ABSTRACT Aim Climate change is a major driver for the wave of species extinctions predicted for this century. Extinction risk assessments are often estimated from climate‐driven reductions in the geographic ranges of species. However, the empirical relationship between extinction risk and range change remains unclear because few extinctions are observed in the Holocene. Using fossil evidence, this study investigates how extinction risk changes with the percentage change in geographic range of a genus over multi‐million‐year timescales. Location Fossil evidence is derived from the Paleobiology Database, which documents global fossil occurrences. Time Period We examine Neogene to Holocene (23–0 Ma) occurrences of marine genera. Major Taxa Studied The dataset includes six marine classes—Bivalvia, Echinoidea, Gastropoda, Mammalia, Anthozoa and Chondrichthyes—encompassing 1215 genera. Methods A Bayesian hierarchical weighted generalized additive model was implemented to analyse 2600 range change observations on a multi‐million‐year timescale. Results Results show that extinction risk increases non‐linearly with range loss and differs between taxa. Responses vary broadly between taxonomic classes. Model comparisons indicate extinction risk is best explained by the interaction between absolute range and range change. With constant absolute range, however, extinction risk does not consistently increase with range loss. Alternative binning and weighting schemes revealed systematic effects on extinction risks. Main Conclusions Extinction risk increases with range loss, but this relationship is strongly modulated by absolute range and varies across taxa. For a given absolute range, the effect of range loss on extinction risk largely disappears, suggesting greater vulnerability in small‐ranged taxa (and v.v.). However, model comparison indicated that extinction risk is best explained when both absolute range size and range change are included, implying both capture complementary aspects of vulnerability. By linking extinction dynamics to both absolute range and range change, our results provide a deep‐time perspective on the mechanisms underlying genus persistence and vulnerability.
Straube et al. (Sun,) studied this question.