XIE Rui

College of Harbor, Coastal, and Offshore Engineering, Hohai University, Nanjing 210098, China Shanghai Institute of Waterways, Shanghai 200120, China, E-mail: xierui_12@163.com

WU De-an, YAN Yi-xin

College of Harbor, Coastal, and Offshore Engineering, Hohai University, Nanjing 210098, China

ZHOU Hai

Shanghai Institute of Waterways, Shanghai 200120, China

FINE SILT PARTICLE PATHLINE OF DREDGING SEDIMENT IN THE YANGTZE RIVER DEEPWATER NAVIGATION CHANNEL BASED ON EFDC MODEL*

XIE Rui

College of Harbor, Coastal, and Offshore Engineering, Hohai University, Nanjing 210098, China Shanghai Institute of Waterways, Shanghai 200120, China, E-mail: xierui_12@163.com

WU De-an, YAN Yi-xin

College of Harbor, Coastal, and Offshore Engineering, Hohai University, Nanjing 210098, China

ZHOU Hai

Shanghai Institute of Waterways, Shanghai 200120, China

(Received September 2, 2010, Revised October 25, 2010)

Based on the Environmental Fluid Dynamics Code ( EFDC), a three-dimensional baroclinic model for lower reaches of the Yangtze River, Yangtze River Estuary , Hangzhou Bay and the adjacent sea areas was established with the sigma-coordinate in the vertical direction and the orthogonal curvilinear coordinates in the horizontal direction. The fine silt particle tracers were injected during different tides with the calibrated three-dimensional numerical model in the Yangtze River Estuary to simulate the migration tracks of the tracers in the following three days, and thereby study the migration behavior of the fine silt particles of dredged sediment after dumping at different locations in the navigation channel in the Yangtze River Estuary. The study shows that the particle paths of the tracer particles basically migrate to the open sea along the middle axis of river course in the form of reciprocating oscillation, the migration and diffusion directions of tracer particles are basically parallel to the navigation channel in guide levee, similar to long belt distribution, and the tracer particles migrate to the southeast out of the guide levee. The particle which is on the southern and northern sides enter groin and cross the navigation channel in the process of movement, and is easy to deposit in the navigation channel. The migration tracks of the particles released at different layers are basically similar at the same release station, but specifically different, varying at different cross sections. The result can be used for the choice and evaluation of dredging schemes in the Yangtze River Deepwater Navigation Channel.

deepwater channel, EFDC model, tracer, fine silt particles

1. Introduction

It is very difficult to control the Yangtze River Estuary because of large movement of water and sediment in its navigation channel and frequent changes in its riverbed. In recent years, a project has been carried out in three phases, with its Phase III focusing on dredging. The water depth of the navigation channel is required to increase to 12.5 m from 10 m in Phase II. The navigation channel to be dredged is up to 90.8 km long. According to the estimation of the design institution, the construction volume for the dredging project in Phase III reaches as much as 1.7×108m3, nearly three times of that in Phase II, wherein the dredged soil in need of dredging and beach ascending treatment is 66 140 000 m3and the dredged soil by dumping treatment at dumping areas is up to 103 460 000 m3, accounting for 61% of the total amount. The selection of dumping areas for the Phase III Project is the major technical and economic problem concerning the construction and maintenance of the Phase III Project. In order to improve the operating efficiency of the dredgers and ensure the completion of Phase III Project as scheduled, it is necessary to conduct in-depth andmeticulous study on the theory and simulate the movement of dumped fine silt particles in deepwater navigation channel in the Yangtze River Estuary.

The movement of dumped fine silt particles was studied in different aspects, such as modeling study of motion of particle clouds formed by dumping dredged material[1], bottom shear stress and sediment movement for a wave-current coexisting system[2], total sediment density current differential equation[3], tidal current and sediment mathematical model for the Oujiang Estuary and Wenzhou Bay[4]. Although many research achievements have been made on dumped silt movement, there are still problems required for further study. For example, the study on the process and mechanism of dredging dumped silt movement is immature yet, and numerical simulations on diffusion and migration of dumped silt at different dumping starting times in different dumping manners in the dumping areas and on seabed evolution process caused need dynamic diagnosis study.

In response to these problems, a series of studies should be carried out concerning the study subject based on the physical model and field experiment[5]. Numerical simulation study will merely be conducted in this work with respect to the movement of fine silt particles of the dumped silt in the navigation channel.

The use of Lagrangian particle tracking techniques in the marine environment, offers advantages over more traditional methods of solving the advection-dispersion equation. The technique has been used extensively in many marine applications, e.g., analyses of contaminant dispersal, oil spill modeling, sediment transport, etc..

A Lagrangian two-particle model for the relative diffusion and mixing of two reactive species was proposed[6]. A particle-tracking stochastic method was used to simulate the dispersion of non-conservative radionuclides in the sea[7]. The advancing of water-flooding front and edge water was studied by the tracer simulation technology[8]based on historical performance match and the monitor result match of tracer fluid. A new numerical method[9]for groundwater flow analysis was presented to estimate simultaneously velocity vectors and water pressure head. The advection-dispersion process was treated by the Eulerian-Lagrangian approach with particle tracking technique using the velocities at Finite Element Method (FEM) nodes. In view of the advances in environmental hydrodynamics, more and more attention has been attached to pollutants release characteristic from sediments in Estuary area[10]. Based on the COHERENS[11]model, the Lagrangian particle tracking was simulated to provide tracer trajectories for the contaminant. In order to track the oil slick motion, the behavior of the oil slick on the water surface based on particle approach and tidal flow model in the Huangpu River[12].

The fine silt particles[13]have similar hydrodynamic properties to floating and neutrally buoyant drifters and can be simulated by above-mentioned Lagrangian methods[14].

2. Calibration of Environmental Fluid Dynamics Code ( EFDC ) model

2.1 Governing equation[15]

In order to fit the real boundary, using the orthogonal curvilinear coordinates x and y in the horizontal direction, and using σ in the vertical direction

where z?denotes the original vertical coordinate, ζ the instantaneous water level, h the distance from the mean sea level to the seafloor.

According to the Boussinesq approximation and the assumption of water static pressure in boundary layer, we can derive the momentum equation and the continuity equation:

The momentum equations

The continuity equation

The transport equations of salinity and temperature

where u is the horizontal velocity component in the x direction of the original vertical coordinate, v is the horizontal velocity component in the y direction of the original vertical coordinate, and mx, mythe transformation factors in the horizontal coordinate, m=mxmy. After coordinate transformation, the relationship between the velocity component w in the z direction and its original w?before coordinate transformation.

And H represents the water depth, H=h+ζ, p the practical pressure, p=ρ0gH(1?z),ρ0the reference density, f the Coriolis force, Avthe vertical coefficient eddy viscosity, Qu, Qvthe terms of source and sink, Abthe vertical eddy diffusion coefficient. The continuity equation is treated by the inelastic approximation, that is. Assume that the pressure is affected by density, temperature and salinity both in atmosphere and water flow.

2.2 Turbulence closure model

Assume thatvA is the vertical coefficient eddy viscosity andbA is the vertical eddy diffusion coefficient. IfvA,bA anduQ,vQ are given, the Eqs.(2)-(6) form a closed system about u, v, w, p, ζ. The EFDC is used, which a second-order turbulence closure model proposed by Mellor and Yamada, improved by Galperin to get the vertical coefficient eddy viscosity and eddy diffusion coefficient. The vertical coefficient eddy viscosity, eddy diffusion coefficient, turbulence intensity q, turbulence mixing length l and the Richardson numberqR are connected by the following formulae

2.3 Model setting

The Yangtze River Estuary and Hangzhou Bay are marked by complex coastlines and terrains. Besides, the Yangtze River Estuary has strong dynamic functions of runoff and tide. In order to reduce the interference of boundary conditions on the study area, the model scope shall fully take into account the effect from the fluctuation zone at the tidal current boundary and the tide in the Estuary area. The scope covers the whole Yangtze River Estuary area, Hangzhou Bay and the adjacent open sea areas, up to Jiangyin, with the northern boundary to the Lüsi Port up at about 32.5onorth latitude, the southern boundary to the southern side of the Zhoushan Island at about 29.25olatitude, the eastern boundary to the open sea with isobath of 50 m at about 124.5olongitude, with a wide calculation range. The model scope is shown in Fig.1.

In order to better simulate and forecast the dynamic process of the Estuary and coast, it is of great importance whether or not the grid the model used can properly fit coastline and navigation channel. This article adopts orthogonal curvilinear grids to study the dynamic processes of the Yangtze River Estuary, Hangzhou Bay and the neighboring sea areas. The model maps orthogonal curvilinear grid with the software Delft3D[16]. There are 35 868 wet grids in total in the horizontal direction with the maximum horizontal grid spacing being 4 344 m and minimum 39.5 m, the maximum longitudinal grid spacing being 7 671 m and the minimum 26.3 m, and the time step being 10 s. The program can automatically interpolate topography and water depth at the grid points. The underwater topography adopts 1985-data water depth and the water body is divided into six layers vertically on average. The deepwater navigation channel has been reasonably sealed in the upper reaches of the Estuary near the Chongming Island. Figure 2 showes the computing grid. At the western boundary, the tidal levels measured in Jiangyin are used as the water levels for the open boundary forced condition, while the open boundary water level conditions in the south, while situations in the east and north shall be provided by the model of East China Sea. The water level takes the 1985-data as the reference plane. The simulation period with the model is from 0:00 on September 20, 2002 to 0:00 on September 30, 2002. The floodplain process of tidal current in the Yangtze River Estuary is simulated by using moving boundary processing technology with the wet-dry grid methods. The minimum water depth limit Hmin=0.1m is set to make the model eliminate the appearance of negative water depth in the course of floodplain simulation and ensure operating stability of the model. The model conducts technical treatment of “thin-layer screen” on the guide levee and groin in the deepwater navigation channel in the Yangtze River Estuary and simulates the situation of guide levee and groin with the existence of actual flow fields.

2.4 Validation result of the model and its analysis

The model was validated with the hydrological observed data from 0:00 on September 20, 2002 to 0:00 on the September 30. The validated data includethe measured tidal level, the current velocity and directions of spring, intermediate and neap tides. The layout of sites is shown in Fig.3.

2.4.1 Validation result of the tidal level

10 sites were selected for the validation of tidal levels: Tianshenggang located at the upper of the river, Xuliujing at the south, Congmingzhoutou ahead of the Congming Island, Qinglonggang at the north of the north branch, Lianxinggang at the lower segment of the north branch, Yanglin at the middle of the south branch, Liuao at the middle of the north channel, Zhongjun and Luchaogang at the Nanhui nearshore, Hengsha at the upper of the passage. Because of the textual length, Figs.4 and 5 just list the validation curves at the Xuliujing and Hengsha.

The validation shows that tidal level processes in the spring and intermediate tides at most tide stations are consistent with phase positions. However, the validation results for the neap tide are not so satisfactory. There are three possible reasons. First, the minimum limited depth is a little large at the floodplain processing in the model which will break the balance of the kinetic water mass, this phenomenon will become obvious in the neap tide. Second, the roughness choice need debug especially in the neap tide. Third, the tide is affected by the runoff and the wind wave, which will make the errors bigger.

2.4.2 Validation result of the flow velocity and direction

The model was validated for the stratified current velocity and directions (surface, 0.2, 0.4, 0.6, 0.8 and the bottom layers) in the spring, intermediate and neap tides at 9 measuring points. Because of the textual length, Figs.6-9 just list the validation curves (surface and 0.2 layers) at the Z10 and 4 south points in the spring tide.

As is indicated by the calculation results, the current velocity and current directions obtained through simulation at the measuring points consist with the observed values, basically capable of reflecting the current velocity in the Yangtze RiverEstuary and corresponding change process. Due to the bottom friction and the computing accuracy, there are engendered errors in the processing of the velocity variation. Local velocity sometimes becomes higher or lower. As a whole, the numerical calculation results accord well with the observed data, indicating that the flow field simulation with the model in the Yangtze River Estuary is successful and provides a reliable basis for further study on dredged sediment movement in the area.

3. Simulation on the movement of fine silt particles in dumped silt

3.1 Model establishment

Tracers were injected in the spring, intermediate and neap tides with the calibrated three-dimensional numerical model in the Yangtze River Estuary to simulate the migration tracks of the tracers in the following three days, and thereby study the migration rule of the fine silt particles of dredged sediment after dumping in the Yangtze River Estuary, in which Jiangyin tide station data in the upper reaches are taken as the upstream tidal level boundary conditions of the three-dimensional model in the Yangtze River Estuary, while the open sea tidal level boundary conditions will be given by the calculation results of the East China Sea tidal wave model. Partial parameters in the model shall take the foregoing validated model parameters. The navigation channel of the Yangtze River Estuary is shown in Fig.10.

3.2 Selection of tracer particle injection points

The tracer particle movement tracks are related not only to the starting time of particle injection, but also to the initial location. In order to better study the migration rule of the movement of fine silt particles in dumped silt and provide mathematical model reference for the selection of dumping areas of the Phase III Control Project in the deepwater navigation channel in the Yangtze River Estuary, tracer simulation was conducted for the dumping points at different locations in the navigation channel as shown in Fig.11.

According to the northern passage flow field upon the completion of Phase II Project in the Yangtze River Estuary, the flow pattern at the lower section of the northern passage has changed significantly upon the completion of Phase II guide levee. The currents at W3-W4changed from the original strong rotation ones into the main reciprocating ones. The current direction was concentrated during ebb tide and consisted with the navigation channel direction, creating favorable conditions for the lower section of the northern passage to set dumping areas. The flow field data (with physical model and mathematical model) upon the completion of Phase II Project further show that the drop velocity of the lower navigation channel and the navigation channel included angle are very small through the adjustment of Phase II guide levee and groin. Moreover, there is a turning cross section between S6and S7groins, below which the current particle on the southern side of the navigation channel is far from the navigation channel, difficult for sediment to enter the navigation channel, while the northern side is the opposite. In addition, as is indicated by the observed data for the topographic scouring and siltation in the northern passage from May 2002 - May 2005, the area was in scouring state since the implementation of deepwater navigation channel Phase II Project and it is still possible to become deeper in the future. Therefore, the tracer points of dumped silt are better to be chosen as the south between S7and S8groins. Therefore, the tracers were injected at A (397, 88), B (397, 89) and C (397, 90), at the same time, tracers were also injected at D (397, 94) on the northern side of the section and E (388, 88) and F (388, 94) on the southern and northern sides between S5and S6groins in the upper reaches of the navigation channel in order to conduct comparison at different dumping points. In the computing scope, tracer particles were released at each position in 6 layers according to the relative water depth, i.e., surface, 0.2, 0.4, 0.6, 0.8 and the bottom layers, with totally 36 placed. The model ran for 11 d (from 0:00 on September 20, 2002 to 0:00 October 1, 2002). The coordinates of all the tracer particles were recorded every 2 h.

3.3 Simulation results

3.3.1 Horizontal movement of fine silt particles

Due to space limitations, we only analyze the migration tracks of the tracer particles injected in the spring tides. Figures 12 to 23 respectively show different migration processes of 36 tracer particles injected in the spring, intermediate and neap tides.

As is shown in Figs.12-17, because the water channels within the Yangtze River Estuary sand bar are affected by the function of runoff and river course binding, the particle paths of the tracer particles injected at A (397, 88), B (397, 89) and C (397, 90) basically migrate to the open sea along the middle axis of river course in the form of reciprocating oscillation. The directions of migration and diffusion of tracer particles basically are parallel to the navigationchannel in guide levee, similar to long belt distribution, and the tracer particles migrate to the southeast out of the guide levee. The tracer particles mainly migrate in the area parallel to the navigation channel, with small amount entering or crossing the navigation channel within the guide levee. The migration tracks of the particles released at different layers are basically similar at the same release station, but specifically different, varying at different cross sections. Although the paths of particles rejected at D (397, 94) and E (388, 88) and F (388, 94) at the southern and northern sides between S5and S6groins in the upper reaches of the navigation channel (as shown in Figs.18-23) also basically take the shape of long belt and migrate to the open sea in the form of reciprocating oscillation, too many particles enter groin and cross the navigation channel in the process of movement, easy to deposit in the navigation channel. Besides, the migration behaviors of each layer vary at E and D and F. In the surface layer, tracer particles at D and F basically migrate to the open sea along the current in the form of reciprocating oscillation, entering and crossing the navigation channel several times, while the particle tracks at E are much more disordered, since the particles are doing oscillation migration at E and near the groin, and the migration distance is very short three days later. In the 0.2 and 0.4 layers, the particle migration paths at D and F are similar to those in the surface layer, while the particle paths at E are slightly stable relatively to those in the surface layer, but oscillating to and fro near the groin as well. Starting from 0.6-layer, the particle paths at D and F gradually become disordered, showing reciprocating oscillation near the groin and increased times of entering and crossing the navigation channel. At the bottom layer, the particle paths at F almost oscillate to and fro near the groin at the northern side of the navigation channel. As D is below the turning cross section, the particles oscillate to and fro near the groin for quite a long time period at the beginning, but finally entering the open sea.

As is shown by the figures, the tracer particles were injected at 7:00 on September 22 when flood tide just began in the Yangtze River Estuary, and the tracer particles in all the layers had different upward transport processes. However, because the ebb tidal current is larger than flood tidal current in the Yangtze River Estuary and ebb tide lasts longer than flood tide, the tracer particles migrate to the open sea basically in the form of reciprocating oscillation. As a result of strong current near dumping points, the longitudinal transport speed is high. And the northern passage is ebb channel preferential flow at flood season, with the upward transport distance evidently shorter than downward transport distance, and eventually enters the open sea. Then the particle traces exhibit the form of clockwise rotation. Except for the bottom layer, the migration tracks of tracer particles are almost similar in the other layers, all migrating downwards in rotary oscillation in the shape of long belt. There is certain error in current velocity simulation in the bottom layer because of the impact of bottom friction coefficient and calculation accuracy, and the migration tracks of tracer particles show that the migration scope is smaller compared with the other layers with short downward migration distance without long belt distribution.

Under the current action, the tracer particles transport horizontally in addition to the longitudinal migration along the tidal current. According to the analysis of Figs.12-17, the horizontal transport of the tracer particles at A, B and C has the following characteristics. First, the horizontal transport intensity is very weak compared with the longitudinal transport, mainly in the form of reciprocating oscillation caused by the reciprocating movement of the current in the Yangtze River Estuary, second, the horizontal transport is more evenly to both sides, third, it is relatively stable, i.e., the horizontal transport basically remains near the migration tracks and the distances of peak values horizontally distributed on cross section from the navigation channel almost remain stable. However, D, E and F have passage crossings to different degrees, which is not conducive to dumping. In accordance with the tracer particle migration tracks as shown in Figs.12-23, the migration tracks of the particles injected at A, B and C between S7and S8are satisfactory, all migrating to the open sea in the form of reciprocating oscillation along the axis of the navigation channel, with the migration and diffusion directions of the particles almost parallel to the navigation channel within the guide levee, similar to long belt distribution, and migrating to the southeast out of the guide levee. In the guide levee, the tracer particles mainly migrate in the area parallel to the navigation channel, with small amount entering and crossing the navigation channel. Therefore, the dumping areas are better to be chosen between S7and S8.

3.3.2 Field text

At present three research methods ahave been usually selected, that is, the site survey, physical model and mathematical model in the choice of the research field at the Spoil area. And it also need refer to the data of the site survey involving several aspects such as hydrology, sediment, terrain and deposition. In the aspect of the field sediment motion investigation, the tracer sand technology has been always used. Estuarine and coastal science research center carried out the field text with neutron activation of the tracer sand technology, which is used to research the physical process in the dumping sediment motion.

According to the preconcentration research ofdumping sediment, the release Point A is at the south of the central axis channel between S7and S8(corresponding coordinates: 122o11'11.57"E, 31o9'33.52"N). And the water depth is ?8 m. There are 11 sampling sections around point A, and the nearest sections is 1.4 km from the point A. The others are 2.8 km apart for each line. Because of the ebb current is bigger than the flood current, there are 6 sampling sections at the lower (7# - 12#) where are 5 at the upper(2# - 6#). Furthermore, in order to research the circs and probability of the dumping sediment across the channel, there is arranged one sampling point in the channel, and several points at the north of the channel. The points are shown in Fig.24. In order to insure the examination success and survey the tracers, sampling was arranged 3 times: 2 d later, 3 d later and 4 d later. The examination is during the spring tide with no dredging.

According to the data from different time periods, it can protract the diffusion isolines (as shown in Figs.25-27) of the tracer sand at each sampling period. From these figures, we can find that the migration and diffusion directions of tracer particles are basically parallel to the navigation channel in guide levee, similar to long belt distribution. The tracer particles mainly migrate in the area near the axis of the navigation channel, with small amount entering or crossing the navigation channel. The transport velocity of the tracer particles is so fast that the front of the tracer particles movement downward is outside the guide levee after 2 d later (1×10-9g Ir isoline). The downward distance is much larger than the one upward. The tracer particles migrate to the southeast out of the guide levee, with the outer channel extend to the east.

3.3.3 Dynamics interpretation of tracer particle horizontal movement

As is indicated by the simulation results, the water channels within the Yangtze River Estuary sand bar are affected by the runoff and river course binding, the suspended particles migrate to the open sea basically along the middle axis of river course in the form of reciprocating oscillation, with the migration and diffusion directions almost parallel to the navigation channel in guide levee, similar to long belt distribution, and that the tracer particles migrate to the southeast out of the guide levee. According to the simulation results, the suspended particles injected at different positions represent different migration tracks. For example, the injection results at A, B and C show that the tracer particle migration tracks are basically stable, with small reciprocating oscillation near the groin and limited amount entering and crossing the navigation channel. However, the particle migration tracks at D, E and F are much more complex, especially at E and F, that is, the particles not only oscillate to and fro near the groin, but also have a great amount crossing passage. This is due to the passage returning function of the current in deepwaternavigation channel in the Yangtze River Estuary and the existence of a turning cross section between S6and S7groins. Above the cross section, the included angle of current direction and the navigation channel is positive (taking the navigation channel trend as the baseline, ebb tide current direction in the counter-clockwise, the angle is positive, otherwise negative). The current particles on the southern side of the navigation cross the navigation channel, the sediment is easy to enter the navigation channel. While the included angle of current direction and the navigation channel is negative below the cross section, the current particles on the southern side of the navigation channel is far from the navigation channel, and it is difficult for the sediment to enter the navigation channel. The northern side is the opposite. Moreover, the suspended particles on either the southern or northern side of the navigation channel oscillate to and fro near the groin, with passage crossing amount to different degrees. While below the cross section, the suspended particle migration is almost consistent, but because the current mainly moves to the southeast, the suspended particles in different layers on the northern side of the navigation channel also have passage crossing to different degrees from the injection result at D.

As a result of strong current dynamics in the navigation channel in the Yangtze River Estuary, the longitudinal transport speed of the suspended particles is very high and the migration leading peak of particles injected at A, B and C has been out of the guide levee downwards two days later, while because the particles at D, E and F mainly oscillate to and fro near the groin, there is also a considerable part with downward displacement. As the northern passage is ebb channel preferential flow in flood season, the downward migration distance is evidently longer than upward migration distance. After entering the open sea, the particle traces at each point exhibit the form of clockwise rotation. Except for the bottom layer, the particle migration tracks are almost similar in the other layers, all migrating downward in rotary oscillation in the shape of long belt. While due to the terrain impact in the bottom layer, the migration scope of the particle migration tracks is smaller compared with the other layers with short downward migration distance without long belt distribution.

Under the current action, the suspended particles transport horizontally in addition to longitudinal migration along the tidal current. Above the turning cross section of S6- S7groins and at D, E and F on the northern side of the navigation channel, the suspended particles have relatively great horizontal transport and enter and cross the navigation channel. While below the turning cross section, the horizontal transport of the suspended particles at A, B and C on the southern side of the navigation channel is relatively small. The horizontal transport has the following characteristics. First, the horizontal transport intensity is very weak compared with the longitudinal transport, mainly the reciprocating oscillation caused by the reciprocating movement of the current in the Yangtze River Estuary, second, the horizontal transport is more evenly to both sides, third, it is relatively stable, i.e., the horizontal transport basically remains near the migration tracks and the distance of peak value horizontally distributed on the cross section from the navigation channel almost remain stable.

3.3.4 Vertical movement of tracer suspended particles

Figures 28-33 show the vertical transport processes of fine silt particles simulated by tracers. The abscissa T in the figures is the sampling time and the ordinate H is the simulation water depth calculating up from the seabed, with meter as the unit. The 0.2 layers in the figures indicate the water layers0.2H above the sea surface, with 0.2 being the relative depth. It is obvious that the vertical space coordinates change constantly. The rest relative water depth has the similar meaning.

As is shown in Figs.28-33 the tracer particles were injected at 7:00 on September 22 when flood tide just began in the open sea, the tracer particles at all the injection points have different upward displacements, and at the same time, the particles at all the layers changed in vertical displacement as well, and here the most remarkable is the particles in the surface layer. As the water level increases due to the flood tide, the vertical displacement gradually increases, the water level in the water layers where the particles are located increases constantly, reaching the maximum at flood peak, then begin to reduce; at flood slack, the vertical displacement reduces to the minimum, the water level in the water layers where the particles are located reduces gradually, and ebb tide begins in the open sea simultaneously, then the particles at the injection points begin to change periodically in vertical direction as at the flood tide. Except for E and F, the tracer particles at all layers basically fluctuate within the vicinity of the layer, without the occurrence of vertical mixing, while E and F are above the turning cross section between S6and S7, the included angle of current direction and the navigation channel is positive (taking the navigation channel trend as the baseline, ebb tide current direction in the counter-clockwise, the angle is positive, otherwise negative), easy to enter the navigation channel, so the particles often oscillate to and fro near the groin, easy to have vertical mixing. Because the tracer particles on the surface layer injected at E mainly oscillate to and fro near the groin (see Fig.32), the surface and 0.2 layers change acutely. In addition, the particles released at D and F also have reciprocating oscillation, so in Figs.31-33, the particles between the layers have certain vertical displacement. At the same time, the figures also indicate that the water level of the particles in the surface layer at the points increases at a certain level 2 d later, which is mainly due to the high longitudinal transport speed of suspended particles, the migration leading peak of most particles having been out of the guide levee downwards and entered the open sea two days later and the increased water depth. It can be found from the figures that the vertical movement range of the particles at the lower layer at injection points is small, representing the characteristics of laminar motion.

4. Conclusions

Based on the EFDC, a three-dimensional baroclinic model for lower reaches of the Yangtze River, Yangtze River Estuary, Hangzhou Bay and the adjacent sea areas has been established. The simulated water level and current results are fairly consistent with observations. With the calibrated threedimensional numerical model, the migration tracks of the tracers are simulated and the migration rule of the fine silt particles of dredged sediment after dumping in the deepwater navigation channel are studied and analyzed.

The study just analyzes the migration behavior of the movement of fine silt particles in dumped silt by the tracer technology. In fact, the composition in dumped silt is diversiform, the study of dumped movement need numerical model of total sediment under multiple hydrodynamic forces such as waves and currents. But the mechanism and complicated dynamic interaction process of dumped silt are still not too explicit. This is the foundation of numeric simulation of dumped silt.

Acknowledgement

This working was supported by the Startup Fund of Hohai University (Grant No. 40801107).

[1] YING Xin-ya, AKIYAMA J. and MAO Gen-hai. Modeling of motion of particle clouds formed by dumping dredged material[J]. China Ocean Engineering, 2001, 15(1): 85-96.

[2] KONG Ling-shuang, CAO Zu-de and JIAO Gui-ying. The bottom shear stress and sediment movement for a wave-current coexisting system[J]. Journal of Hydrodynamics, Ser. A, 2003, 18(1): 93-97(in Chinese).

[3] BAO Wei-min, ZHANG Shu-ming and QU Si-min. Total sediment density current differential equation: I-Theory[J]. Journal of Hydrodynamics, Ser. A, 2005, 20(4): 497-500(in Chinese).

[4] LU Yong-jun, LI Hao-lin and DONG Zhuang et al. Two-dimensional tidal current and sediment mathematical model for Oujiang Estuary and Wenzhou Bay[J]. China Ocean Engineering, 2002, 16(1): 107-122.

[5] QI Ding-man, WU Hua-lin and WAN Yuan-yang et al. Application of tracer sediment in studying sediment movement at north passage of Yangtze Estuary[J]. The Ocean Engineering, 2010, 28(1): 64-69(in Chinese).

[6] CRONE G. C., DINAR N. and Van DOP H. et al. A Lagrangian approach for modelling turbulent transport and chemistry[J]. Atmospheric Environment, 1999, 33(29): 4919-4934.

[7] PERIANEZA R., ELLIOTT A. J. A particle-tracking method for simulating the dispersion of non-conservative radionuclides in coastal waters[J]. Journal of Environmental Radioactivity, 2002, 58(1): 13-23.

[8] ZHOU Wei, WANG Tao and WU Yi-ming. Application of tracer simulation to study on the water-flooding front of edge water reservoir: Example from Hade Oil Fields[J]. Journal of Jilin University (Earth Science Edition), 2010, 40(3): 549-556(in Chinese).

[9] ZHANG Qian-fei, LAN Shou-qi and WANG Yan-ming et al. A new numerical method for groundwater flow and solute transport using velocity field[J]. Journal of Hydrodynamics, 2008, 20(3): 356-364.

[10] LI Bin, ZHANG Kun and ZHONG Bao-chang. An experimental study on release of pollutants from sediment under hydrodynamic conditions[J]. Chinese Journal of Hydrodynamics, 2008, 23(2): 126-132(in Chinese).

[11] LI Ning, MAO Zhi-hua and ZHANG Qing-he. Numerical simulation of pollutant transport and accumulation areas in the Hangzhou Bay[J]. Transactions Tianjin University, 2009, 15(6): 400-407.

[12] WU Zhao-chun, WANG Dao-zeng. Simulation of the oil slick movement in tidal waterways[J]. Journal of Hydrodynamics, 2010, 22(1): 96-102.

[13] FRYXELL G. A. Survival strategies of the algae[M]. Cambridge, UK: Cambridge University Press, 1983, 69-136.

[14] CHEN Xiu-hua, ZHU Liang-sheng and ZHANG Hong-sheng. Numerical simulation of summer circulation in the East China Sea and TTS application in estimating the sources of red tides in the Yangtze River Estuary and adjacent sea areas[J]. Journal of Hydrodynamics, Ser. B, 2007, 19(3): 272-281.

[15] ZOU R., CARTER S. and SHOEMAKER L. et al. An integrated hydrodynamic and water quality modeling system to support nutrient TMDL development for Wissahickon Creek[J]. Journal of Environmental Engineering, 2006,132(4): 555-566.

[16] ZUO Shu-hua. Introduction of numerical simulation software Dleft3D and its application on offshore area of Aojiang Estuarine[J]. Journal of China Hydrology, 2007, 27(6): 55-58(in Chinese).

10.1016/S1001-6058(09)60114-1

* Project supported by the Public Welfare Special Scientific Research Project of China Ministry of Water Resources (Grant No. 200701026), the National Natural Science Foundation of China (Grant No. 50709007).

Biography: XIE Rui (1985-), Male, Master

WU De-an,

E-mail: wudeian@163.com