Target-strength, length, and tilt-angle measurements ... - Oxford Journals

1212

Target-strength, length, and tilt-angle measurements of Pacific saury (Cololabis saira) and Japanese anchovy (Engraulis japonicus) using an acoustic-optical system Kouichi Sawada, Hideyuki Takahashi, Koki Abe, Taro Ichii, Kazutoshi Watanabe, and Yoshimi Takao Sawada, K., Takahashi, H., Abe, K., Ichii, T., Watanabe, K., and Takao, Y. 2009. Target-strength, length, and tilt-angle measurements of Pacific saury (Cololabis saira) and Japanese anchovy (Engraulis japonicus) using an acoustic-optical system. – ICES Journal of Marine Science, 66: 1212 – 1218.

Pacific saury and Japanese anchovy generally congregate in dense groups or schools. An acoustic-optical system (the Japanese Quantitative Echosounder and Stereo-video Camera System or J-QUEST) has been developed to measure accurately the target strength (TS) of fish in a dense school. J-QUEST comprises a quantitative, 70 kHz, split-beam echosounder and a stereo-video camera. It was deployed from a research vessel to collect concomitant measures of TS and stereo images of in situ Pacific saury and Japanese anchovy. The stereo-video camera provides estimates of the fish lengths (L) and tilt-angles corrected for J-QUEST motion. In this way, empirical models of TS vs. log(L) were derived for Pacific saury and Japanese anchovy and compared with theory. Keywords: acoustic-optical system, Japanese anchovy, Pacific saury, pitch-angle, target strength. Received 8 August 2008; accepted 9 February 2009; advance access publication 8 April 2009. K. Sawada, H. Takahashi, K. Abe, K. Watanabe, and Y. Takao: National Research Institute of Fisheries Engineering, FRA, 7620-7 Hasaki, Kamisu, Ibaraki 314-0408, Japan. T. Ichii: National Research Institute of Far Seas Fisheries, FRA, 2-12-4 Fukuura, Kanazawa, Yokohama, Kanagawa 236-8648, Japan. Correspondence to K. Sawada: tel: þ81 479 44 5948; fax: þ81 479 44 6221; e-mail: [email protected].

Introduction Pacific saury (Cololabis saira) are widely distributed in the northwestern Pacific Ocean, from subtropical to Subarctic regions. They have closed swimbladders and typically inhabit surface waters to depths of 50 m (Furusawa, 1989; Honda, 1994). Pacific saury are important commercial pelagic fish in Japanese waters, with a total catch of 240 000 t being taken in 2007. Although three acoustic surveys of Pacific saury were conducted near Chishima Island, Japan (Honda, 1994), the results from trawl surveys before each fishing season (Ueno et al., 2004) have been used, since 2001, for Japanese stock assessments of Pacific saury in the northwestern Pacific Ocean. Japanese anchovy (Engraulis japonicus) are also widely distributed in the coastal waters of the western Pacific Ocean. They too are important commercial, pelagic fish species in Japanese waters, with a catch of 420 000 t being made in 2007. Since 2005, acoustic-survey results have been used as supplementary information to estimate the population numbers of Japanese anchovy in the area outside the fishing grounds. To estimate fish density from volume scattering, the mean target strength (TS) of anchovy within the study area must be known (MacLennan and Simmonds, 1992). Because fish length (L), behaviour, depth, and physiological conditions, i.e. sex, fat content, and gonad stage, can all affect target strength (TS), depending on the species, variations in these parameters must be considered to estimate TS vs. log(L) relationships accurately and precisely (Ona, 2003). In the absence of information specific to Japanese anchovy, an empirical

relationship for clupeids recommended by Foote (1987) has been used to estimate their TS. Both Pacific saury and Japanese anchovy congregate in dense schools. It is difficult to measure the TS of fish spaced closely, because their unresolvable echoes may be erroneously interpreted as echoes from individual targets (Sawada et al., 1993; Soule et al., 1995). To mitigate this problem, the transducer can be positioned closer to the fish to obtain higher resolution and to resolve single echoes (Traynor, 1996; Ona, 2003). Adding visual equipment, such as a still camera (Huse and Ona, 1996), a laser line-scan camera (Ermolchev and Zaferman, 2003), or a stereo-video camera, to the acoustic system allows identification of the fish species and their lengths and behaviour (Sawada et al., 2006). The Japanese Quantitative Echosounder and Stereo-video Camera System (J-QUEST) was designed to be deployed from a drifting research vessel to measure fish acoustically and optically from above or within a school. An initial sea-trial indicated that the TS from in situ fish could be measured by the split-beam method (Sawada et al., 2004), and the fish L and orientation could be measured by the stereo method (Takahashi et al., 2004). The aims of this study were to use data from J-QUEST to develop empirical equations for distributions of tilt-angle (u) and TS vs. log(L) for Pacific saury and Japanese anchovy, and to compare these empirical models with theoretical models and historical data (Miyanohana et al., 1990). In addition, the effects of beam-angle thresholding on the TS and the derived models of TS vs. log(L) were explored.

# 2009 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved. For Permissions, please email: [email protected]

1213

TS of Pacific saury and Japanese anchovy

Methods J-QUEST J-QUEST was described by Sawada et al. (2004) [There is an error in Equation (7) in Sawada et al. (2004). The corrected equation is: ffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rC ¼ x0 sin uC cos fC þ x02 sin2 uC cos2 fC 1 þ r 2 :]. An echosounder (Kaijo Sonic Corp., KFC5000) with a modified, 70 kHz, split-beam transducer (beam width = 11.88) and a stereo-video camera were installed in a pressure housing (Figure 1). Motion, depth, and temperature sensors were attached to the housing. The echosounder, stereo-video camera, and sensors were controlled and monitored aboard the ship through a 300 m, opticalpower cable. To calibrate the acoustic system, a 38.1 mm diameter, tungstencarbide sphere was tethered below the transducer at all times during the experiments, using two or three thin monofilament lines. The distance between the transducer and the sphere was 6 and 5 m, respectively, during the Pacific saury and Japanese anchovy experiments. Transmitting and receiving responses were measured using sphere echoes detected within 1 dB of the beam axis. For all experiments, the pulse duration was 0.6 ms. The stereo-video camera comprised two highly sensitive, black-and-white, high-gain, avalanche-rushing, photoconductor (HARP) cameras. These allowed observations of fish under conditions of low light intensity, with a resolution of at least 700 lines. The two HARP cameras had parallel views with a separation

(baseline) of 30 cm. To match the sampling volume of the acoustic system, the lenses had a focal length of 23 mm (158 field of view). The direct linear-transformation method (Adbel-Aziz and Karara, 1977) was used to detect three-dimensional positions (x, y, and z) of fish. Motion-corrected positions were obtained using the rotation matrix proposed by Furusawa and Sawada (1990) and 60-s averages of J-QUEST motion. Details of the L and u measurements were described by Takahashi et al. (2004). From the results of a tank calibration, the root-mean-squared measurement error at a distance of 300 cm was 1.8 +1.37 cm (Takahashi et al., 2005).

TS analysis The algorithm for measuring echoes from individual fish was given by Sawada and Furusawa (2001). Echoes are rejected if their durations measured at one-half and one-fifth of the peak-amplitude lay outside the ranges 0.5 –0.7 ms and 0.7 –0.9 ms, respectively. Echoes are rejected if the difference between the maximum and minimum electrical phase angles within the echo is .208. Echoes are also rejected if they do not pass the following beampattern threshold (Traynor and Ehrenberg, 1979). The number of single targets N is proportional to the cutoff solid angle V (N ¼ kV), if there is no bias towards stronger echoes. The relationship is expressed as N ¼ kV. The slope k is a function of the density of single echoes, the start and end depths of the analysis layer, and the total number of pings (Sawada et al., 1993). The differential slope Dk = DN/DV, and DV = 0.28, was used for this study. Both k and Dk are considered to be good indices of the linearity between N and V. The maximum V, where k is considered a constant value, is a suitable cutoff angle for TS analysis. The beam pattern at this cutoff angle is the beam-pattern threshold.

Pacific saury The first trial of J-QUEST was conducted from 16 to 20 December 2003, and acoustic and optical data were collected on board RV “Kaiyo-Maru”. J-QUEST was deployed at a depth of 10 m off Kashima-Nada on the east coast of Japan (3686.10 N 14182.20 E), between 19:44 (JST = GMT + 9 h) and 21:12 on 19 December 2003. The light on J-QUEST was switched off for approximately the first 30 minutes of the deployment; then its intensity was incrementally increased every five minutes until fish were observed.

Japanese anchovy

Figure 1. Geometry of the J-QUEST transducer beam width and stereo-video camera field-of-view.

Acoustic and optical data were collected on RV “Shunyo-Maru” during a survey of the forage fish of neon flying squid (Ommastrephes bartramii) conducted from 14 to 30 July 2004 in the North Pacific Ocean. The survey started from the southernmost station and progressed to the north. After the biological and oceanographic data had been collected, the J-QUEST survey started from the northernmost station and moved south, covering four stations in intervals of 90 nautical miles. J-QUEST was lowered with an armoured cable and winch. At Station 10 (4381.50 N 175836.20 E), it was deployed at a depth of 12 m from 00:44 to 01:15 h, and at a depth of 25 m from 01:17 to 03:28 h on 30 July 2004. Approximately five nautical miles away, at depths of 60 –90 m, on 23 July 2004, samples of Japanese anchovy were obtained with a midwater trawl (22:18 –22:34 h, ship mean time = GMT + 11 h).

1214

K. Sawada et al.

TS model of Pacific saury Although the targets were not identified with net samples, birds dropped eight fish on the deck of RV “Kaiyo-Maru”. These “birdsampled” Pacific saury were frozen and taken to the National Research Institute of Fisheries Engineering in Hasaki, Japan, where they were measured for total length (LT) and fork length (LF), and digitized X-ray images were made of their swimbladder shapes in lateral- and dorsal-aspect. Further detail is provided by Sawada et al. (1999) who used the same measurement and imaging methods. The backscattering vs. u, or scattering directivity patterns (SDPs), of a Pacific saury (LF = 22.2 cm) were calculated at frequencies ranging from 20 to 120 kHz, using the vacant, prolatespheroid model (PSM) with Dirichlet’ boundary condition (Furusawa, 1988) and the measured swimbladder shapes, as well as the liquid, deformed-cylinder model (DCM; Ye et al., 1997). The sound speed and density of the gas in the swimbladder were assumed to be 340 m s21 and 1.29 kg m23, respectively. The measured sound speed and density of the seawater at the average depth of the TS measurements (22 m) were 1498 m s21 and 1025 kg m23, respectively. The SDPs were averaged in the linear domain over a probability density function of u and expressed in dB (TSA ). Assuming that TSA is proportional to L2, the y-intercept of the linear relationship is given by b20 ¼ TSA 20 logðLÞ:

ð1Þ

Sawada et al. (1993) established that Equation (1) is equivalent to 2 b20 ¼ TSA 10 log L þ s:d:ðLÞ2 ;

ð2Þ

if L is normally distributed with mean L. Miyanohana et al. (1990) measured SDPs of ex situ Pacific saury (eight specimens) at four frequencies (25, 50, 100, and 200 kHz). In our experiment, 51 SDPs were measured of ex situ Pacific saury, including repeated measurements of seven specimens (LF = 29.7 –32.1 cm) at three frequencies (25, 50, and 100 kHz). Estimates of b20 were obtained from the SDPs of ex situ and in situ Pacific saury and estimates of their normally distributed L and in situ u using Equation (1) for ex situ and Equation (2) for in situ data.

Results TS of in situ Pacific saury Pacific saury were observed on stereo images 25 minutes after the light was turned on (Figure 2a and b). They were distributed from close to the surface to a depth of almost 30 m. Values of TS were measured at ranges r = 5 –15 m from the transducer. The TS generally increased with estimated fish density, because multiple targets were accepted as individuals (Sawada et al., 1993; Soule et al., 1995). Figure 3 shows TS measured in a region with the lowest fish density. The N and k vs. V are displayed in Figure 3a. The k decreases slightly as V increases to 0.04 sr. Figure 3b shows the TS vs. the beam-pattern threshold. There is an inflection point near 23 dB corresponding to V = 0.03 –0.04 sr. The vertical dotted lines in Figure 3a and b correspond to the selected V. Figure 3c shows the TS histogram for different V values. The largest of several modes is near 235 dB. The TS varies from 264 to 229 dB. TS =239.7 dB (n = 323) at r = 9.3 +2.2 m for

Figure 2. Images of Pacific saury from 19 December 2003 at 20:46:00 JST taken with the left (a) and right (b) cameras; and Japanese anchovy and a Boreopacific gonate squid (G. borealis) from 30 July 2004 at 02:21:43 SMT taken with the left (c) and right (d) cameras. Both cameras imaged the same Pacific saury (black arrows). The estimated LT of the saury and squid were 21.8 and 37 cm, respectively. V = 58; TS =239.9 dB (n = 460) at r = 9.4 +2.2 m for V = 68; and TS =239.8 (n = 586) at r = 9.3 +2.2 m for V = 78. The effect of V on TS was ,0.2 dB.

1215


Figure 3. Cutoff solid angle V, number of echoes N, and slope k (a); beam-pattern threshold and average TS (b); and TS histogram at different V values for Pacific saury (c). The TS and ranges from the transducer r were: 239.7 dB at 9.3 +2.2 m (n = 323) for V = 5.08; 239.9 dB at 9.4 +2.2 m (n = 460) for V = 6.08; and 239.8 at 9.3 +2.2 m (n = 586) for V = 7.08.

Figure 4. Cutoff solid angle V, number of echoes N, and slope k (a); beam-pattern threshold and average TS (b); and TS histogram at different V values for Japanese anchovy (c). The TS and ranges from the transducer r were: 246.5 dB at 9.3 +2.1 m (n = 601) for V = 2.88; 246.2 dB at 9.3 +2.2 m (n = 1151) for V = 4.08; and 245.9 dB at 9.3 +2.2 m (n = 2311) for V = 6.08. Three simulation results are displayed (blue lines).

TS of in situ Japanese anchovy

Measurements of u and LT with stereo images

Japanese anchovy were attracted to the light of J-QUEST at Station 10. At a J-QUEST depth of 25 m, stereo images were obtained of a Boreopacific gonate squid (Gonatopsis borealis) preying on a Japanese anchovy. With tentacles leading, the squid moved from the right to the left side of J-QUEST and caught an anchovy; then, with fins leading, it returned to the right side (Figure 2c and d). Figure 4a shows N and k vs. V. As for Pacific saury, the Dk varied vs. DV. The k was larger for Japanese anchovy than for the Pacific saury because the densities of Japanese anchovy were higher. Figure 4b shows TS vs. the beam-pattern threshold. An inflection point of TS was observed at 20.3 dB corresponding to V = 0.003 –0.004 sr (1.8– 2.08). The selected V = 2.88 because the TS values are reasonably stable there (Figure 4b). The selected V corresponded to a beam-pattern threshold =20.65 dB (Figure 4a and b). Figure 4c shows TS histograms for various values of V. The TS varied from 261.5 to 237.5 dB. TS =246.5 dB (n = 601) at r = 9.3 +2.1 m for V = 2.88; TS =246.2 dB (n = 1151) at r = 9.3 +2.2 m for V = 4.08; and TS =245.9 (n = 2311) at r = 9.3 +2.2 m for V = 6.08. TS less than 255.5 dB generally decrease with increasing V, and TS is modulated by V.

During measurements of Pacific saury, J-QUEST experienced 60-s-average pitch ranging from 212.7 to 24.58 and 60-s-average roll ranging from 0.8 to 3.58. Average pitch = 2.28 and average roll =21.78 were observed during measurements of Japanese anchovy. Positive pitch is bow up, and positive roll is port up. Figure 5a and b shows the measured u for Pacific saury and Japanese anchovy, respectively, corrected for the motion of J-QUEST. The distributions were normal: u = N(mean =21.18, s.d. = 15.48) and u = N(21.38, 20.88), respectively. The fish were measured by the stereo-video camera at ranges r = 1.7 –3.4 and 1.4 –2.9 m, respectively. The measured LT were 25.7 +3.0 cm (n = 24) and 12.2 +1.3 cm (n = 20), respectively.

Modelled TS of Pacific saury From measurements of the “bird-sampled” Pacific saury, LT = 29.1 +4.4 cm (n = 8) and LF = 0.943 LT (r2 = 0.9972, n = 8). The major and minor radii of the swimbladder of a single Pacific saury (LF = 22.2 cm) were 35.1 and 4.1 mm, respectively. These values were used to calculate the SDPs using the PSM. Substituting TS =239.9 dB and LF = N(24.3, 2.9 cm) into

1216

K. Sawada et al. =23.1 dB at V = 68 was in the range 22 to 24 dB (one way) proposed by Traynor and Ehrenberg (1979). Moreover, there were no significant differences in the TS around that threshold (Figure 3b). The LT = 29.1 cm (n = 8) of the Pacific saury in the bird samples was larger than the LT = 25.7 cm (n = 24) measured by the stereo method (t-test, p, 0.03). The LT from the birdsampled saury might not have been representative of the fish measured acoustically, because the sample size was small and birds might have been selective samplers. From Figure 6, the TS predicted by the PSM and DCM models agreed closely with each other, and the measured TS was equal to 267.6 dB at 70 kHz. The TS measurements of ex situ Pacific saury were also similar to the model predictions.

Japanese anchovy

Figure 5. Distributions of measured tilt-angle u of Pacific saury and Japanese anchovy.

Figure 6. Comparison of b20 for Pacific saury from in situ (grey circle) and ex situ (plus, cross, asterisk, upward- and downward-pointed triangles, square, diamond) measurements and PSM (thick line) and DCM (thin line) model predictions (TF = 22.2 cm). Different symbols for the measurements of ex situ fish indicate the average values of the repeated measurements of each specimen, and the vertical lines indicate the ranges from the maximum and minimum values. Equation (2), b20 =267.6 dB for Pacific saury at 70 kHz. Figure 6 compares b20 estimated from model predictions with measurements of ex situ and in situ fish.

Discussion Pacific saury The large variation in Dk (Figure 3a) was a result of the small N within each DV. Likewise, the variation in TS was large for a beampattern threshold ,0.5 dB (Figure 3b). A V = 68 (0.0344 sr) was selected for the TS analysis, because the beam-pattern threshold

Figure 4c shows a bimodal-TS distribution similar to that observed by Barange et al. (1996) for the South African anchovy (E. japonicus). They assumed that the weaker peak was generated by single targets from the edge of the scattering layer, whereas the stronger peak was a consequence of multiple echoes from its centre. To investigate the possibility of a true bimodal distribution of Japanese anchovy TS, values were simulated using measured distributions of u and SDPs predicted by the PSM (see Williamson and Traynor, 1984, for details of the method). The model parameters for the PSM in this study were obtained from regressions of fish standard length (LS) vs. both dorsal-aspect, swimbladder length (LSL) and width (LSw) developed by Abe et al. (2007): LSL = 0.364 LS 2 0.0352 (r = 0.94) and LSw = 0.046 LS 2 0.048 (r = 0.82); valid for LS from 4.7 to 29.9 cm. For the Japanese anchovy in this study, LS = 10.6 +1.1 cm using LS = 0.874 LT (r2 = 0.9953, from n = 90 fish with LT = 4.729.9 cm). Therefore, for LS = 10.6, the estimated LSL and LSw were 3.85 and 0.44 cm, respectively. To simulate TS distributions, a distribution of LS = 10.6 +1.1 cm was used, assuming that swimbladder sizes (LSL, LSw) increased proportionally with LS. Simulated TS distributions are presented in Figure 4c for three inclination angles between the swimbladder and body axes (2108, 08, 108). There are three modes and the strongest two near 254 and 242 dB resemble the measurements. Therefore, a bimodal TS distribution is theoretically predicted for Japanese anchovy with normally distributed LT. The weaker third mode near 258.5 dB was not observed in the measurements, perhaps because weak echoes might have been rejected in schools with high fish densities. Traynor and Ehrenberg (1979) established that a smaller beam-pattern threshold is appropriate when the fish density is higher. Therefore, that threshold must be matched to the density of the fish school. It is noted that the estimated average TS values by the three simulations (2108, 08, 108) were 246.6, 246.0, and 246.4 dB, and they were very close to the measured value of 246.5 dB (V = 2.88). For LT = 12.2 +1.3 cm and TS = 246.5 dB, b20 =268.3 dB at 70 kHz. For 38 kHz, Zhao et al. (2008) proposed TS = 20 log(LT)2 20/3 log(1 + z/10)2 67.6, where z is the depth of the fish. Solving this equation with the average depth of fish in this study (21.2 cm) resulted in b20 =270.9 dB. In addition, for 38 kHz, Kang et al. (2009) proposed TS = 20 log(LT)2 65.8 based on measurements of ex situ Japanese sardine at a r = 4 m. Compensating for depth [i.e. 220/3 log(1 + z/10)], the latter equation becomes TS = 20 log(LT)2 68.1. This b20 for 38 kHz is approximately equal to that for 70 kHz derived from the J-QUEST measurements.


Uncertainties The motion of J-QUEST contributed two types of TS measurement error. First, it contributed error in the process of beam compensation (Furusawa and Sawada, 1990). TS estimation error increased with decreasing beam width, especially below 108, and with increasing r. The J-QUEST transducer had a beam width of 11.98, and r was ,15 m for all TS measurements. Therefore, the error attributable to beam compensation was probably negligible. Second, J-QUEST motion contributed error in sampling the SDP of the fish. The perceived u was therefore the result of the convolution of the orientations of J-QUEST and the measured fish. In future, the J-QUEST transducer should be horizontally stabilized. Variability in the empirical TS vs. log(LT) model might have been because of small sample sizes for the LT measured by the stereo-video camera (Pacific saury, n = 24; Japanese anchovy, n = 20). In future, the J-QUEST measurements should be complemented with additional net sampling. The latter could validate the fish species composition and size distributions, as well as physiological conditions, such as sex, fat content, and gonad stage, which are all sources of TS variation (Ona, 2003). The J-QUEST light could also have influenced fish behaviour and thus their TS. Therefore, the estimated distribution of u might not have been representative of natural fish behaviour in darkness. Because J-QUEST had two cameras with high sensitivity, it was possible to observe the fish at several tens of metres in daytime, in summer, without lights.

Conclusion J-QUEST is an effective tool for identifying fish species and for measuring their TS, length, and tilt-angle distribution. Using data from J-QUEST, empirical TS models for 70 kHz were developed for Pacific saury and Japanese anchovy: TS = 20 log(LF)2 67.6 and TS = 20 log(LT)2 68.3, respectively. The normal distributions of fish lengths and tilt-angles were 25.7 +3.0 cm (n = 24) and 21.1 +15.48 (n = 297), and 12.2 +1.3 cm (n = 20) and 21.3 +20.88 (n = 92) for the two species, respectively. To refine these TS models and develop more for other species, further studies with J-QUEST, including net sampling, should be conducted for schooling fish of different size compositions and at different depths. Furthermore, the effect of lights on fish behaviour should be studied.

Acknowledgements We thank the sea-going staff and crew of RVs “Shunyo-Maru” and “Kaiyo-Maru” for their dedicated assistance. This article was greatly improved by anonymous referees and David Demer.

References Abe, K., Amakasu, K., Sadayasu, K., and Takao, Y. 2007. Standardization of the body and swimbladder shapes of Japanese anchovy for target strength estimation. 2007 Proceedings of a Meeting of the Marine Acoustics Society of Japan, pp. 13 – 16 (in Japanese). Adbel-Aziz, Y. I., and Karara, H. M. 1977. Direct linear transformation from comparator coordinates into objects space in close-range photogrammetry. Symposium on Close-Range Photogrammetry, American Society of Photogrammetry, Falls Church, VA, pp. 1– 18. Barange, M., Hampton, I., and Soule, M. 1996. Empirical determination of in situ target strengths of three loosely aggregated pelagic fish species. ICES Journal of Marine Science, 53: 225 – 232.

1217 Ermolchev, V. A., and Zaferman, M. L. 2003. Results of experiments on the video-acoustic estimation of fish target strength in situ. ICES Journal of Marine Science, 60: 544 – 547. Foote, K. G. 1987. Fish target strengths for use in echo integrator surveys. Journal of the Acoustical Society of America, 82: 981 – 987. Furusawa, M. 1988. Prolate spheroidal models for predicting general trends of fish target strength. Journal of the Acoustical Society of Japan, 9: 13 – 40. Furusawa, M. 1989. Bubbles and underwater sounds—swimbladder and sound scattering by fish. Journal of the Marine Acoustic Society of Japan, 16: 181– 197 (in Japanese). Furusawa, M., and Sawada, K. 1990. Effects of transducer motion on quantifying single fish echoes. Nippon Suisan Gakkaishi, 57: 857– 864. Honda, S. 1994. Acoustic survey of Pacific saury. Nippon Suisan Gakkaishi, 60: 809– 810 (in Japanese). Huse, I., and Ona, E. 1996. Tilt angle distribution and swimming speed of overwintering Norwegian spring spawning herring. ICES Journal of Marine Science, 53: 863 – 873. Kang, D., Cho, S., Lee, C., Myoung, J-G., and Na, J. 2009. Ex situ target-strength measurements of Japanese anchovy (Engraulis japonicus) in the coastal Northwest Pacific. ICES Journal of Marine Science, 66: 1219 – 1224. MacLennan, D. N., and Simmonds, E. J. 1992. Fisheries Acoustics. Chapman and Hall, London. Miyanohana, Y., Ishii, K., and Furusawa, M. 1990. Dorsal aspect target strength functions of eight species of fish at four frequencies. Technical Report of the National Research Institute of Fisheries Engineering, 10: 65– 154 (in Japanese). Ona, E. 2003. An expanded target-strength relationship for herring. ICES Journal of Marine Science, 60: 493 – 499. Sawada, K., and Furusawa, M. 2001. Computer simulation study of in situ target strength estimation method. Proceedings of the International Symposium on Advanced Techniques of Sampling Gear and Acoustic Survey for Estimation of Fish Abundance and Behaviour (ACOUSTGEAR), pp. 11 – 16. Sawada, K., Furusawa, M., and Williamson, N. J. 1993. Conditions for the precise measurement of fish target strength in situ. Journal of the Marine Acoustics Society of Japan, 20: 73– 79. Sawada, K., Takahashi, H., Abe, K., and Takao, Y. 2006. In situ measurement of target strength, tilt angle, and swimming speed of Boreopacific gonate squid (Gonatopsis borealis). Journal of the Acoustical Society of America, 120: 3107. Sawada, K., Takahashi, H., Takao, Y., Watanabe, K., Horne, J. K., McClatchie, S., and Abe, K. 2004. Development of acoustic-optical system to estimate target strengths and tilt angles from fish aggregations. Oceans04/Tech-Ocean04 Proceedings, pp. 395– 400. Sawada, K., Ye, Z., Kieser, R., McFarlane, G. A., Miyanohana, Y., and Furusawa, M. 1999. Target strength measurements and modeling of walleye pollock and Pacific hake. Fisheries Science, 65: 193 – 205. Soule, M., Hampton, I., and Barange, M. 1995. Evidence of bias in estimates of target strength obtained with a split-beam echo-sounder. ICES Journal of Marine Science, 52: 139 – 144. Takahashi, H., Sawada, K., Takao, Y., and Abe, K. 2005. Inspection of the three-dimensional measurement accuracy of J-QUEST Stereo TV-camera system. Technical Report of the National Research Institute of Fisheries Engineering, 27: 81 – 92 (in Japanese). Takahashi, H., Sawada, K., Watanabe, K., Horne, J. K., McClatchie, S., Takao, Y., and Abe, K. 2004. Development of a stereo TV camera to complement fish school measurement by a quantitative echo sounder. Oceans04/Tech-Ocean04 Proceedings, pp. 409 – 414. Traynor, J. J. 1996. Target-strength measurements of walleye pollock (Theragra chalcogramma) and Pacific whiting (Merluccius productus). ICES Journal Marine Science, 53: 253 – 258. Traynor, J. J., and Ehrenberg, J. E. 1979. Evaluation of the dual beam acoustic fish target strength measurement method. Journal of the Fisheries Research Board of Canada, 36: 1065– 1071.

1218 Ueno, Y., Suyama, S., Kurita, Y., and Kumazawa, T. 2004. Design and operation methods of a mid-water trawl for quantitative sampling of a surface pelagic fish, Pacific saury (Cololabis saira). Fisheries Research, 66: 3 –17. Williamson, N. J., and Traynor, J. J. 1984. In situ target-strength estimation of Pacific whiting (Merluccius productus) using a dualbeam transducer. Journal du Conseil International pour l’Exploration de la Mer, 41: 285– 292.

K. Sawada et al. Ye, Z., Hoskinson, E., Dewey, R. K., Ding, L., and Farmer, D. M. 1997. A method for acoustic scattering by slender bodies. 1. Theory and verification. Journal of the Acoustical Society of America, 102: 1964–1976. Zhao, X., Wang, Y., and Dai, F. Q. 2008. Depth-dependent target strength of anchovy (Engraulis japonicus) measured in situ. ICES Journal of Marine Science, 65: 882 – 888. doi:10.1093/icesjms/fsp079