EDUKATE SINGAPORE · GOLD STANDARD PRACTICAL CASE STUDIES · CASE 57
An accurate biodiversity map can do real harm when every location is published without checking its sensitivity.
A volunteer nature survey celebrates a public map that includes every observation, date and precise location. It is excellent for transparency, but several records reveal where vulnerable birds are nesting. The team's success at sharing data has created a risk to the very wildlife it wants to protect. This case follows the management decisions between collecting good scientific evidence and publishing it responsibly.
Every survey, species, observation and operational decision here is fictional. No precise wildlife location, real nest site or directions to sensitive species are supplied. Public references describe biodiversity data stewardship, not site-specific wildlife regulations or proof that all occurrence data should be restricted.
Chapter index
- 1. The map is launched to applause
- 2. Data quality and publication suitability are separate tests
- 3. The hidden location can still leak through other fields
- 4. What GBIF and iNaturalist actually recommend
- 5. Protect the affected record without destroying scientific history
- 6. Decide when openness supports conservation and when it undermines it
- 7. Make the public dataset explain its own limitations
- 8. Test the entire release process, not just the map
- 9. Trust grows when the team can explain the corrected decision
- 10. Practice: four sensitive records among 240 observations
1. The map is launched to applause
A fictional volunteer network has gathered 240 biodiversity observations in a suburban nature corridor. Students help identify broad habitat types while experienced volunteers check the records. A publicly accessible map allows visitors to click any point to see the observation date, image and location. The dashboard states that all 240 verified records are visible.
During a presentation, a teacher notes that the map appears complete. Visitors can see the survey route and compare sightings across months. For most ordinary sightings, this is the sort of transparency that supports community science: records become useful to educators, planners and people who want to appreciate living systems.
Then a conservation volunteer notices that four records concern an active nesting location regarded by the project's specialist adviser as sensitive. The map has not simply displayed that the species occurs in the area. It has provided enough precise information, including photographs with recognisable surroundings, to narrow down an area that should not have been publicised so exactly. The four records are accurate. Their publication is the failure.
2. Data quality and publication suitability are separate tests
A biological record has scientific value when its identity, observation evidence and associated context can be assessed. Publishing that record without alteration is one possible use, not the only acceptable use. The collector's detailed original may be appropriate for authorised conservation assessment while an openly accessible version needs less precise geography.
Data quality asks whether the observation is authentic and correctly described. Publication suitability asks who should see which fields, how the information might be used, what harms are reasonably foreseeable and which approved restrictions apply. Passing the first test does not automatically pass the second.
Consequently, 'all records visible' is an inappropriate Gold Standard metric without a disclosure review. An ethical dataset can be complete for authorised researchers while intentionally generalising or withholding selected public details. The system must preserve the distinction instead of treating openness and completeness as identical.
3. The hidden location can still leak through other fields
Removing one coordinate from a public map may be insufficient. A photograph can contain recognisable landmarks. Free-text descriptions can mention a precise feature. A URL or cached map tile can retain location clues even after the visible marker is moved. Photographs and supporting files may also contain metadata that needs checking before release.
The fictional team learns that its location policy was applied only to the latitude and longitude columns. The location described in image captions was never reviewed, and a small number of exports had already been downloaded by other users. A tidy public map now conceals some information that remains visible elsewhere.
The Gold Standard lesson is that sensitive information flows across a publication package, not just one spreadsheet cell. Reviewers should inspect how different fields and products combine. They must also recognise the limits of remediation once data are copied and avoid making promises that a prior release can be completely recalled.
4. What GBIF and iNaturalist actually recommend
The Global Biodiversity Information Facility publishes best-practice guidance for generalising sensitive species occurrence information. It recognises that precise records can facilitate disturbance or exploitation in some circumstances, and recommends documented, risk-proportionate controls rather than an indiscriminate ban on sharing biodiversity data.
GBIF's guide to publishing survey data discusses fields such as dataGeneralizations and informationWithheld, which allow a publisher to communicate that data were generalised or deliberately withheld. iNaturalist's geoprivacy guidance likewise explains obscured and private locations, while warning that landmarks or descriptions may still reveal clues.
These are real reference points for data governance, not a universal law requiring every bird location to be secret. Whether a particular record merits restriction depends on ecological sensitivity, local conditions, competent expert advice and applicable institutional rules. The public explanation should be clear enough to preserve trust without undoing the restriction itself.
5. Protect the affected record without destroying scientific history
The programme lead temporarily suspends the exposed map view and the affected public downloads while the conservation adviser assesses what needs immediate restriction. A public correction can explain that the location presentation is under review without repeating the precise location. The team avoids circulating screenshots of the sensitive points simply to demonstrate its problem.
Authorised staff preserve the original observations in a controlled environment, along with a traceable record of what was published and when. Those source records may be needed for legitimate ecological work and to understand the disclosure. They should not be casually erased in a panicked attempt to remove evidence.
The lead assigns a clear owner for contacting any partner organisations that received the affected export and for assessing whether wider incident-response or reporting obligations apply. No volunteer is told to pressure people to disclose or surrender unrelated private information. The repair begins with containment, expert judgement and accurate communication.
6. Decide when openness supports conservation and when it undermines it
The network designs a simple decision structure. Ordinary, low-risk observations may remain public at useful precision. Potentially sensitive observations are reviewed by the nominated ecological adviser and data steward. For higher-risk information, the public record can retain species-level or broad habitat information without revealing the exact location, where that approach remains scientifically appropriate.
Research use of precise coordinates should follow a justified, authorised route with defined purpose, access, protection and any necessary agreements. Generalisation should not be so sweeping that it makes all data scientifically meaningless, nor so permissive that a reader can reconstruct a vulnerable location.
An important governance question is who decides sensitivity. A communications volunteer should not make that decision from the map's popularity. The conservation lead needs authority to assess ecological risk, and the data custodian must translate the decision consistently into the published product. Responsibility cannot rest with the last person to click 'export'.
7. Make the public dataset explain its own limitations
The revised data dictionary differentiates the precise source location, the approved public location representation and the reason for any generalisation. The public dataset includes a transparent description of the modification without disclosing the underlying sensitive point. A broad mapped area must not be represented as an exact sighting coordinate.
Good records also preserve taxonomic uncertainty, dates, review status and source references. If an observation cannot yet be verified, it should not be silently presented as confirmed merely to maintain a full-looking map. Similarly, removing the most sensitive observations from a public layer must not be misrepresented as evidence that those species are absent.
GS-R08 concerns the integrity of evidence, while GS-R10 concerns learning from review. The Gold Standard is achieved through faithful public meaning and controlled access, not by a larger count of dots on a screen.
8. Test the entire release process, not just the map
A new publication test uses fictional training records: one ordinary sighting, one sensitive nesting observation, one photo with a recognisable landmark, one exported spreadsheet and one previously approved public record whose sensitivity is later revised. The expected treatment is documented before the team runs its checks.
The publication reviewer confirms that approved public fields are consistent in the map, downloads, captions, thumbnail previews and partner data feeds. If any preview can reconstruct the prohibited detail, the release fails despite the main map looking safe. The team also checks how to withdraw, supersede or flag old material when a protection decision changes.
These tests do not prove that sensitive data can never be inferred. They show whether the designed controls work under a credible set of scenarios and whether remaining limits are understood. A Gold Standard review describes what was tested and what continues to require expert judgement.
9. Trust grows when the team can explain the corrected decision
After the repair, the map displays useful general habitat information but no longer publishes the sensitive location at the same precision. Most of the 240 records retain their earlier public treatment, while the small reviewed group follows the documented sensitivity decisions. The team does not claim that every record is now equally public.
A short editorial note says which types of information may be generalised, why sensitive wildlife locations sometimes need different handling and how legitimate researchers can enquire about appropriate access. The note does not name an exact location or describe a route to it.
The organisers can still celebrate the community's careful observations. They simply stop equating maximal disclosure with maximal scientific responsibility. Good conservation data should enable learning while managing foreseeable risks to the organisms being studied.
10. Practice: four sensitive records among 240 observations
A fictional dataset contains 240 verified observations, including four whose precise location disclosure has been assessed as sensitive. The project manager proposes publishing all points unchanged because the observations are accurate and the grant required an open map. What information needs to remain accurate, what must be reassessed before public release and who should approve the treatment?
A strong answer preserves the source records, documents the ecological risk review, considers generalised or withheld location data where justified and checks photographs, captions and exports as well as coordinates. It makes the access decision by competent authority and explains the public limitations without revealing sensitive detail. It also verifies what the grant's actual open-data requirement permits; the exercise does not assume that 'open' overrides every safeguard.
Try writing an acceptance criterion for the next release: 'Every public occurrence has an approved disclosure classification, and the published fields and derivative files conform to it under the tested conditions.' That is a better Gold Standard promise than 'Every exact point is on the map'.
Source reading and related Gold Standard guidance
- GBIF · Current Best Practices for Generalizing Sensitive Species Occurrence Data
- GBIF · Guide for Publishing Biological Survey and Monitoring Data
- iNaturalist · Geoprivacy and Obscured Observations
- The Gold Standard of Urban Biodiversity
- eduKate · Why Science? Environmental DNA and Biodiversity Evidence
- The Gold Standard of Coastal Ecosystem Restoration
- Case 37 · Anonymous Survey and Small-Group Identification
- Gold Standard · Core Requirements and Evidence
- Gold Standard · Casebook and Decision Practice
- The Gold Standard · Master Hub
This fictional case illustrates risk-based biodiversity data stewardship. Exact protections depend on scientific advice, the species and circumstances, the rules governing the data and any applicable law. The GS-R01–GS-R12 labels refer to eduKate's educational management reference system, not a third-party conservation certification.