Identifying the demographic pathways linking environmental covariates to population dynamics in an avian migrant.

Understanding and predicting the effects of climate change on populations requires linking the environmental conditions to demographic rates and the demographic rates to population-level consequences, but often this complete demographic pathway is not studied. Integrated population models (IPMs) incorporate demographic data into a single analytical framework, allowing for the inclusion of environmental covariates to test hypotheses considering how the environment influences demographic rates, and consequently, to which demographic rates population growth rate is most sensitive. In birds, there is strong evidence that environmental conditions impact population growth, and that long-distance migrant avian species with short phenological windows are at greatest risk of population decline due to changing environmental conditions. We built a Bayesian IPM with over 40 years of mark-recapture, fecundity, and nest box occupancy data and incorporated environmental covariates hypothesized to be driving the observed changes in two populations of a fast-lived long-distance migrant, the European pied flycatcher. Using variance decomposition methods, we identified the demographic pathways through which environmental covariates were acting. While several environmental covariates impacted fecundity and survival, only precipitation acted via apparent juvenile and adult survival contributed significantly to variation in population growth rate. Increased precipitation during the nest initiation, incubation, and hatchling stages had negative carry-over effects on juvenile survival during the post-fledging and overwintering period, and increased precipitation negatively impacted adult apparent survival, likely due to the increased energetic demands of caring for eggs and hatchlings in challenging conditions and reduced availability of aerial prey. We show that linking environmental covariates to demographic rates does not sufficiently explain or predict population-level consequences, and that decomposing variation along the complete demographic pathway is a necessary step to appropriately identify how covariates influence population dynamics.