Frédéric Dulude-de Broin
Department of Biology
Pierre Legagneux (Regular member)
IntroductionPredation is central to the regulation of many arctic ecosystems. The movement of predators, influenced by the physical landscape and resource distribution, creates spatial variation in predation risk for the prey. This heterogeneity of risk and the way prey respond to it could be one of the key mechanisms that shape local biodiversity in the arctic tundra. The physical landscape such as refugia or restricted availability of certain habitats could also influence the distribution of prey through its effect on predator movement and prey habitat use. ObjectivesThe main objective of my project is to assess how predation risk and the physical landscape influence arctic biodiversity distribution in a terrestrial community. Specifically, I will i) model the spatial variation in predation risk using an agent-based model approach, ii) assess how snow melt patterns influence the distribution of the prey community by restricting available nesting habitats, and iii) evaluate the relationship between predation risk and prey distribution while accounting for the potential unavailability of safe habitats due to physical constraints like snow cover. Study sitesMy project will take place on Bylot Island in Nunavut and benefit from the long-term ecosystem monitoring program started 25 years ago. In this system, arctic fox predation drives the cyclic abundance of lemmings, the reproduction of geese and the population dynamics of more than 30 accidental prey species. Foxes are equipped with high resolution GPS collars and we assess predation risk and prey distribution in the field through artificial nest experiments and biodiversity surveys.Material and methodsI will use fox GPS collars to model spatial variation in predation risk. The challenge is to assess risk for the entire ecosystem, including places where not all foxes are collared. I will use reinforcement learning, a subfield of artificial intelligence, to train a model that mimics fox movements. Based on this model, I will simulate the movement of individual foxes in the landscape to record the intensity of use per spatial unit (25mx25m cells). Second, I will use daily satellites images available since 2010 to evaluate how snow melt patterns affects nest distribution. Finally, while accounting for the effect of the physical environment on nest sites availability, I will assess how predation risk shapes the distribution of avian biodiversity across the landscape. Expected resultsThis project will provide a better understanding of arctic biodiversity distribution and contribute new tools for the study of animal movement by developing an innovative approach using reinforcement learning.
Clermont, J., Grenier-Potvin, A., Duchesne, É., Couchoux, C., Dulude-de Broin, F., Beardsell, A., Bêty, J., Berteaux, D., 2021. The predator activity landscape predicts the anti-predator behavior and distribution of prey in a tundra community. Ecosphere, 12(12), e03858. DOI: 10.1002/ecs2.3858.
LeTourneux, F., Grandmont, T., Dulude-de Broin, F., Martin, M.-C., Lefebvre, J., Kato, A., Bêty, J., Gauthier, G., Legagneux, P., 2021. COVID19-induced reduction in human disturbance enhances fattening of an overabundant goose species. Biological Conservation, 255, 108968. DOI: 10.1016/j.biocon.2021.108968.
Dulude-de Broin, F., Hamel, S., Mastromonaco, G.F., Côté, S.D., 2020. Predation risk and mountain goat reproduction: Evidence for stress-induced breeding suppression in a wild ungulate. Functional Ecology, 34(5): 1003-1014. DOI: 10.1111/1365-2435.13514.
Dulude-de Broin, F., Côté, S.D., Whiteside, D.P., Mastromonaco, G.F., 2019. Faecal metabolites and hair cortisol as biological markers of HPA-axis activity in the Rocky mountain goat. General and Comparative Endocrinology, 280: 147-157. DOI: 10.1016/j.ygcen.2019.04.022.
© 2023 Centre for northern studies - All rights reserved