Figure captions

A Survey of Studies on the South China Sea Upper Ocean Circulation Guohong Fang¹, Wendong Fang², Yue Fang¹and Kai Wang¹ (received 1998/2/5, revised 1998/4/14, accepted 1998/4/16) Abstract

Advances in understanding the upper ocean circulation of the South China Sea (SCS) since the works of Dale (1956) and Wyrtki (1961) are reviewed. The focus is on the major features of the circulation pattern. The circulation in the northern SCS is driven mainly by monsoon winds and Kuroshio intrusion, and secondly by surface heat flux. Major components include the Kuroshio intrusion through the Luzon Strait, the SCS Branch of the Kuroshio, the Northwest Luzon Cyclonic Gyre, the Northwest Luzon Cyclonic Eddy, the Northwest Luzon Coastal Current, the SCS Warm Current and the Guangdong Coastal Current. The circulation in the southern SCS is driven mainly by monsoon winds. In winter, the SCS Southern Cyclonic Gyre occupies most of the southern SCS. A weaker anti-cyclonic gyre may exist southeast of the main gyre. Along the border of these two gyres a strong upwind current called the Natuna Off-Shelf Current flows northeastward. In summer, the SCS Southern Anti-Cyclonic Gyre occupies most of the southern SCS. Its northern edge is a very strong off-shore current jet called the Southeast Vietnam Off-Shore Current located at a latitude of about 11° N. The circulation in the central SCS is governed by monsoon winds and the interaction between the circulation systems in northern and southern SCS. The basic features of the circulation in the central SCS are not yet well understood, though some results based on dynamic calculation and numerical simulation have been presented.

(Key Words: South China Sea, Circulation, Upper Ocean)

¹Institute of Oceanology, Chinese Academy of Sciences, 7 Nanhai Rd., Qingdao 266071 CHINA
²South China Sea Institute of Oceanology, Chinese Academy of Sciences, 164 West Xingang Rd., Guangzhou, 510301 CHINA Introduction The South China Sea (SCS) occupies an area of about 3.4 x 10⁶ km², in the northeast of which is situated a deep basin with an area of about 1.5 x 10⁶km². The deep basin connects with the Pacific Ocean through the Luzon Strait. The Kuroshio flows from south to north through the eastern part of the Luzon Strait and tends to intrude into the SCS. A quite strong northeasterly monsoon wind in winter, and a relatively weaker southwesterly monsoon wind in summer, drive the surface water to form accordant circulation patterns.

The complicated circulation pattern in the SCS has recently attracted many oceanographers' interest, and a number of new findings and concepts have been presented. In the present review we will focus on the major features of the upper ocean general circulation.

Early Studies The early studies were mostly based on observational drift current data. These results are still valuable for comparison with the circulation systems presented in later studies. Dale (1956) presented monthly schematic surface current charts for the SCS. Though the data used were very old, the basic patterns do not differ from the recently published atlas. The charts for January and July are shown in Fig.1. Fig.1 Surface current directions drawn by Dale (1956). (a) January. (b) July. Wyrtki (1961) made a comprehensive study of the physical oceanography of southeast Asian waters. He presented bimonthly charts for surface currents and volume transports. The SCS portion of his charts for February and August are shown in Fig.2. Fig. 2 Transports of surface circulation drawn by Wyrtki (1961). (a) February. (b) August. Several major features can be observed from their current patterns: (1) the Kuroshio intrusion through the Luzon Strait takes place in winter but not in summer; (2) the currents in the Taiwan Strait and in the northern and western SCS basically follow the monsoon winds; (3) upwind currents may appear in the central and southeastern SCS and form gyral circulation by combining with the wind-following current.
MAIN COMPONENTS OF THE CIRCULATION The Kuroshio Intrusion Through the Luzon Strait and the SCS Branch of Kuroshio

Dale's surface current charts showed that the Kuroshio intruded into the SCS through the Luzon Strait from October to April but not during the rest of the year. Wyrtki's charts had the same features (Figs.1 and 2). Because the prevailing wind directions are towards the southwest in winter and the north in summer, the above features of the surface currents are quite understandable in view of Ekman transport behavior.

Since the Kuroshio also tends to intrude into the SCS, the intrusion in winter seems undoubted. Guo et al. (1985) proposed that the intrusion could even form a branching current called the SCS Branch of Kuroshio (SCSBK) based on dynamic calculation (Fig.3). This is consistent with the feature presented in Dale’s and Wyrtki’s charts. The intrusion and the SCSBK can also be observed in numerical results (Shaw and Chao, 1994; Li et al., 1994; Li et al., 1996; Fang et al., 1996; Metzger and Hurlburt, 1996) and satellite images (Farris and Wimbush, 1996).

Fig.3 Sea surface dynamic heights (in dyn.cm) relative to 500 db in the northeastern South China Sea. (a) Winter (Dec. 1981-Jan. 1982), from Guo et al. (1985);(b) Summer (June-July, 1979), from Qiu et al. (1984). However, the Kuroshio intrusion in summer, especially for the subsurface layer, is still a controversial issue. The dynamic calculation made by Huang (1983) showed that the westward flow in the Luzon Strait appeared around 20° N in summer for all 4 years examined (1940, 1942, 1965 and 1966). Guo and Fang (1988) gave a westward transport of 11-12 Sv (1 Sv=10⁶m³/s). Qiu et al. (1984) presented a sea surface dynamic height map (Fig.3) and proposed that the SCSBK exists in the northern SCS in summer. Zhong (1990) also presented the dynamic height maps for surveys conducted in 1984, as shown in Fig.4, and the SCSBK can be easily observed in his winter map. A westward flow can also be seen in the summer map, but it appeared as a component of a couple of warm eddies. By examining the ADCP measurements, Pu et al. (1992) concluded that the Kuroshio intrudes into the SCS in all four seasons. However, Guan (1990) and Shaw (1991) proposed that the Kuroshio does not intrude into the SCS in summer.

Xu et al. (1995), Wang and Chern (1996), Xu and Su (1997), Wang and Chern (1997) and Li et al. (1997) made profound studies based on hydrographic data obtained from the mainland China-Taiwan joint surveys of March 1992 and August-September 1994. All these studies suggested that there was no Kuroshio branch current, not only in summer but also in spring. Xu and Su(1997) emphasized the independence the SCS and Kuroshio waters from each other. Li et al. (1997) suggested that the detachment of warm rings from the Kuroshio might be the main mechanism for transporting the Kuroshio water into the SCS. Wang and Chern (1997) found that the region between the Dongsha Islands and the northwest of Luzon was always occupied by cyclonic systems and proposed that the entrainment process was responsible for the water exchange.

Fig.4 Sea surface dynamic heights (in dyn.cm) relative to 1000 db in the northern South China Sea (from Zhong, 1990). (a) December 1984, (b) August 1984.
These studies demonstrated the existence of important dynamics in the anticyclonic (Li et al., 1997) and cyclonic (Wang and Chern, 1997) eddies. From the study of Li et al. (1997), we may further estimate that one warm eddy has a volume roughly equal to (p /4) (150km)²(1km) ~ 1.77 x 10¹³m³. Supposing the shedding frequency to be equal to the average frequency occurring in the Gulf of Mexico, that is, once per 11.5 months (Li et al., 1997), then the equivalent volume transport is 0.6 Sv. This value is small compared with the results from other studies.

The dynamic calculations have shown great variability in the currents in the Luzon Strait and northern South China Sea. Furthermore, the internal waves are very active in this area and can seriously disturb pressure fields (Fang et al., 1995b). Many investigators have used the salinity maximum and minimum in the subsurface and intermediate layers, respectively, as tracers for studying circulation pattern. The maximum salinity generally exists at a depth of 100-200m. Though the distributions of salinity in the 150 m layer have often been used, they are actually inferior to the distribution of maximum salinity because the former are affected by internal waves, upwelling, and many other dynamic processes, and important features may be lost. The distributions of the maximum salinity of the subsurface layer presented by Wang and Chern (1996) and Wang and Chern (1997) are shown in Fig.5. By examining the positions of the high salinity Kuroshio intrusion waters in the vicinity of the Luzon Strait, Wang and Chern (1996) noticed a retrocession process between March and June. The data they used were collected in different years. If one compares the values of maximum salinity in the region away from the Luzon Strait, a conclusion in favor of existence of the SCSBK may be drawn. For most of the area, the maximum salinity in summer is significantly higher than in spring. This suggests that the water supply from the Kuroshio is needed during the period from March to August. The distance between the spring isohaline of 34.74 and the west reach of the isohaline of the same salinity in summer is about 400 km (Fig.6), indicating that a westward flow with a speed of at least 3 cm/s is needed. The weak point of the above argument is that the data were obtained in different years. It would be valuable to compare the maximum salinity distributions in sequential months within the same year.

Fig.5 Distribution of the subsurface maximum salinity in the northeastern South China Sea (from Wang and Chern, 1997). (a) Spring cruise, March 8-27, 1992; (b) Summer cruise, August 28-September 10, 1994. Fig.6 The maximum salinity isolines of 34.74 for spring (solid) and summer (dashed), showing westward spreading of the Kuroshio water. From Fig.5 we can see that two branches seemingly exist to the west of the Luzon Strait. The northern one originates from the loop southwest of Taiwan and tends to flow first northwestward and then turns to the northeast. The southern one branches from the central Luzon Strait and flows toward the west and southwest.

Some direct current measurements, though not systematic, are also in favor of the existence of a westward current. They can be found in the reports by Qiu et al. (1984), Fang et al. (1995b) and Huang et al. (1997). By using the current data measured for 14 months at a site about 100 km west of the Dongsha Islands, Fang et al. (1995b) found that the averaged current was towards the southwest with speeds of about 10 cm/s in the upper layers (0-50m ) and towards the northwest with speeds of about 5 cm/s in the lower layers (60-200 m).

The Northwest Luzon Cyclonic Gyre, Northwest Luzon Cyclonic Eddy and Northwest Luzon Coastal Current

As mentioned previously, from field measurements Wang and Chern (1996) and Wang and Chern (1997) found cyclonic flows in the area northwest of Luzon. Shaw et al. (see Chao et al., 1996) found that it was also an area of upwelling. Also, the numerical results of the barotropic model of Fang et al. (1996) also indicate such a system. Because the system reproduced by numerical simulation is much more complete, the model-produced stream functions are thus shown in Fig. 7. In winter, the cyclonic gyre develops very well and can be called the Northwest Luzon Cyclonic Gyre (NWLCG). There are two cores in this gyre. One is located immediately to the northwest of Luzon, and we will call it the Northwest Luzon Cyclonic Eddy (NWLCE) and further discuss it later. The other is located at about 17° N and 116° E. The northern portion of the NWLCG corresponds to the SCSBK enhanced by the northeasterly monsoon. It starts from the central Luzon Strait and flows in a WSW direction. At a longitude of about 115° E, the northern fringe of the SCSBK branches to the north and then separates into two flows. The east flow forms a part of the SCS Warm Current (SCSWC) and the west flow turns to the west again and is combined with the wind-driven coastal current. These features can be evidenced by an observed snapshot map of dynamic heights of Zhong (Fig.4) and the climatological map of Xu (1982), see Fig. 8a.

In summer, the NWLCG is greatly reduced by the southwesterly monsoon (Fig. 7b). However, the NWLCE still exists. This eddy was probably first indicated by Nitani (1972).

From Fig.7 we can conclude that the NWLCE is a quasi-permanent feature. Accompanying this feature, a near-shore current west of the northern Luzon Island flows to the north regardless of any change in monsoon direction. This coastal current can be observed from Dale’s current charts for all months (Fig.1). We will call it the Northwest Luzon Coastal Current (NWLCC). The permanence (or quasi-permanence) of the NWLCE and NWLCC suggests that there is a persistent forcing mechanism. We have done a series of numerical simulations and found that they were generated by the momentum advection (Fang et al. 1995a). Thus, this system can be regarded as an inertial recirculation of the Western Boundary Current, the Kuroshio. This further indicates that the Kuroshio intrusion should be at least a quasi-permanent phenomenon and play at least an important role in the formation of the NWLCG.

Fig. 7 Stream functions obtained from the barotropic numerical model by Fang et al. (1996). (a) January, (b) July. The contour interval is 0.510⁶ m³/s. Fig. 8 Sea surface dynamic heights (in dyn. m) relative to 1200 db in the South China Sea based on climatological data, from Xu et al. (1982). (a) Winter (Dec.- Feb.). (b) Summer (June-Aug.). The South China Sea Warm Current and Guangdong Coastal Current

The South China Sea Warm Current (SCSWC) was first named by Guan and Chen (1964). It flows from southwest to northeast. The existence of this current in summer is undoubted. Its behavior and forcing mechanism are still not well understood. Guan and Chen (1964) divided the annual variation of the SCSWC into three stages, that is, strong period (June to September), declining period (October to January) and recovering period (February to May). Even in the declining period, some measured currents were still towards the northeast, when simultaneously measured winds were from the east or northeast. Later, Guan (1978) found that the geostrophic currents were directed to the northeast in winter in an area 200 to 400 km away from the shore. Guo et al. (1985) and Zhong (1990) also found that SCSWC existed in the offshore area (Figs. 3 and 4). But the path and origin given by different authors shows great diversity, which in turn raised a variety of explanations for the generation of the SCSWC (Guan, 1998).

It seems clear that most of the SCSWC water will rejoin the loop current southwest of Taiwan and further the Kuroshio. A smaller part can flow into the Taiwan Strait. The monsoon winds also have great influence on the currents in the Taiwan Strait. Although there have been various estimates of the volume transport in the strait, there is no doubt that the annual average is toward the north, which is opposite to the direction of the annual mean wind stress. Fang (1995) attributed this fact to the sea surface height drop from the northeastern SCS to the Oyashio area. This surface slope was believed to be caused by the ocean-wide wind stress curl and the north-south difference in heating by the sun.

To the north of the SCSWC, there is the less saline Guangdong Coastal Current, which simply follows the monsoon winds.

The South China Sea Southern Cyclonic Gyre and Natuna Off-Shelf Current in winter

In the southern SCS, an eddy-like current pattern can be observed from Dale's and Wyrtki's winter charts (Figs. 1a and 2a), and they can be called the South China Sea Southern Cyclonic Gyre (SCG). Wyrtki’s winter transport chart indicates it as a narrow cyclonic gyre centered at about 12° N, 110° E. The dynamic height map of Xu et al. (1982) showed that this eddy occupies the central area of the southern SCS, centered at about 6.5° N, 109° E (Fig. 8a). The model-produced SCG by Fang et al. (1996) is centered at about 6° N, 111° E (Fig. 7a). This gyre has been further confirmed by a field survey carried out by the SCS Institute of Oceanology. The geopotential topography map of Fang et al. (1997) does show a cyclonic gyre in the southern SCS (Fig. 9a). The center is located at almost the same site as that calculated by Fang et al.(1996). Both the simulation by Fang et al. (1996) and the observation by Fang et al. (1997) reveal a strong north-by-northeastward jet-like current. The current originates from the area near the Natuna Island and crosses the Sunda Shelf edge flowing in a direction opposite to the northerly monsoon. This distinguishable current may be called the Natuna Off-Shelf Current (NOC). From the measurements by Fang et al. (1997), the NOC is not only the southeast part of the SCG, but also a part of an anti-cyclonic gyre to the east of the SCG. The numerical results from Fang et al. (1996) also show a weak anti-cyclonic flow, but located further north.

The South China Sea Southern Anti-Cyclonic Gyre and the Southeast Vietnam Off-Shore Current in Summer

An anti-cyclonic flow in the SCS in summer was noticed by Dale (1956) and Wyrtki (1961). The dynamic height chart of Xu et al. (1982) presented a basin-wide gyre-like pattern (Fig. 8b). This gyre has been further confirmed by the numerical simulation (Fig. 7b) by Fang et al. (1996) and the field observation (Fig. 9b) by Fang et al. (1997). The existence of this gyre seems certain, and can be called the South China Sea Southern Anti-Cyclonic Gyre (SAG). The center computed from the numerical model agrees well with the observed one.

A most striking feature of the SCS summer circulation is an offshore current jet situated at a latitude of about 11° N. For convenience, we call it the Southeast Vietnam Off-Shore Current (SEVOC). This current was well recognized as early as Dale’s time (Fig. 1b). Wyrtki estimated the transport of the SEVOC to be 4 Sv. Shaw and Chao's (1994) numerical model was able to reproduce this current, and it showed that the current was located at latitudes between 11° and 14° N. The current computed by Fang et al. (1996) was present at about 11° N with a volume transport of 8 Sv. The dynamic height map by Fang et al. (1997) indicates that the current is situated at 10° N to 12° N, and a volume transport of 34 Sv was estimated based on dynamic calculation.

Fig. 9 Sea surface dynamic heights (in dyn.m) relative to 1500 db in the southern South China Sea (from Fang et al. 1997). (a) Winter (Dec. 1993). (b) Summer (Sept. 1994). A part of the SEVOC turns first toward the southwest at a longitude of about 114° E, flowing along the northwestern margin of the Nansha Islands, and then it turns toward the northwest, finally forming the SAG.

It has not been fully clarified whether, on the north of the SEVOC, there is a cyclonic gyre as produced by the numerical model from Fang et al. (1996). Dale's (1956) current charts, especially for September, did show a southward coastal current east of central Vietnam. Since the summer monsoon is toward the northeast, the appearance of a southward coastal current may imply the existence of a cyclonic gyre off the coast of central Vietnam, as predicted by Fang et al. (1996). If this cyclonic gyre does exist, the SCS summer circulation pattern is comparable with the ocean circulation of the Pacific. The SEVOC is analogous to the Kuroshio Extension -- to the south of the SEVOC, the SAG corresponds to the subtropic gyre; and to the north of the SEVOC, the cyclonic gyre corresponds to the subarctic gyre. Just like the oceans, the wind stress curl is also an essential factor in the formation of the above features of the SCS summer circulation (Metzger and Hurlburt, 1996).

Circulation in the Central SCS

It is clear that the circulation in the central SCS is a result of monsoon forcing, surface heat budget, and interaction between the northern and southern SCS circulation systems. However, the results based on observation (e. g. Dale, 1956; Wyrtki, 1961; Xu et al., 1982) and those from numerical simulation (e. g. Shaw and Chao, 1994; Li et al., 1994; Fang et al., 1996) are greatly disparate. It is almost impossible to find common feature from these results, except that a year-round southward current is likely to exist off the coast of central Vietnam.

A Synthesis of the Basic Pattern of Winter and Summer Circulation

To demonstrate the circulation components discussed above, we present two schematic upper ocean circulation charts for winter and summer (Fig. 10). These charts, of course, mainly reflect the authors? thought and may differ from other results. There are some dashed arrows in the central SCS, indicating that they are very uncertain.

Concluding Remarks The circulation in the SCS is driven by the monsoon winds, the Kuroshio intrusion, and sea surface thermal forcing. The southern SCS circulation is basically governed only by monsoon winds and thus the current pattern is relatively well understood. Consistency among the results of various authors (Dale, 1956; Wyrtki, 1961; Xu et al., 1982; Li et al., 1994; Shaw and Chao, 1994; Fang et al., 1996; Fang et al. 1997) can be identified. Monsoon winds and the Kuroshio intrusion are the major forcing for the northern SCS circulation. The Kuroshio intrusion is in turn dependent on the intensity and path of the Kuroshio in the Philippine Sea of the Pacific Ocean. And furthermore, the thermal forcing also plays a definite role. Thus the circulation in the northern SCS becomes variable and complicated. The circulation in the central SCS is governed by monsoon winds and the interaction between the northern and southern SCS circulation systems, and to some extent, is influenced by sea surface thermal forcing. Therefore, the circulation in the central SCS would be even more variable. Furthermore, field surveys in this area have seldom been reported. The results so far presented have shown great diversity.

A number of circulation features are recognizable and are shown in Fig.10 (thus, they will not be discussed here). However, there is a large amount of observational and numerical results that may differ from the patterns presented in this paper. Also, many points of view regarding the

driving mechanism for various circulation components have been proposed. In conclusion, we may say that there is still a long way to go to fully understand even the upper ocean general circulation patterns and their dynamics in the SCS.

Fig. 10 Schematic diagram of the South China Sea circulation patterns (a) Winter. (b) Summer. 1. Kuroshio, 2. Loop current, 3. SCS Branch of Kuroshio, 4. NW Luzon Cyclonic Gyre, 5. NW Luzon Cyclonic Eddy, 6. NW Luzon Coastal Current, 7. SCS Warm Current,8. Guandong Coastal Current, 9. SCS Southern Cyclonic Gyre,10. Natuna Off-Shelf Current, 11. SCS Southern Anticyclonic Gyre,12. SE Vietnam Off-Shore Current.

Acknowledgment The authors wish to thank Ms. Xinyi Wang for typing the text and Ms. Meishan Du for drawing the figures. This study was supported by the China National Natural Science Foundation, No. 49576280, and the National Special Key Project, No. 97-926-05.
References

Chao, S.-Y., P.-T. Shaw and S. Y. Wu (1996) Deep water ventilation in the South China Sea. Deep-Sea Research, 43(4), 445-466.

Dale, W. L. (1956) Wind and drift currents in the South China Sea. The Malaysian Journal of Tropical Geography, 8, 1-31.

Fang, G. (1995) The structure of the Taiwan-Tsushima-Tsugaru Current System and its relation to the Kuroshio. Marine Sciences. 4, 43-48 (in Chinese, English abstract).

Fang, G., K, Yu and Y. Fang (1995a) Numerical experiments on dynamics of circulation in Taiwan adjacent waters. Program and Abstracts of the Marine Science Conference in Taiwan Adjacent Seas. Institute of Oceanography, National Taiwan University, Taipei, 50-51.

Fang, G., J. Zhao, J. Bole, J. P. Mathisen, P. Schjolberg and H. Pan (1995b) Circulation, internal tides and solitons at the shelf break of the northern South China Sea-An analysis of measured currents. Program and Abstracts of the Marine Science Conference in Taiwan Adjacent Seas. Institute of Oceanography, National Taiwan University, Taipei, 51-52.

Fang, W., Z. Guo and Y. Huang (1997) Observation and study on the circulation in the southern South China Sea. Chinese Science Bulletin, 42(21), 2264-2271 (in Chinese).

Fang, Y., G. Fang and K. Yu (1996) ADI barotropic ocean model for simulation of Kuroshio intrusion into China southeastern waters. Chinese Journal of Oceanology and Limnology, 14(4), 357-366.

Farris, A. and M. Wimbush (1996) Wind-induced Kuroshio intrusion into the South China Sea. Journal of Oceanography, 52, 771-784.

Guan, B. (Kwan, P.) (1978) The warm current in the South China Sea. Oceanologia et Limnologia Sinica, 9, 117-127 (in Chinese, English abstract).

Guan, B. (1990) Features of the summer circulation in the Bashi Strait and adjacent area. Journal of Oceanography of Huanghai & Bohai Seas, 8(4), 1-11 (in Chinese, English abstract).

Guan, B. (1998) A review of studies on the South China Sea Warm Current. Oceanologia et Limnologia Sinica (to appear).

Guan, B. and S. Chen (1964) The current regime of the China adjacent seas. In: Report of the National Ocean Comprehensive Survey, 5, Chap. 6, 1-85 (in Chinese).

Guo, Z. and W. Fang (1988) The Kuroshio in Luzon Strait and its transport during September, 1985. Tropic Oceanology, 7 (2), 13-19 (in Chinese, English abstract).

Guo, Z., T. Yang and D. Qiu (1985) The South China Sea Warm Current and a southwestward current on its right side in winter. Tropic Oceanology, 4(1), 1-8 (in Chinese, English abstract).

Huang, Q. (1983) Variations of velocity and transport of the Kuroshio in the Bashi Channel. Tropic Oceanology, 2(1), 35-41 (in Chinese, English abstract).

Huang, Q., W. Wang, S. Fu, R. Chen, Y. S. Li and K. W. Chu (1997) A westward current that flows through the north of the Dongsha Islands in summer. Tropic Oceanology, 16(2), 58-66 (in Chinese, English abstract).

Li, L., J. Su and J. Xu (1997) Detached Kuroshio rings in the South China Sea. Tropic Oceanology, 16(2), 42-57 (in Chinese, English abstract).

Li, R., Q. Huang and W. Wang (1994) Numerical simulation of the upper layer currents in the South China Sea. Acta Oceanologica Sinica, 16(4), 13-22 (in Chinese).

Li, R., D. Guo and Q. Zeng (1996) Numerical simulation of interrelation between the Kuroshio and the current of the northern South China Sea. Progress in Natural Science, 6(3), 325-332.

Metzger, E. J. and H. Hurlburt (1996) Coupled dynamics of the South China Sea, the Sulu Sea and the Pacific Ocean. Journal of Geophysical Research, 101 (C5), 12331-12352.

Nitani, H. (1972) Beginning of the Kuroshio. In: Kuroshio, Its Physical Aspects (eds. H. Stommel and K. Yoshida), University of Tokyo Press, Tokyo, 129-183.

Pu, S., H. Yu and S. Jiang (1992) Branchings of Kuroshio into Bashi Channel and South China Sea. Tropic Oceanology, 11(2), 1-8 (in Chinese, English abstract).

Qiu, D., T. Yang and Z. Guo (1984) A westward current in the northeastern part of the South China Sea in summer. Tropic Oceanology, 3(4), 65-73 (in Chinese, English abstract).

Shaw, P.-T. (1991) The seasonal variation of the intrusion of the Philippine Sea water into the South China Sea. Journal of Geophysical Research, 96(C1), 821-827.

Shaw, P.-T. and S.-Y. Chao (1994) Surface circulation in the South China Sea. Deep-Sea Research, 41, 1663-1683.

Wang, J and C.-S. Chern (1996) Some aspects on the circulation in the northern South China Sea. La Mer, 34, 246-257.

Wang, J. and C.-S. Chern (1997) On the trajectory of subsurface and intermediate waters in the northeastern South China Sea. Tropic Oceanology, 16(2), 24-41 (in Chinese, English abstract).

Wyrtki, K. (1961) Physical Oceanography of the Southeast Asian Waters. Naga Report, 2, 195pp.

Xu, J., J. Su and D. Qiu (1995) Hydrographic analysis on the intruding of Kuroshio Water into the South China Sea. Proceedings of Symposium of Marine Sciences in Taiwan Strait and Its Adjacent Waters, China Ocean Press, Beijing, 30-44.

Xu, J. and J. Su (1997) Hydrographic analysis on the intrusion of the Kuroshio into the South China Sea, II. Observational results during the cruise from August to September in 1994. Tropic Oceanology, 16(2), 1-23 (in Chinese, English abstract).

Xu, X., Q. Zhang and H. Chen (1982) The general descriptions of the horizontal circulation in the South China Sea. Proc. of the Symposium of the Chinese Society of Marine Hydrology and Meteorology, Chinese Society of Oceanology and Limnology, Science Press, Beijing, 119-127 (in Chinese, English abstract).

Zhong, H. (1990) Structures of the density circulation. In: Report of Decadal Hydrographic Series Survey of the Shelf and Adjacent Waters of the Northern South China Sea (eds. Y. Ma et al.), Ocean Press, Beijing, 215-241 (in Chinese).