Assessing a dual-frequency identification sonars’ fish-counting accuracy, precision, and turbid river range capability

Abstract

Accurately assessing migrating salmon populations in turbid rivers with hydroacoustics is challenging. Using single, dual, or split-beam sonars, difficulties occur fitting acoustic beams between the river's narrow boundaries, distinguishing fish from nonfish echoes, and resolving individual fish at high densities. To address these issues, the fish-counting capability of a dual-frequency identification sonar (DIDSON), which produces high resolution, video-like images, was assessed. In a clear river, fish counts generated from a DIDSON, an echo counter, split-beam sonar, and video were compared to visual counts from a tower, a method frequently used to ground-truth sonars. The DIDSON and tower counts were very similar and showed the strongest agreement and least variability compared to the other methods. In a highly turbid river, the DIDSON's maximum detection range for a 10.16 cm spherical target was 17 m, less than absorption and wave spreading losses predict, and 26 m in clear water. Unlike tower and video methods, the DIDSON was not limited by surface disturbances or turbidity. DIDSON advantages over other sonars include: better target resolution; wider viewing angle; better coverage of the water column; accurate direction of travel; and simpler to aim and operate.