The critical challenge of nonstationarity in ecological modelling

Abstract

Ecological systems are dynamic and shaped by complex interactions that vary across space, time, and levels of biological organization (Fischer et al., 2025). Therefore, most ecological systems display confounding spatial and temporal trends, such as nonstationarity, at some scale. Nonstationarity (abrupt change in the mean or variance of a system property in space or time) can result from variation in environmental conditions (e.g., topography) or occurrence of a disturbance (e. g., a flood). Models of systems exhibiting nonstationarity that are based on non-spatial and equilibrium representations of system properties can produce misleading inferences and predictions. The potential consequences include misidentification of habitat priorities, underestimation of extinction risks, and the development of conservation strategies that ignore critical local contexts (Alagador and Cerdeira, 2020). Therefore, we must aim to incorproate nonstationarity in ecological models and address its impact on processes occurring on the multiple scales of time, space and organisation relevant to management (Wang et al., 2023).