Differences in use between associatedOccurrences and associatedOrganisms

[EstebanMH-SiB]

We should like to know what are the different practices that nodes or data managers have with regard of this two “similar” darwinCore terms.

Our understanding is that associatedOccurrences applies when you have a relationship between two records that belongs to separated taxonomic groups, for example a host (mammal) and a parasite (bacteria), we also recommend using this field together with associatedTaxa, where you can put the nature of the relationship.

Meanwhile, associatedOrganisms applies when you have a relationship between two records that belongs to the same taxonomic group, for example a mother-son or a pack of animals recorded together.

We hope to open a discussion about the good practices documenting this fields and have several points of view, so we can do better recommendations to our data publishers.

As an addendum, we have somehow a similar situation with individualCount and organismQuantity, we use the former only when all the records belongs to individuals and the later when there are different types in the same dataset (e.g. %biomass, cel/ml, individuals, etc). But we want to be sure we are doing properly.

Thanks in advance for your answers.

[PaulGeerders]

Very important issue, especially with a view of recovery of data sets in the future.
Standardization of metadata terms and their meaning and usage, is essential, to be applied to the complete system.
This is where, in my opinion, GBIF could play an excellent role and give a valuable contribution with world-wide and long lasting benefits.
Paul Geerders.

[neilcobb]

For associatedTaxa we just put the scientific name but not necessarily the nature of the relationship

[TimvdStap]

Interesting, I’ll follow this discussion. The way I’ve been doing it is using both associatedTaxa and associatedOccurrence, i.e.:

associatedTaxa: “Pathogen found on” : “Coho salmon”
associatedOccurrence: Pathogen found on: [occurrenceID]

But so from my understand you would omit from including the nature of the relationship from the associatedOccurrence column?

[EstebanMH-SiB]

@neilcobb We always try to put the relationship, because sometimes is not obvius at all. In some cases when the publisher do not have that information, we left the scientificName alone. But we try to avoid that.

@TimvdStap Yes, we omit the relationsip in the column associatedOccurrence, mostly because is “duplicate” information. But it can be a good practice, maybe some people do not always check both fields, and is better to always have the relationship explicit.

[neilcobb]

I was wrong, we follow TDWG strictly and our system scan-bugs.org makes it easy for people to follow

Associated Occurrences (no relationship type, just a list of URLs): Darwin Core quick reference guide - Darwin Core
Associated Taxa (relationship type + taxon): Darwin Core quick reference guide - Darwin Core

[tuco]

I have created an issue (Change term - associatedOccurrences · Issue #324 · tdwg/dwc · GitHub) for the term dwc:associatedOccurrences, which is inconsistent between its definition and its placement in the Organism class, which happened in 2014 with the advent of the Organism class. It would be good to consolidate the discussion around dwc:associatedOccurrences in that issue.