| dc.description.abstracteng | The Red Kite (Milvus milvus) is a European near-endemic raptor species and due to its limited distribution and small global population size in the focus of national and international conservation efforts. Even within Europe, the distribution of the Red Kite is strongly restricted and the majority of the species’ breeding and wintering populations, especially in Germany and Spain, showed declines over the last decades. In contrast, some reintroduced or formerly marginal Red Kite populations, especially in in the UK, Switzerland and Sweden, showed increasing populations in recent years. Red Kites are opportunistic scavengers in the agricultural landscape and the species has seen a changeful history of population increase and decline, strongly linked to human cultivation of farmlands and anthropogenic mortality. In comparison to other raptors and considering its intermediate size, the Red Kite is an extremely long-lived and highly social bird species – which strongly affects its demography and distribution. Although mortality of adult breeding individuals has by far the strongest impact on Red Kite population growth rates, demographic data and studies on age-specific survival rates are lacking – limiting the understanding of the species’ population trends. Also, the impact of apparent strong philopatry on breeding occurrence and population development of Red Kites as well as the consequences of a propensity for delayed recruitment (‘floating’ behaviour) are currently not well understood.
The aim of this doctoral thesis is to synthesise data on Red Kite demography and distribution from Germany with current analytical methods, to increase the understanding of central demographic rates and to describe crucial influencing variables for the species’ breeding occurrence. The results are expected to inform conservation and research on the Red Kite in Germany and to contribute to a scientific basis for evidence-based management in the future.
Chapter 2 addresses the lack of recent age-structured survival estimates for the Red Kite in Germany. Using a long-term dataset of nearly 30,000 Red Kites marked with metal rings and about 1500 dead recoveries of these individuals, we estimate age-specific survival probability over nearly 50 years in a major part of the German Red Kite population. With a multinomial ring-recovery model, we consider age-dependent recovery probability, based on separate datasets of birds marked as nestlings and as adults/immatures, and thereby estimate juvenile, subadult and adult survival probability over time. The results showed a substantial long-term decline in Red Kite juvenile survival of more than 40 % from the 1970s until today. Furthermore, from years 1974-2014 adult survival probability showed a consistently decreasing trend (-0.26 % p.a.). The recovery probability for dead Red Kites in the first year (as juveniles) was estimated as being two times lower than for birds that reach subadult/adult age classes, which could be related to differing causes or locations of death in the first year of life.
The spatiotemporal patterns in juvenile Red Kite recoveries suggested an increase in mortality at the breeding grounds, but in >60 % of the cases the cause of death was unknown or not reported. To understand which factors are driving changes in survival of the Red Kite, age-dependent causes of mortality need further study. This work lays the foundation for further analyses of Red Kite population viability in this thesis and it allows to study annual variation in Red Kite survival and its potential drivers in the future.
Chapter 3 examines how environmental factors and local correlation patterns shape the breeding distribution of the Red Kite in Germany. Based on a national-scale population survey, with more than 6,000 breeding occurrences and high-resolution data on land use, habitat structure and climate variables, it decomposes environmental variability and spatial correlation and derives predictions of habitat suitability and probability of occurrence for the Red Kite in Germany. To account for spatial autocorrelation in the distribution data, a hierarchical model was used which corrects the model estimates using random effects (‘Gaussian random fields’). The model results showed very good predictive accuracy (AUC = 0.89) and explained 64.6 % of the variability in the distribution data, of which more than half was attributable to the environmental variables. Local occurrence of the Red Kite was strongly influenced by agricultural use and habitat diversity, but also by human disturbance. A high proportion of grassland in the surroundings, but also arable fields paired with woody margins (groves, hedges etc.) strongly increased the probability of a Red Kite nest being present. In addition, the results showed a substantial negative effect of agricultural intensification on the occurrence of the Red Kite, measured by the density of livestock farming. Model predictions of habitat suitability confirmed the need of a spatially comprehensive approach for protection in Germany, while the actual distribution of the Red Kite in some areas also deviated substantially from the model predictions. The work underlines and extends previous evidence on important habitat characteristics for the Red Kite, which is necessary for well-targeted habitat improvements, but also opens new possibilities to identify essential habitats for effective spatial protection measures.
Chapter 4 investigates the total population dynamics of the Red Kite in Germany using an age-structured demographic model. Based on the survival estimates from chapter 2 and published data on reproduction, we reconstructed key features of the population development over nearly 40 years in different age classes, including also non-breeding (‘floater’) individuals. Because recruitment age is a key demographic factor and varies substantially across Red Kite populations, we also considered a simple density-dependent function for age of first reproduction based on the literature in different model scenarios. The simulations showed a drastic decrease of juvenile and non-breeding individuals in the Red Kite population over time – driven both by declining vital rates and a density-dependent shift towards a younger age of first breeding.
This process was however not visible when judged by the size of the breeding population, which our model estimated to be of similar size today as in the 1980s. The total Red Kite population, including also non-breeding and juvenile individuals, was reduced to nearly 50 % of its former size in our reconstruction. Comparing the different model scenarios with existing estimates of breeding population size in Germany suggested that age of first reproduction for the Red Kite most likely varies non-linearly with density. Such a general feedback mechanism has largely been overlooked in previous studies but should be considered for improving the robustness of demographic simulations. The model reconstruction also highlights that conservation assessments for long-lived bird species, with a propensity for floating behaviour, profit from further demographic data and modelling procedures to avoid overlooking potentially cryptic population declines.
| de |