Density of reef sharks estimated by applying an agent-based model to video surveys
School of Natural Sciences / Centre for Ecosystem Management
Policies on harvesting and conservation are developed in response to information about trends in the abundance of species, so making accurate estimates of abundance is important. However, estimating the abundance of sparsely distributed species is challenging, especially where direct observations are difficult. We collected observations of blacktip reef sharks Carcharhinus melanopterus by using remote underwater video cameras, and developed an agent-based model to generate estimates of the density of sharks from the frequency of observations made using the video. We augmented these observations with diel patterns in detections in different habitats of C. melanopterus with surgically implanted acoustic transmitters. Median estimates of density ranged from 2-9 ind. km-2 at noon to 20-90 ind. km-2 at dusk, depending on whether modelled movement paths were random or directional. These estimates suggest that individuals might exhibit diel patterns in movement, with directional movement to the reef flat during dusk. Data from tagged individuals supported this hypothesis, with more detections recorded from reef flat habitat during early evening and early morning than at other times of day. Estimates of density were among the highest reported for C. melanopterus. The agent-based model approach is flexible, and can be extended to simulate a range of behaviours and other types of observations.