Demographic rates of Adélie penguins from Approximate Bayesian Computation and time-lapse cameras

The underlying mechanisms by which Pygoscelis penguin populations are changing cannot be fully understood without knowledge of how stressors influence specific phases of their life histories. We developed a method to estimate key vital rates of Adélie penguins- age-specific survival, breeding success, and breeding propensity- using Approximate Bayesian Computation and data from time series of abundances and breeding success samples. Our goals were to (1.) estimate demographic vital rates without tracking individuals, (2.) determine the spatial autocorrelation of breeding propensity, and (3.) assess whether auxiliary estimates of breeding success, such as from time-lapse cameras, can improve estimates of penguin demographic rates. We estimated demographic rates for eight Adélie penguin populations in the Antarctic Peninsula region and found that time series of abundances yielded the strongest inference on average adult survival, which our model assumed was shared by all populations, and mean breeding propensity rates for two populations. Adult survival and the highest and lowest cases of breeding propensity were likely discernible because they are strong drivers of trend and interannual variability, respectively. Breeding propensity rates varied by population, though were more similar to nearby populations. This points to breeding propensity being driven primarily by local environmental factors. Surprisingly, the inclusion of auxiliary breeding success data did not improve our estimates of demographic transitions or model performance. This study demonstrates that while some demographic insights remain limited without mark-recapture data, time series of abundances can provide key information on survival and breeding propensity that is readily and cost effectively scaled up.