Aarhus University Seal / Aarhus Universitets segl


We know that marine mammals react to noise, that they are affected by changes in food availability because of overfishing and some even get stock in fishing gear. To understand how all of these factors act together, it is necessary to use different types of statistic and dynamic models. Our modelling toolbox spans modelling approaches at the individual level, e.g. individual-based modelling (IBM), dynamic energy budget (DEB) modelling, physiologically-based pharmacokinetic (PBPK) modelling and stable isotope mixing (SIM) modelling, at population level, e.g. population-effect modelling, and at ecosystem level, e.g. ecological network analysis (ENA). Moreover, we focus on scaling individual-level outcomes in health or behaviour to population and ecosystem responses.

Applied models

Dynamic Energy Budget (DEB)

These models study the flow of energy through living organisms as it relates to physiological processes such as growth, development, reproduction and maintenance. DEB theory provides a mechanistic-based framework to model interactions between animal ecology, toxicology, and metabolic processes throughout the entire lifecycle of any organism. We use this type of modeling to study the physiological mechanisms and interactions of natural and anthropogenic stressors and how these impact individuals and populations.

More information

Classical approaches in ecotoxicology rely on descriptive techniques to define observed effects (e.g. EC50, NOEC, …). The major problem with these descriptive approaches is that they ignore important real-world interactions between ecology and toxicology, and they cannot provide insight into the processes underlying toxic effects. Dynamic energy budget (DEB) theory resolves many of the shortcomings of descriptive toxicology by using a process- or mechanism-based modelling approach that accounts for interactions between animal ecology, toxicology, and metabolic processes throughout the entire lifecycle of any organism. DEB models study the flow of energy through living organisms as it relates to important individual and population relevant physiological processes such as growth, development, reproduction and maintenance. The adaptability of DEB theory to any species and its ability to incorporate natural and anthropogenic stressors underlines why we have adopted it in our research of wildlife health. 

The application of DEB models for wildlife health is a relatively new field, and we have just begun exploring this avenue in our research. We use DEB to model the growth, development and reproduction of animals over their lifetime and then use this health baseline to study the effects of stressors. We incorporate toxicokinetics in order to model the accumulation of contaminants over time and toxicodynamics to model how accumulated contaminants impact specific pathways and parameters in the model. When toxicity modes of action are unknown, we use pattern-oriented analyses to determine which pathways are most likely to cause observed effects. Future work will combine this dynamic contaminant exposure modelling with other stressors as well as link DEB to population-effect models.

Stable Isotop Mixing (SIM)

Stable isotopes, particularly of carbon and nitrogen, are established biogeochemical tracers to infer trophic interactions. Unravelling their mixing dynamics in the biosphere is important to identify the direction of interactions and species involved in the trophic network. In our research, we employ this modelling technique to reconstruct individual trophic interactions and food web dynamics for energy, contaminants and zoonoses.

More information

Stable isotopes of the same element take part in the same biochemical reaction at a slightly different speed, but, as a consequence of their mass difference, fractionate [word explanation?]  in a predictable way. This predictable fractionation allows trophic ecologists to use ratios of specific stable isotopes, especially of carbon and nitrogen, as biogeochemical proxies for individual variation in feeding ecology. At the same time, trophic feeding continuously mixes sources with distinct stable isotope ratio values, thus obscuring the relative dietary importance of these species. Stable Isotope Mixing Models (SIMMs) are the mathematical solution to this trophic ecological problem, which has applications in studying the trophodynamics of contaminants and zoonoses, as well as the maintenance of wildlife health.

We typically use stable isotope mixing models to reconstruct individual dietary habits of Arctic, Baltic and northern wildlife with a special focus on apex predatory marine mammals, such as polar bear (Ursus maritimus), narwhal (Monodon monoceros) and killer whale (Orcinus orca), as well as birds, such as white-tailed eagle (Haliaeetus albicilla) and tawny owl (Strix aluco). Using carbon and nitrogen, we focus on expanding the routine dual approach to a triple-isotope strategy, also including sulphur, as well as exploring compound-specific (amino and fatty acid) stable carbon and nitrogen isotope data to better understand trophic relationships and, ultimately, improving our ecological network analysis.

PROJECTS: BONUS BaltHealth, EcoStress, EnviStress, EiderHealth, NOW, NØG, UNEXPECTED

Ecological Network Analysis (ENA)

An ecological network is a representation of the abiotic and biotic interactions in a food web or ecosystem. Ecological network analysis offers a theoretical framework in which an ecological system can be quantitatively described, thus allowing for a mechanistic framework to investigate stressor-induced changes on the ecosystem level.    

More information

An ecological network is a representation of the biotic interactions in an ecosystem in which species (nodes) are connected by interactions (links). In our research, we focus on trophic ecological networks, i.e. food webs, that are directed (fluxes rather than simply relationships) and weighted (including the intensity of the flux). As such, we reconstruct the spatial and temporal variation in the flux of energy, contaminants and zoonoses through the food web. Of particular focus is how the network structure and dynamics may be influenced by anthropogenic activities and a changing abiotic environment, such as global climate change. These research questions are specifically applied for the Arctic and Baltic ecosystems as these are relatively well-understood ecosystems within highly variable, highly stressed (naturally and anthropogenically) and rapidly changing environments. Finally, the construct of the Arctic and Baltic networks use triple- and compound-specific stable isotope mixing models.

CONTACT: Igor Eulaers (ie@bios.au.dk)

Agent-/Individual-Based modeling (IBM)

We use agent-based models to analyse the impacts of anthropogenic disturbances on marine mammal populations. In these models, the movements, energetics and survival of animals are simulated explicitly. We have used these models to investigate how harbour porpoise populations are affected by construction of offshore wind farms and by bycatch in fishing nets.


Agent-based modeling

Populations udvikling


Population-effect modelling is a tool to extrapolate effects we observe on an individual level to the more conservation-relevant population scale. Many types of models exist for this and all offer a framework to model stressor influences on demographic parameters important for population growth. We employ population-effect modelling in our research to evaluate the risk of contaminants on high exposure marine mammal populations in the North-Atlantic and Baltic.

More information

Population models are commonly used to evaluate the growth of populations of animals of interest given a certain set of observed demographic parameters. These models provide invaluable insight into how demographic parameters shape population dynamics over time. Though much data exists on how certain natural and anthropogenic stressors influence molecular, cellular and even individual level fitness, much less is available on how these effects translate to the population scale, which is more relevant for regulatory agencies and conservation efforts.

We use population models to extrapolate the effects we observe at multiple levels of organization during field wildlife health assessments, controlled feeding studies or in vitro experimentation. Our focus is on how contaminants influence certain physiological processes that have population-level consequences. We work with collaborators at the University of St-Andrews Sea Mammal Research Unit to implement novel population-effect models, which incorporate reproductive toxicity, immunotoxicity and pathogen exposure in high exposure marine mammal species. We also work in the Baltic, where multiple stressors, including contaminants, pathogens and environmental change, act in concert to influence the dynamics of local populations. 

CONTACT: Jean-Pierre Desforges (jpd@bios.au.dk) and Rune Dietz (rdi@bios.au.dk)

Movement Models

These models are used for describing how variation in animal behaviour is related to environmental conditions or to the state of the animals. Movement speeds and the convolutedness of movement tracks are, typically, related statistically to environmental parameters. Alternatively variations in home range sizes are used to describe variations in space use among animals. We have used this kind of models to describe how polar bear home range size is related to environmental variation and contaminant loads (Full article at Springer doi.org/10.1007/s00300-015-1876-8).

Spatial Distribution Models

These models are used for describing spatial variation in the distribution of species or in how they utilise resources. Known relationships between the presence (or absence) of a species and various environmental parameters are typically used to statistically predict the presence of the species in other areas. We have used spatial distribution models in a range of projects, including a project mapping the distribution of porpoises in the North Sea and to describe variations in seal foraging behaviour.

Physiologically-Based Pharmacokinetic (PBPK)

The PBPK model is a structural mathematical model comprising the tissues and organs of the body of mammals, each perfused by the blood circulatory system. The transport of chemicals in the body is described by mass balance differential equations that incorporate blood flows, compartment partitioning, and tissue volumes. We use this model to estimate critical body residues and critical daily doses for persistent organic pollutants to predict and assess the risk of health effects (immune, reproductive, cancer) from a mixture of environmental contaminants in birds, marine mammals, polar bears and humans.