Theoni Photopoulou will present the Department of Statistical Science seminar with a talk entitled, "Modelling movement patterns of great white sharks using acoustic detections".  

Theoni Photopoulou is a Postdoctoral Researcher under SEEC, Department of Statistical Sciences, University of Cape Town and Institute for Coastal and Marine Research, Nelson Mandela University, South Africa

Abstract: Acoustic monitoring is being used more and more in the marine environment to gather information on species occurrence, animal density and ranging patterns. An array of acoustic receivers was deployed in False Bay, South Africa, in 2005 as part of a tagging programme to equip great white sharks with acoustic transmitters to monitor their presence in the Bay. The land area adjacent to False Bay is the very developed suburban sprawl of Cape Town, a major South African city, and is heavily populated. The beaches along the north coast of False Bay are used throughout the year with a peak over the summer months. White sharks are perceived as a threat to water users and a better understanding of shark movement patterns will help both with (1) mitigating conflict between sharks and humans, and (2) improving our knowledge of the ecology of white sharks in False Bay. Acoustic detections of fish have been used to study residence patterns, seasonality and association patterns but have seldom been used to study individual movement patterns. The sequence of acoustic detections can be thought of as a time series and offers opportunities to understand movement patterns. Knowing about individual trajectories may help map the risk of interactions between white sharks and humans as well as provide insight into the way sharks use their environment. From the raw data we know when the sharks are in the vicinity of receivers but have no information on their movements away from receivers. We processed the data into detection events using a set of criteria, and modelled them using a hidden process approach. In this framework, the observations are used to do inference on the underlying process of interest: the sharks’ spatial locations through time irrespective of receiver location. We treat this as a spatial process whereby the underlying states are spatial locations and, to make the problem more tractable, we considered the study area as a grid, where each grid cell is a state. Filling in the gaps in our knowledge of where sharks spend time outside of the range of current receivers may also inform future acoustic monitoring designs to maximise effectiveness of the technique for this population.