Quantifying the acoustics packing density of fish schools with a multi-beam sonar

Abstract

Multi-beam (swath) sonar systems provide the capability to ensonify an entire aggregation of fish in a single pass. However, estimation of abundance and discrimination between species via the use of target strength are considerably more complex than using traditional echosounders, because they ensonify targets at a much wider range of incidence angles. The beam pattern and along beam resolution of multi-beam swaths can produce individual sample volumes that are of similarmagnitude to an individual fish (particularly for large fish, say >1m in length). If individual fish can be resolved, (either as a single fish within a sample, or as multiple contiguous samples that delineate a single fish), and if one assumes that this situation applies to the whole school, acoustic packing density can be determined by dividing the volume of the school by the number of detected acoustic targets. This estimate is proportional to the actual packing density of the fish, defined asthe number of fish per unit volume of water. Acoustic backscatter of fish from a number of schools comprising different species were collected off Perth, in 2005 and 2007, using a Reson Seabat 8125 and 7125 respectively. Nearest neighbour distances of between 1 and 3 body lengths were observed and packing density of acoustic targets showed distinct variation between some species. However, schools of the same species also displayed different acoustic packing densities at differentstages of their growth and development. Such differences were more difficult to observe in schools of fewer fish because the variations in packing density had less impact on the overall volume of the smaller schools associated with fewer fish. Therefore discrimination between species was only deemed possible when surveying two species of different sized fish at the same time. Video ground truth data is recommended to confirm species composition whatever the type of schoolobserved.