Spatio-temporal scaling of environmental selection in pelagic predators
Over the last couple decades, scientists have gathered many tracking datasets of marine pelagic predators. These animals perform large scale, ocean wide movements and migrations. These predators vary greatly, in terms of size, physiology or trophic level. I use datasets collected over a 10-year period to understand the effect that different biological scales (animal size, trophic level...) can have on organism location predictions, and the impact of observational grain size on these predictions. In addition to tracking datasets, I use Eulerian NASA satellite products of the environment (e.g. temperature, light attenuation coefficient, chlorophyll a...) and computed Lagrangian features (Lagrangian Coherent Structures, a measure of aggregation in the oceans) from the HYCOM model. I investigate whether considering Eulerian and/or Lagrangian variables increases our understanding of marine animal movements.
Pelagic fish vision and vertical habitat
This project aims to investigate whether there is a correlation between fish eyes (size, relative growth rate) and their vertical habitat. To do so, we collect lots of fish eye measurements from preserved specimens (so far, mainly at the Smithsonian Museum) and compute physiological parameters for many different species. We can then relate these parameters to the typical habitat of fish in the water column. Preliminary results suggest that the relative eye growth rate increases as animals live deeper (less light means that fish need to invest more in their eyes), but only until a certain threshold before it decreases (when there is too little light available investing in a powerful eye is not worth it anymore). This project is a great way to include undergraduate students in my research, as they can collect all the data and perform the analyses themselves.
Development of new tracking technologies for large pelagic organisms
Because GPS signal and radio waves do not penetrate water, ARGOS or GPS technologies are useless for tracking subsurface animals, making the acquisition of movement data challenging for marine animals. Geolocation, the primary method for tracking underwater animals, uses underwater light measurements to back-calculate latitude and longitude from day length and local noon. This technique is notoriously unreliable and has low spatial resolution. This limitation hinders the use of geolocation tags to track underwater species’ movements. I attach animal geolocation tags (minipats) to underwater gliders to simulate the movements of marine animals. As I control the glider track, I know exactly where my tag is, using it to validate the tag results. I use these data to develop new geolocation algorithms that use Sea Surface Temperature and Diffuse Attenuation Coefficient measured from satellites to increase the accuracy and confidence intervals of geolocations estimates.
In addition, I am also part of a project that aims to develop tags to use sharks as ocean observing platforms. These tags will sample the water column and transmit data back to shore when sharks breach the surface. This means that we will be able to use sharks as near real-time data providers, to e.g. refine our hurricane prediction models (the specific goal of this project).
Vertical Migrations and the biological carbon pump
Diel Vertical Migration (DVM) is the daily movement of organisms up and down the water column. It is believed to be the biggest migration of the planet in terms of biomass. Simply put, organisms hide in darkness during daytime to avoid being eaten, and migrate to the surface at night, where more food is available. The predators of these migrating organisms follow them as well, to maximize their encounter rates. As a consequence, the optimal DVM pattern of each organism depends on its prey, predators and conspecifics in the system but also on their migrating strategy. I use game theoretic principles to explore the optimal migration patterns of all organisms of a food-web at once.
This is not only interesting in itself, but it has also consequences on the world's biogeochemical cycles. Organisms feeding at the surface and migrating down actively bring carbon to depths of 100m or more, where it is respired and excreted by the migrator itself, or consumed by their predators. This represents the active part of the biological pump, which has been estimated to represent between 16 and 30% of the total export flux associated with the biological pump. I was the first one to quantify the amount of carbon sequestered in the oceans by metazoans performing diel vertical migrations and by zooplankton performing seasonal vertical migrations.
Long-distance migrations
For animals migrating across entire ocean basins (sea turtles, whales...), a good management of arrival time and energy expenditure can be crucial parameters affecting their fitness. A good optimization of these parameters is key to reproductive success or good feeding opportunities. As these animals navigate in an environment subject to currents, the most efficient migration route is not necessarily a straight line but a path making a smart use of these currents. The most efficient path varies depending on the importance of an early arrival or the necessity to save energy. I developed a theoretical model that reproduces migration routes when provided with a current field, a time/energy trade-off to optimize and a risk-sensitivity parameter accounting for the level of risk that an individual is willing to take to accomplish a potentially successful migration. Then, I used GPS tag data to reconstruct the trade-offs that animals are optimizing for based on their migration routes. Among other results, we saw that different humpback whales were optimizing for very similar trade-offs, whereas green turtles displayed a wide inter-individual diversity of strategies.
Over the last couple decades, scientists have gathered many tracking datasets of marine pelagic predators. These animals perform large scale, ocean wide movements and migrations. These predators vary greatly, in terms of size, physiology or trophic level. I use datasets collected over a 10-year period to understand the effect that different biological scales (animal size, trophic level...) can have on organism location predictions, and the impact of observational grain size on these predictions. In addition to tracking datasets, I use Eulerian NASA satellite products of the environment (e.g. temperature, light attenuation coefficient, chlorophyll a...) and computed Lagrangian features (Lagrangian Coherent Structures, a measure of aggregation in the oceans) from the HYCOM model. I investigate whether considering Eulerian and/or Lagrangian variables increases our understanding of marine animal movements.
Pelagic fish vision and vertical habitat
This project aims to investigate whether there is a correlation between fish eyes (size, relative growth rate) and their vertical habitat. To do so, we collect lots of fish eye measurements from preserved specimens (so far, mainly at the Smithsonian Museum) and compute physiological parameters for many different species. We can then relate these parameters to the typical habitat of fish in the water column. Preliminary results suggest that the relative eye growth rate increases as animals live deeper (less light means that fish need to invest more in their eyes), but only until a certain threshold before it decreases (when there is too little light available investing in a powerful eye is not worth it anymore). This project is a great way to include undergraduate students in my research, as they can collect all the data and perform the analyses themselves.
Development of new tracking technologies for large pelagic organisms
Because GPS signal and radio waves do not penetrate water, ARGOS or GPS technologies are useless for tracking subsurface animals, making the acquisition of movement data challenging for marine animals. Geolocation, the primary method for tracking underwater animals, uses underwater light measurements to back-calculate latitude and longitude from day length and local noon. This technique is notoriously unreliable and has low spatial resolution. This limitation hinders the use of geolocation tags to track underwater species’ movements. I attach animal geolocation tags (minipats) to underwater gliders to simulate the movements of marine animals. As I control the glider track, I know exactly where my tag is, using it to validate the tag results. I use these data to develop new geolocation algorithms that use Sea Surface Temperature and Diffuse Attenuation Coefficient measured from satellites to increase the accuracy and confidence intervals of geolocations estimates.
In addition, I am also part of a project that aims to develop tags to use sharks as ocean observing platforms. These tags will sample the water column and transmit data back to shore when sharks breach the surface. This means that we will be able to use sharks as near real-time data providers, to e.g. refine our hurricane prediction models (the specific goal of this project).
Vertical Migrations and the biological carbon pump
Diel Vertical Migration (DVM) is the daily movement of organisms up and down the water column. It is believed to be the biggest migration of the planet in terms of biomass. Simply put, organisms hide in darkness during daytime to avoid being eaten, and migrate to the surface at night, where more food is available. The predators of these migrating organisms follow them as well, to maximize their encounter rates. As a consequence, the optimal DVM pattern of each organism depends on its prey, predators and conspecifics in the system but also on their migrating strategy. I use game theoretic principles to explore the optimal migration patterns of all organisms of a food-web at once.
This is not only interesting in itself, but it has also consequences on the world's biogeochemical cycles. Organisms feeding at the surface and migrating down actively bring carbon to depths of 100m or more, where it is respired and excreted by the migrator itself, or consumed by their predators. This represents the active part of the biological pump, which has been estimated to represent between 16 and 30% of the total export flux associated with the biological pump. I was the first one to quantify the amount of carbon sequestered in the oceans by metazoans performing diel vertical migrations and by zooplankton performing seasonal vertical migrations.
Long-distance migrations
For animals migrating across entire ocean basins (sea turtles, whales...), a good management of arrival time and energy expenditure can be crucial parameters affecting their fitness. A good optimization of these parameters is key to reproductive success or good feeding opportunities. As these animals navigate in an environment subject to currents, the most efficient migration route is not necessarily a straight line but a path making a smart use of these currents. The most efficient path varies depending on the importance of an early arrival or the necessity to save energy. I developed a theoretical model that reproduces migration routes when provided with a current field, a time/energy trade-off to optimize and a risk-sensitivity parameter accounting for the level of risk that an individual is willing to take to accomplish a potentially successful migration. Then, I used GPS tag data to reconstruct the trade-offs that animals are optimizing for based on their migration routes. Among other results, we saw that different humpback whales were optimizing for very similar trade-offs, whereas green turtles displayed a wide inter-individual diversity of strategies.
