Calibration of acoustic instruments.

Abstract

Acoustic sampling has long been a standard survey tool for estimating the abundance and distribution of fish, zooplankton, and their seabed habitat (Kimura, 1929; Sund, 1935; Holliday, 1972a; Nielson et al., 1980). Typically, acoustic surveys are conducted using multifrequency echosounders that transmit sound pulses down beneath the ship and receive echoes from animals and the seabed in the path of the sound waves (Simmonds and MacLennan, 2005). Generally, for surveys of animals, the backscatter signal is normalized to the range-dependent observational volume yielding the volume backscattering coefficient, which provides indications of the target type and behaviour. Objects scatter sound if their product of mass density and sound speed is different from that of the surrounding medium. A fish with a swimbladder has a large acoustic-impedance contrast (Foote, 1980), and thus has a large reflecting area, backscattering cross-section. Plankton, e.g. krill and salps, generally have much lower acoustic-impedance contrasts, but can produce large volume backscattering coefficients when they are aggregated in large densities (Hewitt and Demer, 1991, 2000). Under certain conditions, the summed and averaged volume backscattering coefficients are linearly related to the density of the fish or plankton aggregations that contributed to the echoes (Foote, 1983a). The number density can be estimated by dividing the integrated volume backscattering coefficient from an aggregation of target species by the average backscattering cross section from a representative animal (Ehrenberg and Lytle, 1972). An estimate of animal abundance is then obtained by multiplying the average estimated fish density and the survey area. Increasingly, multifrequency echosounder surveys are being augmented with samples from other acoustic instruments such as multibeam echosounders (Gerlotto et al., 1999; Simmonds et al., 1999; Berger et al., 2009; Colbo et al., 2014), multibeam imaging sonars (Korneliussen et al., 2009; Patel and Ona, 2009), and long-range scanning sonars (Bernasconi et al., 2009; Nishimori et al., 2009; Stockwell et al., 2013)(Figure 1.1). Use of these instruments provides information on many more aspects of the biotic and abiotic environment, e.g. bathymetry, seabed classification (Humborstad et al., 2004; Cutter and Demer, 2014), oceanographic fronts (Wade and Heywood, 2001), mixed-layer depths, anoxic regions, internal waves (Lavery et al., 2010a), turbulence (Stanton et al., 1994), currents, and methane seeps, all contributing to a broader ecosystem perspective (Demer et al., 2009a). In each case, the quantitative use of the data requires that the acoustic instrument is calibrated. Calibration of acoustic instruments | 11 Figure 1.1. A conceptual image of acoustic-sampling beams projecting from a survey vessel equipped with multifreque ncy split-beam (green) and multibeam (orange) echosounde rs, multibeam imaging sonar (purple), and long-range scanning sonar (grey). Instrument calibration involves the characterization of measurement accuracy (bias or systematic error) and precision (variability or random error). Sampling with the calibrated instrument involves additional systematic and random error (Demer, 2004). Calibration accuracy is estimated and optimized by comparing measured and assumed values for a standard, and correcting for the difference. The selection and characterization of a calibration standard is, therefore, paramount to the accuracy of an instrument calibration (Foote and MacLennan, 1984). Calibration precision is estimated by comparing multiple measures of a standard. Importantly, the performance of an instrument and thus its calibration accuracy and precision may change vs. time or the environment (Demer and Hewitt, 1993; Brierley et al., 1998a; Nam et al., 2007). Therefore, instruments should be calibrated frequently within the range of environments where they are used to make calibrated measurements (Demer and Renfree, 2008). If this is not possible, account should be made for any changes in the instrument or environment that appreciably affect the calibration accuracy and precision. This report includes general instruction and current best practices for calibrating a selection of acoustic instruments commonly used to conduct fishery science and surveys. It also describes some less developed protocols for other acoustic instruments. For practical reasons, not all fishery acoustic instruments are included. The remainder of Chapter 1 (i) summarizes some of the theoretical principles of acoustic instruments used to conduct fishery research and surveys, (ii) describes some commonly used instruments and their deployment platforms, and (iii) briefly introduces some common methods for calibrating acoustic instruments. Readers seeking only protocols for calibrating echosounders may wish to skip this and other sections and consult the Contents table to access information related to their interest and need. Chapter 2, details the sphere calibration method. Chapter 3 explores uncertainty in sphere calibrations. Chapter 4 describes protocols for calibrating some commonly used echosounders. Chapter 5 describes emerging protocols for some other acoustic instruments. Chapter 6 acknowledges valuable contributions to this CRR by people not included in the list of authors.