Institutions inside the Artsobservasjoner dataset

[dagendresen]

The Artsobservasjoner dataset (and "NBIC other ds") published by NBIC on GBIF includes data from many different institutions that are not credited as data publishers in GBIF.

https://artsdatabanken.no/artskart/bidragsytere?Key=1435226523

Split by institutionCode

	Institution code	Count
	nof	20,102,346	Norsk Ornitologisk Forening
	nbf	2,648,781	Norsk Botanisk Forening
	miljolare	1,161,319	Miljølære.no (Skolelaboratoriet i realfag ved UiB)
	nef	891,511	Norsk Entomologisk Forening
	miljodir	654,097	Miljødirektoratet
	nsnf	441,278	Norges sopp- og nyttevekstforbund
	nzf	216,155	Norsk zoologisk forening
	zmbn	2,783	(UiB Zoologisk museum??)
	gu	1,900	(Norges geologiske undersøkelse???)
	nibio	1,829	NIBIO
	dn	1,714	(Miljødirektoratet ?)
	o	1,505	(UiO NHM O ?)
	bg	644	(UiB Bergen museum ?)
	imr	507	Institute of Marine Research
	ntnu-vm	114	NTNU University Museum
	trom	81	UiT University Museum
	smnh	46	(Sveriges meteorologiska och hydrologiska Institut ???)
	nvi	32	(Norsk vitskapsindeks ???)
	trh	10	(NTNU University Museum ???)

[dagendresen]

Artskart has a list of all institutions

Institusjon	Antall observasjoner	Antall unike arter
Norsk Ornitologisk Forening	20 150 581	885
Norsk institutt for naturforskning	4 157 926	8 137
Naturhistorisk Museum - UiO	3 076 454	26 569
Norsk botanisk forening	2 651 902	6 583
NTNU-Vitenskapsmuseet	1 414 820	20 883
Miljølære.no	1 282 356	3 372
Norsk entomologisk forening	897 997	12 974
Miljødirektoratet	883 266	5 864
MUST	806 008	596
GBIF-noder utenfor Norge	575 791	15 392
Havforskningsinstituttet	451 528	2 948
Norges sopp- og nyttevekstforbund	441 966	5 030
Universitetsmuseet i Bergen, UiB	423 107	12 765
BioFokus	399 760	14 804
Tromsø museum - Universitetsmuseet	299 222	12 704
Norsk institutt for vannforskning	291 348	2 860
NOF-NINA-DN	238 935	261
Agder naturmuseum	222 579	2 764
Norsk zoologisk forening	217 282	407
JBJordal	177 768	2 500
MFU	122 930	2 377
Norges miljø- og biovitenskapelige universitet	113 383	2 773
Fjellstyrene	48 475	5
Molltax	22 803	428
Svalbardflora.net	21 242	214
Rådgivende Biologer AS	14 241	1 220
Ecofact	12 481	1 383
Helgeland Museum - Rana	11 447	906
Arne Fjellberg	10 013	303
Norsk Polarinstitutt	9 671	150
ARC	8 913	93
Arkeologisk Museum, UiS	7 979	1 204
Ento Consulting	4 800	186
Asplan Viak	3 907	599
Bioforsk	3 740	95
Nord Universitet	3 313	19
Norsk Natur Informasjon	3 112	565
Norsk institutt for bioøkonomi	2 736	701
Faun Naturforvaltning AS	2 324	659
Artsdatabanken	1 594	83
Sweco Norge AS	1 112	246
Naturrestaurering AS	777	249
Norges fiskerihøgskole, UiT	680	101
NINA FMNT	465	1
Multiconsult	308	29
Veterinærinstituttet	31	14

[dagendresen]

And some other possible "institutions"

Consultancy companies (private sector)

institution	approximate occurrences	notes
Akvaplan-niva	(ca 115 995 occurrences)
Miljøfaglig Utredning	(ca 36 703 occurrences)
Natur of Samfunn AS	(ca 7 mill occurrences)	including recorded by etc
COWI AS	(ca 289 516 occurrences)
...	...	...
...	...	...

Bird Stations

The NOF (Norwegian Ornithological Society) institution code includes numerous separate bird stations. In particular, Nordre Øyeren bird station has asked for better visibility under their own institution ID.

institution	approximate occurrences	notes
Nordre Øyeren Fuglestasjon	(ca 525 075 occurrences)	Note! registered in GBIF
Kragerø Ringmerkingsgruppe	(ca 853 557 occurrences)
Utsira Fuglestasjon	...	...
Jomfruland Fuglestasjon	...	...
...	...	...
...	...	...

[rukayaj]

Status of this task: We requested that NBIC split up the dataset (so that most of the bigger institutions at least are separate) a few years ago. Since then we've had several discussions and emails about it. The latest one with Knut Anders Hovstad on 2021-02-16:

This is an issue that we are working on and I hope we have solutions quite soon that can target the needs of consultants, organizations etc. that delivers data through Artsobservasjoner. I'll return to you with more information on this.

I volunteered to help them:

You could just make multiple datasets, and for each of them just modify the select statement from the View_Artsobservasjoner2​ to have a WHERE InstitutionCode = "nof", for example. You could even make a spreadsheet of the different institutions and write a script to automate the dataset creation process on the IPT. I haven't done anything like that but the IPT doesn't have a database so I am pretty sure it would just be a case of changing flat files. ​I'm happy to help figure out how it can be done, if that would be useful for you.

But we have not had a response to this.

Next action: I suggest we ask for a status update again in a month or so.

[rukayaj]

Just to keep this issue up to date:

\We received an email - GBIF: Notification of upcoming removal of your organization "Nordre Øyeren Bird Observatory" from the GBIF list of data publishers

Dag sent another email query: Hei Stein og Knut,

Hvordan er status og muligheten for å knytte datasett for fuglestasjoner slik som Nordre Øyeren direkte til egen registrering i GBIF?

(Ihvertfall bryte opp til egne datasett, dersom å splitte opp "publisher" er sensitivt).

[rukayaj]

Linked to #81

[dagendresen]

If we add Nordre Øyeren Bird Observatory (NØF, noef) and Biorehab Klepsland, etc to GRSciColl it should be possible to have the institution code from NBIC ArtsObs linked to the institution code in GRSciColl...

Sorry, just noticed that NBIC ArtsObs code nof must be the entire Norw Ornithological Society :-) --> so updated the institution code for Nordre Øyeren Bird Observatory to noef.

BUT could have worked if NBIC ArtsObs might have used the code noef for Nordre Øyeren Bird Observatory.

[rukayaj]

Do you think it's suitable to add non scientific collections institutions to GRSciColl?

[dagendresen]

I actually think that it makes little sense to include any institutions in the GRSciColl (because I do not believe in DwC triplets)

The original rationale for GRSciColl/GRBio was/is to maintain a list of institution codes and collections codes for the Darwin Core triplet (doi:10.3897/BDJ.4.e10293) - as is described in Darwin Core as a recommended option for building the occurrenceID

And the occurrenceID is the identifier for an Occurrence (notably NOT for identifying a specimen). From this perspective, it would thus make equal sense (or non-sense) to include institutions that hold/publish Occurrence data as for institutions that hold collections. Because the institution code managed by GRSciColl is used to identify Occurrence records, not collection specimens ;-)

[dagendresen]

If background is needed, see Guralnick et al. 2014 The trouble with triplets in biodiversity informatics: a data-driven case against current identifier practices, doi:10.1371/journal.pone.0114069

[dagendresen]

Nordre Øyeren Bird Observatory (Registry) is asking for support with finding their own records on GBIF.org.

They try to remake the following API search on GBIF.org https://api.gbif.org/v1/occurrence/search?datasetname=N%C3%98F-vannfugltellinger

They now have ROR ID https://ror.org/04t667177

Could maybe this ROR ID be published as institutionID through Artsobservasjoner...??

We could also remind NBIC on splitting the large Artsobservasjoner into more appropriate datasets?

[rukayaj]

We could also remind NBIC on splitting the large Artsobservasjoner into more appropriate datasets?

I'll ask Knut about the ROR ID/institiontID thing in the meeting. I'm not feeling very well today, caught a cold (not covid), but I'll also have a look at the datasetName/datasetID search thing later.

[rukayaj]

I asked about ROR/InstitutionID, K will ask Stein about it but is not very optimistic as it's a political issue, but he will try.

[dagendresen]

Thomas Sæther wrote 2024-04-10:

Jeg ble akkurat oppmerksom på at GBIF nå endelig har åpnet opp for strukturert søk på Dataset ID og Dataset Name. For eksempel: https://www.gbif.org/occurrence/download?advanced=1&dataset_id=218 Det betyr kanskje at vi kan få laget noen DOI'er til noen av prosjektene våre og på den måten få Publisher-siden vår på GBIF opp å stå igjen? Eventuelt legge til datasett på GRSciColl-siden vår: https://scientific-collections.gbif.org/institution/cf68b61e-a6e6-4483-a97b-1ad7eb21188b Hilsen Thomas, Nordre Øyeren Fuglestasjon

[dagendresen]

Unfortunately functionality in GBIF to filter for a datasetName or a datasetID is unfortunately not enabling us to make it a dataset with a DOI (from GBIF). When the datapoints are published through Artsobservasjoner we will still need for Artsdatabanken to actually publish the data records as a separate dataset! We can try again to convince Artsdatabanken, but they seem to have concluded on publishing Artsobservasjoner data credited to the biological societies (birds as NOF).

An alternative is for you to publish the data records outside of Artsobservasjoner directly to GBIF by yourself. But this would mean that we do not have any database such as Artsobservasjoner for you to manage your data records. And as I understand not what you want.

I have attached a figure 1 to show that the datasetID 218 is not sufficiently unique to identify your records from NØF. Notice that other institutions use the very same datasetID - nsw, kku, mma, uconn, ... The datasetID 218 is maybe generated (automatically) by Artsobservasjoner?

Your link to your dataset in the GRSciColl Scientific Collections data portal will unfortunately not work for your dataset which include species observations. This particular data portal (a so-called GBIF Hosted Portal) is setup to only display collections specimens. However, we could of course help you to setup a NØF hosted portal - provided we can find an appropriate filter (maybe eg. datasetID = 216 + institutionCode = nof, or similar).

[dagendresen]

Email to Artsdatabanken 2024-04-10:

Nordre Øyeren Fuglestasjon søker fortsatt muligheter for å krediteres i datasiteringer til egne dataposter publisert via Artsobservasjoner. Jeg tror den eneste muligheten per idag er dersom det er mulig å legge ut deres datasett som et eget separat datasett fra Artsdatabankens IPT, men forstår at dette fortsatt er litt politisk krevende? Jeg har forsøkt å overtale GBIF sekretariatet til å legge til funksjonalitet for å telle datasiteringer til dwc:institutionID = deres ROR ID på hver enkelt datapost. Jeg fikk også god støtte til dette ønsket i GBIF mid-term komiteemøte i forrige måned. Men det er selvsagt fortsatt noe som uansett bare ville komme inn på GBIF sekretariatets langsiktige arbeidsplan - som er eksperimentell pilot. (Men de innser at dette er et ønske fra flere). Det jeg umiddelbart funderer på er om det er mulig å legge inn dwc:institutionID = https://ror.org/04t667177 i publisering av Artsobservasjoner til GBIF?