To clean or not to clean: Cleaning open‐source data improves extinction risk assessments for threatened plant species

• DOI: 10.1111/csp2.311
• Publication Date: 2020-11-17T18:13:53Z
• Authors: Connor T. Panter; Rosemary L. Clegg; Justin Moat; Steven P. Bachman; Bente B. Klitgård; Rachel L. White
• Keywords: Herbarium; Data deficient; Occupancy; Global biodiversity; Extinction (optical mineralogy)

Abstract Plants are under‐represented in conservation efforts, with only 9% of described species published on the IUCN Red List. Biodiversity aggregators including Global Information Facility (GBIF) and more recent Botanical Ecology Network (BIEN) contain a wealth potentially useful occurrence data. We investigate influence these data accelerating plant extinction risk assessments for 225 endemic, near‐endemic, socioeconomic Bolivian species. Geo‐referenced herbarium voucher specimens verified by taxonomic experts comprised our control set. Open‐source 77 was subjected to two‐stage cleaning protocol (using an automated R package followed manual clean) threat categories were computed based extent thresholds. Accuracy highest using cleaned GBIF (76%) uncleaned BIEN (79%). Sensitivity (73%) (80%) suggesting essential maximize sensitivity rates. Comparisons between control, sets revealed paucity 148 (66%), 72% which qualified threatened category. Balancing quantity accuracy must be considered when open‐source Filling gaps is priority improve coverage within biodiversity aggregators.