Evaluation of a Doppler sonar system for measurements of fish swimming velocity

Abstract

Good measurements of fish swimming velocity would add another piece of valuable scientific information to the toolkit of the fisheries acoustician; however, previous investigations of Doppler sonar measurements of fish velocity have been relatively sparse, and no previous study has explored the possibilities available with newer coherent Doppler sonar systems. Comprehensive field and laboratory studies were undertaken to characterize the performance of a 250-kHz, 10-kHz bandwidth coherent Doppler sonar when applied to measurements of fish swimming velocity. A computer model was designed to assist in understanding experimental results and broaden the range of parameters investigated. The quality of velocity estimates was quantified by calculating the error and the standard deviation for estimates of Doppler velocity for individual targets. The effect of sonar parameters (pulse length, pulse coding, lag, ping rate, bandwidth) and environmental conditions (target and water velocity, target spacing, signal-to-noise ratio) on the quality of velocity estimates was examined through four experiments with simulated fish targets, three experiments with live fish targets, and the computer model. With signal-to-noise ratio ranging from 0 dB to 75 dB, typical velocity errors of 1--2 cm s -1 and standard deviations of 5--13 cm s-1 were realized with the test instrument, amounting to 4--11% uncertainty if a fish swimming velocity of 30 cm s-1 is assumed. Experimental, theoretical, and model results taken together suggest that useful estimates of fish swimming velocity could be obtained using a coherent Doppler sonar with the following specifications: a transmit frequency of 200--300 kHz with a 16% bandwidth, a beamwidth of 4--6°, and several beams to measure multiple components of fish velocity.