Uncertainty in fish counting using an echo-counting technique as applied to data from a single-beam sonar

Abstract

A model of fish distribution in time and space and a single-beam sonar model are presented. Simulated sonar data are obtained and analyzed using the echo-counting method to determine the estimated number of fish. The results show that (1) when the fish rate is less than 1 fish/s, the error is within plus minus 15% and fish are overcounted more often than undercounted, (2) this method underestimates the number of fish by 57% of the actual number of fish for a fish rate of 5 fish/s, (3) fish counts are dominated by the noise if the threshold is lower than the noise level, (4) by varying the ping rate, the error could be as much as 72% for a fish rate of 10 fish/s and a ping rate of 10 pings/s, (5) by varying the pulse width, the error could be as much as 80% for a fish rate of 10 fish/s and a pulse width of 1.0 ms.