Detection efficiency of fish tracking by autonomous sailboat while underway

Abstract

Background Acoustic telemetry is a fundamental tool for studying aquatic organisms and offers powerful insights into their behavior across habitats. Researchers can choose from numerous deployment methods to suit specific species and habitats. Yet, shallow waters are particularly challenging in part because the tools available are reduced; most mobile tracking platforms cannot be used due to depth requirements. Furthermore, surface vehicle detection efficiency is limited by noise interference from the surface and the vehicle itself, rendering it an underutilized tool. Therefore, this work improved upon the design of sensor placement and the resulting acoustic detection efficiency of a mobile, near-surface receiver. Results An autonomous sailboat was outfitted with custom software, a common acoustic receiver, and a high-performance hydrophone to survey an area of Boston Harbor, USA for acoustically-tagged winter flounder. To enhance detection, the design incorporated a high-transmissivity, flooded cowling. Increased wind- and wave-induced bubble plumes decreased mobile receiver efficiency as compared to stationary receivers from a concurrent study, and the resulting efficiency was quantified over a range of wind speeds. The mobile receiver detected 10.6% of the known tagged population of winter flounder in less than two days, similar to 11.6% detected collectively by multiple stationary receivers in the same area over the same period. In addition, a probabilistic model using the hydrophone data was developed to estimate and map fish positions within the surveyed habitat while incorporating uncertainty. Conclusions The utility of the autonomous sailboat to track fish without the need to limit mobility or attach the receiver by trailing it at depth is demonstrated here. The addition of the sensor cowling minimized drag, shielded the sensors from turbulence, and reduced noise caused by vessel movement, and the hydrophone enabled continuous monitoring of detection efficiency. The fish distribution model has the potential to have greater accuracy of fish positions as compared to standard receiver-based inferences. Limitations still exist depending on sea state among other factors, as high winds greatly impaired detection efficiency and could impact the distribution estimation. Overall, these results provide essential design and analytical guidance for enhancing acoustic telemetry via surface platforms, providing further potential for broader adoption and innovation in the mobile tracking of aquatic organisms.