Measurement of fish velocity using Doppler sonar

Velocity bias in Doppler current profiler data caused by fish movements is a recognized problem. The authors take the alternative perspective and investigate to what degree Doppler profilers can determine fish velocities. Field data collected with an RD Instruments 307 kHz Workhorse Acoustic Doppler Current Profiler (ADCP) demonstrate an ability to measure the speed and swimming direction of large herring schools off the coast of Norway. For smaller fish concentrations, direct velocity measurements are not possible. In order to investigate how the broad-band system responds to discrete targets, a series of tests were undertaken in a tow-tank facility. Styrofoam spheres of from 2.5 to 12.5 cm in diameter were positioned alone and in target arrays of up to 27 spheres. Target locations were identified by combining echo amplitude and signal autocorrelation data. For the system configuration tested, single ping estimates of (horizontal) velocity had a standard deviation as low as 10 cm s^-1. This accuracy is substantially better than the single ping accuracy of ≃40 cm s^-1 expected for the broad-band system as configured and suggests that for distinct, isolated targets, the approach is equivalent to fully coherent Doppler processing