Using high-resolution in situ data from gliders, satellite data of sea level anomaly and geostrophic currents, we presented the detailed structure of an anticyclonic eddy during spring 2015 in the northern South China Sea. The impact depth of the anticyclonic eddy reached about 1000 m and had a maximum temperature anomaly of about 3°C at ∼120 m and maximum salinity anomaly of more than 0.3 psu in the mixed layer. The maximum geostrophic velocities perpendicular to the glider path were about 0.3 m s−1 at 100 m. The estimated radius was about 72 km and the translation velocity was about 5.2 cm s−1. The rotational speed of the eddy was estimated to be 0.35 m s−1. The high temperature and large salinity of the anticyclonic eddy indicated it did not originate from the South China Sea locally. The analysis of water mass indicated the character of the eddy water was similar to Kuroshio water, and the time evolution of the sea level anomaly and surface geostrophic velocity anomaly further validated that it originated from the Kuroshio intrusion as a loop current to the southwest of Taiwan.

Introduction

The northern South China Sea (NSCS) is connected with the western Pacific Ocean (WPO) via the Luzon Strait, and with the East China Sea (ECS) through the Taiwan Strait (Figure 1). The Kuroshio intrusion through the Luzon Strait, together with the seasonally reversed monsoons, largely controls the dynamic processes in the NSCS (Wyrtki, 1961; Chu and Fan, 1999; Shu et al., 2011, 2014; Wang et al., 2014). Under the influences of monsoon winds, Kuroshio intrusion, and complex topography, the NSCS is characterized as an area with high mesoscale eddy activities (Li et al., 1998; Chu and Fan, 2001; Metzger and Hurlburt, 2001; He et al., 2002; Jia and Liu, 2004; Wang et al., 2008a,b; Chen et al., 2011; Nan et al., 2011; Xiu et al., 2010; Zu et al., 2013).

Figure 1.

Bathymetry in the northern South China Sea and its surroundings. Contour lines represent 200 m, 500 m, 1000 m, 2000 m and 3000 m. The solid-dotted line indicates the glider path. Arrow denotes the Kuroshio. Two dashed boxes are selected to represent the typical area of the NSCS water and Kuroshio water.

Figure 1.

Bathymetry in the northern South China Sea and its surroundings. Contour lines represent 200 m, 500 m, 1000 m, 2000 m and 3000 m. The solid-dotted line indicates the glider path. Arrow denotes the Kuroshio. Two dashed boxes are selected to represent the typical area of the NSCS water and Kuroshio water.

Previous studies indicated that the anticyclonic eddies in the NSCS mostly originated from the vicinity of the Luzon Strait. Based on two CTD surveys in March 1992 and September 1994, Li et al. (1998) reported anticyclonic eddies in the NSCS and speculated that these eddies were probably shed from the Kuroshio. Wang et al. (2008a) investigated the origins and evolutions of two anticyclonic eddies in the NSCS based on the observations from satellite sea level anomaly (SLA), sea surface temperature (SST), and buoys, and found one was generated in the interior SCS and the other was shed from the Kuroshio meander. Jia and Liu (2004) studied the eddy shedding at the Luzon Strait from 1992 to 2001 using TOPEX/POSEIDON-ERS satellite altimeter data and model results. They found the Kuroshio bend varied with time and periodically shed anticyclonic eddies, with the most frequent eddy-shedding intervals of 70, 80, and 90 days. Using SLA data, Yuan et al. (2007) concluded that the anticyclonic eddy found by Li et al. (1998) was generating locally west of Luzon, rather than from the Kuroshio. Wang et al. (2008b) found the wind jet and enhanced wind stress curl in the lee of the Taiwan Island was responsible for the generation of eddies in that area. In addition to eddies originating locally or shed from the Kuroshio, other studies found that some eddies could propagate from the western Pacific through the Luzon Strait as nonlinear Rossby waves (Sheu et al., 2010; Zheng et al., 2011; Hu et al., 2012). Though there are different sources of eddies observed in the NSCS, most of them originated by eddy shedding from the anticyclonic Kuroshio loop because of the instability of the Kuroshio (Li et al., 1998; Metzer and Hurlburt, 2001; Jia and Liu, 2004; Wang et al., 2008a; Zu et al., 2013).

Though previous studies have presented sufficient evidence for frequent appearances of anticyclonic eddies in the NSCS, the vertical structures of these eddies are not well understood because of the rarity of in situ data. Using high-resolution observations in space and time derived from a glider, we present some detailed features of an anticyclonic eddy shed from the Kuroshio.

Data

Autonomous, buoyancy-driven gliders have been widely applied to survey spatial variability of oceanic processes in recent years. Gliders collect a variety of measurements as they dive or climb through the water column, depending on the equipped sensors. A Sea-wing underwater glider, developed by the Shenyang Institute of Automation, Chinese Academy of Sciences (Yu et al., 2011; Qiu et al., 2015) was used in this study. The glider crossed the northern SCS from east to west along the continental slope, with the initial site southwest of Taiwan (118.75 °E, 21.61 °N) during a one-month mission from 28 April to 30 May 2015 (Figure 1). The Sea-wing underwater glider was equipped with conductivity temperature depth (CTD). The sampling cycle of the glider included a diving profile and a climbing profile, which took about four hours. The observed maximum depth was 1000 m. During the deployment, 205 sample cycles were finished and 410 high spatial resolution vertical profiles of temperature and salinity were collected. The sample interval of the CTD was 6 s and the mean vertical sample resolution was about 1 m. In order to remove the high frequency signals (e.g. internal waves, tides) and noises, a 50-km horizontal and 10 m vertical filter was used to smooth observed data.

Daily SLA and surface geostrophic velocity anomaly from February to July 2015 were obtained from AVISO gridded dataset to illustrate the time evolution of the eddy. The monthly temperature and salinity from World Ocean Atlas 2013 (WOA13) was used to offer the basic characteristics of the NSCS water and Kuroshio water.

Results

Features of the anticyclonic eddy

The temperature, salinity, and potential density along the glider path are shown in Figure 2. There was an obvious depression for the temperature between 115.3°E and 117.5°E, from 13–23 May 2015, where the 20°C isotherm sunk to about 50 m. Isotherm depression in the vertical is a typical characteristic of warm eddies (anticyclonic eddies). However, between 117.6°E and 118.8°E, the isotherm of 20°C was in the shape of a dome, with subsurface water elevated upwards from 2–12 May, which is a feature of cold eddies (cyclonic eddies). In the anticyclonic eddy, the saltier water was observed above 200 m between 115.3°E and 117.5°E, with the maximum salinity of 34.8 psu at a depth of about 120 m, similar to the site of the 20°C isotherm, which is the typical thermocline depth index used in the tropical sea. Usually, a locally generated anticyclonic eddy is characterized by high temperature and low salinity (Wang et al., 2008a). The observed high temperature and high salinity between 115.3°E and 117.5°E indicates the anticyclonic eddy was not originated in the NSCS. The potential density shown in Figure 2c presents a downward depression structure between 115.3°E and 117.5°E throughout the observed depth.

Figure 2.

Observed temperature (a, unit: °C), salinity (b, unit: psu) and potential density (c, unit: kg m−3).

Figure 2.

Observed temperature (a, unit: °C), salinity (b, unit: psu) and potential density (c, unit: kg m−3).

In order to further investigate the vertical structures of the anticyclonic eddy, the anomalies of temperature, salinity, and potential density are presented in Figure 3. Here, the anomalies are derived by subtracting the averaged values along the glider path. The positive temperature anomaly of the anticyclonic eddy was mainly located in depths of 50–200 m, and could be seen in all depths below 50 m between 115.3°E and 117.5°E, from 13–23 May 2015. The maximum temperature anomaly appeared at thermocline, about 120 m, with values of more than 3°C. However, at the mixed layer, above 50 m, smaller temperature anomalies were observed. Conversely, the maximum salinity anomaly of the anticyclonic eddy appeared in the mixed layer, with the maximum value of more than 0.3 psu (Figure 3b). The salinity anomalies of the anticyclonic eddy were mainly observed above 200 m. The potential density anomaly of the anticyclonic eddy shown in Figure 3c indicates that it was determined by the salinity anomaly above 50 m and by the temperature anomaly below 50 m. In the mixed layer, the potential density anomaly of the anticyclonic eddy was positive, with the maximum value of about 0.4 kg m−3. It was negative below the mixed layer, with the maximum anomaly of more than 0.8 kg m−3. In general, the anticyclonic eddy did not present the ‘warm’ feature in the mixed layer where negative temperature and positive density anomalies were found. The ‘warm’ feature of the anticyclonic eddy appeared below 50 m and reached to the maximum observed depth of 1000 m, with the largest anomaly in thermocline. East of the anticyclonic eddy between 117.6°E and 118.8°E, the anomalies of temperature, salinity, and potential density are opposite to those of the anticyclonic eddy, perhaps indicating a cyclonic eddy in that region.

Figure 3.

Observed temperature anomaly (a, unit:°C), salinity anomaly (b, unit: psu), and potential density anomaly (c, unit: kg m−3).

Figure 3.

Observed temperature anomaly (a, unit:°C), salinity anomaly (b, unit: psu), and potential density anomaly (c, unit: kg m−3).

The geostrophic velocity perpendicular to the glider path can be obtained through the thermal wind relationship,
formula
(1)
where f is the local Coriolis parameter, the geostrophic velocity perpendicular to the glider path, g the gravity acceleration, the reference potential density with the value of 1025 kg m−3, the potential density, and s the coordinate along the glider path. The reference plane is set to be the observed maximum depth of 1000 m, where we assume . In order to ensure the geostrophic approximation, the potential density was smoothed at a horizontal scale of 50 km, which is approximately the first baroclinic Rossby deformation radius in the slope of the NSCS (Cai et al., 2008).

The geostrophic velocity perpendicular to the glider path presented an anticyclonic structure with the positive values west of 116°E and the negative values between 116°E and 117.4°E (Figure 4). The maximum geostrophic velocities perpendicular to the glider path were observed twice at (116.69°E, 19.69°N) and (115.46°E, 19.2°N) on 16 and 22 May, respectively, with a value of about 0.3 m s−1 (Figure 4). If we assume that the eddy consists of solid body rotation within a core (r < R, where R is the radius of the eddy) and decay elsewhere (r > R), the maximum velocity should be observed at the site of the radius (Lilly and Rhines, 2002; Carpenter and Timmermans, 2012). Using the SLA data, we can determine the centers of the eddy on 16 and 22 May. The radiuses of the eddy can be derived with the value of about 67 and 77 km on 16 and 22 May, respectively. Considering that the gridded SLA may introduce errors in eddy center estimation, we used the mean radius of about 72 km next. From 16–22 May, the center of the eddy moved about 27 km. The estimated translation velocity of the eddy during these 6 days was about 5.2 cm s−1.

Now we estimate the eddy maximum rotation speed according to the observed geostrophic velocity perpendicular to the glider path. As shown in Figure 5, when a glider passes an anticyclonic eddy, it observes twice the maximum velocities, and the eddy center moves from O1 to O2. According to the geometry under the assumption that the translation velocity of the eddy remains unchanged and the velocity of the glider is constant,
formula
(2)
formula
(3)
where and are the maximum velocity perpendicular to the glider path derived from Equation (1); and, are the rotational and translational speed of the eddy, respectively; and are the angles between and and between and , as well as the angles between the glider path and the radius (R) at P1 and P2, respectively, which can be derived from the sites of the eddy center and the glider when the glider passes the eddy; and is the angle between and the glider path. Eliminating , we get
formula
(4)
Figure 4.

Geostrophic velocities perpendicular to the glider path calculated from potential density via the thermal wind relationship (unit: m s−1).

Figure 4.

Geostrophic velocities perpendicular to the glider path calculated from potential density via the thermal wind relationship (unit: m s−1).

Figure 5.

Schematic diagram of geometric relationship between rotation velocity (), translation velocity (), and observed maximum velocity perpendicular to the glider path ( and ) when the glider passed a moving anticyclonic eddy.

Figure 5.

Schematic diagram of geometric relationship between rotation velocity (), translation velocity (), and observed maximum velocity perpendicular to the glider path ( and ) when the glider passed a moving anticyclonic eddy.

From Equation (4), the estimated rotation speed of the anticyclonic eddy was 0.35 m s−1. Then, as noted by Lilly and Rhines (2002), we can further estimate the characteristic Rossby number () of the eddy using the relative vorticity magnitude and the Coriolis coefficient as
formula
(5)

The anticyclonic eddy had a Rossby number of 0.20.

Source of the anticyclonic eddy

As described above, the large salinity of the observed anticyclonic eddy in the upper layer indicates that it was not generated locally. The salty water might come from the western Pacific through the Luzon Strait. In order to investigate the source of the water, the temperature-salinity (T-S) scatter diagram is shown in Figure 6. As a reference, the characteristics of the SCS water and the Kuroshio water from WPO are represented by the respective T-S diagrams derived from WOA13 in Boxes 1 and 2 (Figure 1). The characteristics of eddy water were derived from the glider-observed temperature and salinity between 115.46°E to 116.69°E and from 16–22 May 2015. The observed T-S diagrams east and west of the anticyclonic eddy (denoted as ESCSW and WSCSW) are also presented here. As shown in Figure 6, properties of the Kuroshio water are obviously different from those of the SCS water, especially for the upper layers, where the salinity of the Kuroshio water is higher than that of SCS water by about 0.4 psu. The eddy water has similar properties to the Kuroshio water and is obviously different from the SCS water, suggesting that the source water of the eddy is from the Kuroshio. Except for the eddy water, the characteristics of the observed water masses (i.e. EEDYW and WEDYW) are consistent with the SCS water.

Figure 6.

Temperature-salinity (T-S) scatter diagrams for the Kuroshio water (KSW), SCS water (SCSW), eddy water (EDYW), east of eddy water (EEDYW), and west of eddy water (WEDYW).

Figure 6.

Temperature-salinity (T-S) scatter diagrams for the Kuroshio water (KSW), SCS water (SCSW), eddy water (EDYW), east of eddy water (EEDYW), and west of eddy water (WEDYW).

Figure 7 shows the time evolution of the SLA and surface geostrophic velocity anomaly where the data was averaged using the nearest five days. The anticyclonic eddy was originated off the southwest of Taiwan at the end of February 2015, when the Kuroshio intrusion strengthened and formed a loop structure northwest of the Luzon Strait (Figure 7a). After that, the Kuroshio weakened and the loop current disappeared. The anticyclonic eddy developed and shed from the Kurshio until 25 March 2015 (Figure 7b). It then moved southwestward and its intensity slightly weakened from 5–15 April, but strengthened after 15 April (Figure 7c-h), suggesting the strong interactions between the mesoscale eddy and ambient fluids. The intensity of the anticyclonic eddy was strongest on 10 June and then disappeared on 10 July (not shown here). The anticyclonic eddy was sustained for more than four months and its mean translation velocity was about 5.3 cm s−1, which was smaller than that found by Hu et al. (2001) and Wang et al. (2008a).

Figure 7.

Time evolutions of the SLA (unit: m) and geostrophic current (unit: m s−1) from February to May 2015. The glider path is also shown in f, g, and h.

Figure 7.

Time evolutions of the SLA (unit: m) and geostrophic current (unit: m s−1) from February to May 2015. The glider path is also shown in f, g, and h.

Discussion and conclusions

Based on high-resolution in situ data observed by a glider, we analyzed the temporal and spatial evolution of an anticyclonic eddy in the NSCS. The anticyclonic eddy had high temperature and salinity between 115.3°E and 117.5°E, obviously different from the ambient waters, indicating its non-SCS origin. The impact of the anticyclonic eddy reached about 1000 m. The maximum temperature anomaly of the eddy was more than 3°C at ∼120 m and smaller temperature anomalies were observed above 50 m. However, the maximum salinity anomaly of the anticyclonic eddy appeared in the mixed layer, with a value of more than 0.3 psu. The anticyclonic eddy did not present the ‘warm’ feature in the mixed layer where negative temperature and positive density anomalies were observed. The maximum geostrophic velocities perpendicular to the glider path were observed at 100 m, with a value of about 0.3 m s−1. Together with SLA data, we estimated the radius of the eddy to be about 72 km, the translation velocity to be about 5.2 cm s−1, and the rotational speed of the anticyclonic eddy at about 0.35 m s−1.

The T-S scatter diagram shows the character of eddy water was similar to the Kuroshio water and obviously different from the SCS water. The time evolution of the SLA and surface geostrophic velocity further validate that the anticyclonic eddy originated from Kuroshio intrusion as loop current southwest of the Taiwan Island. The eddy firstly weakened and then strengthened after it departed from its origination area. The anticyclonic eddy was sustained for more than four months.

Eddies in the NSCS are mainly of three types (Zu et al., 2013): eddy shedding from the Kuroshio loop (Jia and Liu, 2004), generated locally (Wang et al., 2008), and propagating from WPO as Rossby waves (Hu et al., 2012). The observed anticyclonic eddy in this study was obviously the first type. Moreover, its mean translation velocity was about ∼5 cm s−1, far less than the speed of the first baroclinic Rossby waves of about 10 cm s−1 in the NSCS. The generation of this type of eddy was thought to be due to barotropic and baroclinic instabilities introduced by the Kuroshio intrusion (Zu et al., 2013). Jia and Liu (2004) found that when an eddy shedding event occurs, there are usually two high SLA centers with a positive geostrophic vorticity appearing between them. This may be why a cyclonic eddy was observed east of the anticyclonic eddy (Figures 3–5).

Acknowledgements

This study benefited from the SLA data generated by DUACS and distributed by AVISO.

Funding

This work was supported by the National High Technology Research and Development Program of China (project 2012AA091003), the 100 Talent Program of the Chinese Academy of Sciences, the National Natural Science Foundation of China (projects 41476012, 91228202, 41476013, 41476011 and 41476014), and the project of Guangdong Provincial Department of Science and Technology (2012A032100004).

References

Cai, S., Long, X., Wu, R., Wang, S.,
2008
.
Geographical and monthly variability of the first baroclinic Rossby radius of deformation in the South China Sea
.
J. Mar. Syst.
74
,
711
720
.
Carpenter, J. R., Timmermans, M.L.,
2012
.
Deep mesoscale eddies in the Canada Basin, Arctic Ocean
.
Geophys. Res. Lett.
39
, doi:.
Chen, G., Hou, Y., Chu, X.,
2011
.
Mesoscale eddies in the South China Sea: mean properties, spatiotemporal variability, and impact on thermohaline structure
.
J. Geophys. Res.
116
,
C06018
. doi:.
Chu, P. C., Fan, C.,
2001
.
Low salinity, cool‐core cyclonic eddy detected northwest of Luzon during the South China Sea Monsoon Experiment (SCSMEX) in July 1998
.
J. Oceanogr.
57
,
549
563
, doi:.
Chu, P., Edmons, N., Fan, C.,
1999
.
Dynamical mechanisms for the South China Sea seasonal circulation and thermohaline variabilities
.
J. Phys. Oceanogr.
29
,
2971
2989
, doi:.
He, Z. G., Wang, D., Hu, J.Y.,
2002
.
Features of eddy kinetic energy and variations of upper circulation in the South China Sea
.
Acta Oceanol. Sinica
21
(
2
),
305
314
.
Hu, J., Kawamura, H., Hong, H., Kobashi, F., Wang, D.,
2001
.
3–6 months variation of sea surface height in the South China Sea and its adjacent ocean
.
J. Oceanogr.
57
,
69
78
.
Hu, J., Zheng, Q., Sun, Z., Tai, C.-K.,
2012
.
Penetration of nonlinear Rossby eddies into South China Sea evidenced by cruise data
.
J. Geophys. Res.
117
,
C03010
, doi:.
Jia, Y., Liu, Q.,
2004
.
Eddy shedding from the Kuroshio bend at Luzon Strait
.
J. Oceanogr.
60
(
6
),
1063
1069
.
Li, L., Nowlin, Jr., W., Su, J.,
1998
.
Anticyclonic rings from the Kuroshio in the South China Sea
.
Deep-Sea Res. I.
45
,
1469
1482
.
Lilly, J. M., Rhines, P.B.,
2002
.
Coherent eddies in the Labrador Sea observed from a mooring
.
J. Phys. Oceanogr.
32
,
585
598
. doi:.
Metzger, E. J., Hurlburt, H.E.,
2001
.
The nondeterministic nature of Kuroshio penetration and eddy shedding in the South China Sea
.
J. Phys. Oceanogr.
31
,
1712
1732
.
Nan, F., He, Z., Zhou, H., Wang, D.,
2011
.
Three long-lived anticyclonic eddies in the northern South China Sea
.
J. Geophys. Res.
116
,
C05002
. doi:.
Qiu, C., Mao, H., Yu, J., Xie, Q., Wu, J., Lian, S., Liu, Q.,
2015
.
Sea surface cooling in the Northern South China Sea observed using Chinese sea-wing underwater glider measurements
.
Deep-Sea Res.
1
(
105
),
111
118
.
Sheu, W., Wu, C., Oey, L-Y.,
2010
.
Blocking and westward passage of eddies in the Luzon Strait
.
Deep-Sea Res. II
57
(
19–20
),
1783
1791
.
Shu, Y., Chen, J., Yao, J., Pan, J., Wang, W., Mao, H., Wang, D.,
2014
.
Effects of Pearl River plume on the vertical structure of coastal currents in the Northern South China Sea during summer 2008
.
Ocean Dyn.
64
,
1743
1752
, doi: .
Shu, Y., Wang, D., Zhu, J., Peng, S.,
2011
.
The 4-D structure of upwelling and Pearl River plume in the northern South China Sea during summer 2008 revealed by a data assimilation model
.
Ocean Model
.
36
(
3–4
),
228
241
.
Wang, D., Shu, Y., Xue, H., Hu, J., Chen, J., Zhuang, W., Zu, T., Xu, J.,
2014
.
Relative contributions of local wind and topography to the coastal upwelling in the northern South China Sea
.
J. Geophys. Res.
119
(
4
),
2550
2567
.
Wang, D., Xu, H., Lin, J., Hu, J.,
2008a
.
Anticyclonic eddies in the northeastern South China Sea during winter 2003/2004
.
J. Oceanogr.
64
(
6
),
925
935
.
Wang, G., Chen, D., Su, J.,
2008b
.
Winter eddy genesis in the eastern South China Sea due to orographic wind jets
.
J. Phys. Oceanogr.
38
(
3
),
726
732
.
Wyrtki, K.,
1961
.
Scientific results of marine investigation of the South China Sea and Gulf of Thailand
.
NAGA Report
,
2
.
Xiu, P., Chai, F., Shi, L., Xue, H., Chao, Y.,
2010
.
A census of eddy activities in the South China Sea during 1993–2007
.
J. Geophys. Res.
115
,
C03012
, doi:.
Yu, J., Zhang, A., Jin, W., Chen, Q., Tian, Y., Liu, C.,
2011
.
Development and Experiments of the Sea-Wing Underwater Glider
.
China Ocean Eng.
25
,
721
736
.
Yuan, D., Han, W., Hu, D.,
2007
.
Anti-cyclonic eddies northwest of Luzon in summer–fall observed by satellite altimeters
.
Geophys. Res. Lett.
34
(
13
),
L13610
. doi:.
Zheng, Q, Tai, C.-K., Hu, J., Lin, H., Zhang, R.-H., Su, F.-C., Yang, X.,
2011
.
Satellite altimeter observations of nonlinear Rossby eddy–Kuroshio interaction at the Luzon Strait
.
J. Oceanogr.
67
(
4
),
365
376
.
Zu, T., Wang, D., Yan, C., Belkin, I., Zhuang, W., Chen, J.,
2013
.
Evolution of an anticyclonic eddy southest of Taiwan
.
Ocean Dyn.
63
,
519
531
.