Movements and Habitat Use of Dolphinfish (Coryphaena hippurus) in the East China Sea

To refine the regional and global stock and ecological assessments of dolphinfish in fisheries, it is necessary to have information on the species habitat use, fisheries interaction, migration corridors, and on changes in spatial-temporal patterns over their ontogeny. In order to inform management, pop-up satellite archival tags (PSATs) were deployed on dolphinfish (Coryphaena hippurus) in the Northern East China Sea to get data on the movement patterns and habitat utilization of this species in this location. During October–November 2018, four dolphinfish (94–102 cm fork length) were captured by set-nets and PSAT tagged. Tagged dolphinfish retained PSATs for 5–31 days-at-liberty (total 69 days) and linear dispersion from deployment to pop-up locations ranged from 63 to 204 km. According to most probable tracks, tagged fish made primarily northward movements. Tagged fish reached depths of ~94 m and experienced ambient temperatures from 17.8 to 23 °C. Movements appeared to be limited by a 3 °C change relative to sea surface temperature (SST) and were confined to the mixed-layer. Diel diving patterns indicated dolphinfish spent >80% of daytime activity and 40% of nighttime activity near the surface, where variability in diving patterns was more pronounced. The vertical diving patterns showed pronounced changes during dawn and dusk, where dolphinfish mirrored the movements of diel migrating prey organisms. Our preliminary results provide vertical distribution patterns of dolphinfish in a location that supports important fisheries. This information will be useful for management to develop stock assessments which support the sustainable use of this species.


Tagging Operation
A total of four dolphinfish were captured with set-nets from a complex in Southwest Japan ( Figure 1, Table 1). Fish were hoisted on deck to a platform using nets and placed on a stretcher. Next, during PSAT affixation, a wet towel was placed over the eyes and a seawater hose was inserted into the mouth of the fish to calm it. To prevent infection, tag heads and tethers were disinfected with alcohol and bacitracin-neomycin ointment was applied to the tag head before inserting it into the dorsal musculature.
Each PSAT was rigged with fluorocarbon tethers (200 lb. test) stainless steel crimps, nylon umbrella tag heads, and stainless steel ball-bearing swivels (Sampo no. 5, Barneveld, NY, USA) were placed~6 cm from the tag head to reduce torque and precession. Fork length (FL) was measured to the nearest centimeter.

Data Analysis
Analytical procedures and data partitioning followed methods in [26,28]. Briefly, raw light-based geolocations were provided by the manufacturer [29] and most probable tracks (MPTs) were determined using an SST-corrected (unscented) Kalman filter [30,31]. Local times of sunrise and sunset were determined from the MPTs and data were partitioned into daytime and nighttime. The data distributions were tested for normality using Kolmogorov-Smirnov tests and found to be non-normally distributed. Therefore, differences in depth and temperature between daytime and nighttime were tested by using non-parametric Mann-Whitney W-tests [26,32]. Distribution patterns were tested by examining movements relative to changes in ∆SST (i.e., temperature differences from daily mean sea surface temperature) [28,33,34]. For comparison, the ∆SST analysis also included six tags from a previous study [26]. For presentation, depth and temperature data were aggregated into 5 m and 1 • C intervals, respectively. To look for movement patterns in different areas, the greatest vertical distances between cumulative distribution functions (D N ) among tags (from two-sample Kolmogorov-Smirnov tests) were formatted into an input matrix for UPGMA (unweighted pair-group method using arithmetic averages) clustering using Euclidean distances [28]. Dolphinfish tagging data from the previous [26] and the present study were used in the cluster analysis.

Results
PSATs were deployed on four dolphinfish (94-102 cm FL) with the deployment details, retention periods, and data return rates provided in Figure 1 and Table 1. Straight-line distances from deployment to pop-up locations ranged from 63 to 204 km with speeds of 3.95-12.4 km/day (Table 1). According to the MPTs, fish #64585 moved eastwards and then exhibited a northward movement along the coast, and fish #64587 moved northwards from the tagging area to a point near Tsushima Island ( Figure 1). Straight-line movement suggests #45912 moved northwards ( Figure 1) until it reached the 30-day programmed pop-up date near the Japan Sea. Fish #45914 moved westwards for five days-at-liberty after which the tag detached ( Figure 1). Table 1. Details of tag deployments, pop-up and percentage data received from pop-up satellite archival tags (PSATs) deployed on dolphinfish in the Tachiura set-net (Northern East China Sea). DAL = days-at-liberty; the straight-line distance was measured from deployment to pop-up locations. Data return rate represents the average of depth, temperature, and geolocation data (see text for explanation [29]   The maximum depth attained was 94 m (Figure 2), with temperatures ranging from 17.78 to 23.05 • C. Dolphinfish exhibited significant diel movement patterns (p < 0.05, Figure 3) and spent more than 80% (20-22 • C) of time near the surface (<5 m) during daytime and 40% of their time at the surface during nighttime. More extensive and variable vertical excursions occurred at nighttime but were confined to the mixed-layer depth (MLD) above 20 • C and all fish exhibited pronounced crepuscular transitions ( Figure 4). Figure 5 shows daily SST, minimum temperature, and maximum depth experienced by the tagged fish. Fish #45912 experienced daily temperatures that significantly decreased over time as it moved to a different water mass at higher latitudes ( Figure 5). than 80% (20-22 °C) of time near the surface (<5 m) during daytime and 40% of their time at the surface during nighttime. More extensive and variable vertical excursions occurred at nighttime but were confined to the mixed-layer depth (MLD) above 20 °C and all fish exhibited pronounced crepuscular transitions ( Figure 4). Figure 5 shows daily SST, minimum temperature, and maximum depth experienced by the tagged fish. Fish #45912 experienced daily temperatures that significantly decreased over time as it moved to a different water mass at higher latitudes ( Figure 5).   Cumulative percentage of the temperature readings expressed as ∆SST showed that tagged dolphinfish spent >95% of movements within 3 • C of SST ( Table 2). The aggregated temperature-depth profile indicates residency within the MLD ( Figure 6). The cluster analysis ( Figure 7) suggests a clinal trend in thermal clusters with limited movement or mixing among locations, but the tag retention times were too short to adequately test the assumption of mixing or population connectivity.   Cumulative percentage of the temperature readings expressed as ΔSST showed that tagged dolphinfish spent >95% of movements within 3 °C of SST ( Table 2). The aggregated temperaturedepth profile indicates residency within the MLD ( Figure 6). The cluster analysis ( Figure 7) suggests a clinal trend in thermal clusters with limited movement or mixing among locations, but the tag retention times were too short to adequately test the assumption of mixing or population connectivity.    Figure 7. UPGMA (unweighted pair-group method using arithmetic averages) clustering using the Kolmogorov-Smirnov DN dissimilarity distance matrix among samples constructed from temperature data for dolphinfish in this study and a previous study [26]. Goodness-of-fit between the matrices and dendrograms was measured by the cophenetic correlation (0.965 for daytime; 0.936 for nighttime) with correlations >0.9 considered very good [33].

Discussion
This study provides new insights into the movement behavior of dolphinfish in the Northern East China Sea. To our knowledge, this is the first PSAT tagging experiment with dolphinfish from a coastal set-net fishery. Compared with other studies on dolphinfish [25,26], we observed similar PSAT shedding (~80%) and retention rates [28]. In this study, dolphinfish were captured in set-nets . UPGMA (unweighted pair-group method using arithmetic averages) clustering using the Kolmogorov-Smirnov D N dissimilarity distance matrix among samples constructed from temperature data for dolphinfish in this study and a previous study [26]. Goodness-of-fit between the matrices and dendrograms was measured by the cophenetic correlation (0.965 for daytime; 0.936 for nighttime) with correlations >0.9 considered very good [28].

Discussion
This study provides new insights into the movement behavior of dolphinfish in the Northern East China Sea. To our knowledge, this is the first PSAT tagging experiment with dolphinfish from a coastal set-net fishery. Compared with other studies on dolphinfish [25,26], we observed similar PSAT shedding (~80%) and retention rates [28]. In this study, dolphinfish were captured in set-nets and were considered large enough and in good condition to be PSAT tagged. The premature pop-ups in the study were from detached shed tags and were not indicative of post-release mortality. We checked to ensure that rigged PSATs (i.e., with tether and tag head attached) would float when shed to discriminate a dead sinking fish (with PSAT attached) from a shed, floating PSAT. Many factors are thought to explain premature pop-off in pelagic fishes and sharks affixed with PSATs, including infection and biofouling [28].

Short-Term Movement Patterns
In the Pacific Ocean, dolphinfish are found year-round within latitudes 30 • [35]. In the Sea of Japan, dolphinfish show seasonal movements from summer to autumn when they are targets of commercial fisheries [26]. Dolphinfish move southwards in early summer and northwards in early winter in Taiwan, in association with movements into warm water during autumn and winter, but this pattern was not found in this study. According to SSTs in the East China Sea in November 2018, temperatures in the PSATs largely paralleled the 20 • C isotherm (Figure 1). In this study, we estimated dolphinfish spent >80% of time above 20 • C (Figure 8), which indicates cooler preferences compared to similar distribution patterns found in the Western Central Atlantic (25-28 • C), Gulf of California (28-30 • C) [10,25], and Southeastern Taiwan (27-29 • C) [26]. The preferred thermal habitat of dolphinfish was generally considered >20 • C in the Northern East China Sea [24], which corresponds favorably to our study. In the Northwest Pacific Ocean, the fishing grounds of dolphinfish are the Sea of Japan to the northeast of Taiwan during autumn and winter [17]. A prominent feature, the Tsushima Warm Current through the Tsushima Strait, flows northwards into the Sea of Japan from the East China Sea [36][37][38][39]. It is likely that dolphinfish near Tsushima Island move northwards with the Tsushima Warm Current. Based on foraging studies [40], adults feed mainly on pelagic fish in Japan, especially Japanese anchovy (Engraulis japonicus). On an annual basis, dominant cohorts of Japanese anchovy have been reported in the East China Sea in spring and autumn [24] and dolphinfish probably mirror movements of these recruits. This finding suggests that Japanese anchovy could influence dolphinfish residency patterns in the East China Sea. dolphinfish are the Sea of Japan to the northeast of Taiwan during autumn and winter [17]. A prominent feature, the Tsushima Warm Current through the Tsushima Strait, flows northwards into the Sea of Japan from the East China Sea [36][37][38][39]. It is likely that dolphinfish near Tsushima Island move northwards with the Tsushima Warm Current. Based on foraging studies [40], adults feed mainly on pelagic fish in Japan, especially Japanese anchovy (Engraulis japonicus). On an annual basis, dominant cohorts of Japanese anchovy have been reported in the East China Sea in spring and autumn [24] and dolphinfish probably mirror movements of these recruits. This finding suggests that Japanese anchovy could influence dolphinfish residency patterns in the East China Sea.

Vertical Distribution
Dolphinfish mainly occupied the uniform sea surface (<5 m) and exhibited significant differences in depth and temperature distributions between daytime and nighttime. More extensive

Vertical Distribution
Dolphinfish mainly occupied the uniform sea surface (<5 m) and exhibited significant differences in depth and temperature distributions between daytime and nighttime. More extensive vertical movements were made during nighttime rather than daytime, which corresponds with previous tagging studies using acoustic tags, electronic tags, and PSATs [23][24][25][26]. This diel pattern, however, is contrary to the behavior of large pelagic apex predators such as sharks, istiophorid billfish, bigeye tuna, and broadbill swordfish, which are characterized by deeper diving during daytime with shallow movements at nighttime [28,[41][42][43][44].
The opposite diel patterns exhibited by dolphinfish compared to larger apex pelagic predators may reflect predator avoidance [26] and/or feeding preferences. As mainly visual predators, dolphinfish forage around the clock [2,5,45] and tend to prefer environments with good visibility [2,46]. There is evidence to suggest that there are two cohorts of Japanese anchovy in the East China Sea during spring and autumn [24,47] and during daytime, dolphinfish reside primarily at the surface to presumably forage on these resources. Studies have reported that changes in vertical movement patterns of large planktivores such as whale sharks [48], basking sharks [49], and piscivores such as porbeagle sharks [50] and great white sharks [51] are adaptable and can be altered to pursue vertically migrating prey. This plasticity in foraging behavior can be regarded as a shift in feeding strategy that increases prey encounters whilst minimizing energy expenditure [24,27,48]. Unfortunately, we did not analyze stomach contents to test these relationships, but diel vertical movement patterns varied according to time of day and likely reflect changes in prey distribution. Crepuscular dives are likely related to targeting aggregating prey organisms that adjust their own diel cycles based on ambient light, foraging, predator avoidance, and physiological limitations [44].
Ambient temperature is a critical factor affecting the vertical movement of dolphinfish [12,24,26]. In this study, dolphinfish occupied ambient temperatures within 3 • C of the warmest water available. However, comparable dolphinfish studies indicate different ∆SST patterns: Atlantic (8 • C), Eastern Taiwan (6 • C), and Kagoshima Bay, Japan (2 • C) ( Table 2). Although dolphinfish exhibit the ability to experience a wide range of temperature changes, depth distributions in this study and others appeared to be limited by temperature gradients and occupancy in the mixed layer (Table 3). In the Atlantic, seasonal changes in the thermocline and oxycline have been shown to affect the vertical distribution of dolphinfish [25,27,52,53]. The Northern East China Sea is mostly continental shelf and is relatively shallow, but the vertical temperature structure changes seasonally [54] and mixing occurs from late autumn to spring [27] with a strong thermocline in late spring through summer. The mixing of different water masses with varied temperature gradients is one of the factors thought to affect the vertical movements of the dolphinfish [24], and the poleward expansion of dolphinfish in the Eastern Pacific Ocean [55]. The cluster analysis suggests a clinal trend of distinct thermal clusters with limited mixing among locations, but tag retention times were too short to test the hypothesis of mixing and connectivity.

Future Research and Management
General inferences into the vertical movements of dolphinfish in the Northern East China Sea were presented, which will be particularly useful for management. The horizontal and vertical movement behaviors of dolphinfish provide fishery-independent information to inform stock assessments in Taiwan [21]. Due to similarities in the vertical extent of their habitat, dolphinfish in Southeastern Taiwan and Northern East China Sea appeared to utilize similar habitats and were hypothesized to comprise the same stock [20]. To understand the connectivity and linkages of dolphinfish between Japan and Taiwan, additional research is required to further correlate and validate movement patterns over larger time scales. One of the major outputs provided by PSATs to inform fishery management is the extent of the preferred habitat and habitat utilization patterns, which form an essential component of plans to improve and protect fishery resources. With information on vertical movement patterns, alterations to the depth of fishing gears and deployment times could be made to mismatch diving patterns and susceptibility to fishing gears in order to decrease dolphinfish by-catch. Results from this study were hampered by small sample sizes and, because only large dolphinfish were tagged, results may not be indicative of younger and juvenile dolphinfish. Future research should target different age and size classes with conventional and smaller implantable archival tags to better define the vertical and thermal niche and changes in habitat use to understand population connectivity and movement corridors.