Agent-based models (ABMs) are a useful tool for predicting the impacts of stressors on individual seabirds.
Agent-based (or individual-based) models (ABMs) are a useful tool for predicting the impacts of environmental change on ecological systems. We can use ABMs to simulate various aspects of a species’ ecology, such as their natural movement, behaviour, growth, or reproduction. Then, by changing the environment or input variables of the model, we can simulate how a species may respond to environmental change.
Part of the MARCIS project is focussed on investigating the impact of anthropogenic stressors on seabird populations. As such, we have drawn upon the large resources of tracking data collected within the SEATRACK project to model the non-breeding season movement and behaviour of six seabird species (Atlantic puffin, Brünnich's guillemot, common guillemot, little auk, black-legged kittiwake, and northern fulmar) from multiple populations. Using population averages of these metrics, we will simulate daily movement and behaviour of individuals seabirds (‘agents’) in an ABM throughout the non-breeding season. Within this ABM, we will also simulate interactions with marine stressors to predict the effects that these stressors have on individuals. Below is an example of an agent-based model of common guillemots on day 1, 100 and 200. Each agent is numbered and has an indicator of body mass in orange. In these simulations, Agent 1 survived to day 200; Agent 2 suffered a lethal impact with a stressor on day 27 and died; and Agent 3’s body mass dropped below the critical threshold on day 156, leading to mortality.
An example of an agent-based model of common guillemots.
Marine stressors can have lethal and sub-lethal effects on seabirds. Lethal effects, such as collision with wind turbines or bycatch, are somewhat simpler to model, as they have a clear outcome (mortality). Sub-lethal effects, however, can have energetic impacts such as reduced food intake or increased energy expenditure. These energetic impacts can have subsequent implications for individual survival or future breeding success, so are equally important to model. We will use the simulated behaviour data to indicate energy expenditure and account for the energetic impacts of stressor interactions. We will subsequently convert energy expenditure into daily mass change and if this mass value falls below a critical threshold, we will assume mortality has occurred.
The modelling framework for the agent-based models to be used in MARCIS. Start and end data are in blue ovals, questions are in orange diamonds, and processed are in yellow rectangles.
A key advantage of ABMs is that you can simulate the effects of multiple stressors at once, which enables us to account for the cumulative effects of sub-lethal stressor interactions (i.e. the impact of multiple stressors occurring at once). We will therefore run combinations of different stressors to assess the individual-level impacts of these stressors both in isolation (e.g. offshore wind farms) and in combination with other stressors (e.g. offshore wind farms plus fisheries bycatch), leading to a cumulative effect. By running multiple simulations of each of these stressor combinations, we will be able to scale up these individual effects to determine the population-level effects of each of these stressors, which will greatly improve our understanding of the cumulative impacts of these stressors on seabird populations.
ABMs have previously been used to predict the impacts of stressors on seabird populations during the breeding season, but not yet during the non-breeding season. The non-breeding season is a particularly important period of the annual cycle for polar and temperate seabirds, as they usually experience their greatest levels of mortality at this time. Our approach for developing a non-breeding season ABM has been enabled thanks to the extensive non-breeding season tracking dataset collected by the SEATRACK project. We hope that, as well as answering the core questions addressed by MARCIS, the novel modelling approaches that we develop in the project will be able to be adapted in the future to investigate related questions surrounding the impacts of human-based activities on seabird populations.
References
Pollock, C., 2022. Modelling breeding season foraging and tracking autumn migrations to fill knowledge gaps in gannet ecology relating to impacts of offshore wind farms. Doctoral dissertation.
Searle, K., Mobbs, D., Butler, A., Bogdanova, M., Freeman, S., Wanless, S. and Daunt, F., 2014. Population Consequences of Displacement from Proposed Offshore Wind Energy Developments for Seabirds Breeding at Scottish SPAs (CR/2012/03). Scottish Marine and Freshwater Science, 5(13).
Searle, K.R., Mobbs, D.C., Butler, A., Furness, R.W., Trinder, M.N. and Daunt, F., 2018. Finding out the Fate of Displaced Birds (FCR/2015/19). Marine Scotland Science.