Numerical Simulation and Dynamical Analysis for Low Salinity Water Lens in the Expansion Area of the

China Ocean Eng, Vol. 28, No.6, pp. 777- 790◎2014 Chinese Ocean Engineering SocietyDOI 10.1007/s 13344-014-0060-1，ISSN 0890-5487Numerical Simulation and Dynamical Analysis for Low Salinity Water Lensin the Expansion Area of the Changjiang Diluted W aterZHANG Wen-jing (张文静), ZHU Shou xian (朱首贤)"1, LI Xun-qiang (李训强),RUAN Kun(阮鲲), GUAN Wei-bing (管卫兵)f and PENG Jian(彭剑)"College of Meteorology and Oceanography, PLA University of Science and Technology,Nanjing 211101, Chinab College of Harbor, Coastal and Ofshore Engineering, Hohai University, Nanjing 210098, China。State Key Laboratory of Satelite Ocean Environment Dynamics, Second Instiute of "Oceanography,Hangzhou 310012, China(eceived 23 October 2012; received revised form 30 July 2013; acepted 16 Dcember 2013)ABSTRACTThe low salinity water lenses (LSWLes) in the expansion area of the Changjiang diluted water (CDW) exist in acertain period of time in some years. The impact of realistic river runoff, ocean currents and weather conditions need to betaken into account in the dynamical analysis of LSWL, which is in need of research. In this paper, the POM-0-z model isused to set up the numerical model for the expansion of the CDW. Then LSWL in summer 1977 is simulated, and itsdynamic mechanism driven by wind, tide, river runoff and the Taiwan Warm Current is also analyzed. The simulatedresults indicate that the isolated LSWL detaches itself from the CDW near the river mouth, and then moves towards thenortheast region outside the Changjiang Estuary. Its maintaining period is from July 26 to August 11. Its formation anddevelopment is mainly driven by two factors. One is the strong southeasterly wind lasting for ten days. The other is thevertical tidal mixing during the transition from neap tide to spring tide.Key words: Changjiang diluted water; low salinity water lens; numerical simulation; dynamic mechanism1. IntroductionThe Changjiang River is the largest river in China, and its annual average discharge is about9322.7* 108 m3 (Shen et al, 2001). Its huge freshwater spreads to the Changjiang Estuary (CE), theEast China Sea (ECS) and the Yellow Sea (YS), and forms a strong plume front. The Changjiangdiluted water (CDW) and its plume front make an enormous impact on the current, water mass,sediment movement and environment of the CE, ECS and YS. So they are always the emphases ofocean engineering and physical oceanography research.Observations also indicate that the CDW is not always continuous, and sometimes there are someisolated low salinity water lenses (LSWLes)，which change the plume front of the CDW. Recently,This project was supported by the National Natural Science Foundation of China (Grant Nos. 40906044, 41076048 and 41376012),the Fundamental Research Funds for the Central Universities (Grant No. 2011B05714) and the Doctoral Starting up Foundation ofCollege of Meteorology and Oceanography of the PLA University of Science and Technology, China.Corresponding author. E mail: zhushouxian@vip.sina.com中国煤化工MYHCNMH G778ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790some studies on the mechanism of the LSWL have been made. By analysing the observed data in theeast of Cheju Island, Lie et al. (2003) suggested that CDW patches separated from the shelf shallowarea and then were advected offshore toward Cheju Island as a series of LSWLes were affected bysouth wind in summertime. Under the climatology condition, Chen et al. (2008) used theunstructured-grid, Finite-Volume Coastal Ocean Model (FVCOM) to simulate some LSWLes in thesea west of 123° E, and proposed that the detachment of the LSWL is associated with eddies generatedby the baroclinic instability across the plume front. Using the Regional Ocean Modeling System(ROMS), Moon et al. (2010) simulated the LSWL under the idealized wind, tide and runoff conditions,and suggested that the strengthening of the tidal mixing during spring tide plays an important role inthe formation of the LSWL. The analysis of the salinity observation in August of thirteen years (Pu,2002) showed that LSWLes only existed in seven years. In August of these years, the difference oftides was ltte, but there was a significant difference among runoff, circumfluence (namely the residualcurrent) and atmospheric conditions. Therefore, it is necessary to make further research of the LSWLunder the real runoff, current and atmosphere conditions. In this paper, the real LWSL in summer 1977is simulated to analyze its characteristics and dynamical mechanism.2. The Numerical Model for CDWPOM-σ-Z mode (Zhang et al., 2011) is used in the simulation of the CDW, which is developedfrom POM (Blumberg and Mellor, 1987). As shown in Fig. 1, σ coordinate is for the currentcalculation, and the transform is:z-7 _ z-η_(1)H+η Dwhere H is the bottom topography, η is the surface elevation, and D=H +η is the total waterdepth. As shown in Fig. 2, σ-Z coordinate is for the salinity calculation. The water is divided into twoparts in the vertical direction, and z=-H。 is the interface. σ coordinate is used in the water abovethe interface, and the transform isH +ηH≤H。(2)H。+nH> H。Z coordinate is used in the water below the interface.The computational domain covers the Bohai Sea, the YS and the ECS. The numerical grids areshown in Fig. 3. The length of the smallest grid near the CE is 1540 m, and that of the largest grid faraway from the CE is about 25000 m. In the vertical direction, the current calculation uses 20 σ verticallayers, and the values of σ for these layers are -0.00056, -0.00167, -0.00335, -0.00670, -0.01339,-0.02679, -0.05357, -0.10714, -0.17857, -0.25000, -0.32143, -0.39286, -0.46429, -0.53571,- 0.60714，- 0.67857, - 0.75000， - 0.82143,- -0.89286, and - 0.96429， respectively. The salinitycalculation is in σ-z layers, which have eight uniform σ layers in the upper water and 24 z layers in the中国煤化工MYHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 79079lower water, and the interface of these two parts is at the depth of 25 m. The depths of the 24 z layersare27m,30m,35m,42m,50m,60m,75m,100m,150m,200m,300m,500m,700m,1000m,1500 m, 2000 m, 2500 m, 3000 m, 3500 m, 4000 m, 4500 m, 5000 m, 5500 m and 6000 m,respectively.Fσ=0二2-0Tz=-HFig. 1. σ coordinate for the current calculation.Fig. 2. 0-Z coordinate for the salinity calculation.40°N-8-36 -34 -32 -Fig. 3. Calculation area and grid for CDW.30 -28 -26 -24 t2-118 120 122 124 126 128 130 132°EThe numerical test of CASE77 is made to simulate the LSWL in summer 1977. CASE77 couplesthe tidal current and the circumfluence at the open boundary. The tidal current is driven by theharmonic constant of M2, S2, KI, and O| tidal constituents. The circumfluence is given in somechannels such as influent flux of 2.0 Sv (Sv is 10'm/s) in Taiwan Strait, influent flux of 29.9 Sv inKuroshio Entrance east of Taiwan Island, excurrent flux of 29.4 Sv in Osumi-spit Karma La Strait, andexcurrent flux of 2.5 Sv in Tsushima-Korea Strait. The Changjiang River runoff is offered with the中国煤化工MYHCNMH G780ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790monthly averaged flux in 1977 at Datong Station. The Qiantang River runoff is 2400 m'/s, which is theclimatic average in summer. In order to keep the balance of the total flux, an additional outflow flux isgiven in Tsushima-Korea Strait, which equals the total runoff of the Changjiang River and theQiantang River. The wind force is obtained from the NCEP reanalysis of 6-hourly 10-m wind data in1977. And the climatic average of the salinity and sea temperature in summer is used as the originalfield data. The coupling calculation of the tidal current and the circumfluence begins on June 10, 1977.Then it runs together with the salinity calculation on June 25, 1977. The sea temperature does notchange with time in the calculation. The numerical scheme of CASE77 is similar to that of theprevious study (Zhang et al, 2011), which introduced more details of the model and its validation.3. Analysis of the LSWL's Evolution by Numerical ResultsUsing the same data from the cruise observation isummer 1977, Pu (2002) and Liao et al.(2001) both discovered the LSWL. Fig. 4 is the observed surface salinity drawn by Pu (2002), in whichthere is an isolated LSWL in the northeast outside the CE. The center of the LSWL isat 124° E, 32.9°N, and the value of the center closure isohaline is 24. Fig. 5 is the observed surface salinity drawn byLiao et al. (2001), in which there is also an isolated LSWL. But there is another high salinity waterlens between the LSWL and the CE. The difference of the two drawings may come from theinsuficiency and non- synchronicity of the observation.34°283132 。0 28-一 +335333034)24126128°E20221226Fig. 4. Distribuion of the surface salinity in August 1977Fig. 5. Distribution of the surface salinity in August 1977(Pu, 2002).(Liao et al, 2001).The simulation of CASE77 shows that there was an LSWL in the northeast outside the CE from13:00 July 26 to 11:00 August 11. Fig. 6 gives part of the simulated surface salinity. The CDW spreadmainly towards southeast and its northern plume front was located around 32° N on July 22. At 13:00July 26, the detachment of the LSWL occurred at the interior side of the plume front around 122.2° E,32° N. The size of the LSWL was 30 km, and its value of the center closure isohaline was 13. FromJuly 26 to August 1, the LSWL moved northeastward quickly with its size gradually enlarging and itsvalue of the central salinity increasing. At 9:00 August 1, the center of the LSWL was located around123° E, 32.8°N, its size was about 130 km, and its value of central closure isohaline was 24. From中国煤化工YHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 79081August 1 to 3, the LSWL moved eastward slowly with the size of 100 km, and its central salinityincreased gradually. At 20:00 August 3, when its center arrived around 123.3° E, 32.8* N, its size wasreduced to 80 km, and the value of the central closure isohaline increased to 26. From 20:00 August 3to 11:00 August 11, the central position of the LSWL was nearly fixed, and its size gradually decreasedto zero as its central salinity increased to 27. After August 11, the surface salinity in the northeastoutside the CE had kept a tongue shape until August 25.34°N32 t30 t30211912112312127°E119211251279E(a)at 1:00 July 22(b) at 8:00 July 23230|2819125 127E281192325(C) at 13:00 July 26(d) at 9:00 August 1E|2 t02819古123 12528 L(e) at 20:00 August 3(0 at 11:00 August 11Fig. 6. Simulated surface salinity in CASE77.中国煤化工MYHCNMH G782ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 7904. Analysis of LSWL's Dynamic Mechanism4.1 Effect of Wind4.1.1 Relationships Among the Evolution of Wind, Lagrangian Residual Current and LSWLFig. 7 shows the 5-day averaged wind of NCEP from July 17 to August 15 in 1977. Based on thesimulated results of CASE77, Lagrangian residual current (LRC) was calculated. Firstly, LRC of eachday was calculated by tracing particles in two M2 tidal periods. Then, the 5-day averaged LRC was madebased on the daily LRC. Fig. 8 shows the 5-day averaged LRC of the sea surface. Although the figurers ofwind and LRC from August 16 to 25 were not given, they were also discussed in the following.2°N42°N| 42°N4003886332303C2826424222ii7 1217 121 125 129 1339E 1 1725 129 133°E(a) From July 17 to21(b) From July 22 to 26(c) From July 27 to 3154293634117 121 125 129 133°E17 121 125 129 133°E17 121 125 129 1339E(d) From August 1 to 5(e) From August 6 to 10(I) From Augustl1to 15Fig. 7. Variation of the sea surface wind from July 17 to August 15 in 1977 (the vectors represent the wind diretion .and the color represents the wind speed).For convenience, we only analyze wind and LRC in the expansion area of the CDW, which is inthe range of 121°E- -126°E, 30° N-33.5°N. From July 17 to 21, it was controlled by southeastwardwind with a mean speed of 4 m/s, and LRC mainly pointed to north or northeast with an average speedof 13 cm/s, while LRC pointed to south or southeast in the range between 122° E and 123° E near theestuary. From July 22 to 31, the wind direction still remained southeastward, but the mean wind speedincreased to 6.0- 6.5 m/s, and LRC mainly pointed to northeast or north and its mean speed increased to20 cm/s. From August 1 to 5, it was mainly controlled by southerly wind with a mean speed of 2.5 m/s,but the direction of LRC was not uniform, and it mainly pointed to northeast with a speed of 11 cm/s inthe northeast outside the CE. From August 6 to 10, the wind direction was not uniform with a speed of中国煤化工MYHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790830.5- -4.5 m/s, and the direction of LRC mainly pointed to south with a mean speed of 9 cm/s. FromAugust 11 to 15, it was mainly controlled by the northeasterly wind with a speed of 2.5- -4.5 m/s, andthe direction of LRC was mainly to south near the estuary and to northwest outside the estuary at anaverage speed of 11 cm/s. From August 16 to 25, it was still controlled by the northeasterly wind witha speed of 5.0- -8.5 m/s, but the direction of LRC was mainly to south near the estuary and to westoutside the estuary at an average speed of 20 cm/s.34°N2|20t→20 cm/s2828 L19121123125127°E211279E(a) From July 17 to21(b) From July 22 to 2632 t320上.30→20 cm/s,2811925127°E121 123127°(C) From July 27 to 31(d) From August 1 to 534°N_34°N-028L _11912(e) From August 6 to 10(f From August11 to 15Fig. 8. Distribution of the 5-day mean Lagrangian residual current of the sea surface in CASE77.中国煤化工YHCNMH G784ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790There was a close relationship among the evolution of wind, LRC and LSWL. From July 22 to 31,forced by the strong southeasterly wind with a mean speed of 6.0 -6.5 m/s, LRC pointed to northeast ornorth with a high mean speed of 20 cm/s, thus the CDW mainly extended northward. The detachmentof the LSWL occurred on July 26, and then the LSWL moved towards northeast with its scaleincreasing from July 26 to August 1. Fig. 8c shows LRC from July 27 to 31. Its northward speed near122.5° E was significantly larger than those at its two sides, which can be explained by twomechanisms. Firstly, the wind-induced northerly LRC increased from the deep shelf sea to the shallowestuary. Secondly, the river discharge and density current at sea surface both pointed towards theoutside of the river mouth, and then turned to south by the Coriolis force, which was in the oppositedirection of the wind-induced Ekman flow. As a result, the northerly LRC in the west of 122.50 E wasweak. The nonuniform LRC imposed northward an uneven force on the CDW and cut it off to form theLSWL.4.1.2 Numerical Tests by Changing WindTwo numerical simulations, namely CASE77a and CASE77b, were conducted to test the effect ofthe wind on LSWL.CASE77a used the same model settings like CASE77 except that it got rid of the wind force fromJuly 22. As shown in Fig. 9, the easterly and northeasterly offshore expansion of the CDW inCASE77a was weaker than that in CASE77, and there was not any LSWL.CASE77b also used the same model sttings like CASE77 except that it changed the wind to theclimatic mean southerly wind of 4 m/s from July 22. As shown in Fig. 10, the northeasterly offshoreexpansion of the CDW was also weaker in CASE77b than that in CASE77. There was not any LSWLlasting a long time in the northeast out of the CE. Two small LSWLes happened near the river mouth:the first one was around 122.40 E, 31.8° N with size of 30- 35 km from 5:00 July 28 to 7:00 July 30,and the second one was from 4:00 August 12 to 20:00 August 13 with similar characteristics as thoseof the first one.34°N|34°N3233~3C3(28_1912127E119212325127°EFig. 9. Simulated surface salinity at 20:00 August 3Fig.10. Simulated surface salinity at 4:00 July 30in CASE77a.in CASE77b.中国煤化工MYHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014，-854.2 Effect of TideThree numerical tests, namely CASE77c, CASE77d and CASE77e were conducted to analyze theeffect of the tide on the LSWL.CASE77c used the same model settings like CASE77 except that it switched the tidal force. InCASE77C, from July 22 to July 28, the CDW extended towards north from south. A small LSWLdetached at 3:00 July 27 around 122.50 E, 32° N with size of 30- -40 km, and it had lasted only for 33hours. From July 28 to August 3, a high salinity tongue was formed around 122.80 E and intrudednorthward, cutting off the CDW. As shown in Fig. 11, the LSWL was formed at 1:00 August 3 in thenortheast out of the CE, and its center was around 123° E, 32.9° N with a size of 70 km. It disappearedat 9:00 August 3, and then regenerated at 5:00 August 4 and lasted until 21:00 August 23. At 23:00August 6, the third LSWL was detached near the river mouth with size of 70 km. It disappeared at12:00 August 13, then regenerated at 16:00 August 15 and had lasted until 5:00 August 17.CASE77d used the same model sttings like CASE77 except that in CASE77d only the M2 tidewas considered as the tidal force. In CASE77d, the CDW extended northward from July 22 to July 28,but there was not any LSWL near the river mouth. There were two LSWLes in the northeast out of theCE: one was from 6:00 to 17:00 on August 12 as shown in Fig. 12, and the other was from 21:00August 17 to 15:00 August 18.349N34°N2|2t00|191212325127°E28119Fig. 11. Simulated surface salinity at 1:00 August 3Fig. 12. Simulated surface salinity at 6:00 August 12in CASE77c.in CASE77d.In the northeast outside the CE, the LSWL had remained until August 23 in CASE77c without anytide force, it had remained less than two days in CASE77d with only M2 tide force, and it had remaineduntil August 11 in CASE77 with four tides force. It is clear that the tidal force is unfavorable for themaintenance of the LSWL in the northeast outside the CE, which can be explained by the inhibitoryaction of the tide -induced horizontal mixing. However near the river mouth, the LSWL occurred at13:00 July 26 in CASE77, and then it developed and moved towards the northeast outside the CE. InCASE77, a shot-lived LSWL appeared near the river mouth from 3:00 July 27 to 12:00 July 28, andthere had been not any LSWL in the northeast outside the CE until August 3. In CASE77d, there wasnot any LSWL near the river mouth. So the combined action of the four tidal constituents was中国煤化工MYHCNMH G786ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790important to the detachment and development of the LSWL near the river mouth. The tidal table ofLvhuashan Station at 122.617° E, 30.817° N shows that the time from July 26 to August 2 was thetransition time from neap tide to spring tide. In CASE77b with the southerly wind of 4 m/s, twoLsWLes appeared near the river mouth, one was from July 28 to July 30, and the other was fromAugust 12 to August 13 when the neap tide changed to the spring tide. So it seems that the detachmentof the LSWL near the river mouth was related to the salinity variation during the transition of neap tideto spring tide, which was verified by a new experiment of CASE77e.CASE77e used the same model settings like CASE77 except that it was impacted by southerlywind of 4 m/s and river runoff of 45600 m'/s which were the typical summertime values. CASE77eeliminated the influence of the variation of wind force and river runoff. In CASE77e, there were twoLSWLes near the river mouth: one was from July 28 to 30 as shown in Fig. 13, and the other was fromAugust 12 to 14. A vertical section from 121.97°E, 31.68°N to 123.55° E, 32.64° N was selected toshow the salinity variation from neap tide to spring tide, as shown in Fig. 14. There was clear verticalsalinity stratification with low-salinity at the suface and with high salinity at the bottom on July 25during the neap tide period, and the surface salinity increased from the shallow estuary to the deepshelf sea. As the middle tide arrived, the tide-induced vertical mixing at the shallow water strengthenedfaster than that at the deep water, so the surface salinity in the shallow water increased faster than thatin the deep water on July 28. On July 29, the surface salinity in the shallow water was higher than thatin the transition area between the shallow water and the deep water, and thus the LSWL was formed.On the spring tide, the tide-induced vertical mixing kept on strengthening. The salinity in the shallowwater was nearly uniform in the vertical direction and the surface salinity in the deep water alsoincreased, thus the surface salinity increased from the shallow water to the deep water again. And theLSWL disappeared on August 1.34°N32~打28119121123251279EFig. 13. Simulated surface salinity at 9:00 July 29 in CASE7e.It is well known that there are fronts of sea temperature in the continental shelf in summer, whichure induced by the well-mixed seawater of the shallow nearshore region and the stratified seawater of中国煤化工MYHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 79087the deep offshore region on the impact of tidal vertical mixing (Simpson et al, 1974; Huang et al,1996; Lv et al, 2007). The mechanism of the LSWL detached by the vertical tidal mixing near theriver mouth is similar to that of the tide-induced fronts.- 10 Fell-1一3 1-20自-300-4一50-520406100 120 1400 120 140x (km)k (km)(a) at 12:00 on July 25 (neap tide)(b) at 12:00 on July 28 (middle tide)一1-2节节- 300f- 4(100120 140-6(100 120 140r (km)(c) at 12:00 on July 29 (middle tide)(d) at 12:00 on August 1 (spring tide)Fig. 14. Simulated vertical distributions of salinity along section in CASE77e.4.3 Effect of River RunoffTwo numerical tests, namely CASE77f and CASE77g were made to analyze the effect of riverrunoff on LSWL.CASE77f used the same model settings like CASE77 except that it doubled the value ofChangjiang river runoff. In CASE77f, there was an LSWL forming near the river mouth and moved tothe northeast outside the CE from 13:00 July 26 to 20:00 August 12. This LSWL was similar to the onein CASE77 except that its central salinity was 2- 6 lower. Fig. 15 gives the characteristics of the LSWLat 20:00 August 3. After August 12, the surface salinity maintained tongue shape in the northeastoutside the CE. And two LSWLes appeared in the tongue, one was from 23:00 August 13 to 8:00August 16, and the other was from 8:00 August 18 to 6:00 August 19.CASE77g used the same model settings like CASE77 except that it cut the value of Changjiang riverrunoff in half. In CASE77g, there was an LSWL forming near the river mouth and moved to the northeastoutside the CE from 12:00 July 26 to 17:00 August 6. This LSWL was similar to the one in CASE77except that its central salinity was 2- 5 higher and had lasted for 5 days shorter. Fig. 16 gives thecharacteristics of the LSWL at 20:00 August 3. There was also another LSWL in the northeast outsidethe CE from 23:00 August 13 to 2:00 August 16.中国煤化工MYHCNMH G788ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790In the simulations conducted by Zhu and Shen (1997), it was indicated that the variation of riverrunoff changed only the current speed and the value of salinity outside the CE, rather than the basiccharacteristics of the current and the expansion of the CDW. In CASE77f and CASE77g, the riverrunoff did not change the basic characteristics of the CDW or the LSWL, but affected the centralsalinity and the detachment time of the LSWL.49N+34°N:2t320t30281912125127°E2811912Fig. 15. Simulated surface salinity at 20:00 August 3Fig. 16. Simulated surface salinity at 20:00 August 3in CASE77f.in CASE77g.4.4 Effect ofTaiwan Warm CurrentTaiwan Warm Current (TWC), which mainly flows northeastward parallel to the 50-m isobathsfrom the Taiwan Strait to Tsushima-Korea Strait, is stronger in summer than in winter. In summer, itcan arrive at the expansion area of CDW (Pu, 2002). Two numerical tests, namely CASE77h andCASE77i were made to analyze the effect of TWC on the LSWL.CASE77h used the same model settings like CASE77 except that it doubled the transport of theTaiwan Strait. In CASE77h, there was an LSWL forming near the river mouth and moved to thenortheast outside the CE from 11:00 July 26 to 6:00 August 11, which was similar to the LSWL inCASE77. Fig. 17 presents the characteristics of the LSWL at 20:00 August 3. Moreover, two LSWLesappeared in the northeast outside the CE later, one was from 18:00 August 13 to 1:00 August 17, andthe other was from 2:00 August 22 to 14:00 August 24.CASE77i used the same model settings like CASE77 except that it cut the transport of the TaiwanStrait in half. In CASE77i, there was an LSWL forming near the river mouth and moving to thenortheast outside the CE from 13:00 July 26 to 14:00 August 11, which was similar to the LSWL inCASE77. Fig. 18 provides the characteristics of the LSWL at 20:00 August 3. Moreover, anotherLSWL appeared in the northeast outside the CE from 14:00 August 15 to 11:00 August 17.The main characteristics of the LSWL can be simulated in both CASE77h and CASE77i, so TWCwas not the main dynamic mechanism of the LSWL detachment, and it only slightly affected thecentral salinity value and the detachment time of the LSWL. CASE77h and CASE77i also confirm thatTWC does not make a significant impact on the CDW because its main part is far away from the CEaccording to some previous studies (Zhu and Shen, 1997; Yuan et al, 1982).中国煤化工YHCNMH GZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014，-89电34°N3232 t3030 f28121127°E11125127EFig. 17. Simulated surface salinity at 20:00 August 3Fig. 18. Simulated surface salinity at 20:00 August 3in CASE77h.in CASE77i.5. ConclusionsIn this paper, POM-σ-Z model is used to simulate the LSWL in summer 1977, and the simulatedresults indicate that LSWL was detached near the Changjiang Estuary, and then moved towards thenortheast outside the CE. It remained from July 26 to August 11, its surface size was 30- 130 km, andthe surface salinity of its center varied from 13 to 27. The characteristics of the LSWL are similar tothe observation despite some differences introduced in the third part. The sea temperature in thesimulation did not change with time, which may bring errors to the simulation of the LSWL. On theother hand, there may be some errors of the LSWL described by the observation because the observeddata are so limited, especially the evolution of the LSWL is not clear because continuous cruiseobservation is lacking. So the numerical simulation is a good way to analyze the characteristics of theLSWL, especially to analyze its evolution.According to the analysis of the effects of wind, tide, river runoff and TWC, the wind and the tideplayed key roles in the formation and evolution of the LSWL in summer 1977. The main dynamicprocess of the LSWL is as follows: from July 22 to 26, the CDW expanded from south to north on theimpact of the strong southeasterly wind with a mean speed of 6- 6.5 m/s. And the vertical tidal mixingin the shallow water strengthened faster than that in the deep water from neap tide to middle tide. Thusthe detachment of the LSWL occurred near the river mouth on July 26. From July 27 to 31, thecontinuous southeasterly wind with mean speed of 6- -6.5 m/s remained, which supported the low-salinity water towards north and northeast to push LSWL northeast. Moreover, the nonuniform LRCmainly induced by wind, river runoff and density current was helpful to cut off the CDW, which wasbeneficial to the development of the LSWL. And at this moment, it was still in the transition from neaptide to spring tide, and thus the nonuniform vertical tidal mixing was also helpful for the developmentof LSWL. From August 1 to 10, the wind decreased with a mean speed smaller than 3 m/s. It has nosignificant effect on the enhancement or damage of the LWSL, while the horizontal tidal mixingcaused the decay of the LSWL. Thus the LSWL disappeared on August 11. The hydrodynamics near中国煤化工YHCNMH G790ZHANG Wen-jing et al.1 China Ocean Eng, 28(6), 2014, 777 - 790and outside the CE are very complex, and it is necessary to make further discussion on the dynamicmechanism of LSWLes in other years.ReferencesBlumberg, A. F. and Mellor, G. L., 1987. A description of a three dimensional coastal ocean circulation model, in:Heaps N. (Ed.), Three-dimensional Coastal Ocean Models, Coastal and Estuarine Sciences, Washington:American Geophysical Union, 4, 1-16.Chen, C. S., Xue, P. F, Ding, P. X., Beardsley, R. C, Xu, Q. C., Mao, X. M., Gao, G. P., Qi,J H, Li, C. Y, Lin, H.C., Cowles, G and Shi, M. C, 2008. Physical mechanisms for offshore detachment of the Changjiang DilutedWater in the East China Sea, J. Geophys. Res, 113, C02002, do: 10.1029/2006JC003994.Huang, D. J, Su, J. L. and Chen, Z. Y, 1996. Application of three dimensional shelf sea model in Bohai. - II Theseasonal variation of temperature, Acta Oceanologica Sinica, 4(1): 1-11. (in Chinese)Liao, Q. Y, Guo, B. H. and Liu, Z. P, 2001. Analysis of direction change mechanism of the Changjiang RiverDiluted Water in summer, Journal of Oceanography of Hluanghai & Bohai Seas, 19(3): 19- -25. (in Chinese)Lie, H. J., Cho, C. H, Lee, J. H. and Lee, S., 2003. Structure and eastward extension of the Changjiang Riverplume in the East China Sea, Journal of Geophysical Research, 108(C3): 3077, doi: 10.1029/ 2001JC001 194.Lv, x, Qiao, F, Xia, C. and Yuan, Y, 2007. Tidally induced upwelling off Yangtze River estuary and in Zhejangcoastal waters in summer, Science in China (Series D): Earth Science, 37(1): 133- 144.Moon, J. H, Hirose, N, Yoon, J. H. and Pang, I. C, 2010. Offshore detachment process of the low-salinity wateraround Changjiang Bank in the East China Sea, Journal of Physical Oceanography, 40, 1036- 1053.Pu, Y. X., 2002. The summer salinity distribution types of the 30° N profile in the East China Sea, Donghai MarineScience, 20(1): 1-13. (in Chinese)Shen, H. T, 2001. Material Flux of the Changiang Estuary, Beijing, China Ocean Press, 1-7. (in Chinese)Simpson, J. H. and Hunter, J. R., 1974. Fronts in the Irish Sea, Nature, 250, 404- -406.Yuan, Y. C, Su, J. L. and Zhao, J. Ss, 1982. A single layer model of the continental shelf circulation in the EastChina Sea, La mer, 20, 131-135.Zhang,W. J,, Zhu, S. X., Dong, L. X., Zhang, C. K., 2011. A new hybrid vertical coordinate ocean model and itsapplication in the simulation of the Changjiang diluted water, China Ocean Eng, 25(2): 327- 338.Zhu, J. R. and Shen, H. T, 1997. Expansion Mechanisms of the Changjiang Diluted Water, Shanghai, East ChinaNormal University Press. (in Chinese)中国煤化工YHCNMH G