freshwater-specific extension makes little sense. As pointed out above, many of these terms are relevant to the larger community but could benefit from using an inclusive (freshwater) vocabulary.

As mentioned here #27 I don't think the OrganismGroup makes sense in the context of occurrences and as pointed our above, the information can be deduced from the taxa and Lifestage.

TypeOfContribution and TypeoOfObservation also seem very arbitrary. As discussed here: #12 these categories are also sometimes very difficult to distinguish, plus I would argue that working on improving existing Darwin Core fields like the basis of record would make more sense.

I general, I suspect that a large part of the Freshwater data available on GBIF are part of other, larger datasets which won't likely be able to use freshwater-specific extensions. This means that if users want to get all of the freshwater data on GBIF, they can't rely on a freshwater-specific extension and the relevant information might be scattered across many different fields.

I think that the community would benefit from improving the existing fields or designing a more generic extension with vocabularies that accommodate freshwater data rather than have a specific extension.

[DeboraArlt]

I also agree. I had exactly the same thoughts - that recommending the use of a Freshwater-specific extension makes little sense when many of the terms are relevant to the larger community.
As mentioned in #12 by dschigl about guides - the same applies to extensions: it is quickly becoming problematic to produce separate extentions that are framed as specific to certain data types while such extentions could bu much more useful for the larger community when designed and framed more general, to be able to use for more data types.

[jenlento]

We agree that some of the freshwater extension terms could be more general, as they apply to the broader community.
Biome, EcosystemFunctionalGroup, TypeOfContribution, TypeoOfObservation, Season, SamplingEquipment, SampleProcessing could all be general terms without freshwater at the start. LakeZone and RiverMesohabitat could be subcategories/children of the Biome concept, as suggested.

OrganismGroup should remain as freshwaterOrganismGroup. There are several comments in this thread about end users aggregating this information, or about this information not being necessary. However, this is something that is extremely difficult for many of the freshwater organism groups. We know this both from our own experience and from feedback from users who found GBIF too difficult to use for freshwater because of the lack of such information. The issue arises because those working on freshwater often consider the entire assemblage of, for example, algae from the benthic habitat (including several classes of algae as well as hundreds of species within each class), benthic macroinvertebrates (including multiple phyla, classes, orders, as well as only particular lifestages for those organisms that live outside the water as adults), macrophytes (including species that may be found in freshwater or in freshwater-adjacent habitats, thus adding ambiguity). It's difficult to convey how much of a barrier this can be to using GBIF, unless you have tried it. For example, I led a chapter of the recently-published book Rivers of North America Volume 2, and I tried to rely on GBIF to help put together checklists of freshwater taxa in particular rivers. Wading through the lists of species and trying to determine which were actually freshwater species, and not knowing which observations were in freshwater or terrestrial habitats made this an extremely difficult task.

@kcopas Please add this sentence to the start of section 3.1.1: "Most terms that we suggest are urgently needed for other realms as well."
Please remove "freshwater" from the terms Biome, EcosystemFunctionalGroup, TypeOfContribution, TypeoOfObservation, Season, SamplingEquipment, SampleProcessing in section 3.1.1

[jenlento]

TypeOfContribution and TypeoOfObservation also seem very arbitrary. As discussed here: #12 these categories are also sometimes very difficult to distinguish, plus I would argue that working on improving existing Darwin Core fields like the basis of record would make more sense.

We would argue that these fields are absolutely necessary for any meta-analysis of freshwater data (or other data for that matter). TypeofContribution speaks to the reliability of data and the source of data, which is important for understanding limitations of the data as well as any erroneous observations. Although the distinction between community-based and citizen science data may at first be difficult, we have tried to provide a very clear definition to distinguish these two. TypeofObservation has implications for which data can be combined and how the data can be analyzed to calculate biodiversity indicators. For example, there are different approaches to fish sampling, with some sampling being extremely targeted, sampling a small number of species, and some sampling focusing on the entire fish assemblage. These two sets of data would suggest large differences in biodiversity in the same region, but these differences are an artifact of the type of observation that was made. To not include this field would perpetuate biased meta-analyses of biodiversity patterns.

[jenlento]

Change made

[CecSve]

Could:
typeOFObservation be captured in http://rs.tdwg.org/eco/terms/inventoryTypes? or http://rs.tdwg.org/eco/terms/inventoryTypes?
typeOfContribution be captured in http://rs.tdwg.org/eco/terms/compilationSourceTypes?