Acoustic Target Strengths and Swimbladder Morphology of Chub Mackerel Scomber japonicus in the Northwest Pacific Ocean

Abstract

The Northwest Pacific chub mackerel (Scomber japonicus) is one of the most productive, economically important fishery resources worldwide. Accurately assessing this species and ensuring adherence to total allowable catch limits are crucial owing to fluctuations in their abundance and distribution. Acoustic target strength measurements of S. japonicus were conducted at 38, 70, and 120 kHz using a split-beam echosounder of individuals from nine size groups (mean fork length, 10.8–28.3 cm) swimming freely in a net cage within a seawater tank. An underwater camera was utilized to simultaneously measure swimming angle. Least-squares regression analysis revealed that when the slope was constrained to 20, as per the generally applicable morphometric equation, the resulting values for the constant term (b₂₀) were −67.7, −66.6, and −67.3 dB at 38, 70, and 120 kHz, respectively. S. japonicus mean swimming angle across the groups was −10.5–9.6° (standard deviation [SD], 16.3–33.3°). Furthermore, the ratio of swimbladder height to swimbladder length, the ratio of swimbladder length to fork length, and the tilt angle of the swimbladder (mean ± SD) were 0.191 ± 0.060, 0.245 ± 0.055, and 9.6 ± 3.0°, respectively. These results can be used for the acoustic stock assessment of S. japonicus in the Northwest Pacific Ocean.

1. Introduction

Chub mackerel (Scomber japonicus), belonging to the mackerel family (Scombridae), is widely distributed in temperate marine areas including the Republic of Korea, Japan, the East China Sea, and the eastern Pacific Ocean [1,2,3,4]. It is one of the most productive and economically important fishery resources globally, with annual catches reaching ca. 1.3 million metric tons [5]. The abundance of the Pacific stock of chub mackerel declined in the 1990s and early 2000s. Since 2013, when a strong year class occurred, its abundance significantly increased [6,7]. In the Republic of Korea, chub mackerel has consistently recorded the second-highest catch volume after anchovy, with an average annual catch of 110,640 metric tons between 2018 and 2022 [8]. Consequently, total allowable catch (TAC) limits were implemented for the sustainable management of chub mackerel in the Republic of Korea, Japan, and China [9,10,11,12].

Research on pelagic fish using acoustic techniques involves collecting data from scientific echo sounders installed on ships proceeding along planned acoustic transects in survey areas, while simultaneously conducting sampling via trawling to confirm species identification and length distribution data derived from the echo sounders. This information is then used to estimate the distribution and abundance of the target species [13]. The average acoustic target strength (TS, dB) of the target species is essential in identifying and quantifying target species in a survey area [14]. This parameter is generally expressed as TS = 10·log₁₀σ/4π [15], where σ is the backscattering cross-section of the fish. TS length (L, cm) can be expressed as TS = a·log₁₀(L) + b, where the slope a and the intercept b (dB) are generally assumed to be species-specific constants. When the backscattering cross-section is proportional to the length squared, a is normally close to 20; thus, TS = 20·log₁₀(L) + b₂₀, where b₂₀ is the estimated intercept given a slope of 20.

Pelagic fish TS data are obtained either through ex situ experiments in seawater tanks or in situ at sea, in conjunction with trawling sampling, and are validated by comparison with a theoretical model to derive the TS–L relationship for the target species [16,17,18]. Most ex situ experiments on chub mackerel TS have been conducted using dead specimens [19,20]; however, this may result in inflated values compared to those using living specimens owing to the expansion of body tissues or presence of gas bubbles [21]. Moreover, previous ex situ experiments have been limited to specimens with a total length ≥ 23 cm [19]. Some live mackerel TS experiments have been conducted on adult-sized free-swimming ex situ TS and in situ TS at the common frequencies of 38 and 120 kHz [22,23]. Recently, chub mackerel TS has also been studied using the theoretical Kirchhoff-ray mode (KRM) model [7,24], but deriving an accurate and reliable TS–L relationship requires experiments to be conducted on living specimens of various sizes in a free-swimming state.

Factors affecting fish TS signals underwater include swimming posture angle, detector operating frequency, distribution depth, sound speed ratio, density ratio, and swimbladder presence and morphology. In particular, the TS signal of fish with a swimbladder depends on the large density ratio difference between gas and water, which should be more than 90% [25,26,27]. Therefore, when studying the TS–L relationship in Pacific chub mackerel, the morphology of its swimbladder must be considered during the analysis and interpretation of TS signals. In this study, we measured the ex situ TS at 38, 70, and 120 kHz for various length distribution groups of living chub mackerel and simultaneously filmed the swimming posture angle to verify fish movement. Subsequently, we investigated the characteristics of the swimbladder to validate the TS signal results.

2. Materials and Methods

2.1. Ex Situ Target Strength Measurement

Chub mackerel TS experiments were conducted from April to September 2022 at the Fisheries Resources Research Center’s seawater acoustic tank (5 m length × 5 m width × 10 m depth) (Figure 1) at the National Institute of Fisheries Science in Tongyeong, Gyeongsangnam-do, Republic of Korea. Specimens used in the experiment were caught during a set-net survey off the coast of Tongyeong, Gyeongsangnam-do (34°47.0′ N, 128°26.4′ E) and transported to the laboratory in a seawater tank, where they were acclimated in a circular tank (2.75 m diameter × 0.8 m depth) for 24 h prior to the TS experiment. The set-net fishery to collect S. japonicus samples in is a passive fishing technique that involves setting a fixed net in the path of migrating fish, primarily to catch pelagic species.

The system consisted of a SIMRAD EK80 split-beam scientific echosounder (SIMRAD, Kongsberg Maritime AS, Kongsberg, Norway). The echosounder operated at frequencies of 38 kHz (ES38-10), 70 kHz (ES70-7C), and 120 kHz (ES120-7C), with a transducer utilizing −3 dB beam widths of 10, 7.1, and 7.1° and power outputs of 1500, 750, and 250 W, respectively. The pulse duration and ping rate for all frequencies were set to 0.256 ms and 2 pings/s, respectively. Before the experiments by group, the temperature and salinity of the acoustic tank seawater were measured with a YSI 30M instrument (YSI, Yellow Springs, OH, USA) to determine sound speed from April to September 2022 [28]. Additionally, the transducers were calibrated with a 38.1 mm tungsten carbonate standard sphere [29].

For the TS experiment, a cylindrical cage (2 m diameter × 7 m depth, equipped with a stainless-steel ring and 1.0 cm mesh nylon netting) was installed in the tank to streamline sample replacement and to accommodate the beam width of transducers for each frequency (Figure 1). Transducers were installed in the tank vertically at a depth of 30 cm. For the experiment, visually similar sizes of chub mackerel specimens acclimated in the circular tank were collected in small buckets without exposure to air and transported to the cage installed in the acoustic tank, where TS signals of freely swimming specimens were acquired. During the experiment, an underwater camera was installed on the side of the tank to observe swimming posture angle. The camera was vertically adjustable and recorded as it was positioned to the depth layer where the fish were primarily distributed. After the TS experiment, the fork length (FL) and wet weight (W) of the specimens were measured.

Chub mackerel specimens were placed in the cage divided into nine groups of similar length, and effort was made to visually ensure that lengths between groups did not overlap. Average FL for the groups ranged from 10.8 to 28.3 cm, with a length range within groups of 0.2–1.1 cm (

Table 1. Characteristics of size groups of chub mackerel used in the cage experiments (fork length in cm).

Fish Size Group	I	II	III	IV	V	VI	VII	VIII	IX
Mean	10.8	12.6	14.8	16.0	21.4	23.7	25.4	26.9	28.3
Standard deviation		0.2	0.1	0.9	0.7	0.2	1.1	0.0	0.4
Minimum		12.4	14.7	15.0	20.5	23.5	24.0	26.8	27.4
Maximum		12.6	14.9	17.1	22.4	24.0	26.7	26.9	28.7
Number of individuals in group	1	3	3	3	5	3	3	2	5

).

2.2. Acoustic Data Analysis

Group-specific TS acoustic signal analysis was conducted using Echoview version 11.0 (Echoview Software Pty Ltd., Hobart, Australia). TS signal analysis involved the extraction of individual fish signals using single echo detection (SED) and the detection of fish track modules after removing basic noise like bottom reflection (Figure 1). The single echo detector parameters were set to accept echoes with a threshold value of −65 dB. The SED range was set to minimum and maximum pulse lengths of 0.7 and 1.5, respectively, a maximum beam compensation of 6 dB, and a maximum standard deviation for minor and major axis angles of 0.6°. For detecting fish tracks, the minimum number of single targets and minimum number of pings in a track were each set to 3 and the maximum gap between single targets was set to 1 ping. Average TS values for each size group of chub mackerel were calculated using the SED and detect fish tracks modules, with mean TS = 10·log₁₀ (∑σ/n_i), where σ is the backscattering cross-section of the fish and n is the number of specimens.

2.3. Swimming Angle Measurement

For the analysis of swimming posture angles, video data from an underwater camera were converted into images at 1 s intervals, and the swimming posture angles were measured using ImageJ software version 1.51 (National Institute of Health, Bethesda, MD, USA) based on the center of the fish’s snout and tail fin. Swimming posture angles were defined as positive (+) and negative (−) angles when the fish’s head was pointing upwards and downwards, respectively. Among the video data of the nine groups, group 1 (a single specimen) was not recorded due to a camera depth setting error, so swimming posture angles were analyzed for the remaining eight groups.

2.4. Swimbladder Characteristics

To investigate swimbladder morphology of the chub mackerel used in the TS experiment, samples were subjected to X-ray imaging. After the TS experiment, the samples were measured for FL (cm) and W (g), and then rapidly frozen in a freezer at below −40 °C after placing them in bottles with seawater. The frozen samples were slowly thawed in cold water for 24 h in their bottles to minimize the deformation of the swimbladder’s shape for X-ray imaging [30]. The lateral and dorsal aspects of all individuals used in the TS experiment were photographed using soft X-ray (150 kV, 2mA; SOFTEX M-150W, SOFTECS Corporation, Tokyo, Japan) (Figure 2). Among the 28 chub specimens used in the TS experiment, the swimbladder of one individual (FL = 12.5 cm) from group 2 was damaged and was omitted from analysis. Swimbladder morphology was characterized by tilt angle (SBA), length (SBL), height (SBH), and width (SBW) (Figure 2), and the equivalent radius (α_esr = (a × b × c)^1/3) was calculated using the semi-major axis (a = SBL/2) from the lateral aspect, the semi-minor axis (b = SBH/2), and the dorsal aspect (c = SBW/2) [27].

3. Results

3.1. Environment and Specimens

During the TS experiments, seawater tank temperature and salinity ranged from 16.1 to 23.8 °C and 29.2 to 33.4 psu, respectively, resulting in sound speeds ranging from 1508.3 to 1525.1 m/s. A regression model fit to the relationship between FL and W for the 28 chub mackerels used in the experiment was W = 0.0029·FL^3.4537 (R² = 0.99) (Figure 3), and mean FL showed significant between-group differences (ANOVA, p < 0.05).

3.2. Ex Situ Target Strength and Swimming Angle

The range of TS for the nine groups of chub mackerel was −63.5 to −28.7 dB at 38 kHz, −64.2 to −27.7 dB at 70 kHz, and −64.1 to −29.2 dB at 120 kHz. TS per group mostly followed a Gaussian distribution (Figure 4). The variation in TS per group was 9.2–32.3 dB at 38 kHz, 9.9–30.9 dB at 70 kHz, and 18.9–33.7 dB at 120 kHz. The range of TS variation per group at 38 kHz and 70 kHz increased with FL (R² = 0.56–0.94, p > 0.05), while no correlation was observed at 120 kHz between size by FL and TS (R² = 0.50, p < 0.05). The mean TS per group was −45.8 to −35.8 dB at 38 kHz, −46.6 to −35.9 dB at 70 kHz, and −47.5 to −36.6 dB at 120 kHz, with all frequencies displaying concurrent increases in TS and body size (paired t-test, p < 0.05). Notably, the correlation coefficients of TS by size per frequency showed no significant differences (paired t-test, p > 0.05) (Figure 5).

Least-square regression models of TS vs. log(FL) by acoustic frequency across all groups were composed as follows:

TS_38kHz = 20.2·log(FL) − 67.7 (95% CI: 10.5 to 29.9 and −80.1 to −55.3; R² = 0.75),

TS_70kHz = 18.0·log(FL) − 64.1 (95% CI: 5.7 to 30.4 and −79.9 to −48.3; R² = 0.58),

TS_120kHz = 18.0·log(FL) − 64.7 (95% CI: 4.3 to 28.6 and −78.3 to −51.0; R² = 0.65),

giving the confidence intervals for a and b, respectively.

When forcing the model fit to a slope of 20, as per the standard formula, the following models were estimated:

TS_38kHz = 20·log(FL) − 67.4 (95% CI: −68.7 to −66.2; R² = 0.78),

TS_70kHz = 20·log(FL) − 66.6 (95% CI: −68.2 to −65.0; R² = 0.62),

TS_120kHz = 20·log(FL) − 67.3 (95% CI: −68.7 to −65.9; R² = 0.69)

The group-specific TS of chub mackerel at 38, 70, and 120 kHz frequencies at a = 20 were similar, with b₂₀ within a range of 0.9 dB (highest for 70 kHz and lowest for 120 kHz).

Swimming posture angles of chub mackerel groups showed a wide range of angles from −71.4 to 69.2°, with group means of −10.4 to 9.6° and standard deviations of 16.3–33.3° (Figure 6). The distribution of swimming posture angles was mostly unimodal, with the larger-sized groups 8 and 9 displaying bimodal distributions. The correlation between mean FL and swimming posture angle was negative and differed significantly between groups (R² = 0.61, paired t-test, p < 0.05).

3.3. Swimbladder Morphology

The length of the chub mackerel swimbladders as imaged by X-ray ranged between 1.6 and 8.6 cm (mean ± SD: 5.5 ± 2.3 cm), height between 0.2 and 2.3 cm (1.1 ± 0.6 cm), and width between 0.1 and 2.4 cm (1.1 ± 0.5 cm), and all showed a proportional increase with size (paired t-test, p < 0.01) (Figure 7). The equivalent radius of the swimbladders ranged between 0.20 and 1.76 cm (0.92 ± 0.42 cm) and showed a positive correlation with length. Contrastingly, the tilt angle of the swimbladders ranged from 6.0 to 19.0° (9.6 ± 3.0°; n = 27), displaying a negative correlation with FL (paired t-test, p < 0.01, R² = 0.06).

4. Discussion

4.1. Chub Mackerel TS Sample Specimens

Given the species’ characteristics, chub mackerel TS experiments to date have mostly been conducted using dead specimens due to the difficulty of working with live individuals [19,20]. We were able to conduct TS experiments on various size groups of living chub mackerel by making use of active set-net fishing in Tongyeong, allowing for the capture of live specimens and short transport time to the experimental facility. When multiple specimens are placed in a cage for experimentation, an overlap in group sizes can occur, potentially lowering the reliability of mean size estimation and TS relationship. However, since chub mackerel are a schooling species, they might exhibit abnormal swimming behaviors when isolated in a tank. We therefore placed up to five specimens at a time in a cage to acquire acoustic signals in a stable swimming state. To ensure no overlap in lengths within size groups, specimens were visually selected in the acclimation tank, minimizing group size deviation to 1.1 cm or less and resulting in almost no overlap between groups (one-way ANOVA, F = 244.0, p < 0.05). Cage method experiments on European whitefish (Coregonus lavaretus) have been conducted for five size groups over a length range of 5–59 cm, with a tolerance for a mean size range within groups of 15 cm, showing significant variation [13]. This indicates that a wider size range within a group leads to a relatively wider TS range. By contrast, our study, although covering a smaller range of 10.8–28.3 cm, obtained TS signals for nearly identical size groups within an average size variation of only 2 cm, yielding more precise TS range estimates.

Among methods for fish TS experimentation, the caged method allows the TS of fish swimming relatively naturally to be obtained, compared to that using the tethered method, where fish are fixed with fishing lines and artificially subjected to changes in swimming posture angles [15]. In the present study, we simultaneously acquired TS signals and free-swimming posture angles, resulting in single modes and significant between-group differences for most groups, indicating that the data were obtained in a stable state.

4.2. Chub Mackerel Swimbladder Characteristics

In this study, the morphological parameters of FL, SBL, SBW, and SBH in chub mackerel showed a high correlation. Fish swimbladders are morphologically categorized into five types [31]. The SBH/SBL ratio, SBL/FL ratio, and swimbladder tilt angle (mean ± SD) of chub mackerel used in the experiment were 0.191 ± 0.060, 0.245 ± 0.055, and 9.6 ± 3.0°, respectively, suggesting a classification as round swimbladders (with the standard given as 0.2, 0.282, and 6°, respectively). Chub mackerel swimbladders were previously reported as showing a long extension, bending below the vertebral column, with a volume-to-body ratio of 3.2%, length ratio of 25%, and posture angle of 11.9° [20]. The reported range of body length of 18 specimens was 12.98–22.17 cm, with a mean of 16.08 ± 3.15 cm (mean ± SD); the SBL range was 1.45–6.63 cm, with a mean of 3.64 ± 1.49 cm; and the swimbladder tilt angle was 1.7–14.0°, with a mean of 8.26 ± 3.62° [7]. Although both of these studies lack SBH/SBL results, the SBL/SL findings of our study also classify our specimens as having round swimbladders.

4.3. Chub Mackerel Swimming Posture Angles

In this study, mean swimming posture angles for groups of chub mackerel ranged widely from −10.4 to 9.6°. In comparison, chub mackerel in a smaller rectangular tank (82 cm × 28 cm × 28 cm) displayed narrower variances in swimming posture angles of −4 ± 4° [32]. It seems that many fish with swimbladders have close to horizontal mean swimming angles, as has been found for cod (−4.48 ± 16.28° [33]), capelin (3.88 ± 18.48° [34]), caged saithe (−0.98 ± 5.48° [14]), hoki (11.88 ± 29.98° [35]), in situ Pacific saury (−1.1 ± 15.4°), and Japanese anchovy (−1.3 ± 20.8° [36]). A range of swimming posture angles among chub mackerel that was broader than in other species was observed in this study. It was notable that angles changed sharply when fish ascended or descended within the cage. Previous reports suggest that even small juveniles might not show natural swimming behavior within a 3 m³ cage [37]. These results suggest that natural swimming posture angles that would be encountered in the field are difficult to represent in captivity, suggesting that further observations and comparative analysis using underwater cameras are required.

4.4. Chub Mackerel TS

Major factors in TS versus fish length regressions are tilt angle, acoustic frequency, fish length, and depth at the time of measurement [38,39,40,41], which adhere to the following order of influence: tilt angle > frequency > fish length > depth [42]. Caged measurement method can result in variance in swimming posture angles due to restricted movement within the narrow cage. In this study, group-specific TS values mostly showed a unimodal distribution, ranging widely between 10 and 30 dB. Additionally, changes in swimming posture angles became more pronounced as group size increased. TS measurements of European whitefish using the cage method also showed a wide distribution of 17–19 dB for larger sizes and 10–13 dB for smaller sizes, as well as a widening range of TS differences as group size increased [13]. Variations in TS using the tethered method and models showed significant differences even under the same alterations in swimming posture angles as size increased [43]. When comparing average group-specific TS values converted to b₂₀ based on changes in swimming posture angles, the highest b₂₀ values were observed at an average swimming angle of −4.1° across all frequencies, with b₂₀ decreasing as variation in swimming angles increased. While the tethered method allows for control over swimming posture angles while measuring TS, thereby providing clear TS values for each swimming angle, in this study, we compared mean swimming posture angles and b₂₀ across groups. Therefore, the variation in the swimming angle must be examined to correctly interpret TS data.

Previous studies on S. japonicus TS were all conducted using live or dead specimens using ex situ, in situ, and model methods (

Table 2. Comparisons of TS–FL relationship in chub mackerel by acoustic frequency and experimental method in the present and previous studies.

Method	Condition	Freq. (kHz)	FL (cm)	b20 (dB)	Reference
Ex situ	Live	38	10.8–28.3	−67.4	This study
		70		−66.6
		120		−67.3
	Live	38	17.4–34.0	−67.9	[22]
		120		−69.2
	Dead	25	23.0–26.8	−64.1	[19]
		100		−65.5
	Dead	50	26.2–38.3	−67.2	[20]
		75		−69.9
		120		−66.9
		200		−71.1
In situ	Live	38	20.0–30.0 (L)	−72.8	[23]
		120		−73.6
Theoretical model	Dead	38	15.4–26.2	−66.02	[22]
		70		−66.50
		120		−66.00
		200		−67.35
	Dead	38	12.04–22.17 (BL)	−73.27	[7]
		70		−73.56
		120		−74.18
		200		−73.46

FL, fork length; BL, body length; L, mean fish length; b₂₀, regression intercept given a forced slope of 20.

). Reported TS measurements at b₂₀ using the cage method are −67.9 dB at 38 kHz and −69.2 dB at 120 kHz for lengths of 17.4–34.0 cm [22] and for the tethered method included −64.1 dB at 25 kHz and −65.5 dB at 100 kHz for lengths of 23.0–26.8 cm [19] and −67.2 dB, −69.9 dB, −66.9 dB, and −71.1 dB for lengths of 26.2–38.3 cm at 50, 75, 120, and 200 kHz, respectively [20]. The b₂₀ values of live and dead individuals were approximately similar, within about 2 dB. However, the b₂₀ of chub mackerel estimated using the in situ method were approximately 5 dB weaker than in this study at both frequencies [23]. This is likely due to the fact that the TS values obtained in the field are different from the results of the present study because the length of the fish in the target signal is not known, and they swim in schools, resulting in fluctuations in the swimming angle. A previously reported KRM model used for frequencies of 38, 70, 120, and 200 kHz for S. japonicus TS (FL = 15.4–26.2 cm) b₂₀ yielded −66.02, −66.50, −66.00, and −67.35 dB, respectively, within 1 dB of the present study’s results [24], while another study reported values 5–6 dB lower, indicating significant outcome differences based on model parameters [7].

We discovered a high correlation between group-specific TS and length for chub mackerel across all three tested acoustic frequencies, with the optimal model slope consistently being close to 20. This indicates that the horizontal cross-sectional area of the swimbladder, as that in the generalized fish model, is proportional to the square of the length in this species, affecting TS values [15]. Additionally, the decrease in slope with increasing frequency suggests that the variation in TS due to swimming behavior becomes more pronounced at higher frequencies.

In conclusion, we successfully derived a TS–L relationship for various sizes of live chub mackerel at 38, 70, and 120 kHz, and chub mackerel swimming behavior and natural body tilts. We have shown that TS taken in consideration of fish orientation can be very helpful for understanding acoustic data. These findings provide foundational data for evaluating the density and stock of chub mackerel.