Well-intentioned initiatives hinder understanding biodiversity conservation: Cloaked iNaturalist information for threatened species

Abstract

To protect threatened species, iNaturalist, one of the most important repositories of geographic information on species´ geographic distributions, deliberately adds uncertainty to otherwise precise occurrence records of threatened species. This “clouded” or “cloaked” information is shared with other biodiversity repositories, such as the Global Biodiversity Information Facility (GBIF). Although a mechanism exists by which users can obtain unbiased information, it requires the user to ask each data owner for the exact location of the record, which is difficult as many data providers may have participated in assembling the currently available data. To test for effects of adding uncertainty in estimation of the ecological niche and the potential distribution, we modeled the climatic requirements of three amphibian species using five sets of records: (i) a set of records at full accuracy and precision, and (ii) the same records but with 25 %, 50 %, 75 % and 100 % of biased occurrences induced by applying the iNaturalist protocol for introducing biases into biodiversity records. The three species present contrasting distribution sizes (from micro-endemic to widely distributed in the Sierras Pampeanas Centrales in Argentina). By employing ecological niche models, we showed that inducing bias generates misleading estimates of geographic distributions and ecological requirements even in situations in which the percentage of biased information is “low.” We also evaluated the percentage of occurrence records of IUCN-threatened vertebrates in GBIF provided by iNaturalist, and found that, for 2011 to present, the majority of GBIF records comes from this citizen science repository. These results could have relevant implications for conservation strategies. This effect accentuates the Wallacean and Hutchinsonian shortfalls, which are already some of the most significant impediments to developing effective conservation strategies. The iNaturalist data-cloaking initiative certainly will aid in protecting some threatened species, but could prove harmful to others. We encourage researchers to consider the precision and accuracy of the occurrence data used in their analyses, and we urge iNaturalist to simplify procedures by which credible researchers can access the full-accuracy data.