Numerical simulation is performed using a high-resolution ocean general circulation model to investigate seasonal variations of the Kuroshio transport. The simulated velocity profiles of the Kuroshio agree surprisingly well with ADCP observations. The model transport variations relative to 700 m across the PN-line near the Nansei Islands are almost the same as the geostrophic transport variations based on the hydrographic observation. The transport shows a weak maximum in summer and a weak minimum in winter, which is 1800 out of phase with that of non-topographic Sverdrup transport. Active eddy formation along the continental slope near the Nansei Islands in summer plays a key role in the increase of the Kuroshio transport.
A simple two-layer planetary geostrophic model with a continental slope explains the seasonal transport variations qualitatively. During the period of weak winds in summer, the transport is much larger than the non-topographic Sverdrup transport. This is because the joint effect of baroclinicity and bottom relief (JEBAR) drives the barotropic flow with releasing the available potential energy accumulated in winter. In winter, the transport is much smaller than the non-topographic Sverdrup transport. This is mainly due to the topographic beta-effect. The barotropic flow excited by the strong winds in winter impinges on the bottom slope to induce the baroclinic flow so that the wind-driven barotropic flow is converted into the available potential energy to be released in summer. Thus the role of JEBAR is to make the total transport of the Kuroshio relatively insensitive to seasonal changes of winds as observed. This mechanism may be called JEBAR rectification" for simplicity.
1.INTRODUCTION
The Kuroshio, one of the most dramatic western boundary currents in the world oceans, has been studied intensively, particularly to determine the volume transport because of its relation to the peculiar bimodal paths south of Honshu as well as to the climate of the North Pacific. As is well-known, direct measurement of the transport across a fixed section for a long period is still very expensive.
Therefore, we have few direct data across particular sections long enough to derive transport variations of the Kuroshio. However, regular hydrographic observations have been carried out routinely along two fixed lines, so that we can estimate geostrophic transports based on dynamic calculations . One section is located west of Okinawa and is called the PN-line by the Nagasaki Marine Observatory; another is located south of the Ku Peninsula and is called the G-line by the Kobe Marine Observatory. The geostrophic transport across the G-line relative to 1000 db is subject to the seasonal variation with a maximum value in summer (46.5 Sv) and with a minimum value in winter (38.5 Sv) [Minami et al.,1979; Taft, 1972]. Similarly, the former section shows a weak maximum in August or September for the geostrophic transport referred to 700 db.
Sekine and Kutsuwada [1994], using a primitive 2-layer numerical model with a rather coarse grid size of 1° x 1° including simplified (despite the authors' claim of realistic") bottom topography, have recently reported that the model Kuroshio transport south of Japan shows a maximum in winter and a minimum in summer. They have tried to account for the phase discrepancy between the model and the above observations by suggesting existence of a barotropic flow east of the Nansei Islands in winter. This remedy, however, contradicts the long-term mean geopotential anomaly field of Hasunuma and Yoshida [19781]. On the contrary, Hasunuma and Yashida [1978] suggest the existence of southwestward countercurrent east of Nansei Islands.
