Supplementary material to: "Demographic causes of the pesticide crash in the peregrine falcon"

Abstract Population crashes in many avian predators during the 1950–70s, caused by organochlorine pesticides, belong to the most spectacular cases in the history of conservation and ecotoxicology. Negative effects of DDT on eggshell thickness are well documented and led to egg breakage and drastic declines in the annual number of young produced per pair (productivity). In addition, cyclodiene pesticides such as Dieldrin were strongly suspected to contribute to crashes by increasing mortality, but this hypothesis could never be tested owing to a lack of early enough data and suitable analyses. We studied a large population of the peregrine falcon (Falco peregrinus) in the Jura mountains over 60 years (1964–2023), from crash to full recovery. We combined in an integrated population model data on ring- (a.k.a., band-) recovery, productivity, and population counts, tested for a reduction in survival during crash years, and used retrospective analysis to compare the relative importance of survival and productivity for population dynamics. Incidentally, we discovered that for data with unequal sample sizes over time, only an autoregressive time-series formulation properly captured annual trajectories of demographic parameters, while traditional models with unstructured temporal random effects did not. The population crash continued until the early 1970s, and subsequent recovery was not complete until the early 2000s. Productivity was greatly reduced during the crash and increased afterwards. Adult annual survival probability was very low during the crash (0.62 [credible interval 0.46–0.75] in 1964) and increased to 0.85 [0.77–0.91] in 1978. Juvenile survival probability showed a long-term decline. Temporal variance in the population growth rate was primarily explained by adult survival (67%), followed by productivity and juvenile survival with 16% each. Hence, adult survival in this peregrine population had far greater effects on the pesticide crash and the ensuing recovery than did productivity. This is the first time that the hypothesis, that survival was the major cause of the pesticide crash, was tested and corroborated. Our study illustrates well the key role of intensive long-term monitoring schemes: in combination with modern analytics, they can generate critical demographic knowledge with wide conservation implications and thus serve as invaluable environmental early-warning systems.