Bridging the scale gap: Predicting large‐scale population dynamics from small‐scale variation in strongly heterogeneous landscapes

Abstract

Abstract Often, ecologists are challenged with a mismatch of scales: how do we upscale from local variation and available data to landscape‐level models and predictions? We present a general recipe for coarse‐graining from local‐ to landscape‐scale reaction–diffusion equations when spatial heterogeneity is small in extent compared to dispersal of organisms. Our homogenization approach uses the fundamental ecological concepts of Turchin's residence index and Skellam's dynamic level. Our approach opens avenues to new ecological theory that connects different scales, which we illustrate using predator–prey interactions. It also presents opportunities for using the increasingly available small‐scale data for landscape‐level predictions, such as range expansion rates. We find several unexpected nonlinear relationships between the movement behaviour on the local level and the spatially implicit and explicit outcomes at the landscape level, for example, predator spread rate may increase or decrease when predators move faster locally. Our method provides a mechanistic link for population dynamics and data integration across spatial and temporal scales, addressing a fundamental goal of landscape ecology.