See No Weevil: Dark Data Patterns in Repositories and Their Implications for Conservation Planning

Abstract

Reliable species distribution data are essential for conservation planning, yet for many invertebrates, these data are incomplete or inaccessible. Institutional collections hold millions of insect specimens, but most of the data and metadata are undigitised ‘dark data', limiting their integration into global biodiversity platforms. We assessed the effect of completeness of digitisation on estimations of species’ environmental space using Anagotus (Coleoptera: Curculionidae), a genus of large‐bodied, flightless weevils endemic to New Zealand, as a case study. We compiled occurrence records from institutional collections and the Global Biodiversity Information Facility (GBIF), cleaned, georeferenced and linked them to environmental variables from the New Zealand Environmental Data Stack. We analysed the geographic and environmental space occupied in GBIF and non‐GBIF records using choropleth maps and environmental principal component analysis. Prior to our digitisation efforts, only 25.6% of the 2202 specimens held in major institutions had metadata available through GBIF, with large gaps in the New Zealand Arthropod Collection (72.6% undigitised), which also contributed most records (70.7%). The consequences of these absences extend beyond coverage statistics; for the 14 named species we analysed, adding non‐GBIF records recovered two species entirely absent from GBIF and expanded the observed ecological regions by 0%–325% ( µ = 84.90%) for the remaining 12, fundamentally altering our understanding of their geographic range and likely rarity‐matrix classification. Despite this shortfall, GBIF records broadly captured the environmental space occupied by Anagotus , although some climatic outliers and historically or geographically unique localities were not represented. Our analysis suggests that while partial digitisation can be sufficient to support some modelling for conservation planning, comprehensive digitisation remains critical for capturing environmental extremes and historical baselines. For taxa like New Zealand's endemic flightless weevils, the records absent from aggregated repositories could influence whether a species is correctly classified as threatened.