Warm Eddy Movements in the Eastern Japan Sea

Transcription

1 Journal of Oceanography Vol. 50, pp. 1 to Warm Eddy Movements in the Eastern Japan Sea YUTAKA ISODA Department of Civil and Ocean Engineering, Ehime University, Matsuyama 790, Japan (Received 29 October 1992; in revised form 10 April 1993; accepted 12 July 1993) Warm eddy movements and their areal extent in the eastern Japan Sea were described by presenting space-time diagrams for the warm eddy locations and magnitudes. The analyzed data were compiled from Japan Maritime Safety Agency thermal maps at 200 m depth from 1985 to Two to four warm eddies always existed in the eastern Japan Sea and exhibited both interannual and annual signals. We found that warm eddies were generated in spring around Oki Spur at least three times during the analyzed period of eight years, moved eastward, and interacted with neighboring warm eddies, which were involved in coalescences or separations. The warm eddy distributions off Noto Peninsula have clear seasonal preference. Warm eddies moved eastward from Noto Peninsula in winter-spring to North Japan in the next winter, with mean translation speeds of cm s 1. Warm eddies reaching North Japan typically decayed during a few month after splitting into two or three mesoscale warm eddies. 1. Introduction The Japan Sea is one of the marginal seas of the North Pacific Ocean and about 90% of its entire water body is occupied by homogeneous water of low temperature and low salinity called the Proper Water in the Japan Sea. Above the Proper Water, warm and saline water which enters through the Tsushima/Korea Strait flows northeastward and flows out through the Tsugaru and Soya Straits (see Fig. 1). The surface flow associated with warm water is called the Tsushima Current. The seasonal variation of its volume transport is characterized by the minimum in winter-spring, and maximum in summer-autumn (Kawabe, 1982; Toba et al., 1982; Minami et al., 1987). Figure 2 shows the thermal maps at 100 m and 200 m depth in the Japan Sea for June, 1988 from Japan Maritime Safety Agency (JMSA). The features shown are not unusual but typical synoptic patterns. When we consider the inflow and outflow of different water mass through the above three shallow straits being of less than m depth, the dynamical process suggesting the local baroclinic radius of deformation of km becomes important. Nevertheless, the Tsushima Current has largely spread into the interior of the southern Japan Sea, having a northsouth width of km. The seaward extension of warm waters delimited by the polar front contains several warm eddies. It is also found that these eddies can induce relatively intense currents or branches and great heterogeneity of the hydrological characteristics. From these features of water temperature, multiple, possible transient branches of the Tsushima Current have been proposed by some previous studies, e.g. Suda et al. (1932), Uda (1934), Tanioka (1962), Moriyasu (1972) and Naganuma (1977). This implies that warm eddies play an important role in the spreading of the surface waters from the southern coast to the interior region. Further support for the importance of such eddy transport has been also derived from Fukuoka (1961), Moriyasu (1972), Ichiye and Takano (1988) and Isoda et al. (1992). However, many of the

2 2 Y. Isoda Fig. 1. Bathymetric chart of the Japan Sea and the study area which is the eastern Japan Sea indicated by the hatched marks. The Tsushima Current around each strait are shown schematically by an arrow. Fig. 2. The thermal maps at 100 m and 200 m depth in the Japan Sea for June 1988 from Japan Maritime Safety Agency. Broken thick outline in the map at 200 m depth is the area covered by a statistical regions in Fig. 3(a).

3 Warm Eddy Movements in the Eastern Japan Sea 3 fundamental dynamics for warm eddies in the Japan Sea, e.g. eddy generation, propagation and decay, have been little understood. Variability types of warm eddies in the Japan Sea can be identified from the following statistical characteristics. Figures 3(a) and (b) show the relative frequency distributions of warm eddies in 1 1 squares (Isoda and Nishihara, 1992) and annual mean temperature at 200 m depth (JODC, 1975), respectively. These two maps suggest the horizontal distributions of tracks and strengths of warm eddies as a result of long-term trajectory of eddies. It is seen that the warm eddy movements are not random, but those warm eddies are generally found in the following two major regions. (1) [Y], [Kn] and [Ks] areas in Fig. 3(a). This region is characterized by relatively stable locations with influence from the deep bottom topographic conditions, i.e. the Yamato Rise (Isoda et al., 1991, 1992) and the Ulleung Basin (Kim et al., 1991; Isoda and Saitoh, 1993). However, the physical reason for their relationship to the topographic features has not been clarified. (2) [O] and [N] areas in Fig. 3(a). This region is under the direct influence of the incoming Fig. 3. (a) The relative frequency distributions of warm eddies during the statistical period from 1980 to 1990 (Isoda and Nishihara, 1992). [Y], [O], [N], [Kn] and [Ks] indicate the sea areas with a local maximum frequency and suggest the stable existence area of eddies. (b) Annual mean thermal map at 200 m depth, which have been accumulated for about seventy years (JODC, 1975).

4 4 Y. Isoda warm waters which tend to flow northeastward along the Japanese coast. This coastal boundary flow displays unstable features which evolve into synoptic or mesoscale eddies as shown in Fig. 2. Figure 3(b) shows a large warm pool in the eastern Japan Sea, which corresponds to the region from [O] to [N] in Fig. 3(a). This suggests that warm eddies in this region may move parallel to the Japanese coast without an exact boundary between [O] and [N] areas. We have been specifically interested in the eastern Japan Sea because synoptic phenomena appear to be important for the coastal boundary flow along the Japanese coast. In the present study, as a preliminary step to approach the physical mechanism of warm eddies in the eastern Japan Sea, it is necessary to know the characteristic features of their movements. We examine the warm eddy locations in time by using long-term successive hydrographic data. 2. Data and Analysis Method It is not easy to investigate the detailed characteristics of the long-term and continuous trajectories of warm eddies, because almost all hydrographical observations of the Japan Sea in the past were not carried out at a sufficiently large number of observation points. Furthermore, the behavior of warm eddies is complicated by interactions between the synoptic eddies and the mean flow of the Tsushima Current. Based on the statistical results shown in Figs. 3(a) and (b), the following new idea of analysis for the warm eddy movements will be presented. Fig. 4. The thermal map at 200 m depth on July 1989 as an example, detecting the synoptic four warm eddies indicated by black arrows. Dashed thin lines show 16 divided sub-areas with 0.5 degree longitudinal intervals. In the lower panel, the open circles with the diameter proportional to coretemperature indicate the locations and magnitudes of warm eddies. The lateral bars indicate the noobservation areas.

5 Warm Eddy Movements in the Eastern Japan Sea 5 We used the routine thermal maps at 200 m depth issued two times a month by JMSA from 1985 to 1992 to detect warm eddies. Four sources of water temperature data are mainly used in creating this routine map; they are Japan Defense Agency (JDA), Japan Meteorological Agency (JMA), Japan Fisheries Agency (JFA) and JMSA. The typical observation spacing of km significantly resolves the synoptic warm eddies with km diameter, but not the mesoscale eddies with about 10 km diameter. Consequently, the only evidence of mesoscale perturbations will be apparent station to station noise. Figure 1 shows the bottom topography of the Japan Sea and the present study area, i.e. the eastern Japan Sea indicated by the hatched marks. The western boundary of this study area can be clearly defined by the low occurrence band of warm eddies (less than 20%) as shown in Fig. 3(a). It is likely that warm eddies do not enter into this band area, or that if they do enter, they decay rapidly. In order to simplify movements of warm eddies parallel to the Japanese coast, the attention is focused only on the movements in the east-west direction. We divide the study area into 16 sub-areas with 0.5 degree longitudinal intervals. Each sub-area is numbered from 1 to 16 from west to east. Figure 4 shows the thermal map at 200 m depth on July 1989 and the superimposed 16 sub-areas as an example. Its lower panel shows the detected warm eddy locations and magnitudes. An enclosed isothermal contour undoubtedly indicates a warm eddy. We also consider that a meandering contour of relatively high temperature marks a warm eddy with deeper structure. The location and magnitude of warm eddies were compiled by grid number and core-temperature, respectively. It is possible to some degree to estimate the strength of a warm eddy by reference to its magnitude of core-temperature. In Fig. 4, four warm eddies (open circles with the diameter proportional to core-temperature) and two areas of no-observations (lateral bars) are found. 3. Results 3.1 Observed frequency in the study area Figures 5(a) and (b) show the histograms of the observed frequency at each month and each grid number, respectively. The frequency at each grid number is almost the same, ranging from 50 to 60% (Fig. 5(b)), whereas the frequency at each month ranges from 20 to 90% (Fig. 5(a)). In winter season, since cold-air outbreaks frequently occur in the Japan Sea, the observed frequency becomes much smaller (monthly mean ratios of less than 30% for January and December). 3.2 Annual mean features of warm eddy distributions As a basis for understanding the variability of warm eddies, we will start with the annual mean spatial distribution and briefly discuss its relationship to geography in the eastern Japan Sea. Figures 6(a) and (b) show the relative frequency of warm eddies and mean core-temperature, respectively for each grid. Vertical bars in Fig. 6(b) denote twice the standard deviations of coretemperature. The symbol O and N denote the approximate location of Oki Islands and Noto Peninsula, respectively. West of Oki Islands, i.e. grid cells 1 to 3, the numbers of detected warm eddies are extremely small and core-temperatures are relatively cold. This suggests an intense evolution or formation of warm eddies east of Oki Islands. It is also seen that warm eddies appear with almost the same frequency of 20 40% in grid cells 5 to 15. This indicates that at any particular time, two to four warm eddies exist in the eastern Japan Sea. However, mean core-temperature has two local

6 6 Y. Isoda Fig. 5. Histograms in percentage of observed frequency at (a) each calendar month and (b) each grid number. maxima, i.e. the area between Oki Islands and Noto Peninsula and the area off the west coast of North Japan, as was also seen in Fig. 2(b) from historical temperature data. These coretemperature differences will be discussed later in the context of eddy to eddy interactions and the decay processes of warm eddies. 3.3 Space and time variations of warm eddies Figure 7 shows the space-time diagram of warm eddies, plotting eddy locations and coretemperature as a function of time. The remarkable feature of this diagram is that large temporal variations with interannual and annual time scales occur in the warm eddy distributions. By connecting the same order of eddy magnitudes, we clearly see a tendency for individual eddies to move eastward on average. In particular, east of Noto Peninsula, the slope of multiple eddies linked by solid line shows the eastward movements clearly. We evaluated the translation speed of these features to be cm s 1 ( km d 1 ). Warm eddies in the eastern Japan Sea were generated at least three times, G1 to G3 marks, during the analyzed period of eight years and evolved around Oki Islands. Also the interactions between two warm eddies, i.e. the coalescence C1 to C3 marks, occurred in the sea area between Oki Islands and Noto Peninsula. Although the predominant periods of those generations and coalescences vary from year to year, we can see a seasonal cycle where the starts of eastward movement from Noto Peninsula dominate in winter-spring and their decay around North Japan in the next winter, indicated by D1 to D5 marks. It is seen that these seasonal moving eddies often split from another eddy around Noto Peninsula, indicated by S1 to S6 marks.

7 Warm Eddy Movements in the Eastern Japan Sea 7 Fig. 6. Longitudinal distributions of (a) relative frequency of warm eddies and (b) mean core-temperature in the eastern Japan Sea. Twice a standard deviation for each mean core-temperature is shown by a vertical bar. The symbol O and N denote the approximate location of Oki Islands and Noto Peninsula, respectively. Thus, this phase diagram can be regarded as a sharply nonstationary movements of warm eddies. As referred in this diagram, we can easily trace the trajectory of warm eddies using the original thermal maps at 200 m depth Generation of warm eddies The process of their generations from G1 to G3 can be seen in the sequence of the thermal maps of 200 m depth (Fig. 8). Four sequential maps for each generation indicate the temporal variations from spring to summer in 1985, 1987 and This implies that the generation of warm eddies occurred in spring when the volume transport of the Tsushima Current starts to increase. It is found that a meander ridge of the Tsushima Current around Oki Spur grew an isolated warm eddy with diameter of order km. These generations can be characterized in two

8 8 Y. Isoda Fig. 7. Space-time diagram for eddy locations, where the open circles denote eddy locations and coretemperature, and lateral bars denote no-observation area. The marks G, C, S and D indicate the eddy generation, coalescence between two warm eddies, separation and decay, respectively.

9 Warm Eddy Movements in the Eastern Japan Sea 9 Fig. 8. The thermal maps at 200 m depth from G1 to G3, indicating the generation process of warm eddies in spring. ways. One conditioned the trajectories of coastal eddies, G1 and G2 cases. This type eddy was generated at the eastern side of Oki Islands and moved eastward slowly along the Japanese coast. After that, it came into contact with a warm eddy around Noto Peninsula as discussed later. The other, G3 case, deflected seawards and was leading to the formation of a large isolated warm eddy on Oki Spur. After separating from Oki Spur, this type eddy began to move eastward very slowly without the guide of coast line (see Fig. 7 or Fig. 10). Thus, it is suggested that the generation of warm eddies is due to the flow on the bottom topography around Oki Spur Coalescence between two warm eddies Warm eddies situated between Oki Islands and Noto Peninsula moved, accompanied by interactions with neighboring warm eddies. Figure 9 shows the bi-monthly sequence of thermal maps at 200 m depth revealing the interactions between two warm eddies in 1985 (C1), 1987 (C2) and 1988 (C3), respectively. The results suggest that enlargement of a warm eddy was due to the coalescence with an another eddy. The naming convention is that warm eddies formed on March are labeled A, B and C, respectively. Black area shows the warm water more than 5 C. The coalescence between eddies [A] and [B] shows the following common seasonal characteristics. From March to July, eddies [A] and [B] located east of Oki Islands and northwest

10 10 Y. Isoda Fig. 9. Sequential evolution of warm eddies, revealing the coalescence between two warm eddies, deduced from the thermal maps in 1987 (C1), 1987 (C2) and 1988 (C3). Black area shows the warm water more than 5 C. Three eddies on March are labeled by A, B and C, respectively. of Noto Peninsula, respectively. From July to September, eddy [A] came into contact with eddy [B] and had increased in size with a diameter of about 200 km in the sea area between Oki Islands and Noto Peninsula, namely a large warm eddy [A-B] was formed there. Warm eddy [C] in 1987 moved eastward and came into contact with the west coast of North Japan. After that, eddy [C] split into two mesoscale eddies in July to September and three ones in November. On the other hand, eddy [C] in 1985 and 1988 had an almost unchanged position throughout the year. Such stable warm eddies north of Noto Peninsula have been frequently observed as mesoscale mushroom-like feature in satellite IR images Seasonal eastward movement of warm eddies Figure 7 suggests that a warm eddy indicated by G3 mark was born in spring 1989 and had a long lifetime through the two separations (S5 and S6). Figure 10 shows the sequence of thermal maps at 200 m depth from 1989 to Black area shows the warm water more than 5 C. On

11 Warm Eddy Movements in the Eastern Japan Sea 11 Fig. 10. Sequential evolution of warm eddies, revealing the separations and the seasonal eastward movements, deduced from the thermal maps from 1989 to Black area shows the warm water more than 5 C. Two warm eddies on April 1989 are labeled by A and B, respectively.

12 12 Y. Isoda April 1989, warm eddies [A] and [B] were formed around Oki Islands and Noto Peninsula, respectively. From April 1989 to October 1990, warm eddy [A] was nearly stationary northeast of Oki Islands, but moved slightly eastward. On February 1991, two warm water protrusions [A1] and [A2] from eddy [A] were formed around Noto Peninsula (S5). Subsequently, eddy [A2] split into two warm eddies [A21] and [A22] on August 1991 (S6). A warm water protrusion [B2] from eddy [B] was also formed around Noto Peninsula on November Then each separated eddy, [B or B1] in 1989, [B2] in 1990, [A1] in 1991 and [A21] in 1992, moved eastward from Noto Peninsula in winter-spring to North Japan in the next winter. However, such seasonal movement did not occur in 1985 and 1988 (see Fig. 7), when a warm eddy located around Noto Peninsula combined with a warm eddy near Oki Islands (C1 and C3) Decay of warm eddies In Fig. 11, we depict the temporal decay process of warm eddies, according to the marks from D1 to D5 in Fig. 7. Four sequential maps for each decay process indicate the temporal variations from autumn to winter. Concerning a warm eddy reaching North Japan in autumn, the core of warm eddy reduced in size and split into two or three mesoscale eddies. After splitting, the eddy Fig. 11. The thermal maps at 200 m depth from D1 to D5, indicating the decay process of warm eddies from autumn to winter.

13 Warm Eddy Movements in the Eastern Japan Sea 13 structures no longer existed and these eddies would change to flow along the west coast of North Japan. Thus, the eddy energy does seem to be transferred to smaller intrusions in the coastal boundary flow during the winter. 4. Discussion and Conclusion It is inferred that the upper circulations in the Japan Sea may be affected by instability processes, which lead to the generation of several synoptic warm eddies. Figure 12 illustrates the characteristic features of warm eddy movements in the eastern Japan Sea. Space-time diagrams for the warm eddy locations and magnitudes (Fig. 7) show the importance of Oki Spur geometry as the source of warm eddies (G) and a tendency for warm eddies to move eastward on average. Also, eddy to eddy interactions occur in the sea area between Oki Islands and Noto Peninsula, which are involved in coalescences or separations (I). The warm eddy distributions off Noto Peninsula have clear seasonal preference. Warm eddies move eastward from Noto Peninsula in winter-spring to North Japan in the next winter with mean translation speed of cm s 1. The life of such eddies ends when they merge with the coastal boundary flow along the west coast of North Japan. These decay processes occur for few months after splitting into two or three mesoscale warm eddies (D). Considering the movement of an isolated eddy, the most possible mechanism for the observed eastward movement is a mean flow of the Tsushima Current. In order to investigate the Fig. 12. Schematic illustration of the movement process of warm eddies in the eastern Japan Sea. Thin contour lines indicate the horizontal distribution of annual mean temperature at 100 m depth (JODC, 1975).

14 14 Y. Isoda mean flow variations, the seasonal mean temperature fields of the Japan Sea at 100 m depth in winter and summer are shown in Fig. 13. The polar front characterized by temperature ranging from 3 to 7 C stagnates retaining almost the same pattern and locates latitudinally at about 40 N in both seasons. On the other hand, the most remarkable feature for the seasonal flow variations can be seen along the Japanese coast. The strong current found in summer flows from the entrance of the Tsushima/Korea Strait and then flowing northward just seaward of Oki Islands as far north as North Japan. Minami et al. (1987) estimated the seasonal mean geostrophic volume transport across the PM-line (its location is shown in Fig. 13) as 2.3 Sv (Sv = 10 6 m 3 s 1 ) in winter and 3.3 Sv in summer by using the hydrographic data during 20 years. Using these volume transports, we can calculate the eastward mean speed of 1.5 cm s 1 in winter and 2.2 cm s 1 in summer across the Tsushima Current. These values are of the same order as the observed eastward translation speeds. Therefore, the significant eastward movements from winter to summer off Noto Peninsula are possibly due to the increase of eastward mean flow. Also, the generation and coalescence of eddies showed the following seasonal characteristics, although they did not occur every year. The generations of warm eddies around Oki Spur occur in spring when the volume transport starts to increase. West of Noto Peninsula, the coalescences between two eddies occur in summer autumn when the volume transport becomes maximum. Thus, the combined effects of Oki Spur or Noto Peninsula geometry and seasonal variations of eastward mean flow will be of particular physical interest in the Japan Sea. We also should be noted that warm eddies west of Noto Peninsula have a long lifetime, at least more than one year. As a result, air-sea exchange in wintertime will form a surface homogeneous water and modify the cores of these eddies, as well as available potential energy decrease. However, wintertime structural changes of eddies have not been clarified. Therefore, it is an important study in the future to trace a particular warm eddy and delineate the characteristics of the evolution and movement including the internal structure. Fig. 13. Seasonal mean thermal maps at 100 m depth in winter (left) and summer (right), which have been accumulated for about seventy years (JODC, 1978).

15 Warm Eddy Movements in the Eastern Japan Sea 15 Acknowledgements The author greatly thanks Prof. T. Yanagi for the useful discussions and encouragement during this study. Prof. M. Ikeda made a constructive criticism of an earlier version. Thanks are also due to the two anonymous reviewers who read the manuscript carefully and gave thoughtful comments. The data analysis was carried out on a FACOM M770 of Ehime University. Reference Fukuoka, J. (1961): An analysis of the mechanism of the cold and warm water masses in the seas adjacent to Japan. Rec. Oceanogr. Warks Japan, 6, Ichiye, T. and K. Takano (1988): Mesoscale eddies in the Japan Sea. La mer, 26, Isoda, Y. and M. Nishihara (1992): Behavior of warm eddies in the Japan Sea. Umi to Sora, 67, (in Japanese with English abstract). Isoda, Y. and S. Saitoh (1993): The northward intruding eddy along the east coast of Korea. J. Oceanogr., 49, Isoda, Y., S. Saitoh and M. Mihara (1991): SST structure of the polar front in the Japan Sea. p In Oceanography of Asian Marginal Seas, Vol. 54, ed. by K. Takano, Elsevier, Amsterdam. Isoda, Y., M. Naganobu, H. Watanabe and K. Nukata (1992): Horizontal and vertical structures of a warm eddy above the Yamato Rise. Umi no Kenkyu, 1, (in Japanese with English abstract). Japan Maritime Safety Agency ( ): Quick Bulletin of Oceanic Condition. Japan Oceanographic Data Center [JODC] (1975): Marine Environment Atlas: Northwestern Pacific Ocean 1 (Annual mean). Japan Hydrogr. Association, Tokyo. Japan Oceanographic Data Center [JODC] (1978): Marine Environment Atlas: Northwestern Pacific Ocean 2 (Seasonal and monthly mean). Japan Hydrogr. Association, Tokyo. Kawabe, M. (1982): Branching of the Tsushima Current in the Japan Sea, Part I. Data analysis. J. Oceanogr. Soc. Japan, 38, Kim, K., K. R. Kim, J. Y. Chung and H. S. Yoo (1991): Characteristics of physical properties in the Ulleung Basin. J. Oceanogr. Soc. Korea, 26, Minami, H., Y. Hashimoto, Y. Konishi and H. Daimon (1987): Statistical features of the oceanographic condition in the Japan Sea. Umi to Sora, 4, (in Japanese). Moriyasu, S. (1972): The Tsushima Current. p In The Kuroshio, ed. by H. Stommel and K. Yoshida, Univ. of Tokyo Press, Tokyo. Naganuma, K. (1977): The oceanographic fluctuations in the Japan Sea. Mar. Sci. (Kaiyo Kagaku), 9, (in Japanese with English abstract). Suda, K., K. Hidaka and T. Kubo (1932): The results of the oceanographical observations on board R.M.S. Syunpu Maru in the summer of J. Oceanogr. Imp. Mar. Observ., 4, (in Japanese). Tanioka, K. (1962): A review of sea conditions in the Japan Sea (2) On the cold, warm and saline water regions. Umi to Sora, 38, (in Japanese). Toba, Y., K. Tomizawa, Y. Kurasawa and K. Hanawa (1982): Seasonal and year-to-year variability of the Tsushima- Tsugaru Warm Current system with its possible cause. La mer, 20, Uda, M. (1934): The results of simultaneous oceanographical investigations in the Japan Sea and its adjacent waters in May and June, J. Imp. Fisher. Exp. St., 5, (in Japanese with English abstract).