Where does environmental stochasticity most influence population dynamics? An assessment along a regional core‐periphery gradient for prairie breeding ducks

Abstract Aim Geographical disparities in the effect of the environment on population dynamics have been shown to follow a core‐periphery gradient when peripheral populations are less abundant and occur in marginal habitat. Whether the gradient in environmental influence occurs in the absence of a gradient in abundance is not known. We tested whether duck populations on the periphery of their main breeding region were more strongly affected by environmental stochasticity and less abundant than at the core. Location The Prairie Pothole Region of central U nited S tates and C anada Methods We used the N orth A merican W aterfowl B reeding P opulation and H abitat S urvey to model the dynamics of 10 duck species at 1059 sites spanning 1961–2012. We used the N orth A merican B reeding B ird S urvey to measure abundance at the same sites, averaged over the same time span. We used structural equation models to characterize relationships among a site's distance from the regional centre, a site's abundance, and the degree to which a site's ducks were affected by environmental stochasticity. Results Environmental stochasticity never strongly contributed to population dynamics in the region's core. However, we did not find a linear increase in environmental stochasticity with increasing distance from the core. For seven species, we found that environmental stochasticty was stronger in sites of lower abundance. For only two species did average abundance decline from the region's core to periphery. Main conclusions Variability in the magnitude of environmental stochasticity follows coarse spatial differences in wetland productivity. Yet, among peripheral sites, the contribution of the environment to population dynamics varies, possibly reflecting the importance of the local environment and dispersal.