Long-term population dynamics of a human-associated colonial raptor: Case of Lesser Kestrel Falco naumanni

Long-term monitoring is essential for understanding how demographic processes operating across generations, and under environmental stochasticity, shape population change. This is especially true in human-modified landscapes where nesting opportunities and food resources can shift rapidly. We analysed 20 years of standardized monitoring (2005–2025) from a Lesser Kestrel (Falco naumanni) nest-box colony in central Greece, treating annual colony fledgling output as a fecundity-like component of demography and explicitly decomposing it into breeding participation (occupancy) and productivity per occupied box. Using generalized additive models, we modelled non-linear effects of year and laying date on phenology, fledgling counts and complete failure risk, assessed temporal structure in incubation and nestling durations, and included nest-box identity as a random effect where relevant. Basic breeding parameters were stable across years: mean clutch size was 3.96 eggs and mean brood size was 2.63 fledglings per occupied nest, with most losses occurring before hatching (mean hatching success 0.72). In contrast, colony output varied strongly (108–284 fledglings/year) because occupancy fluctuated widely (mean 0.68; range 0.44–0.98) and followed a pronounced non-linear trajectory, whereas fledglings per occupied box showed no detectable temporal trend. Coupling the occupancy and productivity models yielded an annual relative productivity index spanning 0.44–1.59, with a clear mid-series peak. Breeding phenology also varied markedly: mean laying date was 130.3 ± 9.1 (SD) Julian days and shifted by ~18 days among years without a consistent directional trend. At the nest level, reproductive output followed a unimodal seasonal pattern, with the largest broods and lowest failure risk at intermediate laying dates and poorer outcomes toward both phenological extremes. Incubation (27.7 ± 3.8 days) showed only subtle temporal variation, and the nestling period (29.6 ± 5.6 days) was essentially constant. Overall, long-term variation in colony fecundity was driven primarily by changes in breeding participation, with phenological extremes amplifying annual variance, indicating that nest-box conservation should evaluate and manage occupancy alongside per-nest productivity. Because occupancy reflects settlement and recruitment dynamics beyond per-nest performance, tracking box use can flag change and prioritize interventions such as maintenance, disturbance reduction, and overheating mitigation.