Conservation & Fieldwork·Guide
Publishing to GBIF: endorsement first
A public Evidence page proves a sighting to a donor. A GBIF dataset puts the same fieldwork into the scientific literature under a citable DOI, credited to your organization. That second path is real, and it starts with an external gate ArtInStack cannot skip: your institution must be endorsed as a GBIF Publisher before anything you prepare can appear live on gbif.org.
What this guide decides
Four judgments before you chase a DOI
Institution, not individual
GBIF and its Participants publish data from organizations, networks, and societies. Solo photographers without an institutional Publisher affiliation hit a wall no platform can remove.
Endorse in parallel with prep
Registration and Node endorsement can take days. Preparing a Camtrap package does not wait on that clock, and should not be framed as if the platform is the bottleneck.
You own the data either connection
With Instant Connect, Delegated Access, or a prepared zip uploaded to your own IPT, the Publishing Organization stays yours. Connection mode is not a credit transfer.
Honest eligibility beats hopeful copy
Say who this is for early. Hiding the institutional bar is the same failure mode Evidence storytelling warns against: a claim the reader cannot actually complete.
Who this is for, and who it is not
This guide is for NGOs, labs, universities, networks, and other institutions that can be endorsed as a GBIF Publisher. If you photograph for such an organization, you prepare and publish under that organization’s Publisher account. Credit and DOI land on the institution’s name, which is the point of the network.
It is not a path for an individual conservation photographer to personally become a GBIF Publisher with no institutional affiliation. GBIF’s model does not endorse solo publishers the way a portfolio site endorses a creator. If that is your situation, keep publishing Evidence of Impact and Camtrap packages for partners; do not plan a personal gbif.org dataset as the next product step.
What a DOI-backed dataset buys you
Evidence of Impact is for visitors and donors: a redacted, narrative surface on your site. A GBIF dataset is for science infrastructure: occurrences discoverable in global search, citable via DOI, and attributed to your Publishing Organization, not to ArtInStack as the publisher of record.
You still keep the public Evidence page. The DOI path does not replace it. It answers a different ask: partners, funders, and researchers who need a stable citation and machine-readable harvest, not only a beautiful timeline.
Endorsement, demystified
Endorsement is GBIF’s process. A National Node (or equivalent Participant process) reviews and endorses your organization as a Publisher. That step is independent of ArtInStack. It can take real calendar time (often days, not minutes), so experienced publishers prepare data while endorsement is in flight.
Until endorsement completes, sandbox and test registries can prove packaging and crawl mechanics. They do not substitute for live Node endorsement on production gbif.org. Treat “we prepared a valid package” and “we are a live Publisher” as two different statuses, both honest.
Publisher, Installation, Organization (in plain language)
Newcomers stall on three Registry ideas that GBIF docs often leave abstract. Your Organization is endorsed as a Publisher: that is who owns and is credited for the data. An Installation is the technical system that hosts the harvest file (your IPT, or ArtInStack’s hosted Installation). A Dataset is registered under your organization and served from an installation’s endpoint.
Publishing organization and hosting organization can differ. You can own the science while someone else technically hosts the zip. ArtInStack’s product framing collapses that confusion to one sentence: your org owns the data either way; Instant Connect versus Delegated Access only changes how ArtInStack authenticates to your Publisher account when you Publish.
Open standards stay the same
GBIF harvest expects open biodiversity packaging (Camtrap DP / Darwin Core vocabulary), not homemade CSV columns. The why of that choice lives in Why open standards (Darwin Core). This page does not repeat the schema; it assumes you already speak that language in the survey workspace before you ask for a DOI.
Sensitive species: fuzz, withhold, and a different bar than Evidence
On Evidence of Impact, public redaction softens place labels and map precision for visitors. For GBIF, elevated IUCN taxa need an explicit harvest judgment: fuzz coordinates to a declared uncertainty the network can use, or withhold those records from the package entirely. A vague public label is not enough for the Registry’s coordinate expectations.
That judgment sits with the institution that owns the data. Softening for donors and preparing for science are related ethics, not the same control. Use Publishing sensitive species for the public-path reasoning, then apply fuzz-or-withhold when you Prepare for GBIF.
Instant Connect or Delegated Access
Prepare always starts the same way: validate the survey against harvest rules, build a Camtrap package, download it, and stage it for Publish. Your organization remains the Publisher. When you Publish from ArtInStack, Settings → Integrations → GBIF is where you choose how the platform connects to that Publisher account.
Instant Connect (BYO): enter your Publishing Organization key plus a GBIF service account username and password (or token). Self-serve; credentials stay vaulted for Publish. Choose this when your institution can store a service account in ArtInStack and wants Registry registration without waiting on a helpdesk grant.
Delegated Access: enter the organization key only. Ask your Node or GBIF helpdesk to grant editor rights on your org to ArtInStack’s platform username (shown on the Integrations card; copy it for the request). Use this only after that grant is confirmed; until then Publish will fail on rights. Choose it when governance prefers not to store GBIF credentials in a SaaS tool.
Neither mode transfers credit. DOI and publisher of record stay with your endorsed organization. If your institution already runs IPT and wants every technical hop inside your walls, Prepare → Download still gives you the same zip: create a new Camtrap DP resource in IPT by uploading that zip as an **archive** (not metadata-only), then use **Track IPT publish…** in the survey with your GBIF dataset key so status stays in sync. Publishing Organization stays yours. Details: Prepare and publish to GBIF.
An honest status callout
If your organization is not yet endorsed, say so on the program page and in partner materials. “Package prepared; Publisher endorsement in progress” is trustworthy. Claiming “On GBIF” before the Node finishes, or before crawl and index complete, is the same overclaim Evidence storytelling rejects for donors.
If you are an individual without an institutional Publisher, do not market a personal DOI path you cannot complete. Publish Evidence, share packages with the lab or NGO that can endorse, and keep the eligibility bar visible. Specificity about who can finish the journey is part of the product story, not a footnote.
Endorse, prepare, then connect
Start Node endorsement early, Prepare the Camtrap package in parallel, choose Instant Connect or Delegated Access in Integrations, and only claim a live DOI when GBIF has indexed under your organization.
Also in Guides: Evidence storytelling for skeptical donors · Why open standards (Darwin Core) · Publishing sensitive species
