Characterizing uncertainty in target-strength measurements of a deepwater fish: orange roughy (Hoplostethus atlanticus)

The variability of ensemble 38 kHz, target-strength (TS38) estimates for orange roughy (Hoplostethus atlanticus) (4.9 dB, factor of 3.1) in deep water (>600m) limits the use of echo integration for absolute-biomass estimates. Orange roughy are high in oil content, have a wax ester swimbladder, and show an active-avoidance response to sampling gear. The interpretations of ensemble, in situ target strengths of orange roughy (range hTS38i 1⁄4 52:9 to 51.0 dB for standard fish length hSLi 1⁄4 35 cm) are lower than previous model and surface-based measurements (hTS38i 1⁄4 48 dB, SL 1⁄4 35 cm). In situ TS measurements from individuals on the periphery of dense schools were processed to minimize uncertainties from single-target selection criteria, species composition, and active avoidance. Video and acoustic-tracking data quantified the variability in TS measurements arising from the variability in fish orientation. Multi-frequency acoustics and fish tracking are used to quantify in situ TS variability due to species identification and fish density. The Kirchhoff-ray mode backscatter model was used to illustrate the sensitivity of speciesspecific backscatter to assumptions of tilt-angle and material properties (density and soundspeed contrasts). We conclude that a remaining source of uncertainty for in situ TS measurements is the assumption that dispersed targets are representative of the survey population.