Why Taxonomic Splits Matter for Bat Biodiversity and Viral Risk Analyses
| Conference: | Living Data 2025 |
|---|---|
| Location: | Bogotá, Columbia |
| Date and Time: | Thu, 23 Oct 2025 12:10 PM UTC -05:00h |
| Session: | Modern challenges of classic taxonomy, how to connect and keep catalogues up to date |
time: 10:45am (Bogotá)
Presenter: Nathan Upham (ASU)
Authors: Aja Sherman (Bat Eco-Interaction Project), Nancy Simmons (AMNH), Kendra Phelps (UoM), Anna Willoughby (UoG), Quentin Groom (Meise Botanic Garden), Donat Agosti (Plazi), DeeAnn Reeder (BU), Jorrit Poelen (RIIS), Connor Burgin (UoNM)
Taxonomy is generally viewed as external to ecology, a fixed framework rather than a variable that needs tuning to understand its impact on inferences. However, since taxonomy is a human-imposed perspective rather than something innate, biodiversity scientists have the responsibility to quantify its impact on derived knowledge. Here we study how the 35% increase in recognized bat species (Chiroptera) since 1993 has impacted our knowledge of bat-virus interactions and, in turn, inferences of viral spillover risk to humans. We focus on change in the geographic concepts of bat species globally due to taxonomic splits between two periods: (i) 2008-2020 using IUCN range maps (based on Mammal Species of the World, 3rd edition) relative to Mammal Diversity Database (MDD) v1.2 range maps; and (ii) 2020-2023 comparing MDD v1.2 to newly produced v1.11 maps. We then intersect these conceptual changes with known bat-virus interactions from databases to assess their impact upon risks of cross-species viral transmission. We find that taxonomic splits affected 248 bat species across both periods (185 and 63 species, respectively), which has impacted 16.9% of currently recognized bat species globally. Those taxonomic splits have rendered ambiguous 3,249 bat-virus interactions — 21.9% of all digitally known observations — since these data are indexed by host species name rather than observation location, which leads to ambiguity when species are split. We discuss high-throughput solutions for accurately translating the taxonomy of these data, and the impact of inaction upon estimates of viral spillover risk.