Extracting fish and water velocity from Doppler profiler data

Abstract

Doppler current profilers are optimized for measuring water velocities, but have the demonstrated capability to measure fish swimming speeds. This is possible when fish form schools that are large enough for all multiple Doppler sonar beams to sample the fish speeds at the same time. In situations where fish are not present in at least three acoustic beams, it is impossible to extract fish velocity with the data-processing algorithms normally used to extract water velocity. We present an alternative method of analysing Doppler sonar data that treats data from individual acoustic beams independently, so that velocities can be extracted when fish appear intermittently in the sonar beams. The method determines the variance for each velocity estimate so that data averaging can be adjusted to achieve the desired accuracy. The algorithm is applied to extract both water and fish velocities from Doppler profiler observations of overwintering Atlantic cod (Gadus morhua) in Smith Sound, Newfoundland. Currents in this enclosed coastal area are slow (∼10 cm s-1), and the fish appear to move passively with the water much of the time. However, there are times when the fish have velocities different from those of the water, and profiles averaged over 20 d show clear differences in fish and water velocities. © 2009 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved.