Mind the gaps: Harnessing iNaturalist to target natural history specimen collection

Natural history collections serve as critical records for biodiversity monitoring, biosecurity, and climate change research. However, these collections often suffer from geographic and researcher-driven biases, compounded by a global decline in funding for traditional collecting. To address these gaps cost-effectively, a new methodology integrates professional museum expertise with citizen science data.
By utilizing observations from the iNaturalist platform, researchers can identify specific species present in a region that are missing from official herbarium records. A five-year case study in Wellington demonstrated the efficacy of this approach, resulting in the addition of over 200 new plant species to Te Papa’s herbarium—a fourfold increase compared to the previous five-year window. This flexible framework has since been scaled to other regions and applied nationwide across Aotearoa New Zealand to detect newly arriving organisms and fill systemic taxonomic gaps.

* Lara Shepherd and Leon Perrie published a specimen collection method in the journal Biodiversity Informatics.
* The method utilizes citizen science observations from the iNaturalist platform.
* Te Papa’s herbarium is the receiving institution for the collected specimens.
* A case study was conducted over five years in the Wellington region.
* The project began during the 2020 COVID-19 lockdowns.
* Joe Dillon assisted in creating a project to map plant observations in Wellington.
* The method involves subtracting existing herbarium records from iNaturalist observations to identify missing species.
* Over 200 previously uncollected plant species were added to the herbarium during the study.
* The collection rate represents a fourfold increase over the previous five-year period.
* The approach has been expanded to Lower Hutt, the Pohangina Valley, and New Plymouth.
* A New Zealand-wide project exists to identify species not represented in any national herbarium database.

This methodology represents a shift toward "data-driven foraging," where digital crowdsourcing directs physical scientific labor. Using ACADEMIC MODE, the design appears logically sound for its stated purpose: reducing "collector's bias" by substituting the researcher's intuition with a comprehensive digital dataset. However, a peer reviewer would likely flag a significant selection bias inherent in the source data. iNaturalist observations are not random; they are clustered around hiking trails, urban parks, and accessible roads. Consequently, the "filled gaps" may still omit species in inaccessible terrains, meaning the collection is becoming more representative of *observed* biodiversity rather than *actual* biodiversity.
The claim of a "fourfold increase" in specimens is a measure of efficiency, not necessarily a measure of total biodiversity capture. While the method is cost-effective, the reliance on citizen science introduces a variable of identification accuracy. While the authors mention verifying identifications via photographs, the scale of national-level projects may challenge the capacity for rigorous professional verification of every target.
If this approach is scaled, it transforms the role of the museum curator from a generalist explorer to a targeted harvester. This increases the speed of collection but may decrease the serendipitous discovery of species that neither the curator nor the citizen scientist knows to look for.
Bridge Questions:
1. How does the bias of citizen science geography (the "roadside effect") compare to the bias of professional researcher interests?
2. Would integrating satellite imagery or ecological niche modeling alongside iNaturalist data provide a more objective map of collection gaps?
3. What is the error rate of citizen identifications for cryptic species, and how does that affect the efficiency of targeted trips?