Harnessing the potential of integrated systematics for conservation of taxonomically complex, megadiverse plant groups

Abstract

The value of natural history collections for conservation science research is increasingly recognized, despite their well-documented limitations in terms of taxonomic, geographic, and temporal coverage. Specimen-based analyses are particularly important for tropical plant groups for which field observations are scarce and potentially unreliable due to high levels of diversity-amplifying identification challenges. Specimen databases curated by specialists are rich sources of authoritatively identified, georeferenced occurrence data, and such data are urgently needed for large genera. We compared entries in a monographic database for the large Neotropical genus Myrcia in 2007 and 2017. We classified and quantified differences in specimen records over this decade and determined the potential impact of these changes on conservation assessments. We distinguished misidentifications from changes due to taxonomic remodeling and considered the effects of adding specimens and georeferences. We calculated the potential impact of each change on estimates of extent of occurrence (EOO), the most frequently used metric in extinction-risk assessments of tropical plants. We examined whether particular specimen changes were associated with species for which changes in EOO over the decade were large enough to change their conservation category. Corrections to specimens previously misidentified or lacking georeferences were overrepresented in such species, whereas changes associated with taxonomic remodeling (lumping and splitting) were underrepresented. Among species present in both years, transitions to less threatened status outnumbered those to more threatened (8% vs 3%, respectively). Species previously deemed data deficient transitioned to threatened status more often than to not threatened (10% vs 7%, respectively). Conservation scientists risk reaching unreliable conclusions if they use specimen databases that are not actively curated to reflect changing knowledge.