INTRODUCTION

Paleontological approaches have been very useful for understanding changes in environmental conditions through geologic time. Phylum Mollusca is a large fauna mainly comprised of shellfish and arthropods. Mollusca are divided into seven classes: Aplacophora, Polyplacophora, Monoplacophora, Gastropoda, Cephalopoda, Scaphopoda, and Bivalvia. The fossils of Gastropoda, Bivalvia, and Scaphopoda are the main components of shellfish discovered in sedimentary strata from various regions and time. Therefore, paleontologists have used fossils of Gastropoda, Bivalvia, and Scaphopoda to study paleoenvironment, paleogeography, biostratigraphy, paleoecology, and evolution biology.

Age determination of fossil material using radiocarbon decay, along with oxygen and carbon isotope ratios, has provided time and temperature information for understanding the formation timing of sedimentary strata, sea temperature, and sea-level changes, paleogeography, and paleocurrents.

Quaternary deposits are widely distributed on the continental shelf of the Korean peninsula. Mollusc shells are plentiful in the coarse-grained relict sediments of the southern continental shelf. Recent species well suited for the present ocean environments (deep water) are mixed with fossil species that inhabited past ocean environments when the sea level was lower (Lee, 1997a, 1997b, 1998; Yoon et al., 1994). Late Pliocene–Pleistocene fossil species such as Mizuhopecten tokyoensis hokurikuensis, Amussiopecten praesignis, and Turritella saishuensis have been reported in the front sea area on the continental shelf of Geoje Island and Ulsan, Korea (Lee, 1997b, 1998). Molluscan species collected from this study area provided a very useful tool for understanding the paleoenvironment on the continental shelf of the South Sea.

About 15,000 years ago, Korean shorelines were displaced seaward by more than 100 km as sea level dropped because of widespread glaciation. Between 15,000 years ago and the present, sea level has been rising in response to deglaciation. Coarse sediment (sand and gravel) blankets the outer shelf where water depths are deep and bottom and physical energy is minimal. These coarse-grained deposits on the outer two-thirds of the Korean shelf are not consistent with the present low-energy conditions. Such material was termed relict sediment by Emery (1952) because it accumulated at an early time and under very different depositional conditions (Pinet, 2000). Thus, sediment on the outer Korean shelf could be classified as relict because it was deposited under significantly different environmental conditions. Although “relict sediment” is a valuable term used to describe sediment that has been displaced from its original environment of deposition, “palimpsest sediment” is a convenient term for the mixture of relict and modern materials (Swift, Stanley, and Curray, 1971).

The Korean southern shelf has produced faunas or floras from relict and palimpsest sediment. After the last glacial maximum (LGM), the current speed increased continuously as sea level rose around Korea Strait. High current speeds facilitated the reworking/mixing of sediments. Shell valves in palimpsest sediment must have had an autochthonous history to provide valid chronological data for a sea-level curve.

The depth of the sea level lowstand on the continental shelf during the LGM needs to be determined to understand the paleoenvironment of the South Sea continental shelf (Figure 1). The characteristics of this environment include the nature of the land bridge between the Korean peninsula and Japan, the flow of the paleo-Tsushima Current centering around the Korea Strait, and the desalination of the East Sea. Park et al. (2000) and Suk (1989) have conducted radiocarbon age determination of the shell remains from the sea around Tsushima and concluded that the last glacial sea level was approximately 130 m below the present. These results contradict the research results reported for the East Sea desalination at the last glacial period (Feng, 1983; Kim et al., 2000), which indicated the formation of a land bridge between Korea and Japan and disputed the inflow of the Tsushima Current.

The purpose of this study was to understand the variation of sea level and paleoenvironment after the last glacial period. Our main tools were ¹⁴C radiocarbon age determinations of the dominant and character species, and interpretation of species distribution and assemblage composition of the shell remains produced from the surface sediments of the Korean South Sea shelf.

STUDY AREA SETTING

The South Sea coast of Korea is a typical Riatype coast with intricate coastlines and multiple inner bays, and was formed by rising sea level after the last glacial period. Sedimentary deposits primarily consist of recent fine-grained sediments known to be delivered from seas influenced by river sources or tidal currents (Chough, 1983). The present study area (Figure 1) is divided into two sea areas, one of the southeastern continental shelf centering around Tsushima, and the other around Jindo on the southwest coast. The southeastern continental shelf of Korea, the sea area between southeast Korea and southwest Japan, is the gateway where the Tsushima Current, a tributary of the warm Tsushima–Kuroshio Current, flows into the East Sea. At the East Sea, an influx has formed a very steep landform sloping to great depths. The Tsushima islands are located at the center of the continental shelf, whereas the Korea Strait is at the northwest of the island and the Tsushima Strait is at the southeast of the island. The Tsushima sea area is separated into two regions. The Kyusyu coast sea area of Japan has depths of less than 70 m over a gentle undulating surface; the Tsushima coast sea area is distinguished by the relative absence of undulation, where discontinuous and shallow troughs are present between the Tsushima islands (Ujiie, 1973). In the Korea Strait, the Tsushima trough (230 m depth; Park et al., 2000) slopes upward toward the coast of the Korea peninsula. The Tsushima trough is the deepest area in the South Sea continental shelf. During the last glacial epoch, shellfish inhabited the littoral zone in this area (Habe and Kosuge, 1970).

Many tectonic lines with NE-SW and NNE-SSW orientations occur in the southeastern continental shelf (Katsura and Nagano, 1976). These faults are known to be main factors forming the Korea Strait (Ministry of Science and Technology, 1994).

The southwestern sea area of the Korean peninsula is an archipelago with the islands of Hauido, Jangsando, and Anjwado surrounding the island of Jindo. It is a complicated sea area with interlacing networks of large and small channels with rapid, semidiurnal tidal currents. During mean spring tide, the maximum flood and ebb currents are 1.9–3.7 and 1.6–3.9 knots, respectively (Office of Hydrographic Affairs, 1996). Sedimentary deposits are composed of coarse-grained sand and gravel. Numerous shellfish remains, such as Arca boucardi (from the last lowstand), have been found in surface sediments (Lee, 1997a).

METHODS

Surface sediments were collected from 31 localities, 26 in the southeastern continental shelf and 5 in the southwest sea area (Figure 1). Samples were taken with the use of a grab sampler (Van Veen type) and frame dredge (5-minute drags). During drags, an echo sounder (Furuno, model JFV-200) was used to verify and maintain constant depth. Drags were made at constant slow speeds between 1 and 2 knots.

The species of modern shell remains were identified in reference to illustrations and reference plates (Gakken, 1975; Habe, 1977; Habe and Kosuge, 1967; Min Shell House, 2001) and those of the Pliocene–Pleistocene fossils were identified from descriptions in Habe and Kosuge (1970) and Okamoto and Honza (1978). Sediment samples were initially prepared for size analysis by completely eliminating organic materials and carbonates through the sequential addition of 10% hydrogen peroxide and 0.1 N hydrochloric acid. Grain size fractions were determined using sieve analysis for the coarse fraction and a Sedigraph 5100 automatic grain size analyzer for the fine fraction. ¹⁴C radiocarbon ages were determined by accelerator mass spectrometry (AMS) and liquid scintillation counter (LSC). AMS was done on small samples of less than 7 g at the radioactivity laboratory at Waikato University of New Zealand. LSC was done at Chonnam University of Korea on larger samples of more than 14 g. Samples for age analysis were selected from the dominant and characteristic species of shell remains produced from each locality. The degree of shell reworking/mixing was assessed by comparison of present water depth with known habitat depths for the species and preservation state of the shell.

RESULTS

DISCUSSION

Surface sediment of the South Sea shelf of the Korean peninsula is a mixture of recent fine-grained muddy sediments and coarse-grained relict and palimpsest sediments with shell material. Mizuhopecten yessoensis and M. tokyoensis hokurikuensis retrieved from locality nos. 0806 and 0807 were fossil species known to occur during the late Pliocene–Pleistocene. Late Pleistocene fossils such as Amussiopecten praesignis and M. tokyoensis hokurikuensis were reported in water depths of 100∼150 m southwest of Yamaguchi Prefecture, Japan (Okamoto and Honza, 1978; Okamoto and Ibaraki, 1988). Moreover, M. tokyoensis hokurikuensis and Turritella saishuensis were reported from depths of 150 m northeast of Busan and around Geoje Island in Korea (Lee, 1997b, 1998). These fossil species are representative component species of the Seogwipo fauna from the late Pliocene Seogwipo Formation in Jeju, Korea (Yoon, 1988). They are also known as principal component species of the Omma-Manganji and Kakegawa fauna in Japan. In our study area, M. tokyoensis hokurikuensis and M. yessoensis occurred at locality nos. 0806 and 0807. These species appeared north of Tsushima between the Korean peninsula and Japan, making a distributional band in the direction from the southwest to the northeast. The limited and directional distribution toward the northeast correlates with the tectonic movement of the southeastern continental shelf where the northeastern tectonic lines develop.

Fossil species on the South Sea shelf of the Korean peninsula have habitat requirements that may suggest a history of sea level rise in this area. The late Pliocene sedimentary strata comprising shell layers during LGM are exposed on the submarine surface of the continental shelf.

Twelve shell assemblages including GVPA were discriminated from the southeastern continental shelf. Among the assemblages, GVPA, GPPA, GCA, and GVCA appearing in the 11 localities with Glycymeris rotunda were the main species categorized (Table 2). These assemblages were distributed along the southeastern continental shelf. Radiocarbon data indicated that Glycymeris rotunda had a temporal range from 6461 ± 135 YBP to modern times. The assemblages appeared as Glycymeris fauna inhabiting a confined region during the particular period. The main species of Glycymeris fauna was Glycymeris rotunda with Cryptopecten vesiculosus, Nemocardium samarangae, Pitar noguchii, Phacosoma japonicum, Ventricoloidea foveolata, and Paphia amabilis as accompanying species. These are typical fauna after the last glacial period created about 6000 years ago in the area around the southeastern continental shelf and have been preserved to the present. According to Katsura and Nagano (1982), there was a channel-like seaway linking the western North Pacific and the East Sea at the west of Tsushima when the sea level was 130 m below present, at the last glacial period. The peripheral zone of the channel-like seaway was affected first by the rising sea. At the same time, the southern sea area (loc. nos. 0102 and 0706) became a good habitat for Glycymeris fauna along with the rising sea level. In the southwestern sea area, A. boucardi of the Arca fauna inhabited depths between the rocky littoral zone and about 10 m below sea level about 3000 years ago.

Determination of the depression depth of the sea level low-stand at the South Sea shelf during LMG is an important step to developing an understanding of the paleoenvironment, especially the land bridge between the Korean peninsula and Japan, the flow of paleo-Tsushima Current, and the desalination of the East Sea. Feng (1983); Park et al. (2000), and Suk (1989) investigated the sea-level change of the South Sea area after the last glacial period. Results of these investigations have revealed discrepancies in depression depths of the last glacially induced lowstand about 15,000 years ago (Figure 5). Feng (1983) considered the sea-level depression during LMG to be about 160 m below the present sea level, compared with 130 m according to Suk (1989) and Park et al. (2000). In addition, Park et al. (2000) insisted on the formation of a channel to the west of Tsushima during LGM and the inflow of the paleo-Tsushima Current to the East Sea, whereas Kim et al. (2000), on the other hand, asserted the desalination of the East Sea on the basis of the results of oxygen and carbon isotope analyses from Ulleung Basin cores.

The East Sea is a back-arc basin with an average depth of 1680 m and a semi-isolated marginal sea connected with the Okhotsk Sea, East China Sea, and North Pacific by shallow straits (Korea Strait, 130 m; Tsugaru Strait, 130 m; Soya Strait, 55 m; Tatarskiy Strait, 15 m). The Tsushima Current flows into the East Sea through the Korea Strait and flows out mainly through the Tsugaru and Soya straits.

Significantly lower sea levels during the last glacial epoch formed a land bridge between Korea and Japan with Tsushima in the middle. However, opinions disagree over this issue, as with the inflow of the paleo-Tsushima Current to the East Sea (Gorbarenko, 1983). Suk (1989) harvested 24 shell remains and Park et al. (2000) harvested 17 drill samples for age determination by AMS. Feng (1983) also utilized shell remains, but did not discuss issues of time-averaging and autochthonous formation. There is partial disparity in the classification of shells and depth-related habitats. These factors can cause different interpretations of the radiocarbon age data about shells.

In this study, the sea level of the South Sea area was determined from radiocarbon data of the littoral species (Scapharca broughtonii, Macoma calcarea, Callithaca adamsi, and Mya truncate: loc. no. 0704). This suggests that sea level fell by about 150–160 m below present sea level during the LGM.

These specimens had large valves and were well preserved, in the exterior margin (Figure 6). In the marine environment, prolonged posthumous exposure on submarine surfaces degrades the shell material through the serial process of reorientation and transportation, disarticulation, fragmentation, and corrosion. Characteristic aspects of the sequential process are revealed in the pattern of hard part degradation (Brett and Baird, 1986; Martin, 1999). Shells experience separation of attached valves within several weeks after death (Schafer, 1972). Corrosion primarily occurs in the high-energy environment, with coarse, poorly sorted deposits (Driscoll and Weltin, 1973). Littoral zone species like Scapharca broughtonii (which was produced in the present research area) are parautochthonous and are more affected by corrosion by refloating or reworking of the seawater flow, rather than transportational fragmentation. These parautochthonous materials were considered to have formed in the littoral zone–shallow sea environments. Seismic investigation at loc. no. 0704 indicated a former beach environment (Park et al., 2000), and this result was consistent with the present research results.

Radiocarbon data (Table 3) illustrated a variety of ¹⁴C radiocarbon dates among the species within the same locality (loc. nos. 0703 or 0704). At loc. no. 0704, for example, Crenulilimopsis oblonga was dated at 3655 ± 60 YBP, whereas Macoma calcarea had an age of 15,694 ± 328 YBP, giving an age difference between the two species of about 11,939 years. At loc. no. 0703, Ventricoloidea foveolata had an age of 2053 ± 110 YBP and Cyclocardia sp. was dated at 8643 ± 72 YBP, yielding a difference of about 6590 years. At loc. no. 0706, Ventricoloidea foveolata was dated at 1258 ± 118 YBP and Cryptopecten vesiculosus at 8779 ± 63 YBP, yielding a difference of about 7521 years. At loc. no. 0808, Glycymeris rotunda yielded a date of 676 ± 97 YBP and Trapezium bicarinatum had a date of 7851 ± 61 YBP, producing a difference of about 7175 years. Thus, the age differences between the species sampled from the same locality ranged from 6590 years to 11,939 years.

According to Walker and Bambach (1971), “time-averaging” is defined as “accumulate from the local community during the time required to deposit the containing sediment.” It is interpreted broadly to mean not only the age differences but also ecological combination of unique species or accumulation of their remains. Fujiwara and Kamataki (2003) explained that the mixture of fossil species having different ages in one stratum is an example of the phenomenon of time averaging. This phenomenon appears in shell samples of ages ranging from several decades to more than 3500 years in the littoral zone (Flessa, Meldahl, and Cutler, 1990) and from several hundred years to about 1500 years in cheniers of the Yellow River delta, China (Saito et al., 1999). The phenomenon inhibits the accurate judgment of sedimentation time because shells are occasionally delivered (or remain) to an area after death and are subsequently found as buried fossils in different environments from their native habitat. This can occur when shells are buried after death but exhumed by erosion and later transported to a new site. Radiocarbon age differences in species produced at the same locality and the ecological blending of the littoral zone species in the present research are considered to reflect the phenomenon of time averaging. However, instead of using an “average time” to describe the surface of the continental shelf, we sorted the shells into depth-sensitive habitats and have found that shells with similar depth constraints had similar dates. This allowed us to reconstruct water-depth estimates on the shelf at different periods of time between 15,000 years ago and the present.

CONCLUSIONS

Our coupled age and habitat data indicated that the sea level of the South Sea area was about 150–160 m below present sea level about 15,000 years ago. This drop allowed a land bridge to form between the Korea peninsula and Japan, which remained above sea level for about 6000 years. The sea level rose to approximately 60 m (below present sea level) about 9000 years ago, remained at this depth until about 4000–5000 years ago, and then rose to 10–20 m below the present level about 3000–4000 years ago.

Glycymeris fauna were created in the sea area of the southeastern continental shelf about 6000 years ago, while Arca fauna prospered in the southwestern sea area about 2000–3000 years ago when the sea level was 10–20 m below present sea level.