Integrating demographic data: towards a framework for monitoring wildlife populations at large spatial scales

Landscapes are becoming increasingly intensively managed resulting in greater anthropogenic disturbance of ecosystems. Effective policies for conservation and management of wildlife populations will require a mechanistic understanding of the processes underlying the population responses to these changes. Detailed demographic studies are often of individual populations in relatively stable habitats, whereas there is a need to characterize demographic variation at larger scales in wider landscapes that are subject to environmental change. We use long-term (18-46 years) data sets, gathered by volunteers across Britain, to parameterize Bayesian state-space demographic population models for 17 widely occurring bird species. As some demographic processes were not quantified directly, we introduce a scaling parameter (ρ) to account for this unmeasured variation, and assess the extent to which this influences our ability to understand the processes driving population change. The population models provided a good fit to the population trajectory in all but one species (the nocturnal tawny owl Strix aluco). The scaling parameter generally showed little annual variation and did not make a major contribution ( < 12%) to population change among passerine species. For five non-passerine species, the contribution was still not large (11-21%), but did indicate the importance of unmeasured demographic variables in these species. We were able to retrieve known demographic signals both within and across species. The importance of the nest failure rates was greater for species that construct nests in open rather than cavity situations. We also confirmed the demographic causes of population decline in two species of conservation concern, song thrush and lapwing, as changes in juvenile survival and reproductive success respectively. Furthermore, we show that population dynamics of declining species are generally driven more by recruitment than increasing species (for which variation in adult survival is more important). These analyses demonstrate that volunteer-based monitoring programmes can provide insight into population changes at large spatial scales across a range of species. By facilitating the combination of data sets on different demographic parameters, the integrated approach adopted here allows important demographic parameters to be identified, even if they are not measured directly