1 Introduction

The Kuroshio Extension is an ocean current that separates from the Pacific Western Boundary Current at the Japan coast (35°N, 140°E) and enters the open basin of the North Pacific Ocean (Fig. 1). It flows eastward, has properties of an inertial jet, and is characterized by the largeamplitude meanders (Stommel and Yoshida, 1972; Kawabe, 1995; Qiu, 2003). Many mesoscale energetic pinched-off eddies exist along the meandering path of the current (Kawai, 1972; Yasuda et al., 1992). The Kuroshio Extension region is widely defined as 32° – 38°N, 140°E – 180°, which can be generally divided into three segments. The first segment spans from the Japan coast to the Izu Ridge and features two large-amplitude meanders (axes at 144° and 150°E; Mizuno and White, 1983). The second segment crosses the Shatsky Rise oceanic plateau (at 159°E). Within this area, the current bifurcates into two branches: the main current continues moving eastward, and a branch current runs toward NE and eventually becomes part of the subpolar circulation (Hurlburt and Metzger, 1998). The last segment is located east of 171°E and passes the Emperor Seamount Chain, where the main current flows relatively wide and exhibits the pattern of the mean flow (Joyce and Schmitz, 1988). The Kuroshio Extension region is unique because it is one of the areas with the largest magnitude of mesoscale eddies, strongest submesoscale oceanfronts, and highest kinetic energies (Scharffenberg and Stammer, 2008; McWilliams, 2016; Su et al., 2018). Thus, it is a key region in the mid-latitude areas for studies on mesoscale and submesoscale oceanic processes. Furthermore, mesoscale oceanic eddies have been shown to affect the atmosphere and oceanic circulations through ocean-atmosphere interactions (Qiu, 2002; Small et al., 2008), eddy-mean flow interactions (Qiu and Chen, 2010; Waterman et al., 2010), and the subtropical-mode water formation (Qiu et al., 2007; Nishikawa et al., 2010); moreover, the eddies are crucial for understanding both the oceanography and climate dynamics in the mid-latitude areas.

Mesoscale oceanic eddies within the Kuroshio Extension region are intensively developed. Based on satellite altimetry data of the last 20 years, statistical characteristics of surface eddies show approximately 3000 cyclonic and 2900 anticyclonic eddy trajectories, with mean lifetimes of 10 and 11 weeks and mean radii of 70 – 72 km, respectively (Hu et al., 2018). Cold cyclonic eddies usually occur along the axis of the Kuroshio Extension and toward the south of the axis, which is surrounded with warm waters, whereas warm anticyclonic eddies often exist toward the north of the axis, where cold waters dominate (Itoh and Yasuda, 2010) (Fig. 1b). Most eddies propagate westward at a migration speed of 1 – 7 cm s⁻¹ with little meridional movement (Ichikawa and Imawaki, 1994; Ebuchi and Hanawa, 2001). Moreover, seasonal variations of mesoscale eddies within the Kuroshio Extension region indicate that the quantity, vorticity, kinetic energy, and energy intensity are relatively high in summer (and spring) (Hu et al., 2018). The mechanisms proposed to explain the formation of mesoscale eddies in the Kuroshio Extension region include barotropic and baroclinic instability induced by the horizontal and vertical shear of the mean current (Tai and White, 1990; Qiu and Chen, 2005; Xu et al., 2011). The meanders and bifurcation of the Kuroshio Extension lead to intensive energetic mesoscale eddies pinched-off from the main current. Although the aforementioned surface or near-surface characteristics of mesoscale eddies have been well known, owing to the limitation of deep-ocean hydrological data, properties of the subsurface eddy vertical structure, including temperature, salinity, current, and vorticity, remain poorly understood.

Marine seismic reflection survey is conventionally used to image sedimentary and crustal structures below the seafloor and to explore oil, gas, gas hydrate, and other geological resources. In the past two decades, the success of this technique for imaging the structure of the water column above the seafloor has been demonstrated (Holbrook et al., 2003; Ruddick et al., 2009; Tang et al., 2019). By studying the seismic images of ocean waters, a cross-discipline subject called seismic oceanography has been developed. Seismic oceanography therefore provides an important additional method to reveal the water structure information in vertical view, which is often sparsely obtained from traditional hydrographic oceanographic observations. Marine seismic data are acquired at a relatively fast speed (usually 2.5 cm s⁻¹) and with high lateral resolution (typically 6.25 m) and rapidly (approximately 1 – 2 d) produce continuous profiles in tens to hundreds of kilometers width and hundreds to thousands of meters depth. These advantages of seismic data allow us to capture detailed crosssection pictures of mesoscale or subme-soscale oceanographic features, for example, mesoscale oceanic eddies (Biescas et al., 2008; Buffett et al., 2009; Menesguen et al., 2012; Tang et al., 2014, 2020), and internal waves (Holbrook and Fer, 2005; Krahmann et al., 2008; Tang et al., 2015, 2016).

In this paper, we present four two-dimensional (2D) multichannel seismic (MCS) reflection lines collected in August 2010 and April 2012 over the Shatsky Rise oceanic plateau in the Northwest Pacific Ocean (Fig. 1), where the Kuroshio Extension crosses and bifurcates, likely resulting in the extensive formation of eddies. After seismic data were reprocessed, focusing on shallow waters above the submarine plateau, the water column reflections in the seismic images were correlated to oceanic finescale structures, i.e., eddies and fronts. Characteristics of seismic reflectors within the shallow waters were analyzed to infer the eddy properties, such as the vertical structure, classification, shape, scale, edge, and trajectory. These observations in seismic profiles were compared with satellite data, including sea level anomaly (SLA), geostrophic current anomaly (GCA), and sea surface temperature anomaly (SSTa), to further imply the dynamic process of the mesoscale and submesoscale features.

2 Data and Methods

In summer 2010 (cruise MGL1004) and spring 2012 (MGL1206), the Shatsky Rise oceanic plateau was visited twice by the R/V Marcus G. Langseth to acquire marine MCS reflection data. During the two cruises, 3350 km of deep-penetrating 2D MCS lines were collected to investigate the submarine volcanism and tectonics of the Shatsky Rise (Korenga and Sager, 2010; Korenaga, 2012). Because the original target of the two cruises was to explore the oceanic crustal structure, in this study on seismic oceanography, we only selected several MCS lines that clearly show mesoscale and submesoscale structures in the water column above the rise, including Lines 1 and 2 acquired on August 27 – 29, 2010 and Lines 3 and 4 acquired on April 1 – 3, 2012 (Figs. 2, 5).

For the MCS data acquisition, the seismic source was a 36-airgunarray with a volume of 108.2 L, which was shot every 50 m (with 2.5 m s⁻¹ vessel speed). The receiving array was a single, 6 km-long, 468-channel streamer with a 12.5 m channel spacing towed at 9 m. The raw data were sampled at a rate of 2 milliseconds and filtered to a frequency range of 2 – 206 Hz. In this study, 120 near-source traces and the first 2 seconds of the data were selected for water column imaging. The MCS data were reprocessed using the prestack depth migration method (Liu and Bleistein, 1995; Tang and Zheng, 2011), FK filtering, and automatic gain control. A constant water velocity model of 1480 m s⁻¹ was applied during the prestack depth migration and the time-depth conversion. The final seismic images were in 6.26 m lateral resolution (technically the common depth point interval) and approximately 10 m vertical resolution (frequency-dependent on a dominant frequency of 40 Hz).

The high-quality MCS reflection data effectively image the shallow water structure (< 1000 m depth) over the Shatsky Rise oceanic plateau (Figs. 2, 5). Layering and features within the shallow waters are visible seismically due to contrasts between water layers with differing physical properties. Seismic reflections occur where the transitions between water layers that could feature changes in temperature, density, or salinity occur (Holbrook et al., 2003; Tang et al., 2019, 2020). Hence, the seismic reflection signals mark the interfaces of water layers and correspond to thermoclines, isopycnals, or isosalines in hydrology. By recognizing and tracing the seismic reflectors, which can sometimes be connected to infer longer features, i.e., horizons, we can interpret the seismic layering, facies, and patterns as the water structure. Specifically, sharp contrasts between water layers result in high-amplitude seismic reflections, and horizontally continuous interfaces in distances lead to long reflectors or seismic horizons. When water layering is homogenous in the lateral scope, smooth and flat reflections are usually well observed. On the contrary, when a complex structure exists in the water column, rough reflections are often seen, such as curved, rugged, chaotic, irregular, and scattered reflectors. In addition, fronts in oceanography can be identified by sets of dipping reflectors.

Satellite data were compared with the observations from the MCS images, although the in situ hydrographic profiles on our study area or the nearby regions were few. The satellite-observed data for the SLA and SSTa on August 28, 2010, and April 2, 2012, were downloaded directly from the AVISO and AVHRR websites. The SLA data are referenced to the climatological mean of sea surface height in the period of 1993 – 2012, while the SSTA data were referenced to the climatological mean of SST in the period of 1971 – 2000. The GCA data were computed from the SLA data according to geostrophic balance, to show the surface features corresponding to the MCS Lines 1/2 and 3/4 (Figs. 3, 4, 6, 7).

3 Results

In August 2010, MCS Line 1 captured a subsurface mesoscale eddy, and Line 2 captured a strong front occurring at the transition between two SLA highs. In April 2012, MCS Lines 3 and 4 documented images of oceanfronts and distinct water structures between an SLA high and an SLA low.

Line 1 (Fig. 2a) was a seismic profile that crossed the center of an SLA high on August 28, 2010. The signature feature on this line was a bowl-like structure at 0 – 100 km width and 400 – 900 m depth. This feature was generally a left-right symmetric shape, but its surface at 400 m depth was slightly oblique and dipped quite gently (< 0.5°) toward the west. The structure was 100 km wide and 400 below the sea surface and had a maximum thickness of 500 m at its symmetric center and 50 m in the horizontal. The left-right symmetric and up-down asymmetric lenslike structure was similar to the mesoscale eddy structure observed from elsewhere in the oceans (e.g., the Gulf Stream, Mirshak et al., 2010; the South of Hokkaido, Yamashita et al., 2011; the Gulf of Alaska, Tang et al., 2014, 2020). The GCA data show that this subsurface eddy was anticyclonic (clockwise geostrophic rotation, Fig. 3). Positive values of the SSTa on maps (Fig. 4) indicate that it was also a warm eddy. One notable discrepancy between this eddy and many other surface eddies is that the surface of the captured eddy structure was not at the sea surface, but instead at 400 m below the sea surface, which implies that it was a subsurface eddy.

Although the quality of the seismic image on Line 1 was not as good as those in the aforementioned studies (Mirshak et al., 2010; Yamashita et al., 2011; Tang et al., 2014, 2020), an onion-like internal structure of the eddy with alternative reflective and blanking zones was similarly observed. That is, the eddy core (ranging from 10 – 70 km at 500 m depth) was characterized by moderate to weak reflectors and surrounded by bands of strong reflectors, separated by acoustically transparent zones tens of meters thick. The inner fringe reflectors were stronger than the outer ones but had narrower transparent zones. The base of the eddy is interpreted as the deepest visible reflections, but it is not clearly imaged because some disrupted, blurred, or absent reflectors occur. This seismic structure of the eddy is analogous to features seen on other subsurface eddies (Song et al., 2011; Menesguen et al., 2012) and comparable to the hydrographic observation (Ladd et al., 2005) and the schematic model (Faghmous, 2012) of mesoscale eddies. Additionally, the alternating reflective zones wrapped around the eddy core indicate the mixing between water masses with differing physical properties (e.g., temperature, salinity, density), as well as the mixing of their interfaces, causing seismic reflectors to vary in amplitude (weak or strong) and shape (flat or oblique).

The west end of Line 1 was located in the transition from an SLA high to an SLA low. The transition is expected to have a frontal structure that shows the contact relationship between different water masses, and oceanfronts are interpreted on seismic images by strong sets of dipping reflectors (Okkonen et al., 2003; Ladd et al., 2005; Tang et al., 2014). On Line 1, the west end (> 200 km) displayed several sets of dipping reflectors at depths from 300 to 600 m, which we interpret as some frontal features in the transition from the eastern SLA high to western SLA low. However, Line 1 was not the best seismic image in our dataset to demonstrate the oceanfront structure; an MCS line continuing westward, Line 2, was a much better profile to illustrate the structure.

Line 2 (Fig. 2b) continued from the west end of Line 1 and ran toward the west with a slightly different direction heading WNW. The seismic section of Line 2 passed through a narrow SLA low in between two SLA highs. The narrow SLA low is deemed to be the transition zone between the two broad SLA highs. Line 2 is likely to capture the frontal structure that shows the contact relationship between the two SLA highs. In the seismic image of Line 2, at least two sets of high-amplitude westward dipping reflectors were observed at 90 – 140 km laterally and 200 –1000 m vertically. The front-related reflectors dipped significantly (> 0.5°), compared with the shallow slopes (< 0.5°) of the strong water striae to the west ranging from 110 – 200 km and 100 – 600 m depth on Line 2 and to those of layers inside the observed eddy stratification on Line 1 above.

Another notable observation on Line 2 is the difference in reflection amplitude between the western water mass (100 – 200 km) and the eastern one (0 – 100 km). The western water mass had high-amplitude reflections and the layering of its structure was clear, whereas the eastern one showed weak, ragged reflections and a vague structure. The discrepancy is probably due to two factors. One is that the eastern and western water masses differed in the real structure; that is, the western one had water layering with sharper changes in physical properties, and the interfaces between water layers had better continuity, resulting in clearer stratification. Although the two water masses were SLA high and anticyclonic GCA (Fig. 3), the SSTa data showed that they had different temperatures. The western water mass (actually the edge of a warm eddy) featured a small-temperature anomaly, whereas the eastern one (the heart of the abovementioned warm eddy) featured high, positive SSTa (Fig. 4). The differences in temperature between the two water masses may cause discrepancies in water stratification and seismic reflectivity. In addition, seismic sections in this study captured different oceanic features. The fine filaments represent submesoscale oceanic fronts generated along the edge of mesoscale oceanic eddies, and the relatively smooth part was the oceanic eddy interior. The other factor is about the seismic imaging condition. Reflections in the eastern water mass were more contaminated by vertically striped noise of seafloor multiples from previous shots or by any form of noise during seismic acquisition.

Line 3 (Fig. 5a) was slightly similar to Line 2 in that it also crosses two water masses. However, Line 3 ran from an SLA high in the west to an SLA low in the east, and it was collected in 20 months later than Line 2, and its location was 200 km to the north. Although the mesoscale feature would evolve and change frequently after such a long period, Line 3 featured some characteristics of Line 2. On Line 3, a narrow zone (slightly wider than that of Line 2) existed between the western SLA high and the eastern SLA low (Fig. 6). Likewise, the frontal feature is expected to be seen on this line. Around 50 – 100 km in the horizontal and 200 – 600 m depth, two sets of dipping reflectors (2°– 3°) corresponded to the front boundaries between the two water masses. The western water mass (0 – 100 km) was characterized by high-amplitude, continuous, and gently tilted reflections, analogous to those of the western water mass under the SLA high on Line 2, probably implying that the water mass interior had strong stratification due to intensively alternating layers with differing physical properties (e.g., anticyclonic GCA in Fig. 6 and high, positive SSTa in Fig. 7). In contrast, the eastern water mass on Line 3 (110 – 260 km) showed different reflections with weak amplitude, less continuity, more flat-layering, and shallower base (corresponding to cyclonic GCA and negative SSTa). Hence, the discrepancy in seismic reflection facies and patterns may imply the difference in structure and hydrology of the two water masses beneath the SLA high and SLA low.

Line 4 (Fig. 5b) was a profile connected with the east end of Line 3 but trending toward the south. This seismic section passed through the edge of the SLA low, then into the center of an SLA high. It also crossed the transition between the SLA low in the north and the SLA high in the south. Like Line 3, the water mass underneath the SLA low on Line 4 displayed similar seismic reflections, which are identified as discontinuous and rugged reflectors (accordingly, cyclonic GCA in Fig. 6 and negative SSTa in Fig. 7). The deepest reflectors show that the seismically visible base of the northern water mass was less than 400 nm depth. At its edge (200 km), the frontal feature was also associated with the occurrence of dipping reflectors. The other side of the oceanfront featured the water mass under the SLA high in the south. Analogous to features seen on Line 3, the southern water mass had stronger and slightly inclined reflectors and a deeper base of 800 m depth (corresponding to anticyclonic GCA and positive SSTa).

4 Discussion 4.1 Vertical Structure of Shallow Waters over the Shatsky Rise

Our seismic images show the high-resolution (6.25 m× 10 m bin) vertical structure of shallow waters (< 1000 m depth) over the Shatsky Rise oceanic plateau in the Northwest Pacific Ocean, which was sparsely obtained from conventional physical oceanography survey. One feature observed from our seismic profiles is mesoscale eddy. MCS Line 1 (Fig. 2a) crossed the center of an SLA high and captured a subsurface anticyclonic warm eddy on August 28, 2010, at the center of the Shatsky Rise (158°E, 34°N). The seismic section displayed that the eddy had an overall bowl- or lens-like structure, which was roughly left-right symmetric and up-down asymmetric. The eddy surface was approximately 400 m below the sea surface and somewhat dipped (< 0.5°) toward the west. The eddy core was wrapped by alternating reflective zones with high/low-amplitude and variations of reflector slopes, implying an onionlike internal structure. The eddy bottom reached about 900 m depth. Based on the shape of the eddy, it is inferred to have a radius of 50 km and a maximum thickness of 500 m. The mesoscale eddy image seen in our data is similar to that of prior seismic oceanographic studies (e.g., subsurface eddies, Song et al., 2011; Menesguen et al., 2012; and surface eddies, Mirshak et al., 2010; Yamashita et al., 2011; Tang et al., 2014, 2020) and consistent with hydrographic observations (Ladd et al., 2005) and a schematic model (Faghmous, 2012). The classification and scale of the eddy in our seismic section also fit the statistical characteristics results from Zhang et al. (2013) and Hu et al. (2018), which demonstrated that anticyclonic warm eddies tended to occur in the north of the Kuroshio Extension (> 33°N, compared to 34°N herein, Fig. 1), and the mean radii were 72 km (compared to 50 km herein). Moreover, the observed eddy in this study occurred in summer, which matches the most robust eddy activities period throughout a year, as indicated by the statistical seasonal variations of mesoscale eddies in the Kuroshio Extension region (Zhang et al., 2013; Hu et al., 2018).

Another interesting feature observed in our seismic images is a submesoscale oceanfront. The oceanfront occurred at the boundaries where water masses with different hydrological properties interacted. However, the frontal feature was of submesoscale size, which is hardly observed by traditional hydrological investigations. Usually, the front boundaries were quite narrow, and their finescale structures were poorly observed. In the seismic images, oceanfronts were characterized by several sets of dipping reflectors with degrees of 2° – 3° and higher. The degrees of dipping reflectors were significantly higher than the slopes of water striae inside the water mass (typically < 1°). Hence, the relatively strong dipping reflectors appeared at the boundaries between water masses that could be recognized in places in our seismic sections, for example, around 100 km on Line 2, 110 km on Line 3, and 200 km on Line 4 (Figs. 2b, 5). The reflectors illustrate the contact relationship between water masses beneath SLA highs and lows and demonstrate the evolution from water column stratification to wave-breaking or turbulence, which may clarify the mixing kinetics and dynamics.

In addition, water masses under SLA highs and lows displayed distinct structures. The water mass beneath the SLA highs appeared to have high-amplitude, continuous, clear-layered, and eddy-style reflections, corresponding to anticyclonic GCA and positive SSTa (Figs. 3, 4, 6, 7). In contrast, the water mass underneath the SLA lows was associated with weak, fragmented, blurred reflectors, matching with cyclonic GCA and negative SSTa. Regardless of seismic noise and technical problems, the discrepancy in seismic facies and patterns results from the difference in water mass properties. The water mass under the SLA highs is likely to be related to the formation of mesoscale eddies, which have isolated, non-linear structure, strong anisotropy, and more stratification. Inside the eddies, the transitions from water layers with significant alternating properties (temperature, salinity, density, etc.) generated clear seismic reflective signals observed in our profiles. On the contrary, the water mass under the SLA lows is inferred to be more transportive and characterized by a linear feature, weaker anisotropy, and less stratification due to gentle changes in water properties with depth. The water mass shallow base (identified by the deepest visible reflectors) depth of < 400 m is also consistent with the above description, different from the water mass beneath the SLA highs, which can reach 800 – 1000 m depth.

4.2 Implications for Mesoscale Eddy Formation

Underneath the SLA highs, mesoscale eddies are formed and become an isolated water mass that contains nonlinear structure and high anisotropy. They can maintain their structural and hydrological characteristics and translate water mass, energy, and properties such as temperature, salinity, vortex, nutrient, and plankton. Within the Kuroshio Extension region (Fig. 1), mesoscale eddies are widely considered to be caused by the barotropic or baroclinic instability (Tai and White, 1990; Qiu and Chen, 2005; Xu et al., 2011). The vertical structure is not diagnostic of the proposed mechanisms of mesoscale eddy formation; thus, near-sea surface geophysical data are not suitable to clearly determine the correct mechanism. Nevertheless, the observations from our seismic images can help guide the explanations for mesoscale eddy formation. Any mechanism explaining the formation of mesoscale eddies over the Shatsky Rise must incorporate our observations as follows: First, the eddy surface depth is 400 m below the sea surface, showing a subsurface eddy (Fig. 2a); however, in other cases, for instance, the Gulf of Alaska, the Gulf stream, and the Hokkaido regions, the eddy surface appears to be at the sea surface (Mirshak et al., 2010; Yamashita et al., 2011; Tang et al., 2014, 2020). The cause(s) of the variations in eddy surface depths should be explained by the eddy formation mechanisms. Second, the left-right symmetric structure is not perfect, as a slightly inclined eddy surface and core was observed in our seismic data (asymmetric shape). Whatever hypothesis is required for predicting the mesoscale eddies inclination. Third, the increasing slopes in degrees (implied by seismic reflectors or horizons) from the eddy core to the surrounding, as well as from the internal striae to the front boundaries, show the evolution from eddy formation and stratification, turbulent diffusion, to dissipation. It is important to address the factor(s) influencing the changes in the slopes of water layers and relevant geostrophic or ageostrophic dynamic process by any proposal of the mesoscale eddy formation mechanism. Lastly, within the eddy, seismic facies vary (such as amplitude, continuity, and blanking zones), implying variations in the physical properties of individual water layers (such as temperature, salinity, and density). Also, the factors responsible for eddy kinetic behavior remains to be solved.

4.3 Future Investigations

Seismic oceanography development over the past two decades provides a new scope to reveal mesoscale and submesoscale structures in the oceans; the field of seismic oceanography can be an important supplement to conventional hydrological and oceanological observations, especially in vertical view and with high-resolution on the order of meters both laterally and vertically. This paper presents seismic imaging applications for ocean eddies and fronts structural interpretations (Figs. 2, 5), as well as satellite data showing SLA, GCA, and SSTa (Figs. 3, 4, 6, 7), to accomplish a cross-disciplinary study between marine geophysics and physical oceanography. The implications from our study are significant to understand the formation of mesoscale eddies and submesoscale fronts in the Kuroshio Extension region and other oceans. However, more investigations should be conducted to enhance the seismic applications on oceanography. For example, the inversion of temperature, salinity, and density structure by seismic reflection data might establish a direct method to learn physical oceanographic information from marine geophysics (Holbrook et al., 2003; Biescas et al., 2010; Fer et al., 2010). To achieve such a method, a forward model called synthetic seismogram is needed to examine the seismic reflections in response to the water stratifications (Tang et al., 2014). Meanwhile, in situ measurements from conductivity-temperature-depth casts (if applicable) would serve as well logging in seismic exploration to constrain the synthetic seismogram and the inversion. In situ measurements are rare for open oceans, and even rarer for oceans interacting with seismic transects. Nonetheless, increasing seismic-oceanography joint surveys are being conducted to collect physical oceanographic data together with the seismic data (Vsemirnova et al., 2012; Tang et al., 2015, 2016, 2018). Particularly, in situ measurements of temperature, salinity, chlorophyll, production, nutrient, etc. (e.g., from expendable bathythermograph and acoustic Doppler current profiler) can be compared with seismic profiles. Where in situ data are not available or not dense enough to infer for a large area, satellite data would be the secondary source (e.g., AVISO, MERIS). Furthermore, quantitative estimation from seismic data might be another strong tool. For instance, based on seismic reflection characteristics, such as horizon slope, reflector amplitude, and spectra wavenumber, water mass rotation speed, geostrophic transport, and diffusion rate, can be estimated by assuming the correspondence between seismic reflections and water layers (Hermann et al., 2002; Sheen et al., 2011; Tang et al., 2014, 2020). Seismic data are continuously acquired with time, generating time-lapse data, which might add temporal variations to spatial structures (Tang et al., 2019). As a result, 4D evolution modeling could be established to demonstrate the spatiotemporal dynamic process of water masses in the ocean. With the abovementioned data and methods, significant scientific issues may be better understood, such as the mixing, exchange, and translation of water mass, energy, and properties, as well as the interactions between ocean and atmosphere and between seawater and seafloor.

5 Conclusions

Here, we present four MCS lines to show high-resolution (6.25 m × 10 m) seismic images of shallow waters over the Shatsky Rise, where the Kuroshio Extension crosses and bifurcates. One of our MCS transects ran through the center of an anticyclonic warm eddy on August 28, 2010, proved by SLA, GCA, and SSTa satellite data. The seismic image showed that the eddy vertical structure featured a bowl-like shape and onion-like internal layering. The slightly tilted (< 0.5°) surface of the eddy was approximately 400 m below the sea surface, indicating that it was a subsurface eddy. The eddy was inferred to have a radius of 50 km and a maximum thickness of 500 m, which is within the ranges from prior hydrological observations of this area. Other MCS sections demonstrated the submesoscale structure of oceanfronts, which feature dipping reflectors (> 2° – 3°) at the boundaries between water masses with differing properties. In addition, the discrepancies in SLA, GCA, and SSTa between water masses resulted in distinct seismic reflectivity. The water masses with high SLA, anticyclonic GCA, and positive SSTa exhibited highamplitude, continuous, clear-layered, and non-linear reflections, whereas those with low SLA, cyclonic GCA, and negative SSTa were associated with weak, fragmented, less stratification and more linear reflectors. Thus, the observed seismic facies and patterns reflect the ocean water structures and properties.

Acknowledgements

We thank the captain, crew, and science party of R/V Marcus G. Langseth cruises MGL1004 and MGL1206 for collecting the MCS reflection data. We also thank Prof. Xiaohui Ma from Ocean University of China for discussion and help in terms of physical oceanography. The satellite data are available on www.aviso.oceanobs.com and www. ncei.noaa.gov/products/avhrr-pathfinder-sst. This research was supported by the National Key R & D Program of China (No. 2018YFC0309800), the Guangdong Basic and Applied Basic Research Foundation (No. 2021B1515020098), the Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (No. GML2019ZD0 205), the National Natural Science Foundation of China (Nos. 41776058 and 41890813), the Chinese Academy of Sciences (Nos. 133244KYSB20180029, 131551KYSB20 200021, Y4SL021001, QYZDY-SSW-DQC005 and ISEE 2019ZR01), and the Laboratory for Marine Mineral Resources, Qingdao National Laboratory for Marine Science and Technology (No. MMRZZ201801).