Figure 1 shows the obtained current velocity at 16m, 104m, and 200m in summer [Tang et al., 1996]. The Kuroshio Current flowed basically northeastward after crossing over the I-Lan Ridge with a maximum speed usually could reach over 80 cm s-1. As approaching the stiff shelf break along 25 the Kuroshio was disrupted. Its main stream flowed basically along the shelf break turned to the east. Owing to the sharply turning, part of its water was spilled over the shelf. West of main stream was a branch current which flowed northwestward through the North Mein Hwa Canyon (NMHC) and intruded onto the shelf. Southwest of branch current was an anti-clockwise eddy whose diameter was around 70 km and centered at Mein-Hwa Canyon (MHC). The current flowed roughly along the bottom topography from the north to the south. Clearly, the countercurrent, observed by Chuang and Wu [1991] and Tang and Yang [1993], is a part of this eddy. In the deep layer, part of the eddy is originated from the branch current, while other portions came from the recycled water of itself. The eddy returned to the Kuroshio at northern tip of Taiwan or further south. The colder and less salty water mixes with the warmer and salty Kuroshio. Then, it re-intruded onto the shelf.
Comparing the measured current velocity with the numerical model result, as shown in Figure 2, the entire flow pattern is seen. The measurements and model results are agreed well in general. The Kuroshio Current could be divided up to three parts as impinging against the steep shelf break of south ECS. One is the main stream which was bent and flowed along the shelf break to the east. The other is a branch current, west of the main stream, carried the Kuroshio water intruding onto the ECS and formed clockwise eddy of diameter around 200
