Identifying the source rookery of green turtles (Chelonia mydas) found in feeding grounds around the Korean Peninsula

Introduction

Management of sea turtle breeding and feeding grounds under the regional management unit framework is an effective approach to conserving endangered sea turtles (Wallace et al., 2010). The management unit (MU) exhibits distinct demographic processes, such as genetic composition and life history, and is a functionally independent unit for turtle conservation (Mortiz, 1994). Various conservation efforts have been conducted in breeding populations, such as protecting breeding sites, releasing captive-breeding turtles, and running educational programs (Hamann et al., 2010; Barbanti et al., 2019). Considering the frequent site fidelity of sea turtles for breeding and feeding grounds over their lifespan, the protection and management of both these sites are crucial for sea turtle conservation (Hamann et al., 2010). Sea turtles from different rookeries often gather on feeding grounds, making it important to protect multiple rookeries (Nishizawa et al., 2013; Piovano et al., 2019). Therefore, conservation efforts for feeding grounds need to be increased. Understanding the interaction between breeding and feeding grounds can be difficult due to sea turtles’ wide distribution and complex life cycle. However, recent advancements in genetic analyses have enabled researchers to gain information about the turtles’ genetic variability, composition, and origin. To assess these, both genetic analyses and mixed stock analysis (MSA) have been conducted on turtles captured or incidentally collected (e.g., bycatch or stranded) (Nishizawa et al., 2013; Piovano et al., 2019). Previous studies on feeding grounds were mainly conducted in easily accessible or abundant areas where breeding grounds were nearby (Shamblin et al., 2012; Read et al., 2015). Among the seven known sea turtles, green turtles (Chelonia mydas) are the most common and are found across tropical and subtropical oceans worldwide (Seminoff et al., 2015). The species is listed as an endangered species in the IUCN Red List of Threatened Species (IUCN, 2023) and is highly migratory. A recent study (Jensen et al., 2019) identified 11 phylogenetic clades of green turtles worldwide; it showed that clade VIII, which includes most turtles in rookeries in the Indo-Pacific and South Western Indian Ocean MU, has the widest distribution range in the Pacific and Indian Oceans. In the Western Pacific Ocean, six clades (III, IV, V, VI, VII, and VIII) were defined, and clades VII, VIII, III, V, and VI have main haplotypes for Japan, Indo-Pacific (IP), Central Western Pacific (CWP), South Western Pacific (SWP), and Central South Pacific (CSP) MU, respectively (Jensen et al., 2019). The regional boundaries of these genetic clades aligned well with the existing MU of green turtle rookeries (Wallace et al., 2010), probably because of their high level of fidelity to breeding and feeding grounds (Nishizawa et al., 2011). Some turtles in the Japan MU also had genetic components belonging to clades III and VIII (Hamabata, Kamezaki & Hikida, 2014), which were found in the IP and CWP MU, while those in the IP MU also had components of clades III and VII. To understand the demographic structure of sea turtle populations in breeding and feeding grounds, and their connectivity with nearby MUs, studies on their phylogenetic origins are also necessary.

In the northwestern Pacific, several green turtle rookeries exist, namely, Ogasawara (OGA), Central Ryukyu Island (CRI), and Yaeyama Island (YI) in the Japan MU, and Taiwan, Hong Kong, Lanyu Island, and Xisha Island (XI) in the IP MU (Fig. 1). According to Okuyama et al. (2009), the dispersion of turtles that hatch in these rookeries occurs through passive transport facilitated by the Kuroshio Current, Kuroshio branches, and other components of the northwestern Pacific Gyre. The green turtles observed at the feeding grounds located in the East China Sea, specifically the YI and CRI, originated from the YI, Southeast Asia, Micronesia, and Marshall Island rookeries in the Western Pacific region (Nishizawa et al., 2013). In contrast, turtles found at feeding grounds located in the northwestern region of mainland Japan, such as Nomaike, Muroto, Kanto, and Sanriku, originated mainly from the OGA, although some originated from the CRI and YI (Hamabata et al., 2015). The Sanriku feeding ground in the northernmost region also reported the presence of some Hawaiian turtles (Nishizawa et al., 2014). Within the northwestern Pacific region, there are additional important feeding grounds for green turtles, namely the Northeast China Sea, which spans Japan, China, and Korea, including the West (Yellow), South, East (Japan), and Jeju Seas. Nevertheless, the current understanding of the genetic composition and source rookeries of turtle populations in this region remains elusive.

Chelonia mydas is the most frequently observed species in Korean waters (Kim et al., 2017). Juvenile and adult green turtles are observed at an average rate of 5–10 turtles per year. Ecological and satellite tracking studies have been conducted to gain insights into the physical attributes, habitat preferences, and movement and migration patterns of green turtles (Jang et al., 2018; Kim et al., 2022, 2024). Such studies have shown that green turtles actively use the sea around the Korean Peninsula and migrate to breeding grounds in both the Japan MU and the IP MU. Nevertheless, the current understanding of the genetic composition, genetic diversity, and source rookeries remains limited and requires immediate investigation. Most sea turtles inhabiting the marine ecosystems surrounding the Korean Peninsula are acquired through the processes of by-catch and stranding (Kim et al., 2017). Local feeding populations of these turtles are jeopardized by a variety of issues, including fishing, construction, and pollution (Moon et al., 2009; Kim et al., 2017). Understanding the source rookeries and MUs of green turtles, for which the Republic of Korea has implemented various government conservation efforts (Kim et al., 2022, 2024; Moon et al., 2022), is crucial for the long-term conservation of sea turtles. The inclusion of relevant studies was crucial and required immediate attention.

To evaluate the genetic vulnerability of the Korean population, we verified genetic composition and genetic diversity, as well as mixed stock analysis (MSA) to infer the source rookery of green turtles caught as bycatch or found stranded in feeding grounds around the Korean Peninsula. Our results provide crucial insights into the rookeries and MUs that are the focus of conservation efforts for endangered green turtles in the northwestern Pacific Ocean.

Materials and Methods

Results

We collected samples from 40 turtles, comprising 21 adults, 15 juveniles, and 4 of unknown maturity. Among these, 24 were female, 3 were male, and the sex of the 13 turtles remained undetermined. Of the 40.9% that were sampled between June and October, the highest number we collected in August (25%). Seasonally, 23 turtles were sampled in summer (June–August), 14 in fall (September–November), 2 in spring (March–May), and 1 in winter (December–February) (Table 1). Geographically, the East Sea accounted for 14 turtles, followed by 13 in the Jeju Sea, 12 in the South Sea, and 1 in the West (Yellow) Sea (Tables 1 and S1).

Genetic composition analysis

Ten distinct CR haplotypes (GenBank accession numbers PP691472–PP691481) were identified from 41 polymorphic sites (15 singleton variable sites and 26 parsimony-informative sites) (796 bp, Table 1). All haplotypes were concordant with previously documented haplotypes in this species (Hamabata, Kamezaki & Hikida, 2014; Hamabata et al., 2015, 2020). Orphan haplotypes were not observed. Haplotype Cmp39.1, found in 18 turtles (45.0%), exhibited the highest frequency, followed by haplotype Cmp50.1, detected in six turtles (15.0%). Each of the three turtle individuals (7.5%) had haplotypes Cmp49.1, Cmp121.1, or Cmp128.1. Haplotypes Cmp53.1 and Cmp127.1 were found in two turtles (5.0%), and Cmp54.1, Cmp70.1 and Cmp79.1 were found in one turtle (2.5%) (Table 1). The haplotype diversity was 0.771 ± 0.060, whereas the nucleotide diversity was 0.01338 ± 0.00190.

Within the haplotype network, six haplotypes (CmP39.1, CmP70.1, CmP79.1, CmP121.1, CmP127.1, and CmP128.1) were assigned to the Japan clade (Fig. 2). The other three haplotypes (CmP49.1, CmP50.1, and CmP53.1) were associated with the IP clade but were exclusively found in the Japanese population except the CmP49.1 haplotype that widespread in the IP MU. The last haplotype (CmP54.1) grouped with the CWP clade and was previously found only in the Japanese population (Fig. 2).

Mixed stock analysis

In the flat MSA results, rookeries (YI, CRI, and OGA) in the Japan MU accounted for 87.7% of the 40 turtles examined (Fig. 3A, Table S4). Rookeries (IP and XI) in the IP MU contributed 7.4% of the total. A closer look at the Japan MU, OGA, and CRI rookeries contributed 41.9% and 37.8% of the 40 turtles examined, respectively. When considering rookery size, the contribution ratio changed slightly to 46.1% for CRI and 38.7% for OGA.

In the flat MSA results of the season, OGA and CRI contributed 48.0% and 17.5% of the 23 turtles in the summer, respectively, whereas CRI, OGA, and YI contributed 38.6%, 14.9%, and 14.7% of the 14 turtles in the fall, respectively (Fig. 3B, Table S4). When considering rookery size, the contribution of OGA increased to 51.0% in summer, whereas CRI increased to 57.2% in autumn. In the comparison of specific feeding grounds, OGA mainly explained the turtles in the Jeju Sea (35.0%), whereas CRI did so in the South Sea (32.5%) and East Sea (20.8%) (Fig. 3C, Table S4). Specifically, 20.4% of the 14 turtles in the East Sea came from the Indo-Pacific rookery, and 17.8% came from the YI rookery. When considering rookery size, the contribution of OGA increased to 40.6% in the Jeju Sea and 22.6% in the East Sea. The contribution of the CRI increased in all feeding grounds. In addition, we conducted MSA based on the maturity groups of the juvenile and adult. The contribution patterns were similar between juveniles and adults and overall to the combined data (Fig. 3D, Table S4). In the flat MSA results, OGA and CRI contributed 29.8% and 19.8%, respectively, in juveniles and 33.9% and 36.2%, respectively, in adults. When considering rookery size, the contribution of CRI increased to 36.5% in juveniles and 44.5% in adults (Fig. 3D, Table S4).

Discussion

Most green turtles in the feeding grounds around the Korean Peninsula have a genetic composition belonging to the Japan clade, although some have sources from the IP and CWP clades. This result is consistent with that of a previous phylogenetic study (Jensen et al., 2019). Haplotype network analysis revealed that six haplotypes were associated with the Japan clade, whereas the other four haplotypes were grouped with either the IP or CWP clade. However, three of these four haplotypes have been identified previously only in the Japanese population. Only one haplotype (Cmp49.1) was found across the populations in Japan and the IP MU. Overall, 9 of the 10 haplotypes (37 of 40 turtles) were found in the Japanese MU. However, two pieces of evidence suggest that some turtles also originate from either the IP or CWP MU. First, as previously described, one haplotype (3 of 40 turtles, Cmp49.1) was found in populations across Japan, IP, CWP, and SWP MUs, and was the main haplotype in the IP clade (Jensen et al., 2019). Second, previous satellite tracking studies have shown that green turtles in Korean waters have migrated to Hainan Island in the Indo-Pacific Ocean (Kim et al., 2022, 2024). The genetic diversity of green turtles (0.77) in the study area was similar to the reported genetic diversity (0.65 – 0.88) in seven Japanese feeding grounds (Nishizawa et al., 2013, 2014; Hamabata et al., 2015), showing that protecting the feeding population in the Republic of Korea could be meaningful to conserve genetic diversity of green turtles in northwestern Pacific Ocean.

Green turtles in the feeding grounds in the vicinity of the Korean Peninsula originated from two geographically distant rookeries: the OGA, situated in the northern area of the North Western Pacific gyre, and the CRI, located in the Central Western Pacific gyre. The OGA rookery made the largest contribution to the feeding ground, accounting for 41.9% of the total and the CRI had the second-largest contribution at 37.8%. Both the use of oceanic currents for migration and active movements between rookery and feeding grounds could explain such results. It is well known that ocean currents are major factors in sea turtle migration (Okuyama et al., 2009; Nishizawa et al., 2013). The OGA is well-linked to the study area through the Kuroshio recirculation in the northwestern Pacific (Hurlburt et al., 1996). The main Kuroshio Current and its branch currents offer good transportation options for the CRI (Park et al., 2013; Zhong et al., 2021). On the other hand, green turtles also actively move between their rookery and feeding grounds. Most turtles in feeding grounds along the southeastern coast of mainland Japan originated from close OGA rookery while turtles in the southern Ryukyu feeding grounds come from geographically close southern Ryukyu or the Indo-Pacific regions (Nishizawa et al., 2013; Hamabata et al., 2015).

The contribution of rookeries to feeding grounds was influenced by the season and specific feeding grounds in the study area. First, there was a clear seasonal pattern in the number of turtles collected and the influx from different rookeries. The majority of the turtles were collected between June and November. Conversely, only three turtles were collected between December and May, indicating that the majority of turtles came from summer and fall, likely because of low temperatures during winter (Kim et al., 2017). It is well known that the feeding activity of sea turtles largely depends on water temperature (Reisser et al., 2013). In summer, most turtles originated from the OGA. The Kuroshio recirculation, generated by the Kuroshio Current in the OGA region, may facilitate the movement of green turtles toward the west or north direction (Hurlburt et al., 1996). This allows them to reach the Kuroshio Current, which flows into the northern parts of the Ryukyu Islands or the southeastern sea of mainland Japan. The turtles then arrive at feeding grounds in the Republic of Korea. In particular, given that the dispersion of post-breeding green turtles at OGA occurs in summer (Kondo et al., 2017), there is a potential for augmentation in the westward and northward migrations of these turtles.

In contrast, the CRI rookery accounted for the largest proportion of turtles in the fall. When considering rookery size, it increased by more than 50%. There were two potential variables for this pattern. First, during this time of year, the Kuroshio Current, which aligns with the northward surface wind direction and monsoon activity in the area, increases in speed and volume (Isobe, 1999; Zhong et al., 2021). Moreover, the current ran closer to the shores of the Ryukyu Islands. Such changes could increase the chance of green turtles moving northward owing to currents in the East China Sea. The Kuroshio Current is widely recognized for its substantial influence on the migration patterns of diverse marine organisms in the Western Pacific (Andres et al., 2015). Second, typhoons are often generated in July and August (Choi, Cha & Kim, 2012). Typhoons originating in the southern Pacific Ocean travel northerly and pass across the southern and central Ryukyu Islands (Choi, Cha & Kim, 2012). Typhoons could potentially enhance the chances of green turtles migrating to the feeding grounds in the study area. Previous studies have shown that typhoons transport diverse flora and fauna from the East China Sea to the Korean Peninsula (Osozawa et al., 2021; Lee et al., 2023). Our findings suggest that the contributions of the OGA and CRI rookeries to the study area have distinct seasonal patterns, largely based on the activity of the Kuroshio Current and its branches.

The influx of green turtles into the study area fluctuated depending on the specific feeding grounds within the study area. The OGA rookery primarily contributed to the Jeju Sea regardless of rookery size, whereas the CRI rookery explained more turtles in the South and East Seas than the OGA. In particular, the YI and rookeries in the IP MU contributed some of the turtles (46.2%) in the East Sea in the flat model. The observed pattern could also be attributed to regional water influx by ocean currents, such as the Kuroshio Current and its branch currents, including the Tsushima Current, as well as the Kuroshio recirculation near the OGA. The Kuroshio branch currents, such as the Tsushima Current, receive water from the Kuroshio Current and, in autumn, transport the water more directly to the East Sea (Isobe, 1999). This flow of water may bring turtles to the South Sea and East Sea from rookeries such as the YI and CRI on the Ryukyu Islands and/or the IP MU. From the OGA to the Jeju and South Seas, the southeastern coastal migration route of mainland Japan has the potential to transport turtles to the feeding grounds studied. The Kuroshio recirculation may first transport OGA turtles to the southeastern shore of mainland Japan (Hurlburt et al., 1996) and then follow the coastline westward to reach the Jeju Sea. On the other hand, in our previous satellite tracking study, green turtles released in the Jeju Sea in the Republic of Korea used across close feeding grounds including the Jeju Sea, the East Sea, and the South Sea in the Republic of Korea as well as sea nearby Kyushu Island in Japan (Kim et al., 2024). These areas are abundant in seaweed, which is food for green sea turtles (Kim et al., 2021). Also, as mentioned above, Hamabata et al. (2015) and Nishizawa et al. (2013) showed active migration of green turtles between close rookery and feeding grounds. These active movements could also affect the overall appearance of green turtles, in particular, in specific feeding grounds around the Korean Peninsula.

Conclusions

In summary, the feeding grounds in the study area were utilized by both juvenile and adult green turtles, originating primarily from the Japan MU and partially from the IP and CWP MU in the northwestern Pacific Ocean. The turtle population in the study area consisted of a distinct combination of two main geographically distant rookeries, the OGA and the CRI. The influx from these two rookeries differed based on the season and specific feeding grounds in the area. In addition, we suggest that the Kuroshio Current and its branches are crucial for the migration of green turtles to the northwest Pacific Ocean. The results of our study provide crucial insights into rookeries and MUs, which are the focus of conservation efforts in the Republic of Korea. They also support the collaboration between local governments and national parties in demographic information exchange and recovery projects to conserve green turtles effectively.

Supplemental Information