INTRUSION OF THE KUROSHIO WATER INTO THE SCS
The SCS is connected to the East China Sea to its north through the Taiwan Strait and to the Java Sea to its south through the Karimata Strait (Fig. 1). The sill depths of these two Straits are, respectively, about 60m and 30m. Persistent northward volume transport has been observed in the Taiwan Strait, likely due to the pressure field associated with the North Pacific Subtropic Gyre (e. g., see Su, 1998). The SCS is also connected to the Sulu Sea through the Mindoro Strait which has a sill depth over 400m. Metzger and Hurlburt (1996) found that a southward transport of around 4Sv throughthe Sulu Sea is required from the SCS if their model is to simulate correctly the characteristics of the low latitude western boundary currents of the North Pacific. If this is proven to be correct, it should have global climate significance because this SCS outflow would contribute directly to the Indonesian Throughflow.
By far the most important water exchange for the SCS is with the WPS through the wide Luzon Strait which has a sill depth of about 2400m. The bottom water of the SCS apparently is derived from the deep water of the WPS and has a renewal time of about 100years (e. g., see Su, 1998). Above this depth the T-S curve of the SCS water has the similar S-shape as that of the WPS water, although the SCS water has lower salinity maximum and higher salinity minimum (Nitani, 1972). Thus the SCS water from the surface down to the salinity minimum also originates principally from the WPS.
Since the Kuroshio flows northward just east of the Luzon Strait water exchange through the Strait of, say, the upper 1000m must cross its frontal zone. However, how this exchange takes place is far from understood. Available hydrographic data in the past do not support any direct intrusion of the Kuroshio water into the SCS, either as a direct branch or as a loop (e. g., see Su, 1998). These include two extensive synoptic hydrographic surveys of the SCS north of 18｡? conducted by us jointly with oceanographers from the National Taiwan University, one in March 1992 (Xu et all, 1996) and the other in August/September 1994 (Xu and Su, 1997). However, water with T-S properties close to the Kuroshio was often observed just southwest of Taiwan. Furthermore, in summer 1994 an intense anticyclonic eddy, extending down to more than 1000m and with at least one station having T-S properties close to those at the Kuroshio front, was observed inside the SCS over its northern continental slope near the Dongsha Islands (Fig. 2). The satellite infrared imageries in December 1996 processed by the Research Cneter of the Hong Kong University of Science and Technology revealed the existence of a warm eddy at about the same location (M. Fang and W. Huang, private communication). However, it cannot be ascertained whether this eddy in fact originated from the Kuroshio front. It is natural to inquire whether eddy intrusion is an important mode by which the WPS water enters the SCS.
Under the South China Sea Monsoon Experiment Program we have undertaken, jointly with oceanographers from Taiwan, two synoptic surveys of the entire SCS in spring 1998, one before and one after the onset of the Asian Monsoon. Again no direct intrusion of the Kuroshio into the SCS was found (Fig. 3). Furthermore, although several eddies with great vertical extent can be identified from the hydrographic data, none possessed the T-S characteristics close to those at the Kuroshio front. It is noted that in Fig. 3 a warm eddy was also present at about the same location near the Dongsha Islands as was the strong warm eddy in summer 1994 (Fig. 2). Whether the plateau near the Dongsha Islands is favorable to entrap warm eddies warrants further study.
