Back to search

MARINFORSK-Marine ressurser og miljø

COEXIST - Condition and energy expenditure estimates from free-ranging marine mammals

Alternative title: COEXIST - Estimering av kroppskondisjon og energiforbruk hos frittlevende sjøpattedyr

Awarded: NOK 3.2 mill.

Project Manager:

Project Number:

234411

Application Type:

Project Period:

2014 - 2017

Partner countries:

Recent development and miniaturization of instruments that can be attached to marine mammals to collect and transmit information about their movements and diving ability has revolutionized our understanding of these important top predators in marine ecosystems. Together with estimates of their population sizes, food consumption and diet, such instruments have the potential to provide important information allowing us to understand their role in marine ecosystems in general, and their impacts on prey popul ations in particular. However, this potential has not been fully realized, due to limitations in instrument capability and analytical sophistication. This project built on recent improvements in onboard instrumentation, involving 3D activity sensors (accelerometers) and underwater cameras, to develop approaches for measuring the activity patterns and body condition of animals, while simultaneously obtaining images of their captured prey. We have conducted a series of captive experiments on harp seals to provide controlled validation data that are now being used to develop onboard algorithms to track changes in body condition and energy expenditure from accelerometer-based activity patterns. We have also initiated the collection of onboard video footage of live prey capture attempts, simultaneously to collecting accelerometer data. These combined datasets will now form the basis for refining acceleration-based detections of prey capture attempts that are currently being implemented on animal-borne instruments. We have also started constructing a library of images of potential prey species that will be used during a subsequent project to develop automated image classification and object recognition systems that can be implemented directly onboard the animal-borne instruments. Together, our contributions are proving crucial for the development of sophisticated animal-borne instrument capable of collecting, processing and compressing large amounts of data on activity patterns and prey capture from free-living marine mammals in their oceanic environment, to provide continuous information about their body condition, energy expenditure and diet via satellite from important feeding areas.

Through a combination of methodology refinement and development of novel and innovative approaches, we will develop the analytical tools needed to estimate the impact of marine predators on fish stocks and the effects of variations in foraging success on the condition of the predators themselves, with greater accuracy. While the proposal focuses on the Northeast Atlantic harp seal (Pagophilus groenlandicus), the methods developed here will have wide application to the study of energetics and prey composit ion of top trophic marine vertebrates in a range of ecosystems. The approach is based on continuous real-time monitoring of energy expenditure, body composition, prey capture rates and diet composition of individuals while at sea. It relies on combining a nimal-borne acceleration and camera data loggers, and focuses on the development of novel and innovative approaches for deriving indices of energy expenditure, prey capture and diet composition from these instruments. This requires detailed studies of ene rgy expenditure, swimming and diving behaviour and prey capture in a captive controlled setup, in combination with developing an underlying bioenergetic model to be used to determine energy allocation and changes in condition from acceleration data. We wi ll also develop indices of prey capture events from acceleration data, capable of triggering the camera at times of interest. In order for the next-generation animal-borne instrument to relay relevant indices in real-time via limited bandwidth satellite s ystems, the high-volume data provided by accelerometers and cameras will have to be compressed into simple energy expenditure and prey classification indices. This study will lay the necessary groundwork by providing such indices for energy expenditure, a nd by initiating the collection of known-species prey images that can be used for subsequent development of appropriate object detection and classification algorithms.

Publications from Cristin

No publications found

No publications found

No publications found

No publications found

Funding scheme:

MARINFORSK-Marine ressurser og miljø