Validation of fish length estimations from a high frequency multi-beam sonar (ARIS) and its utilisation as a field-based measurement technique

Abstract

High frequency multi-beam sonars enable the determination of fish size characteristics and behaviour in highly turbid and low-light conditions when more traditional methods, such as underwater cameras, are unsuitable. We sought to validate this technology, commonly referred to as imaging sonar, by determining its accuracy and precision during the measurement of small-bodied fish typical of coastal and estuarine environments. Validation was conducted in three ways: (1) a comparative tank-based validation of an Adaptive Resolution Imaging Sonar (ARIS) and stereo-camera measurements using two species of live fish, and synthetic targets; (2) assessment of the imaging sonar’s utility when measuring four different fishes possessing a range of swimming modalities, including labriform, sub-carangiform, and carangiform; and (3) comparison of field-based sonar sampling with conventional baited trap methodology. Our results showed that the angle at which targets are presented to the sonar beams caused measurement error during tank-based determinations. When this was factored into measurement protocols there was no significant interactive effect of swimming mode on measurement accuracy, with the sonar results between 0.3 and 9.6% in all four species investigated. Sonar derived measurements were less accurate and precise when compared to stereo-camera techniques. Field-based sampling outcomes were broadly comparable between the imaging sonar and baited trap sampling methodology. The suitability of the imaging sonar is discussed in the context of use in fish sampling surveys, particularly diurnal and nocturnal sampling regimes, and in turbid environments. Results validate the usefulness of imaging sonars for use in diverse fields of fisheries research.