My main research interests are:
|
- Elucidating the distributions of marine megafauna species
Marine megafauna species are large high-trophic level predators characterized by vast geographical ranges. These species often associate with eddies, fronts, and upwelling systems characterized by enhanced biological productivity. During my PhD at the University of La Rochelle and my postdoc at Duke University, I investigated patterns in megafauna distribution across tropical and temperate ecosystems. For example, in a recent study published in Ecography, my coauthors and I found that cetaceans' relationships with habitat may not hold between the two sides of the North Atlantic Ocean. A possible reason for these differences is spatial population structure that may lead to different habitat specializations between regions.
|
Contrasted relationships between fin whale’s probability of presence and log primary productivity in the western (red) eastern (blue) North Atlantic.
|
- Predicting marine megafauna densities in sparsely surveyed regions
Conducting visual surveys to estimate the abundance of megafauna populations is resource-intensive and challenging, especially far from the coasts. Yet, abundance and density estimates are crucial to assess the impacts of human activities on megafauna populations in the high seas. During my postdoc at Duke University, I developed a habitat modeling approach to geographically extrapolate cetacean densities in the North Atlantic high seas. These modeling results are being used by the U.S. Navy to quantitatively assess the impacts of military sonars on cetacean populations. This study is published in Conservation Biology and all accompanying results are freely available through the OBIS-SEAMAP repository.
|
Predicted densities of striped dolphin extrapolated from a habitat model in the North Atlantic high seas
|
- Automating the monitoring of marine megafauna
Image-based surveys and artificial intelligence are bringing great promise for the automated detection of wildlife. Together, they have the potential to revolutionize megafauna abundance estimation by providing a cost-efficient, reproducible, and accurate framework for monitoring populations. A core aim of my Marie Curie project is to automate marine megafauna monitoring by combining image-based aerial surveys and deep learning algorithms. My pilot study is taking place at a lagoon in New Caledonia (Southwest Pacific).
|
Automated detection of a manta ray with a deep learning algortihm
|