Quantifying fish avoidance of small acoustic survey vessels in boreal lakes and reservoirs

Abstract

AbstractMobile hydroacoustic surveys are increasingly used to assess the distribution and abundance of freshwater fish; yet, fish may avoid moving vessels, potentially introducing bias in these assessments. In this study, avoidance in boreal lakes and reservoirs was quantified by developing a simple method based on paired drift:drift (D:D) and drift:motor (D:M) transects. Two systems in eastern Manitoba, Canada were studied: Lac du Bonnet reservoir and Nopiming. Acoustic data were collected using a digital DTX echosounder (BioSonics, Seattle, WA, USA), with a downward facing 200‐kHz split‐beam transducer, deployed from 5.5‐m vessels (Boston Whalers) modified for acoustic research. An avoidance coefficient (Ac) was developed based on comparisons of acoustic fish densities while the vessel moved over the same transects by drifting, and by motoring at survey speeds. Ac did not differ significantly from 1 (no avoidance) at Nopiming (median of 0.81, n = 13), but did at Lac du Bonnet (median of 0.51, n = 31, P < 0.05). Variability in Ac was as high in transect pairs and was unrelated to fish depth (mean 6.9 m at Lac du Bonnet; 13.1 m at Nopiming) or survey speed (up to 3.70 m·s−1, 7 knots). Results indicated that fishes did not dive in the presence of the motoring vessel, nor was avoidance size‐based. We did not detect any evidence of fish attraction to our drifting vessel. Our results suggest that boat avoidance during acoustic surveys of shallow boreal lakes may vary in relatively similar water bodies but can be quantified experimentally.