UC San Diego

The western North Atlantic is a dynamic region characterized by the Gulf Stream western boundary current and inhabited by a diverse host of odontocete, or toothed whale, top predators. Their habitats are highly exploited by commercial fisheries, shipping, marine energy extraction, and naval exercises, subjecting them to a variety of potentially harmful interactions. Many of these species remain poorly understood due to the difficulties of observing them in the pelagic environment. Their habitat utilization and the impacts of anthropogenic activities are not well known. Over the past decade, passive acoustic data has become increasingly utilized for the study of a wide variety of marine animals, and offers several advantages over traditional line-transect visual survey methods. Passive acoustic devices can be deployed at offshore monitoring sites for long periods of time, enabling detection of even rare and cryptic species across seasons and sea states, and without altering animal behaviors. Here we utilized a large passive acoustic data set collected across a latitudinal habitat gradient in the western North Atlantic to address fundamental knowledge gaps in odontocete ecology. I approached the problem of discriminating between species based on spectral and temporal features of echolocation clicks by using machine learning to identify novel click types, and then matching these click types to species using spatiotemporal correlates. I was able to identify novel click types associated with short-beaked common dolphins, Risso’s dolphins, and short-finned pilot whales in this way. Next I characterized temporal patterns in presence and activity for ten different species across our monitoring sites at three different temporal scales: seasonal, lunar, and diel. I observed spatiotemporal separation of apparent competitors, and complex behavioral patterns modulated by interactions between the seasonal, lunar, and diel cycles. Finally I investigated the relationships between species presence and oceanographic covariates to predict habitat suitability across the region, and explored niche partitioning between potentially competitive species. The insights gained here significantly advance our understanding of toothed whale ecology in this region, and can be used for more effective population assessments and management in the face of anthropogenic threats and climate change.