A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope

Transcription

1 Journal of Oceanography Vol. 51, pp. 267 to A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope TOSHIMICHI ITO 1 *, ARATA KANEKO 2, HIROHITO FURUKAWA 3 **, NORIAKI GOHDA 2 and WATARU KOTERAYAMA 4 1 Graduate School of Biosphere Sciences, Hiroshima University, Higashi-Hiroshima 724, Japan 2 Faculty of Engineering, Hiroshima University, Higashi-Hiroshima 724, Japan 3 Faculty of Engineering, Hiroshima Institute of Technology, Saeki-ku, Hiroshima , Japan 4 Research Institute for Applied Mechanics, Kyushu University, Kasuga 816, Japan (Received 10 November 1993; in revised form 16 October 1994; accepted 19 October 1994) The ADCP on an advanced towed fish with controllable main and tail wings, called DRAKE measured a detailed sectional structure of the Kuroshio flowing to the NE along the East China Sea shelf slope west of Okinawa. At the observation period, a countercurrent directed to the SW formed in near-bottom water on the shelf slope. The horizontal flow perpendicular to the stream axis of the Kuroshio constructed a convergence zone around the boundary between the Kuroshio and the countercurrent. An intensive upwelling with the maximum velocity of 2.8 cm s 1 was found to distribute on the shelf slope around the convergence zone. A dynamic cause of this intensive upwelling is discussed carefully. 1. Introduction The interaction between the western boundary current and the shelf sea has attracted great concern of oceanographers because of the complicated mixing and exchange processes between both waters (Nitani, 1972; Lee et al., 1981). The Kuroshio flowing northeastward along the East China Sea shelf slope is known to generate a meandering thermo-haline front at the shelf break and a countercurrent directed to the SW in near-bottom water above the shelf slope (Sugimoto et al., 1988; Qiu et al., 1990; Gohda et al., 1992). From the water-mass and mixing-depth analyses of silica, Ito et al. (1994) provided us an evidence that the intermediate water of the Kuroshio in the depth range of m is upwelled into the East China Sea continental shelf. However, the dynamic cause of the upwelling was not clarified by them. The ADCP on towed fishes (hereinafter called the towed ADCP) has often been operated to measure the Kuroshio in the East China Sea west of Okinawa (Kaneko et al., 1990; Kaneko et al., 1993). The overall velocity structure and volume transport of the Kuroshio were successfully measured by the towed ADCP but its detailed velocity structures on the shelf slope still remain as a target to be elucidated. This paper also presents the Kuroshio data obtained in the East China Sea west of Okinawa by the towed ADCP with controllable main and tail wings. Special attention is focused on velocity structures above the shelf slope. The velocity fields are examined not only for the horizontal velocities but also for the vertical velocity. *Present address: Faculty of Engineering, Hiroshima University, Higashi-Hiroshima 724, Japan. **Present address: Department of Managerial Information Science, Kure Womens College, Aga- Minami, Kure 737, Japan.

2 268 T. Ito et al. 2. Methods The advanced ADCP fish with controllable main and tail wings, called the DRAKE was operated on 2 3 September 1991 at the F-line of length 130 km crossing the Kuroshio west of Okinawa (Fig. 1). An ADCP of frequency 150 khz (RD Instruments; RD-SC0150) was stored inside the DRAKE to look downward. The DRAKE was towed by the tugboat WAKASHIO- MARU (493 t) through a cable of length 800 m and diameter 1.3 cm. The ADCP observation was done in a strategy to sample data every minute with the depth bin of 8 m. The data sampled were further smoothed through a window of depth interval 32 m and time interval 20 minutes for removing small-scale variabilities such as internal waves. Refer to the previous paper (Kaneko Fig. 1. Location map of the observation site. The rough position of the Kuroshio is drawn with a thick solid arrow. The observation site is also shown in a reduced scale on top right. Table 1. Time table for the towed ADCP observation. Forward survey F1 F2 F3 F4 F5 F6 F7 F8 Date 9/2 9/2 9/2 9/2 9/2 9/2 9/2 9/2 Time 8:57 9:52 10:46 11:39 12:33 13:27 14:21 15:10 Return survey F1 F2 F3 F4 F5 F6 F7 F7 F8 Date 9/3 9/3 9/3 9/2 9/2 9/2 9/2 9/2 9/2 Time 4:36 2:53 1:11 23:25 21:35 19:39 17:46 17:08 15:42

3 A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope 269 et al., 1993) for more detailed information on the DRAKE system. The ADCP observation was carried out to make a round-trip course between stations F1 and F8. The forward survey from station F1 to F8 started at 08:57/2 September and ended at 15:10 of the same day. The return survey from station F8 to F1 was done during 15:42/2 September 04:36/3 September (Table 1). At the forward survey, the fish speed and the submerging depth were 5.6 m s 1 and 50 m, respectively. At the return survey from station F8 to F7, the fish speed was reduced to 2.7 m s 1 keeping the same submerging depth as that with the forward survey. In the deep sea from station F7 to F1, the fish was submerged to the depth of 202 m at the same speed as that with station F8 to F7. When the fish was operated in high speeds as done at the forward survey, the observation range of the ADCP became half of the usual range about 400 m due to increased ship noises. As a result of the round-trip survey, the overlapped areas of data acquisition were constructed in the depth range of m between stations F1 and F7 and in the depth range of m between stations F7 and F8 (Fig. 2). The accuracy of velocity observation was checked by comparing the data obtained at the overlapped area between stations F1 and F6 where tidal currents may be weak (Kaneko et al., 1990). Both the data obtained at the overlapped area had the same mean velocity when outputs from the flux gate compass inside the ADCP were modified by 3.9. The standard deviation of the velocity differences resulted in ±6.2 cm s 1. This compass correction was applied to all the velocity data acquired. The data for the vertical velocity were also analyzed in the present study. The RD Instruments four-beam ADCP can measure two kinds of vertical velocities w 1 and w 2. The vertical velocity w is determined as a mean of w 1 and w 2. When there is a big difference between w 1 and w 2, it is natural to consider that the velocity observation is done less accurately. The error velocity Fig. 2. Truck of the towed fish. The areas of data acquisition are indicated with survey and for the return survey for the forward

4 270 T. Ito et al. Ve = w 1 w 2 ( 1) may be used as an index to quantify the reliability of velocity observation. The CTD and XBT casts were done prior to the ADCP observation; stations M1 M3 for the CTD casts and stations F1 F8 for the XBT casts. The stations M1 M3 close to the F-line are the places where currentmeter moorings designed by the Research Institute for Applied Mechanics (RIAM), Kyushu University are running in parallel with the present observation. 3. Water Properties and Horizontal Currents The temperature (T)-salinity (S) diagram is constructed from the CTD data obtained at stations M1 M3 (Fig. 3). The data for stations M2.5 and M3 located at the center of the Kuroshio drew a curve of inverted S-shape. The data obtained in the surface water of station M1 on the shelf slope are remarkably deviated from the curve because of the less saline water originated from the East China Sea shelf water. The data lead us to obtain the following least-squares fits S = T T T T 4 (2) for stations M2, M2.5 and M3, and S = T T T T 4 (3) for station M1. The temperature section crossing the Kuroshio is shown in Fig. 4(a) by using the XBT data at stations F1 F8 and the CTD data at stations M1 M3. The salinities at stations F1 F8 where the XBT casts were done, were determined from the above T-S relationships, Eqs. (2) and (3). The resulting salinity section is shown in Fig. 4(b). The isotherms of 8 16 C and the isohalines of psu slopes steeply down to the right (offshore) between stations F4 and Fig. 3. T-S diagram constructed with the CTD data. The least-squares fits of the data are drawn with the solid lines. The dashed lines show the equi-σ t lines. The marks,, and are used for the data at station M1, M2, M2.5 and M3, respectively.

5 A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope 271 (a) (b) Fig. 4. Contour plots for the hydrographic data. The bottom topography tracked by the ADCP is indicated as an upper boundary of the hatched region. The dashed line shows the boundary of the near-bottom layer where the ADCP could not obtain the data. (a) Temperature. (b) Salinity.

6 272 T. Ito et al. F8, implying the existence of the strong current of the Kuroshio. A less saline wedge intrudes into the surface water between stations F6 and F8, forming a haline front. There is a core of high salinity water in depths m on the offshore side of station F6. The raw ADCP data for the horizontal velocities obtained every minute are shown with the vector diagram in Fig. 5. Nakajima et al. (1992) reported that the main stream of the Kuroshio across the F-line flowed to the direction rotated clockwise by 40 from the north, following the shelf slope. We shall divide the horizontal velocities into two components of the main stream (v) of the Kuroshio and the current (u) perpendicular to it. The vertical velocity (w) is taken to be positive upward. The horizontal velocity sections smoothed with the window of 32 m 20 minutes are presented in Fig. 6(a) for v and in Fig. 6(b) for u. The NE current with positive v forms a maximum velocity of 80.0 cm s 1 at depth 94 m between stations F5 and F6. The NE current is continuously decreased with increasing depth at the layer deeper than 100 m except an isolated core at 250 m depth and closed to zero at depth 600 m. The countercurrent of the Kuroshio with negative v directed to the SW forms in near-bottom water on the shelf slope between stations F6 and F8. In contrast with the NE current, the SW current has a weak vertical shear and is about 23 cm s 1 at 570 m depth. The boundary between the Kuroshio and the countercurrent is parallel to the bottom which slopes down to the SE. The slope of the boundary is much steeper than that of the Fig. 5. Vector diagram of the horizontal velocity obtained at the forward and return surveys. The bottom topography tracked by the ADCP is indicated as an upper boundary of the hatched region. The dashed line shows the boundary of the near-bottom layer where the ADCP could not collect the data.

7 A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope 273 (a) (b) Fig. 6. Horizontal velocities along (a) and perpendicular to (b) the Kuroshio smoothed through the window of 32 m 20 minutes. (a) v. (b) u.

8 274 T. Ito et al. corresponding isotherms and isohalines (see Fig. 4). This boundary is in rough agreement with the boundary for u especially at the depth range of m. The current at this depth range flows to the SE (u > 0) on the left-hand side of the boundary and to the NW (u < 0) on its righthand side. Such a velocity distribution forms a convergence zone around the boundary. We can see another convergence zone for u above the slope deeper than 400 m. The water with negative u at the upper 200 m is directed to intrude into the continental shelf over the whole section. 4. Upwelling The distribution of w smoothed through the window of 32 m 20 minutes are shown with the stick diagram in Fig. 7. The upward current (w > 0) forms around the boundary between the Kuroshio and the countercurrent. The downward current (w < 0) is seen in the main stream of the Kuroshio between stations F1 and F5 and at the upper 200 m of the near-shelf region. The maximum and minimum values of w are 1.7 cm s 1 at the depth 494 m between stations F7 and F8 and 2.4 cm s 1 at the depth 78 m of station F3. The error velocity (Ve) is also shown with the stick diagram in Fig. 8. The Ve is small compared to w over the whole section. Furthermore, there is no clear correlation between the distributions of Ve and w. These evidences provide us confidence on accuracy of the vertical velocity data. The averaged profiles of w between stations F1 and F4 far away from the shelf slope are Fig. 7. Stick diagram for the vertical velocity (w) obtained at the forward and return surveys and smoothed through the window of 32 m 20 minutes. The boundary of the countercurrent is indicated with the thin dashed line.

9 A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope 275 Fig. 8. Stick diagram for the error velocity (Ve). The data are smoothed through the window of 32 m 20 minutes. The boundary of the countercurrent is indicated with the thin dashed line. Fig. 9. Averaged vertical profiles of w for the data between stations F1 and F4; a solid line for the forward survey and a dashed line for the return survey.

10 276 T. Ito et al. Fig. 10. Stick diagram for w obtained by removing the biases from the data presented in Fig. 7. The boundary of the countercurrent is indicated with the thin dashed line. calculated for the forward and return surveys (Fig. 9). The averaged w has negative values at the entire depth and is decreased with the distance form the ADCPs. Because the existence of the strong downward flows in the open ocean between stations F1 and F4 is unrealistic, it is reasonable for us to consider that all vertical velocity data are biased by the negative values profiled above. When the biases are removed from the data presented in Fig. 7, the vertical velocity field changes from Figs. 7 to 10. The region of upwelling is localized in the near-slope region deeper than 200 m. The maximum value of w is 2.8 cm s 1 at 270 m depth between stations F6 and F7. The vertical flows in the remaining region are almost diminished. 5. Discussion The intensive upwelling as found in the present study has never been reported in the deepsea. The up-slope flow can be constructed by horizontal flows toward the shelf slope but their magnitude is too small to generate the intensive upwelling with velocities greater than 2 cm s 1. The NOAA AVHRR images for sea surface temperature showed that the Kuroshio in the East China Sea forms a meandering front of wavelengths km and period days (Qiu et al., 1990). The countercurrent of the Kuroshio may take place in the region where the flow along the front curves cyclonic (Fig. 11). The cyclonic curves of the flow can make not only u but also v converge in the bottom Ekman layer on the shelf or the shelf slope. The subsurface

11 A Structure of the Kuroshio and Its Related Upwelling on the East China Sea Shelf Slope 277 Fig. 11. Sketch of the meandering shelf-break front above the East China Sea shelf slope. The currents taking place around the region where the flow along the front curves cyclonic are drawn with a group of arrows. A thick solid line crossing this region indicates the rough position of F-line. The ADCP data are obtained in the vertical section including the thick solid line. L and C denote the wavelength and phase speed of the meandering front. boundary of the front may slope down to the offshore above the shelf slope keeping a horizontal shape similar to that at the surface. Such an extension of the front is also supported from the fact that the countercurrent of the Kuroshio has been observed in near-bottom water above the shelf slope (Sugimoto et al., 1988; Gohda et al., 1992). The cyclonic curves of the flow along the front also has a chance of spinning up cyclonic eddies (Lee et al., 1981). Both the cyclonic curves of flow and the cyclonic eddies can produce a convergence of horizontal velocities in the underlying bottom Ekman layer. The converging flow turns upward and induces an upwelling because of the bottom restriction. The intensive upwelling observed here may dynamically be understood only in the consideration of converging horizontal flows. The Kuroshio section along the F-line has often been measured by the towed ADCP since The volume transport of the Kuroshio at the upper 700 m is estimated as m 3 s 1 for the present 1991 observation. However, it was 27.2 and ( ) 10 6 m 3 s 1 for the 1987 and 1990 observations, respectively (Kaneko et al., 1990; Kaneko et al., 1993). The transport is remarkably decreased in September We know that the Kuroshio current in the East China Sea usually slacked in autumn (Guan, 1988; Nakajima et al., 1992). It should be noted that the Kuroshio south of Honshu took a nearly straight path only for the 1991 observation. Further study is required to elucidate the year-to-year variability of Kuroshio transport at the observation site. Acknowledgements We wish to express our sincere thanks to Profs K. Kawatate and H. Honji of Kyushu University and to Profs H. Mitsuyasu and S. Mizuno of Hiroshima Institute of Technology for

12 278 T. Ito et al. their best consideration to successfully finish this study. we also thank the captain and crew of the chartered tugboat Wakashio-Maru for their aid in shipboard work. This study was performed as part of the Ocean Research Project of the Research Institute for Applied Mechanics (RIAM), Kyushu University, which was supported by the Japan Ministry of Education, Culture and Science. References Gohda, N., A. Kaneko, S. Mizuno, K. Kawatate and T. Yamaguchi (1992): Measurement of the Kuroshio on the East China Sea shelf slope by means of the ADCPs. Bull. Fac. Eng. Hiroshima Univ., 41, 7 16 (in Japanese). Guan, B. (1988): Major features and variability of the Kuroshio in the East China Sea. Chin. J. Oceanol. Limnol., 6, Ito, T., A. Kaneko., H. Tsubota and N. Gohda (1994): The characteristic distribution of silica over the East China Sea shelf slope. J. Oceanogr., 50, Kaneko, A., W. Koterayama, H. Honji, S. Mizuno, K. Kawatate and R. L. Gordon (1990): Cross-stream survey of the upper 400 m of the Kuroshio by an ADCP on a towed fish. Deep-Sea Res., 37, Kaneko, A., N. Gohda, W. Koterayama, M. Nakamura, S. Mizuno and H. Furukawa (1993): Towed ADCP fish with depth and roll controllable wings and its application to the Kuroshio observation. J. Oceanogr., 49, Lee, T. N., L. P. Atkinson and R. Legeckis (1981): Observation of Gulf Stream frontal eddy on the Georgia continental shelf, April Deep-Sea Res., 28, Nakajima, H., A. Kaneko, N. Gohda, K. Kawatate and S. Mizuno (1992): One-year mooring of an ADCP in the central region of the Kuroshio west of Okinawa. Proc. PORSEC- 92 in Okinawa, 1, Nitani, H. (1972): Beginning of the Kuroshio. p In Kuroshio: Its Physical Aspects, ed. by H. Stommel and K. Yoshida, University of Tokyo Press, Tokyo. Qiu, B., T. Toda and N. Imasato (1990): On Kuroshio front fluctuations in the East China Sea using satellite and in situ observational data. J. Geophys. Res., 95, Sugimoto, T., S. Kimura and K. Miyaji (1988): Meander of the Kuroshio front and current variability in the East China Sea. J. Oceanogr. Soc. Jpn., 44,