Interaction between the East China Sea Kuroshio and the Ryukyu Current as revealed by the self organizing map

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115,, doi: /2010jc006437, 2010 Interaction between the East China Sea Kuroshio and the Ryukyu Current as revealed by the self organizing map Baogang Jin, 1,2 Guihua Wang, 2 Yonggang Liu, 3 and Ren Zhang 1 Received 28 May 2010; revised 13 July 2010; accepted 28 July 2010; published 18 December [1] The self organizing map (SOM) is used to study the linkage between the two western boundary currents, the Kuroshio Current and the Ryukyu Current, through the Kerama Gap. Four coherent ocean current patterns, extracted from a numerical model output for the Kerama Gap area, are used to describe the variability of the Kuroshio current axis and the mesoscale eddies in the Ryukyu Current system. The temporal evolution of these four patterns shows a robust cycle with an average period of 120 days. The shift of the Kuroshio current axis is found to be a dominant factor in determining the water exchange in the Kerama Gap, and the eddies associated with the Ryukyu Current are also important in affecting the strength of eddies in the Kerama Gap. The interaction between the Kuroshio Current and the Ryukyu Current through the Kerama Gap as revealed by the SOM provides new insights in understanding the water exchange between the East China Sea and the northwest Pacific. Citation: Jin, B., G. Wang, Y. Liu, and R. Zhang (2010), Interaction between the East China Sea Kuroshio and the Ryukyu Current as revealed by the self organizing map, J. Geophys. Res., 115,, doi: /2010jc Introduction [2] The Ryukyu Islands act as a barrier of deep water exchange between the northwest Pacific and the East China Sea (ECS). The Kerama Gap lies between the Okinawa and Miyako islands with a maximum depth of about 1800 meters. As a major deep channel, it plays an important role in water exchange through the Ryukyu Island chains (Figure 1). [3] Separated by the Ryukyu Islands are two western boundary currents of the North Pacific Subtropical Gyre: the stronger ECS Kuroshio Current and the weaker Ryukyu Current (Figure 1). Generally, the Kuroshio is a relatively stable current system [e.g., Qiu, 2001], but the shift of the Kuroshio current axis from the area east of Taiwan island to the Kerama Gap can still be seen from many previous studies [e.g., Li, 1993; Su and Yuan, 2005; Ma et al., 2009] while the Ryukyu Current is a northeastward current in the mean state with large temporal and spatial variability [e.g., Yuan et al., 1998; Liu et al., 2000]. The Ryukyu Current velocity and volume transport strongly depend on the mesoscale eddies propagated from the east [e.g., Nakano et al., 1998; Zhu et al., 2003, 2004]. [4] Studies of water exchange between the Kuroshio Current and the Ryukyu Current started in early 1970s. Nitani 1 Institute of Meteorology, PLA University of Science and Technology, Nanjing, China. 2 State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, SOA, Hangzhou, China. 3 College of Marine Science, University of South Florida, St. Petersburg, Florida, USA. Copyright 2010 by the American Geophysical Union /10/2010JC [1972] pointed out that the Pacific water may flow into the ECS through the Kerama Gap. Yu et al. [1993] found that the low salinity water core of the Kuroshio at the Pollution Nagasaki (PN) line to the northwest of Okinawa Island is related to the intrusion of the Ryukyu Current. Current meter data also suggest there is a net mean flow through the Kerama Gap into the ECS [Morinaga et al., 1998]. Zheng et al. [2008] found that the inflow from the Ryukyu Current system through Kerama Gap can make the downstream Kuroshio volume transport larger. Some studies also indicated that the northwestern Pacific mesoscale eddies can propagate into the East China Sea through the Kerama Gap [Ichikawa, 2001; Andres et al., 2008]. However, few argued that the variability of the Kuroshio can also affect the Ryukyu Current through Kerama Gap. The purpose of this paper is to study the variability of the current patterns near the Kerama Gap and further to explore the interaction between the Kuroshio Current and the Ryukyu Current, on the basis of the self organizing map (SOM) [Kohonen, 1982, 2001] analysis of a numerical model output. The sea surface height (SSH) data from both altimetry and drifter observations are used to support the SOM results. 2. Data [5] Ocean general circulation model (OGCM) for the Earth Simulator (OFES) model [e.g., Masumoto et al., 2004] output, forced by the National Centers for Environmental Prediction (NCEP) wind field from 1992 to 2006, is provided by the Japan Agency for Marine Earth Science and Technology. The output has 0.1 degree spatial resolution, 3 day temporal resolution, and 54 vertical layers with varying resolutions from 5 m at the surface to 330 m near the bottom (maximum depth is 6065 m). The model was validated with observations, and it was found to be useful in investigating the large scale 1of7

2 Figure 1. Map of Ryukyu Islands and the mean current field (black arrows) of OFES at 150 m from 1992 to The isobaths in color with 1000 m interval are shown. The Ryukyu Islands are the islands between east of Taiwan and Amami Ohshima Island. I, Taiwan Island; II, Miyako Island; III, Okinawa Island; IV, Amami Ohshima Island; V, Kerama Gap. Inserted map shows the SOM classification area and the Kerama Gap. ocean circulation and mesoscale eddies [e.g., Masumoto et al., 2004]. [6] SSH anomaly (SSHA) data from 1993 to 2006 are downloaded from AVISO website ( oceanobs.com/). The sea level anomalies are defined as differences between the observed SSH and the first 7 year mean sea level. Note the period of SSHA data is different from that of modeling (from 1992 to 2006). The year 1993 is chosen because it is the starting year when we have a complete whole year SSHA data set. This slight difference is not a problem since we are interested in the intraseasonal variability of the current patterns near the Kerama Gap. These 1/4 1/4 gridded data are merged from all available altimeter missions (Topex/Poseidon, Jason 1, ERS 1/2, and ENVISAT) and sampled at a 7 day interval [Le Traon et al., 1998; Ducet et al., 2000]. Anomalous geostrophic currents over the deep ocean (water depth greater than 200 m) are calculated by using the SSHA data. [7] Argos satellite tracked drifters drogued at a nominal depth of 15 m since 1979 are provided by Atlantic Oceanographic and Meteorological Laboratory. In this analysis, the drifter data are interpolated to six hourly time series prior to the construction of the drifter tracks. Among the available records, 11 drifters crossed the Kerama Gap. 3. Method [8] On the basis of unsupervised neural network, the SOM is an effective method for feature extraction and classification [Kohonen, 1982, 2001]. The method can be considered a type of clustering technique, and it is very similar to k means clustering [e.g. Bação et al., 2005]. On the basis of minimum Euclidean distance, the patterns become ordered on a twodimensional grid with each unit assigned to a best matching pattern. From those assignments, a best matching unit (BMU) time series is obtained. The frequency of occurrence for each pattern shown in Figures 2, 3, and 4 is calculated from the BMU time series. The SOM has been demonstrated to be more powerful than the conventional methods, e.g., the empirical orthogonal functions, for feature extraction purposes [e.g., Liu and Weisberg, 2005; Reusch et al., 2005; Liu et al., 2006]. The software package of the SOM toolbox used in this study can be downloaded from the website of the Helsinki University of Technology, Finland ( hut.fi/projects/somtoolbox/). SOM applications in oceanography and climate can be found in the literature [e.g., Richardson et al., 2003; Risien et al., 2004; Liu and Weisberg, 2005; Leloup et al., 2007, 2008; Johnson et al., 2008; Liu et al., 2008; Iskandar et al., 2008]. [9] Prior to the neural network training process, tunable SOM parameters need to be specified. A practical method for choosing among the SOM parameters was given by Liu et al. [2006]. Parameters such as lattice, map shape, initialized weights, training method, and neighborhood function are chosen according to Liu et al. [2006]. The map size, which defines the number of SOM patterns, is specified subjectively according to the variability of the shift of Kuroshio current axis and the strength of the Ryukyu Current. The shift of the 2of7

3 Figure 2. The 2 2 SOM patterns derived from the OFES current fields at the depth of 150 m from 1992 to The colors show the bathymetry and the arrows indicate the currents at the 150 m in the SOM patterns: (a) P1, (b) P2, (c) P3, and (d) P4. The frequency of occurrence is shown in the upper right corner of each panel. Figure 3. Same as Figure 2 but for the 530 m depth. 3 of 7

4 Figure 4. The 2 2 SOM patterns derived from the altimeter geostrophic current anomaly fields for the water deeper than 200 m from 1993 to 2006: (a) P1; (b) P2; (c) P3, and (d) P4. The frequency of occurrence is shown in the upper right corner of each panel. Kuroshio current axis can generally be classified into two types: the Kuroshio is close to or far from the Ryukyu Islands, and the strength of the Ryukyu Current can also be classified into two types: strong or weak. Thus, a map size of 2 2 is chosen. Note that the general conclusions of the study are not sensitive to the choice of map size. We run a series of tests with different map sizes such as 3 3, 4 4, and 5 5. The current variability patterns with these larger map sizes are very similar to the ones with the map size of 2 2 although the patterns in the larger SOM provide more details about the current variability. [10] The velocity data at different depths from OFES output are used in the SOM analysis. To support the model results, we also apply the SOM method to the altimetry data. Note that the long term mean is removed for each data point in the two data sets. Thus, the resulting quantities in the following analyses are referred to as the anomalies. 4. Results 4.1. Spatial Variability [11] The velocity data at the 150 m layer are analyzed first, mainly because 150 m is a good representation of the thermocline depth for both the Kuroshio Current and the Ryukyu Current [e.g., Su and Yuan, 2005]. Fifteen year mean values are removed from each grid point prior to the SOM training. Four coherent current patterns are identified in the current anomaly field from 1992 to [12] In Pattern 1 (P1), the current axis of the Kuroshio is close to the Kerama Gap while the Ryukyu Current is intensified and tends to flow from the Kerama Gap to the Pacific (Figure 2a). A cyclonic anomaly circulation is seen northwest of the Kerama Gap, with northeastward anomaly currents near the Kerama Gap and southwestward currents on the ECS side. Combining the anomaly field with the mean current in Figure 1, the current axis of the Kuroshio in this pattern is closer to the Kerama Gap than the mean current. Southeast of the Kerama Gap, there is a strong northeastward anomaly current, indicating an enhanced northeastward Ryukyu Current. In this pattern, the Kuroshio water flows to the Pacific from the ECS through Kerama Gap with an anticyclonic circulation in the Kerama Gap. [13] In Pattern 2 (P2), the current axis of the Kuroshio stays away from the Kerama Gap while the Ryukyu Current is intensified (Figure 2b). Northwest of the Kerama Gap is an anticyclonic anomaly circulation (southwestward currents near the Kerama Gap and northeastward currents on the ECS side); thus, the current axis of the Kuroshio is farther away from the Kerama Gap. It is interesting to note that there is a strong cyclonic eddy in the Kerama Gap. The water flows to the ECS from the Pacific through Kerama Gap. [14] Pattern 3 (P3) mirrors P2: the current axis of the Kuroshio is close to the Kerama Gap while the Ryukyu Current is weakened (Figure 2c). A cyclonic anomaly circulation is seen to the northwest of the Kerama, and a southwestward anomaly current southeast of the Kerama Gap. Generally, the water flows to the Pacific from the ECS through Kerama Gap with a prominent anticyclonic eddy in the Kerama Gap. 4of7

5 Figure 5. Drifter tracks in the Kerama Gap. (a) Six drifters moved from the ECS to the Pacific; (b) five drifters moved from the Pacific to the ECS. The solid circles (arrows) represent where the drifters entered (left) the research area. [15] Pattern 4 (P4) generally mirrors P1: the current axis of the Kuroshio stays away from the Kerama Gap and the Ryukyu Current is weakened (Figure 2d). The water flows to the ECS from the Pacific through the Kerama Gap with a cyclonic circulation in the Kerama Gap. [16] We also identified the current patterns at other depths using the SOM method. These four patterns described above can be seen from the surface to a depth of 250 m. The currents at the 250 to 600 m layer are similar to those above the 250 m layer. However, as shown in Figure 3, the Kuroshio becomes weaker in this layer, and the eddies in the Kerama Gap are also not seen because the Kerama Gap becomes very narrow (the distance from south to north is around 30 km). The four current patterns are not well identified below 600 m. [17] Anomalous geostrophic currents from satellite altimetry are used to verify the circulation pattern near the Kerama Gap. The patterns from both altimetry observation and OFES model output bear much resemblance although the former does not show those eddies in the Kerama Gap (Figure 4). The diameter of these eddies is usually less than 50 km; thus, it cannot be resolved by the altimetry. The drifters tracks are also used to support the SOM results (Figure 5). Among the 11 drifters that crossed the Kerama Gap, about half (six drifters, 55%) were drifted from the ECS to the Pacific, which is consistent with the flow patterns and the total frequency of occurrence of P1 and P3. The other five drifters moved in the opposite direction (from the Pacific to the ECS), which is consistent with the cases of P2 and P Temporal Evolution [18] On the basis of the BMU time series from 1992 to 2006 shown in Figure 6, the histograms of consecutive days for Figure 6. Time series of the Best Matching Unit (BMU) for the four patterns as shown in Figure 2. The red lines indicate the pattern evolution longer than a cycle and in the sequence of P2 P1 P3 P4 P2. 5of7

6 Figure 7. (a d) The histograms are the number of occurrences in terms of consecutive days for P1, P2, P3, and P4, respectively, on the basis of the BMU time series in Figure 6. each pattern are presented in Figures 7a, 7b, 7c, and 7d. The average consecutive days are 24, 25, 26, and 28 for P1, P2, P3, and P4, respectively. They all have wide ranges from 15 days to more than 75 days. However, the high frequency signal is very robust for each pattern (Figure 7), which is consistent with the presence of high frequent variations of the current axis of the Kuroshio [Qiu et al., 1990; James and Wimbush, 1999; Takahashi et al., 2009]. [19] The temporal evolution of these four patterns can also be seen from the BMU time series (Figure 6). The cycle characterized as a counterclockwise trajectory in the SOM space is very noteworthy. We take P2 P1 P3 P4 P2 as an example: first, the Kuroshio stays away from the Kerama Gap while the Ryukyu Current is intensified (P2); next the Kuroshio shifts closer to the Kerama Gap (P1); then the Ryukyu Current is weakened (P3); and finally the Kuroshio shifts away from the Kerama Gap again (P4). Note P2 as the initial pattern in the sequence is arbitrary. We define the occurrence of the preferred cycle by the counting occurrences of counterclockwise trajectories that pass through all four patterns at least once. The percentage of time that follows the preferred cycle (red lines) to the total time analyzed is 61.5% (Figure 6). The period of this cycle varies from 21 days to 192 days with an average of 120 days. The period of 120 days is consistent with the observed Kuroshio and the Ryukyu Current variability [Zhu et al., 2004]. [20] The evolution of these four patterns may be explained as follows: as the Kuroshio is close to the Kerama Gap, it will induce a negative pressure gradient anomaly, which drives a current from the ECS to the Pacific through the deep part (north) of the Kerama Gap. When a cyclonic (an anticyclonic) eddy approaches the Kerama Gap from the Pacific, a southwestward (northeastward) current in the southeastern Kerama Gap will induce a strong (weak) anticyclonic eddy in the Kerama Gap like P3 (P1) because a stronger (weaker) negative vorticity is induced by the southwestward (northeastward) current [Liu and Su, 1992]. In contrast, if the Kuroshio shifts away from the Kerama Gap, a positive pressure gradient anomaly is induced and further drives a current from the Pacific to the ECS through the deep side. When an anticyclonic (a cyclonic) eddy approaches the Kerama Gap, a northeastward (southwestward) current in the southeastern Kerama Gap will induce a strong (weak) cyclonic eddy in the Kerama Gap like P2 (P4). 5. Summary and Discussion [21] Ocean current patterns of the ECS Kuroshio and the Ryukyu Current variability around Kerama Gap were extracted from the OFES model output from 1992 to 2006 using the SOM. Four coherent patterns were used to summarize the variability of the shift of the Kuroshio current axis and the mesoscale eddies in the Ryukyu Current system. The evolution of these four patterns shows a robust cycle characterized as a counterclockwise trajectory in the SOM space, with an average period of about 120 days. [22] The shift of Kuroshio Current is a dominant factor in determining the water exchange in the Kerama Gap. As the Kuroshio shift to (away from) the Kerama Gap, the water will flow from the ECS (Pacific) to the Pacific (ECS). The eddies in the Ryukyu Current system have an influence on the strength of eddies in the Kerama Gap. A cyclonic (an anticyclonic) eddy in the Ryukyu Current system can enhance the anticyclonic (cyclonic) eddy in the Kerama Gap as the water 6of7

7 flows from the ECS (Pacific) to the Pacific (ECS), which can be explained with the vorticity conservation theory. [23] The characteristic patterns of the currents and their temporal evolution as revealed by the SOM provide new insights to the variability of the Kuroshio and the Ryukyu Current as well as the water exchange through the Kerama Gap. Detailed dynamic and thermodynamic mechanism studies with numerical models and more in situ observations warrant future work. [24] Acknowledgments. We thank the group of Earth Simulator Center of JAMSTEC and Hideharu Sasaki for providing the OFES data. We also thank Nat Johnson for his constructive comments on our paper. Two anonymous reviewers comments were also helpful to improve our manuscript. This study was done at State Key Laboratory of Satellite Ocean Environment Dynamics and was supported by National Basic Research Program in China (2007CB816003), NSFC grants ( and ), and the International Corporation Program of China (2008DFA22230). References Andres, M., J. H. Park, M. Wimbush, X. H. Zhu, K. I. Chang, and H. Ichikawa (2008), Study of the Kuroshio/Ryukyu Current system based on satellitealtimeter and in situ measurements, J. Oceanogr., 64, Bação, F., V. Lobo, and M. Painho (2005), Self organizing Maps as Substitutes for K Means Clustering, Lecture Notes Comput. Sci., vol. 3516, edited by V. S. Sunderam et al., pp , Springer, Berlin. Ducet, N., P. Y. Le Traon, and G. Reverdin (2000), Global high resolution mapping of the ocean circulation from TOPEX/POSEIDON and ERS 1/2, J. Geophys. Res., 105, 19,477 19,498, doi: /2000jc Ichikawa, K. (2001), Variation of the Kuroshio in the Tokara Strait induced by mesoscale eddies, J. Oceanogr., 57, Iskandar, I., T. Tozuka, Y. Masumoto, and T. Yamagata (2008), Impact of Indian Ocean Dipole on intraseasonal zonal currents at 90 E on the equator as revealed by self organizing map, Geophys. Res. Lett., 35, L14S03, doi: /2008gl James, C., and M. Wimbush (1999), Kuroshio meanders in the East China Sea, J. Phys. Oceanogr., 29, Johnson, N. C., S. B. Feldstein, and B. Tremblay (2008), The continuum of Northern Hemisphere teleconnection patterns and a description of the NAO shift with the use of self organizing maps, J. Clim., 21, Kohonen, T. (1982), Self organized formation of topologically correct features maps, Biol. Cybern., 43, Kohonen, T. (2001), Self Organizing Maps, 3rd ed., Springer Ser. Inf. Sci., vol. 30, 501 pp., Springer, New York. Leloup, J. A., Z. Lachkar, J. P. Boulanger, and S. Thiria (2007), Detecting decadal changes in ENSO using neural networks, Clim. Dyn., 28, , doi: /s Leloup, J., M. Lengaigne, and J. P. Boulanger (2008), Twentieth century ENSO characteristics in the IPCC database, Clim. Dyn., 30, , doi: /s Le Traon, P. Y., F. Nadal, and N. Ducet (1998), An improved mapping method of multisatellite altimeter data, J. Atmos. Oceanic Technol., 15, , doi: / (1998)015<0522:aimmom>2.0. CO;2. Li, K. (1993), Climate of the China Seas and Northwestern Pacific (in Chinese), China Ocean Press, Beijing. Liu, X., and J. Su (1992), A reduced gravity model of the circulation in the South China Sea (in Chinese), Oceanolog. Limnologia Sinica, 23(2), Liu, Y., and R. H. Weisberg (2005), Patterns of ocean current variability on the West Florida Shelf using the self organizing map, J. Geophys. Res., 110, C06003, doi: /2004jc Liu, Y., Y. Yuan, T. Shiga, H. Yamamoto, S. Tsubaki, and M. Imai (2000), Circulation southeast of the Ryukyu Islands, in Proceedings of China Japan Joint Symposium on Cooperative Study on Subtropical Circulation System, pp , China Ocean Press, Beijing. Liu, Y., R. H. Weisberg, and C. N. K. Moors (2006), Performance evaluation of the self organizing map for feature extraction, J. Geophys. Res., 111, C05018, doi: /2005jc Liu, Y., R. H. Weisberg, and Y. Yuan (2008), Patterns of upper layer circulation variability in the South China Sea from satellite altimetry using the self organizing map, Acta Oceanol. Sin., 27(suppl.), Ma, C., D. Wu, and X. Lin, (2009), Variability of surface velocity in the Kuroshio Current and adjacent waters derived from Argos drifter buoys and satellite altimeter data, Chin. J. Oceanol. Limnol., 27, Masumoto, Y., et al. (2004), A fifty year eddy resolving simulation of the world ocean Preliminary outcomes of OFES (OGCM for Earth Simulator), J. Earth. Simulator, 1, Morinaga, K., N. Nakagawa, K. Osamu, and B. Guo (1998), Flow pattern of the Kuroshio west of the main Okinawa Island, in Proceedings of Japan China Joint Symposium on Cooperative Study of Subtropical Circulation System, 1 4 December 1997, Nagasaki, Japan, pp , Seikai Natl. Fish. Res. Inst., Nagasaki, Japan. Nakano, T., T. Kuragano, and Y. Liu (1998), Variations of oceanic conditions east of the Ryukyu Islands, paper presented at Japan China Joint Symposium of CSSCS, Fish. Agency of Jpn., Nagasaki, Japan. Nitani, H. (1972), Beginning of the Kuroshio, in Kuroshio, pp , Univ. of Wash. Press, Seattle. Qiu, B. (2001), Kuroshio and Oyashio Currents, in Encyclopedia of Ocean Sciences, pp , Academic, San Diego, Calif. 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(C10), 18,191 18,204. Reusch,D.B.,R.B.Alley,andB.C.Hewitson(2005),Relativeperformance of self organizing maps and principal component analysis in pattern extraction from synthetic climatological data, Polar Geogr., 29, Richardson, A. J., C. Risien, and F. A. Shillington (2003), Using selforganizing maps to identify patterns in satellite imagery, Progr. Oceanogr., 59, , doi: /j.pocean Risien, C. M., C. J. C. Reason, F. A. Shillington, and D. B. Chelton (2004), Variability in satellite winds over the Benguela upwelling system during , J. Geophys. Res., 109, C03010, doi: /2003jc Su, J., and Y. Yuan (2005), Hydrology in China Seas (in Chinese), China Ocean Press, Beijing. Takahashi, D., X. Guo, A. Morimoto, and S. Kojima (2009), Biweekly periodic variation of the Kuroshio axis northeast of Taiwan as revealed by ocean high frequency radar, Cont. Shelf Res., 29, , doi: /j.csr Yu, H., J. Su, and Y. Miao (1993), The low salinity water (LSW) core of Kuroshio in the East China Sea and intrusion of western boundary current (WBC) east of Ryukyu Islands (in Chinese), in Essays on the Investigation of Kuroshio, vol. 5, pp , Ocean Press, Beijing. Yuan, Y., A. Kaneko, J. Su, X. H. Zhu, Y. Liu, N. Gohda, and H. Chen (1998), The Kuroshio east of Taiwan and in the East China Sea and the currents east of Ryukyu Islands during early summer of 1996, J. Oceanogr., 54, Zheng, X. T., Q. Y. Liu, H. B. Hu, Y. Miyazawa, and Y. L. Jia (2008), The study of temporal and spatial characteristics of western boundary current East of Ryukyu submarine ridge and the transport of Kuroshio in East China Sea (in Chinese), Acta Oceanol. Sin., 30(1), 1 9. Zhu, X. H., I. S. Han, J. H. Park, H. Ichikawa, K. Murakami, A. Kaneko, and A. Ostrovskii (2003), The northeastward current southeast of Okinawa Island observed during November 2000 to August 2001, Geophys. Res. Lett., 30(2), 1071, doi: /2002gl Zhu, X. H., H. Ichikawa, K. Ichikawa, and K. Takeuchi (2004), Volume transport variability southeast of Okinawa Island estimated from satellite altimeter data, J. Oceanogr., 60, B. Jin and G. Wang, Second Institute of Oceanograhy, Room 1102, No. 36 Bao Chu Bei Rd., Hangzhou, Zhejiang, , China. Y. Liu, College of Marine Science, University of South Florida, 830 First St. South, St. Petersburg, FL 33701, USA. R. Zhang, Department of Oceanography and Space Environment, Institute of Meteorolgy, PLA University of Science and Technology, Nanjing, Jiangsu, , China. (guihua_wanggh@yahoo.com.cn) 7of7