Census data aggregation decisions can affect population-level inference in heterogeneous populations.

摘要: Conservation and population management decisions often rely on models parameterized using census data. However, the sampling regime, precision, sample size, methods used to collect data are usually heterogeneous in time space. Decisions about how derive population-wide estimates from this patchwork of complicated may bias estimated dynamics, with important implications for subsequent decisions.Here, we explore impact site selection aggregation pup survival estimates, downstream derived matrix (MPMs), a long-term dataset grey seal (Halichoerus grypus) southwestern Wales. The spatiotemporal methodological heterogeneity fairly typical ecological it is, therefore, good model address topic.Data were collected 46 locations (sites) over 25 years, handling by varying years combined parameterize population-level MPMs. We focus because abundant high-quality available developmental stage.We found that probability was highly variable most variation being at level, poorly correlated among sites. This could generate marked differences predicted dynamics depending strategy. size required confident estimate also varied markedly geographically.We conclude populations vital rates sub-populations, important. In particular, including peripheral or less frequently areas can introduce substantial into estimates. is likely be context-dependent, but these choices, use appropriate weights when summarizing across areas, should explored ensure actions successful.