Population-level, state-dependent response as a trait predicting species redistribution under climate change

Abstract Species change their population sizes and distributions in response to fluctuating environments. Predicting such changes across space, particularly distributional shifts under climate change, is a central challenge in ecology and conservation 1 . Traditionally, static traits such as habitat preferences, or performance traits such as abundance–environment relationships, have been used to characterize species’ responses to the environment 2,3 . However, these approaches assume fixed relationships between species and their environments, overlooking the inherently state-dependent nature of ecosystems. Here, we show that population-level, state-dependent responses of species to environments can serve as a novel trait to better represent population dynamics in nature, which we term a dynamic response trait. Using nonlinear time-series analysis 4,5 and long-term marine fish community data collected from Kyoto, Japan 6 , we identify causal influences of water temperature on about one hundred fish species. Species with higher latitudinal centers generally show negative dynamic responses to warming, whereas those with lower latitudinal centers show positive ones. Intriguingly, these dynamic response traits quantified at a single location explain the fish species’ poleward range shift velocities estimated from public biodiversity records; species with negative dynamic responses to warming shift poleward more rapidly, whereas those with positive ones tend to remain. Our findings establish dynamic response traits as a new dimension of trait-based ecology, capturing state-dependent species’ responses. By linking local population dynamics to broad-scale distributional shifts, this approach provides a powerful basis for guiding ecology and conservation under climate change.