Beaked whale group deep diving behavior from passive acoustic monitoring
Abstract
Beaked whales are of particular interest to the Navy, being common on Navy testing ranges and potentially affected by human activities, having been associated with stranding events following sonar use. They occur typically in small groups and spend a large proportion of time at depth, undertaking long, deep dive foraging dives during which distinct echolocation clicks are produced. Due to their elusive behaviour, knowledge about their behaviour at depth is not possible to obtain using conventional visual observation methods. Extensive information on sound production and dive behaviour of individual animals has been obtained via tagging experiments. However, the way animals within a group interact and coordinate at depth is unknown. This has clear implications to understand and interpret individual tag data, which in turn means that it is unclear how effects at an individual level might be propagated to group and population level effects. Blainville’s beaked whales (Mesoplodon densirostris) echolocation clicks are routinely detected on the AUTEC range, Bahamas, showing that despite high Navy activity levels, animals do use the area year round. Understanding this process seems fundamental, as it would help to tease apart whether AUTEC animals are less sensitive to operations (e.g. due to habituation), if there is any advantage that overthrows the disadvantage of being exposed to sonar (e.g. abundant prey), or if effects are actually less than one might predict based on current reports. This makes AUTEC an ideal real world laboratory for understanding potential impacts. Here we propose to combine click detections with state-of-the-art association, localization and tracking algorithms to provide a privileged window of observation on beaked whale group behaviour at depth, and to conceptualize and parameterize models to predict group size and group movement from passive acoustic monitoring data. We will develop click association procedures (task1) that will allow us to routinely associate click detections on AUTEC hydrophones into group deep dives. These sets of associations will then be used as inputs for two distinct analysis: (1) they will be used within multi-hypothesis tracking algorithms, which have shown promising preliminary results to obtain positions of multiple animals at depth (task 1) and (2) they will be used jointly with some independently verified group sizes to create a model that will predict group size from the acoustical footprint of a group (task 2). The individual animal localizations will be used to develop a model for the movement of animals at depth, by extending to 3D methods for movement of animal within a group (task 3). Outputs of the model for estimating group sizes from deep dive acoustical footprints will allow better support for AUTEC range activities by providing near real time group size information and location, and will further allow a mechanism for constant density estimation on the range via group/dive counting procedure. Visually verified group sizes that can be associated with AUTEC range detected deep dives are rare, and hence we propose dedicated additional field work (optional task 4) dedicated to obtain such observations. These will provide a much stronger basis for inferences for task 2 and 3.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2016
- Source ID
- N000141512648
Entities
People
- Len Thomas
Organizations
- Office of Naval Research
- United States Navy
- University of St Andrews