Seabirds are valuable ecological sentinels. Simulating their movements and behaviour and understanding their biological state are essential lines of research for analysing their adaptation and resilience in the face of environmental upheaval. In this context, Noémie Muquet, a doctoral student at the University of Montpellier, is developing a digital twin dedicated to seabirds.
By Carole Saout-Grit et Laurie Henry
Cover photo: Masked booby (Sula dactylatra) © Sophie Lanco
Present in all the world’s oceans, from coastal areas to the open sea, seabirds occupy a key position in the marine food chain. Their diet consists of prey at various levels of the food chain, making them particularly sensitive to variations in ocean productivity. Their biology is characterised by a dependence on land for reproduction and on the sea for food, making them privileged witnesses of the interactions between the two environments.
Protecting seabirds in the face of an ocean under pressure
Despite their importance, seabirds are among the most endangered groups of species. Worldwide, 31% of species are classified as threatened and almost half are declining in number. This worrying decline can be explained by an accumulation of anthropogenic pressures: scarcity of prey due to overfishing or changes in marine ecosystems, accidental capture by fishing gear, and plastic pollution, which seriously impacts their health, development and sometimes their survival. Added to this is the introduction of invasive species on breeding sites, disrupting nesting and threatening the survival of chicks, and the emergence of new marine infrastructures such as offshore wind farms that alter their habitats and migratory routes.
In this context, protecting these species has become a matter of urgency to ensure their survival and conservation over time. Easy to observe and equip, seabirds are now at the heart of vast bioloding programmes* that allow scientists to track their movements, dives and energy expenditure using miniaturised on-board sensors.
Biologging, AI and ecological modelling for monitoring seabirds
Planned for the period 2023-2026, the JUNOM thesis led by Noémie Muquet, a doctoral student at the University of Montpellier, aims to develop a digital twin dedicated to seabirds. This innovative project combines biologging, artificial intelligence and ecological modelling to design deep learning models and simulate the behaviour of seabirds in the face of human pressures and climate change.
These three years of research place Noémie at the intersection between marine sciences and artificial intelligence. With an initial background in fundamental physics, she is using her skills to develop a new vision of the issues involved in ecological modelling, with the aim of developing their applications for the conservation of the ocean and its biodiversity.
© Noémie Muquet
The species monitored as part of JUNOM reflect their diversity and vulnerability: several species of booby such as the masked booby (Sula dactylatra), the brown booby (Sula leucogaster), the red-footed booby (Sula sula) and the variated booby (Sula variegata); two species of tropicbird, the red-billed tropicbird (Phaethon aethereus) and the yellow-billed tropicbird (Phaethon lepturus), or the Bougainville cormorant (Leucocarbo bougainvillii).
15 years of data and innovative ecological modelling
All these species have been the subject of intensive monitoring in Brazil and Peru, during fifteen years of field campaigns. The data collected covers a wide range of information, from GPS trajectories that trace their daily movements, to accelerometry data that reveals their flight and diving behaviour and their energy expenditure.
While this monitoring provides a wealth of valuable information, the data is often heterogeneous, collected at varying frequencies depending on the species, the years or the technologies deployed. Some periods or species are under-documented due to a lack of field campaigns, which complicates overall analyses. There is also currently no unified framework for fully exploiting this mass of data and understanding behaviours in their entirety, by linking them to environmental conditions and anthropogenic pressures. This fragmentation prevents the availability of a coherent predictive vision, which is nevertheless essential for anticipating the future trajectories of these fragile populations.
To meet this need for forecasting and scenario planning, Noémie’s thesis is based on the latest deep learning technologies. The digital twin will realistically simulate the trajectories of birds at sea, integrating not only their horizontal movements, but also their diving behaviour, flight strategies and rest cycles. These behaviours will be associated with a precise estimate of their energy expenditure, in order to better understand the bioenergetic constraints that determine their survival and reproduction.
Brown bobby (Sula leucogaster) © Sophie Lanco
An innovative thesis to simulate the future behaviour of seabirds
Noémie is writing her thesis at the University of Montpellier, within the MARBEC Joint Research Unit and the UMMISCO International Joint Unit. She is co-supervised by Sophie Lanco (IRD) and Jean-Daniel Zucker (IRD). Funding by the Priority Research Programme (PPR) Ocean & Climate, it contributes to the challenge of developing innovative, multi-disciplinary, multi-parametric, multi-scale and multi-stakeholder observation and modelling programmes.
JUNOM aims to go far beyond simple historical reconstruction. By integrating environmental variables such as surface temperature, marine currents and prey availability, as well as anthropogenic pressures such as fishing zones and offshore wind farms, the digital twin will be able to simulate the future behaviour of seabirds in response to different climatic and socio-economic scenarios. This scenario-building capacity will make it a valuable tool for managers of marine protected areas, who will be able to assess the effects of their management decisions on the conservation of populations. The tool will also provide valuable assistance to research in movement ecology, by offering an integrated approach capable of linking behavioural dynamics, energy constraints and environmental contexts.
By combining empirical data from fifteen years of monitoring, artificial intelligence simulation models and prospective scenario planning, JUNOM represents a real methodological innovation in the service of marine conservation. This digital twin of seabirds thus paves the way for a new generation of tools to anticipate, understand and manage the interactions between biodiversity and global change.
* Biologging: the use of tags attached to animals to record aspects of their behaviour, physiology or environment.
Référence : Noémie Muquet, « JUNOM – Vers un jumeau numérique des oiseaux marins : modèles d’apprentissage profond pour la scénarisation des impacts du changement climatique et des pressions anthropiques multiples », thèse 2023-2026
Contact : noemie.muquet@ird.fr
3 Questions to Noémie Muquet
Why did you want to do a thesis in marine sciences?
I was always attracted by the idea of doing a thesis, as much for the intellectual challenge as for the personal enrichment it represents. However, this choice marked a real turning point in my career: having trained in fundamental physics, I was not initially destined for marine sciences. But a combination of circumstances changed things, just as I was thinking about a change of direction that would allow me to better align my professional activity with my convictions, particularly in terms of ecology. In fact, it was rather the marine sciences that came to me!
What made you want to apply for this thesis subject? What were your motivations?
Having been passionate about birds since I was a child, I was immediately excited about this thesis subject, even if the intersection between marine sciences and artificial intelligence is far from my initial field of training. Sophie Lanco and Jean-Daniel Zucker decided to put their trust in me, especially after I explained my motivation to use my background in mathematics and physics to bring a new perspective to the problems of ecological modelling and then develop its applications to the conservation of the ocean and its biodiversity.
How do you see your future after this thesis?
Ideally, I would like to continue after my thesis with a post-doctorate, if the opportunity arises. I am also very drawn to scientific mediation and open science, as well as to questions of governance. My goal is to pursue a career in scientific research while finding a balance between theory, concrete applications and the transmission of knowledge to the general public.