Environmental stochasticity in dispersal areas can explain the "mysterious" disappearance of breeding populations.

We develop two individual‐based models using a large and detailed data set (information gathered over more than a century) on a population of a longlived and territorial predator, the Spanish imperial eagle. We investigated the relationship between survival and predator pressure, prey behaviour and patch availability (i.e. settlement areas). Survival of dispersing individuals was highly dependent on the number of available settlement areas, mediated by prey availability. Changes in prey behaviour due to predation pressure (e.g. shifting from diurnal to nocturnal activity) can decrease their availability for predators even if the density significantly exceeds the predator needs. Environmental stochasticity had a strong influence on population viability when it occurred in a synchroneous way between breeding and settlement areas, and an increase in floater mortality negatively influenced stability and dynamics of the breeding segment of populations in reproductive areas. Our simulations demonstrated the link between the dynamics in settlement and breeding areas: factors affecting floater survival also influence whole population dynamics. Moreover, model outputs provided insights into the relationship between environmental stochasticity and population dynamics.

Environmental stochasticity refers to unpredictable spatiotemporal fluctuation in environmental conditions. The term is often used in the literature on ecology and evolution. Unpredictability is defined as an inability to predict the future state precisely such that only its distribution can be known. The environment is typically defined as any set of abiotic (e.g. temperature and nutrient availability) and biotic (e.g. predator, competitor and food) conditions that organisms experience. Environmental stochasticity influences how population abundance fluctuates and affects the fate (e.g. persistence or extinction) of populations. In an evolutionary timescale, environmental stochasticity also affects the life history strategy of organisms. Environmental stochasticity is included in population models using univariate difference equations, stochastic matrix population models, stochastic differential equations and partial differential equations. Ecological data are analysed to determine the effect of environmental stochasticity using methods such as spectral analysis, capture–recapture analysis, state‐space analysis, generalised linear models and multivariate statistical analyses.Key ConceptsEnvironmental stochasticity is unpredictable spatiotemporal fluctuations in environmental conditions.Observed population dynamics consist of fluctuation due to environmental stochasticity, but it is often confounded with other factors such as observational errors, deterministic fluctuation and demographic stochasticity.Environmental stochasticity is reflected in the fluctuations in ecological processes and affects their fate (e.g. extinction or persistence of populations).Environmental stochasticity plays an important role in the evolution of life history strategies of organisms by affecting their fitness.Stochastic discrete‐time models, stochastic matrix population models, stochastic differential equation models and partial differential equation models are the four basic population models that include environmental stochasticity.Spectral analysis, state‐space model analysis, capture–recapture analysis, generalised linear models and multivariate statistical analysis are commonly used for separating the effects of environmental stochasticity in data.

Environmental stochasticity is unpredictable spatiotemporal fluctuation in environmental conditions, and it is one of the main sources of fluctuation in ecological processes. The term is often used in the ecology and evolution literature. Unpredictability is defined as a lack of ability to predict the future state precisely so that only its distribution can be known. Environment is typically defined as any set of physical, chemical and biological conditions that organisms experience, such as temperature, nutrient availability and the abundance of predators. Environmental stochasticity influences how population abundance fluctuates and affects the fate (e.g. persistence or extinction) of populations. In an evolutionary time scale, environmental stochasticity also affects life history strategy of organisms. Incorporating environmental stochasticity into analysis requires some care. Stochastic equations sometimes do not have explicit solutions so that simulations are required, and statistical analysis must separate other confounding factors such as stochastic sampling errors and demographic stochasticity.Key ConceptsEnvironmental stochasticity is unpredictable spatiotemporal fluctuations in environmental conditions.Correlation between time‐lagged environmental states is one statistical measure of its predictability.Environmental stochasticity is reflected in the fluctuations in ecological processes and affects their fate (e.g. extinction or persistence of populations).Observed population dynamics consist of fluctuation due to environmental stochasticity, but it is often confounded with other factors such as sampling errors and demographic stochasticity.All of the fluctuations in ecological data are often erroneously attributed to environmental stochasticity alone.Evolutionary dynamics are also affected by environmental stochasticity.Autoregressive moving average (ARMA) models are an example of statistical models incorporating environmental stochasticity.Using stochastic differential and difference equation models, mechanistic ecological models incorporating environmental stochasticity can be built.

Capsule Features of the breeding population and temporary settlement area influence the behaviour of Eagle Owls Bubo bubo prospecting for breeding sites during natal dispersal.Aims To understand how prospecting behaviour during natal dispersal is affected by (i) the main characteristics of the breeding and dispersing portions of the population and (ii) main prey availability.Methods We explored the ten-year dynamics and characteristics of radio-tagged breeders and dispersers of an Eagle Owl population.Results During the first years following natal dispersal there was little prospecting behaviour of nesting sites and birds remained mainly within non-breeding settlement areas, bordering the sector occupied by the breeding population. Settlement areas had an abundant food supply, and low intraspecific competition and mortality. We suggest that these features of the settlement areas may reduce the willingness of individuals to search for breeding sites and may have the potential to impact on the viability of breeding populations.Conclusion From a conservation perspective, the lengthy use of the temporary settlement areas by juvenile Eagle Owls suggests that the sites should be considered as important as the breeding areas when planning conservation strategies. Reducing juvenile mortality in settlement areas may represent an overlooked conservation strategy for long-lived species and may have a crucial effect on the viability some animal populations.

Annual movements have been widely described for birds migrating across the Americas and between Eurasia and Africa, yet relatively little information exists for intra-African migrants. Identifying the areas used throughout a species annual cycle by understanding migratory patterns and settlement areas during breeding and non-breeding seasons is essential for conservation initiatives. Here, we describe for the first time, the migratory patterns and settlement areas of an endangered raptor endemic to Southern Africa, the Black Harrier (Circus maurus). From 2008 to 2015, thirteen breeding adult Black Harriers were trapped in south-western South Africa and fitted either with a GPS-GSM or with a PTT tracker device. Adults were monitored for 365 ± 198 days (range: 56–819 days) revealing great individual variability in annual movements. Most Black Harriers performed an unusual West-East migration from their breeding areas, but routes of all migrating individuals covered the entire southern land area of South Africa and Lesotho. The distance travelled averaged 814 ± 324 km, but unlike many other species, migrants travelled faster during post-breeding (i.e. austral summer) (207.8 ± 113.2 km.day-1) than during pre-breeding (i.e. austral winter/spring) migrations (143.8 ± 32.2 km.day-1). Although most marked individuals displayed movements similar to those that bred following pre-breeding migrations, only two of thirteen were confirmed as breeders the year after being tagged. This suggests that individuals may sometimes take a sabbatical year in reproduction, although this requires confirmation. Most tagged birds died on migration or during the non-breeding season. Adults frequently returned to the same non-breeding settlement areas, and often used up to 3 different locations an average of about 200 km apart. On the other hand, there was wide variation in distance between subsequent reproductive events. We discuss the implications of our study for the conservation of Black Harriers and more broadly for intra-African bird migrants.

10 - Extinction Models for Local Populations

Identifying suitable habitat for dispersal in Bonelli’s eagle: An important issue in halting its decline in Europe

Inbreeding is common in nature, and many laboratory studies have documented that inbreeding depression can reduce the fitness of individuals. Demonstrating the consequences of inbreeding depression on the growth and persistence of populations is more challenging because populations are often regulated by density- or frequency-dependent selection and influenced by demographic and environmental stochasticity. A few empirical studies have shown that inbreeding depression can increase extinction risk of local populations. The importance of inbreeding depression at the metapopulation level has been conjectured based on population-level studies but has not been evaluated. We quantified the impact of inbreeding depression affecting the fitness of individuals on metapopulation persistence in heterogeneous habitat networks of different sizes and habitat configuration in a context of natural butterfly metapopulations. We developed a spatial individual-based simulation model of metapopulations with explicit genetics. We used Approximate Bayesian Computation to fit the model to extensive demographic, genetic and life-history data available for the well-studied Glanville fritillary butterfly (Melitaea cinxia) metapopulations in the Åland islands in SW Finland. We compared 18 semi-independent habitat networks differing in size and fragmentation. The results show that inbreeding is more frequent in small habitat networks, and consequently, inbreeding depression elevates extinction risks in small metapopulations. Metapopulation persistence and neutral genetic diversity maintained in the metapopulations increase with the total habitat amount in and mean patch size of habitat networks. Dispersal and mating behaviour interact with landscape structure to determine how likely it is to encounter kin while looking for mates. Inbreeding depression can decrease the viability of small metapopulations even when they are strongly influenced by stochastic extinction-colonization dynamics and density-dependent selection. The findings from this study support that genetic factors, in addition to demographic factors, can contribute to extinctions of small local populations and also of metapopulations.

We analysed migratory connectivity between different winter quarters and breeding sectors in the circumpolar tundra region for arctic shorebirds, in relation to migratory distances and ecological barriers. Total distances and barriers were calculated and measured for all potential migratory orthodrome links between 10 selected winter regions and 12 breeding sectors. The migratory segment between the northernmost stopover site and the breeding area, associated with the entry to and exit from the tundra during spring and autumn migration, respectively, was also identified and measured for each potential link. The analysis indicated that the evolution of migratory links among arctic shorebirds is constrained not by distance as such but by distance across ecological barriers, possibly because of the complex adaptations required for barrier crossing and extensive detour migration (and in a few cases because barrier distances exceed the birds' theoretical flight range capacity). A particularly pronounced barrier effect of the Arctic Ocean, as apparent from a sharp decline in migratory connectivity between the opposite sides of the Arctic Ocean, may reflect a crucial importance of favourable entry and exit conditions for successfully occupying different sectors of the tundra breeding area by shorebirds from winter regions situated at widely different total distances in both the southern and northern hemispheres.

A selection harvesting algorithm for use in spatially explicit individual-based forest simulation models

A machine learning approach to investigate the reasons behind species extinction

Kin competition and the evolution of dispersal in an individual-based model

Despite the success of recent global malaria control efforts, which have halved global malaria mortality since 2000, malaria is still one of the world’s most deadly diseases causing an estimated half a million deaths, mostly among African children, and around a quarter of billion clinical episodes every year as reported in 2014. Drug resistance is one of the most important challenges to malaria elimination. To contain drug resistance, many efforts have been put forth including improvement of surveillance systems and mass treatment in order to stop or slow down the transmission of the resistant strain. To find out whether a population-level treatment strategy can have any benefit in containing drug resistance, mathematical models are an appropriate approach to this problem and individual-based models allow us to have a better understanding of the effect of individual heterogeneities on the outcome. The first part of the thesis is about building and validating an individual based microsimulation. The model is implemented as an individual-based discrete-time event simulation model in C++. The behaviors and the state changes of human individuals are determined by relevant events and mathematical formulas. This integrated model combines components that reproduce the most important features of malaria transmission and epidemiology: the infectiousness of human populations; clinical model of acute illness; heterogeneities in individuals’ age, biting-rate level, drug absorption, drug action, multiple parasite populations, and human immunity. To validate this individual-based model, two types of validation have been done. The model’s parameters were obtained from field or clinical data were used directly in the model. For those parameters that cannot be obtained directly from literature review, sensitivity analysis has been done to find how variation in parameter values affects certain key features of malaria epidemiology. The second part of the thesis focused on the comparison between population-scaled treatment strategies. The results showed that using multiple first-line therapies (MFT) results in a lower number of treatment failures compared to other strategies where a single first-line ACT is recommended. This result is robust to various epidemiological, pharmacological, and evolutionary features of malaria transmission. In addition, including non-ACT therapy in an MFT strategy seems to have a significant benefit in reducing the pressure on artemisinin-resistance evolution, delaying its emergence and slowing its spread. The third part of the thesis focused on individual-level treatment strategies to combat artemisinin resistance. The results showed that lengthening an ACT course or using multiple courses of ACT can reduce the long-term number of treatment failures significantly. The work reported here introduces a novel individual-based simulation that includes drug resistance evolution and the ability to be scaled up to millions number of individuals. The challenge that remains is to evaluate the feasibility of these novel treatment strategies given that they will need to be implemented in the real world of malaria control programs, their operations, human behavior, and economic realities.

Individual-based model of population dynamics in a sea urchin of the Kerguelen Plateau (Southern Ocean), Abatus cordatus, under changing environmental conditions

Current population genetic and population dynamic models are inappropriate to judge the risk of extinction of small populations due to the combined effects of inbreeding, genetic drift, demographic stochasticity, and environmental stochasticity. Instead, a model based on the aggregated fates of individuals is advocated. The unequal distribution of resources over individuals is an essential part of this model. The model allows the incorporation of the mutation-selection dynamics of alleles leading to inbreeding effects and to fixation of slightly deleterious mutations as a result of genetic drift. The slightly deleterious mutations lower the conversion of resources into offspring. Whereas lethal alleles are rapidly eliminated by selection in small populations, the selection against mild deleterious effects depends strongly on effective population size and on the social system, that is, on the division of resources among individuals. The model allows for the study of rates at which processes occur while far away from equilibrium, which is crucial in understanding the extinction risks of threatened populations. One example of the latter is illustrated in simulations in which small populations become extinct between approximately 100 and 200 generations after they became small populations, due to a gradual accumulation of mildly deleterious mutations.