Morphodynamic Characteristics and Medium-Term Simulation of the North—South Passage Under the Impact of the Yangtze Estuary Deepwater Navigation Channel Project

The morphological evolution characteristics of the North—South Passage area since the construction of the Yangtze Estuary Deepwater Navigation Channel Project (DNCP) are analyzed on the basis of the measured data. A two-dimensional morphodynamics numerical model of the Yangtze Estuary is established to verify the morphological evolution of the North—South Passage under the influence of the DNCP and to predict the future evolution in the next 40 years. Data analysis shows that the North Passage has experienced rapid adjustment stages and adaptive stages after the construction of the DNCP. Slow erosion occurred along the main channel, and slow siltation could be observed in the area between the groins. The South Passage showed a state of upper section erosion and down section deposition. At present, the whole South Passage is in a slight erosion state. According to the numerical model, the eroding and silting speed of the North Passage will slow down in the future. The present state that erosion occurs in the main channel and siltation occurs between the groins will continue. The South Passage will still maintain upper section erosion and down section deposition in the future. Due to the main channel erosion of the North Passage and siltation of the South Passage, the sediment division ratio of the North Passage will increase in the future but still be smaller than 50%. After morphological evolution of 40 years, the direction of residual sediment transport caused by M2 and M4 tidal components in the North Passage has not changed, but the transport rate will decrease. It is considered that the morphological evolution of the North—South Passage could reach a relatively stable state after 40 years.


Introduction
The tidal estuary is located at the intersection of river and ocean. The estuarine morphology has important social and economic values for human life as well as ecological environment in the estuarine delta (Syvitski and Saito, 2007). At present, the research methods of morphodynamics include field measurement (Ding and Wang, 2008), physical model (Grasso, 2009), numerical model (de Vriend, 1991) and so on. Although field measurement is a traditional way, it cannot be used for predicting the future morphological evolution (Prandle, 2009). As computer performance greatly improved and mathematical model simula-tion technology developed, it is possible to predict the morphological evolution in medium-and long-time scale (Lesser, 2009). For example, van der Wegen et al. (2011) did a great deal of work on the long-term morphodynamics of the Western Scheldt Estuary in the Netherlands. The Delft3D software is used to simulate the long-period evolution of the estuary. It is found that by considering only non-cohesive sediment and the morphological tide, the numerical model could obtain the shoal−channel interaction pattern which is similar to the measured topography.
In recent decades, with the economic and social development and population growth, channel regulation projects such as dike, groins, dredging etc., have been carried out on many estuaries around the world in order to meet the navigation needs. The large-scale regulation project directly leads to the change of the hydrodynamics and sediment transport in the estuary. For instance, the training wall 'Kugelbake' of Elbe Estuary changed the tidal velocity and direction (Plüß and Milbradt, 2014). The change of hydrodynamic and sediment conditions caused by the project further changed the morphological evolution characteristics of the estuary, for example the regulation project outside the Mersey Estuary and the channel dredging project made the estuary go through three stages of siltation-slight erosion-stability within 70 years. Dastgheib (2012) simulated the changes of hydrodynamics and morphodynamics before and after the construction of the Afsluitdijk (Waden Sea Dam) in the Netherlands, and the results showed a balance between tidal forces and geomorphology.
The Yangtze River is the largest river in China, which is affected by the strong interaction of tidal current and runoff. There is a mouth bar at the mouth of the Yangtze Estuary, which hinders the development of the shipping of the Yangtze River (Chen et al., 2001). In order to meet the navigation demand, the Deepwater Navigation Channel Project (DNCP) was carried out in three phases in the North Passage of the Yangtze Estuary. Owing to the massive spatial scale of the Yangtze Estuary, the researchers are mostly concentrated in short term bed level change of certain areas, such as siltation of the Deepwater Channel (Li et al., 2018), Jiuduan Shoal (Hu et al., 2009), South Branch  and tidal flats (Kuang et al., 2013). In terms of medium to long term morphological change, the researchers mainly focused on the morphodynamics in short time scale. Pan et al. (2012) analyzed the morphological evolution of the North Passage under the influence of the project by using the measured topography from 1998 to 2008. And the morphological evolution of the North Passage was simulated by mathematical model. Guo et al. (2015) simulated the evolution of the Yangtze Estuary on the millennium scale by the Delft3D mathematical model, and put forward the controlling effect of different dynamics on the morphology of the Yangtze Estuary. Luan et al. (2017) predicted the morphological change of the Yangtze Estuary for 20 years by using the mathematical model, and they pointed out that the role of cohesive sediment in the long-term evolution of the Yangtze Estuary should not be ignored. Most of these studies start from the model itself and analyze the morphodynamic mechanism of the Yangtze Estuary. However, insufficient consideration is given to the role of the project.
It has been more than nine years since the completion of the Yangtze Estuary DNCP. There is no clear conclusion on the future morphological evolution trend of the North− South Passage and whether it will reach equilibrium under the existing conditions.
In the present study, the Delft3D numerical model is used to establish a two-dimensional morphodynamic model. The model results are compared with the measured topographical changes in the North−South Passage area before and after the construction of the Yangtze Estuary DNCP (1998−2016). This study finds out the future development trend of the North−South Passage under the influence of the training project, and predicts the morphological evolution of the South−North Passage of the Yangtze Estuary in the next 40 years.

Yangtze Estuary
With a total length of 6300 km, the Yangtze River is the third longest river in the world. Abundant runoff carries a large amount of sediment into the estuary, shaping a wide estuarine delta under the interaction of runoff and tide. River below the Xuliujing Station is the estuary section, which is divided into South and North Branches by the Chongming Island. The South Branch is divided into South Channel and North Channel by the Zhongyang Shoal and the Changxing Island. And the South Channel is divided into South and North Passages by the Jiuduan Shoal. Therefore, the Yangtze Estuary shows a river regime pattern of "three-stage bifurcation and four outlets entering the sea" (Fig. 1).
The discharge and sediment concentration of Datong Station in the lower reaches of the Yangtze River (about 531 km away from Xuliujing) basically represent the characteristics of water and sediment in the Yangtze Estuary. According to the hydrological data from 1950 to 2016, the inflow in the upper reaches of the Yangtze Estuary shows no obvious change trend. The annual runoff is more than 8.00×10 8 m 3 , and the average discharge is basically maintained at about 30000 m 3 /s.
There is an obvious distinction between the flood season and the dry season in the Yangtze Estuary. The discharge in the flood season is much larger than that in the dry season. Since the impoundment of the Three Gorges Project in 2003, it has played a certain role in regulating the monthly runoff of the Yangtze River. However, the flood and dry seasons can still be clearly distinguished, and the monthly runoff varies largely in different years, especially in flood seasons. Since 1950, the measured sediment transport amount from Datong Station has experienced two significant decreases, i.e. in the 1980s and after the impoundment of the Three Gorges Project in 2003. The average sediment transport amount at Datong Station decreased by 57.9% compared with the multi-year average before water impoundment (1985−2003). The Yangtze Estuary is a medium tidal dominant estuary with irregular semi-diurnal tide. According to the statistics of the tide level of Lvhuashan Station in 2003 and 2016, the multi-year average tidal range is 2.63 m, the average high tide level is 3.39 m, and the average low tide level is 1.88 m.
Owing to the interaction of runoff and tidal currents, mouth bar often exists outside the estuary. In order to construct the deepwater channel, the Yangtze Estuary DNCP has been implemented in the North Passage in three phases since 1998. The project is composed of diversion port project, south and north training wall project, spur dike group and dredging project (Fig. 2). The lengths of the south and north dikes are 48.08 km and 49.2 km, respectively. The spur dikes are arranged vertically with the training wall. Nine spur dikes are arranged on the north side of the south training wall and 10 spur dikes are arranged on the south side of the north training wall. The silt reduction project was carried out after the end of the third phase of the project, including lengthening part of the spur dike, heightening the training wall and constructing the sand retaining dike.
This research collects the bathymetry of the South− North Passage of the Yangtze Estuary in 1998, 2002, 2011 and 2016, which correspond to the topography before the project, after the completion of its first phase, after the end of its third phase, and after the implementation of the silt reduction project, respectively. Therefore, figures of 1998− 2002, 2002−2011 and 2011−2016 respectively correspond to the morphological changes during the construction period of the first phase, the second to the third phase, and after the third phase of the project.
2.2 Analysis of the morphological evolution of the North− South Passage since DNCP After the completion of the Yangtze Estuary DNCP, the diversion port of the South and North Passages are basically controlled. The morphological evolution of the North Passage is controlled by guide dikes and spur dikes instead of the natural evolution. The morphological changes in the North−South Passage area of the Yangtze Estuary from 1998 to 2016 are shown in Fig. 3.
After the construction of the Yangtze Estuary DNCP, the evolution characteristics of the North−South Passage are analyzed in this section.

Morphological characteristics of the North Passage
(1) Rapid morphology adjustment stage When the DNCP started to construct in 1998, the North Passage was affected by the regulation project and was in  the stage of rapid adjustment. In terms of the upper section of the North Passage, the main channel is eroded and the area between the groins is deposited. The siltation on the north side is much larger than that on the south side. In terms of the lower part, the water depth changes in stages. At the beginning of construction, the sediment washed down from the upper section of the North Passage formed a northwest−southeast sedimentation zone at the original deep trough in the down section. After the completion of the project, with the help of the dredging project, the downsection of the North Channel is gradually deepened. At the same time, the area between the groins on both sides was silted up. The deposition in the south side is larger than that in the north side.
(2) Self-adaptive stage After the third phase of the project, the North Passage is in the post-project adaptive stage from 2011 to 2016. The change of erosion and deposition is obviously weaker than before. In the past 5 years, there has been a small amount of siltation in the area between the groins and a small amount of erosion in the main channel. Generally speaking, while maintaining a shoal-channel stability, the North Passage shows the changing characteristics of slow erosion and volume expansion in the main channel, and slow siltation and volume decrease between the groins.

Morphological characteristics of the South Passage
During the construction of the first phase of the DNCP from 1998 to 2002, the closure of the North Jiangya Channel and the construction of the south training wall had a great influence on the south channel.
Siltation occurred at the entrance of the main channel of the South Passage. The upper section continues to scour deep, and the scouring depth is about 3−5 m. The sediment in the upper section of the South Passage moves downwards, resulting in deposition in the middle and lower section. The alluvial thickness of Nanhui Flat is 0−1 m. The scouring range of the north side is larger than that of the south side. After the first phase of the project, the South Passage mainly shows the characteristics of erosion in upper section and deposition in down section, and the averaged erosion and deposition thicknesses are smaller than those of the North Passage. After the third phase of the project, the whole South Passage is in a mild erosion state.

Numerical model
The Delft3D model has been applied, which mainly involves the main modules such as hydrodynamics, sediment transport, morphology and so on. The FLOW module is the foundation of the whole model system, which includes the hydrodynamics, sediment transport and morphology. The basic equation is the hydrostatic pressure approximate difference and shallow water Navier−Stokes equation. The finite difference ADI method is used to discretize the equa-tion, which has the advantages of high stability and accuracy.
The calculation area of the model covers the whole Yangtze River Estuary, Hangzhou Bay and its adjacent sea area. The model grid is orthonormal grid, with 1431 horizontal grids and 163 vertical grids, which is shown in Fig. 4. Upstream boundary is from Datong hydrologic station. The eastern boundary of the out sea is outside the contour line of −50 m, and the farthest is located at 124.24°E. The northern boundary is located at 34.67°N. The farthest part of the southern boundary is 29.33°N latitude.
The grid size in the out sea is relatively large, up to 2 km×2 km. The grids in Yangtze Estuary are locally refined. The minimum grid size is 70 m×60 m. The coordinate is Gauss−Kluc coordinates. The training walls and spur dikes on both sides of the deep channel in the North Passage are schematized by Current Deflection Wall (CDW) in the model that the flow can cross the dike when the water level exceeds the specified elevation. According to the principle of Courant number, the time step is 60 s.

Schematization of boundary conditions
The schematized astronomic tides are used as the outsea boundary conditions. The contribution of different tidal components and their combinations to tidal residual sediment transport at the observation points along the North Passage is analyzed.
The results show that the interaction of M2, S2, K1 and O1 tidal components and their interaction with runoff make the greatest contribution to sediment residual transport. So these four tidal components are adopted as the open sea boundary conditions.
In order to determine the upstream boundary conditions of the numerical model, three typical annual runoff processes of flood season, middle season and dry season are selected alternately through the sensitivity test. And the discharges are 37900 m 3 /s, 23800 m 3 /s and 15900 m 3 /s, respectively. JIAO Jian et al. China Ocean Eng., 2020, Vol. 34, No. 2, P. 198-209 201 3.2 Selection of the sediment parameters Owing to the complex sediment fractions in the Yangtze Estuary, uniform sediment cannot represent the actual situation (Feng et al., 2019). A variety of sediment components are considered in this model, including cohesive sediment and non-cohesive sediment. According to the measured data, three fractions of non-cohesive sediment and one fraction of cohesive sediment are taken into account. The particle sizes of non-cohesive sediment are 70 μm, 100 μm and 300 μm, respectively. The initial sediment particle size distribution on the bed surface is obtained by Bed Composition Generation (BCG) method (ven der Wegen, 2012) method. The steps are as follows: firstly, the proportion of sediment with different particle sizes on the bed surface is given according to the measured values. Then the simulation calculation is carried out in the model without morphological change. The particle size is redistributed and reaches a stable state after a period of simulation. In this model, the simulation time is 2 months. Finally, the percentage of each component is obtained. The results of sediment percentage are applied to the model as the initial bed sediment particle size distribution.
The density of non-cohesive sediment is 1600 kg/m 3 . The dry density of cohesive sediment is 500 kg/m 3 . The erosion rate is 5×10 −5 kg/(m 2 •s). The sediment settling velocity is 0.25 mm/s. The critical scour shear stress is 0.7 N/m 2 according to sensitivity analysis. The critical shear stress for sedimentation is not set in the model, which means that the sediment particles will settle at any time . The flocculation effect of salinity on cohesive sediment is considered in the model. According to Mietta et al.'s (2009) experimental study, a simple flocculation model is adopted in the model. It is considered that the sediment flocculates and settles when the salinity of the water exceeds 8 psu, and the settling rate increases to 0.5 mm/s. The salinity of the outer sea boundary is set to 35 psu, and the salinity of the upstream boundary is 0. The salinity field which has been calculated by the model in Wang et al. (2019) for 60 days till it reaches stable is taken as the initial salinity field in this model.

Morphological accelerate factor
As the current computer operation speed is limited, the morphological accelerate factor (MORFAC) is introduced in order to reduce the calculation time. The result of erosion and deposition of each hydrodynamic time step is multi-plied by the MORFAC, which is then taken as the change of erosion and deposition of the accelerated model. The updated bathymetry will be taken as the bathymetry for the next calculation step. In the process of simulation, the total time of the hydrodynamic process multiplied by the acceleration coefficient is the total time of the morphological evolution process, so as to realize the acceleration of the longterm morphological change simulation. This method makes it possible to mathematically simulate the morphological evolution for decades or even centuries within an acceptable period.
The MORFAC should be selected by sensitivity analysis. In this study, scenarios with acceleration factors of 1 (no acceleration), 5, 10, 20, 30, 50 and 100 are set and the morphological time is 5 years. The root mean square error (RMSE) analysis of erosion/sedimentation thickness is carried out between the results of the accelerated model and the unaccelerated model by each grid unit. The results show that the morphology RMSE increases with the increase of acceleration factor. Especially in the initial stage of simulation, the maximum error of single step is 0.5 m, which is considered unacceptable (Lesser, 2009). Therefore, the acceleration factor in this model should be smaller than 30.
At the same time, due to the impact of the project, the morphological evolution is fast in the early stage and then gradually slowing down after it reaches an equilibrium state. According to the sensitivity test, the error caused by single step after the model acceleration is positively related to the morphological evolution rate within the single step. Therefore, the time-varying acceleration factor method could be used to reduce the error caused by the acceleration factor in the simulation.
The settings are shown in Table 1. The MORFAC is varying from 1 to 30. The total hydrodynamic simulation time is 31.4 months.

Hydrodynamics and sediment transport validation
The hydrodynamics and sediment concentration of the model are verified by the hydrological and sediment transport data of flood season (August 2011) and dry season (February 2013). The observation points are located in the South Channel and North Passage (Fig. 5). Due to space limitation, this paper only shows the verification and comparison of the tidal level of Hengsha Station, the flow velocity, flow direction and sediment concentration of Point  Eng., 2020, Vol. 34, No. 2, P. 198-209 CS0 (Fig. 6).
The root mean square error (RMSE), correlation coefficient (CC) and skill score (SS) are used to analyze the accuracy of the model. The formulas are as follows: 1/2 ; (2) where X is a statistical analysis variable; the subscripts mod and obs represent the calculated and observed values of the model, respectively; the superscript horizontal line represents the time average. SS=1 indicates that the result of the model is perfect and SS>0.5 indicates that the model is good. Statistical analysis shows that the maximum RMSE of five tidal level stations in dry season and 14 tide stations in flood season is 0.21 m with the average value of 0.19 m, and the correlation coefficient is larger than 95% with the average value of 98.3%. The average SS also reaches 0.932. It shows that the tidal level simulation has good accuracy.
The correlation coefficient of flow velocity exceeds 85%. The correlation coefficients of flow direction in August 2011 and February 2013 are 93.5% and 82.8%, respectively. The root mean square error range of velocity is 0.07− 0.26 m/s, and their average values of the two months are 0.21 m/s and 0.16 m/s, respectively. The root mean square error, SSC and correlation coefficient of sediment concentration are good enough for the simulation, indicating that the simulation is ideal.

Morphological change validation
The schematized boundary conditions are applied in the model. The parameter settings are consistent with the validation model. The topography from 2002 to 2011 is used to validate the bed level change in the South and North Passage area of the Yangtze Estuary. The simulation results of the model are shown in Fig. 7a and the measured erosion and deposition distribution are shown in Fig. 7b.
The upper, middle and lower sections of the North Passage and the South Passage are selected (Fig. 8) to calculate the measured and simulated erosion and sedimentation volume from 2002 to 2011 (Fig. 9), respectively.
Compared with the measured morphological evolution, it can be seen that the calculated siltation thickness of the area between the groins in the North Passage Deepwater Navigation Channel Project is smaller than the measured value. But the scouring depth is larger than the measured value.
In terms of different sections, the measured net erosion and deposition in the upper and middle sections of the North Passage is positive, which means the amount of sedimentation is larger than the amount of erosion. However, the simulation result is negative, and the specific reason will be discussed later. The calculated sedimentation amount in the down section of the North Passage is almost the same as the measured value. The amount of scour is slightly larger than the measured value.   China Ocean Eng., 2020, Vol. 34, No. 2, P. 198-209 203 The calculated erosion amount in the upper section of the South Passage is larger than the measured value, and the calculated deposition amount is almost the same as the measured value. The calculated erosion and deposition amount in the middle section of the South Passage are al-most the same as the measured value. The calculated erosion amount in the down section of the southern trough is slightly smaller than the measured value. The amount of deposition is larger than the measured value.
In general, the model could well simulate the siltation in the area between the groins of the DNCP and the erosion and deposition characteristics of the North Passage. For the South Passage, the characteristics of erosion in the upper section and siltation in the down section in the measured data are also showed in the simulation results. At the same time, the downward siltation of South Jiangya Shoal and the siltation of East Nanhui Flat could be observed in the model results.  sage of the Yangtze Estuary in the future 40 years (2016−2056) are predicted by the numerical model. And the erosion and deposition pattern is shown in Fig. 10. The patterns with simulation time of 5 years, 10 years, 20 years, 30 years and 40 years are given respectively. In general, the morphological evolution speed of the North Passage will slow down in the future, and the erosion and deposition distribution will maintain the present state. The South Passage will show upper erosion and down deposition characteristics in the future. And the evolution range of erosion and deposition in the North Jiangya Channel and Jiuduan Shoal area is relatively large.

North Passage
According to the erosion and deposition distribution pattern of different time, erosion could be observed at the turn of the North Passage Deepwater Channel in 5 years. The erosion depth is about 1−2 m. There is erosion at the end of the north training wall. The erosion and deposition depth in 10 years is larger than that in 5 years. Siltation occurred in the area between the groins on the north side with a maximum siltation depth of 2 m. Erosion in the main channel and the erosion depth of the down section are obvious and larger than those of the upper section. The scouring depth in the north side is greater than that in the south side.
The trend of erosion and deposition in 20 years is basically the same as that in 10 years, with the erosion depth of the main channel increased, and the maximum being about 4 m. The area between the groins on the north side continues to silt. The south side area between the groins also has siltation, but the siltation thickness is less than that of the north side. Different from the distribution of erosion and deposition in the past 10 years, the entrance section of the North Passage shows a scouring trend. There is no significant change of the erosion/deposition pattern between 20 and 30 years. But the erosion and deposition thickness has increased. The erosion depth in the main channel of the North Passage becomes uniform, most of which is about 5 m. The siltation intensity continues to increase, and the maximum siltation thickness still occurs in the south area between the groins, up to 6 m. The erosion and deposition pattern and amount of the North Passage in 40 years are almost the same as those in 30 years, indicating that the North Passage will reach a relative equilibrium state during the period of 30 to 40 years.
It is worth noting that siltation can be observed at the sea side of the double training wall in the North Passage from the 10-year simulation results. The sediment mainly comes from upstream and local erosion from the North Passage. After 40-year simulation, the maximum siltation thickness could reach 4 m, which may block the waterway. In addition, the local erosion at the tail of the south training wall may lead to the local instability of the structure, which JIAO Jian et al. China Ocean Eng., 2020, Vol. 34, No. 2, P. 198-209 205 should be concerned in the later maintenance.

South Passage
The 5-year simulation results of the South Passage show that the erosion occurs at the entrance and the upper section of the South Passage, with no obvious change in other areas. The 10-year simulation results show that the erosion depth of the upper section of the South Passage increases slightly. The erosion zone extends to downstream. The deposition occurs in the lower part of the South Passage, and the siltation thickness is 1−2 m. There is siltation in the North Jiangya Channel and erosion on the south side of Jiuduan Shoal. After 20 years, the erosion and deposition pattern in the South Passage will still be similar to that of 10 years. However, the intensity of erosion and deposition is obviously higher than that of 10 years ago. The erosion and deposition distribution of the South Passage after 30 years is almost the same as that after 40 years, and the erosion in the upper section of the South Passage and the south side of Jiuduan Shoal have formed a scouring zone. But there is no obvious change in the siltation intensity in the lower part of the South Passage.
In terms of the shoal near the South Passage, simulation shows that there is no significant development of the trench on the head of South Jiangya Shoal. The North Jiangya Channel is not obviously eroded deep. However, the erosion on the south side of Jiuduan Shoal is connected with the North Jiangya Channel. And the south channel forms a new branch, which is disadvantageous to maintaining the channel. At the same time, there is deposition at the tail of Jiuduan Shoal whose main body tends to move slowly to the southeast. It will also affect the stability of the South Passage, which needs to be paid attention to.

Future volume change of the South−North Passage
Statistics are collected on the relationship between the time-varying amount of erosion and deposition in the North−South Passage. The influence of the regulation project of the North−South Passage in the future is analyzed.
The time variation curve of erosion and siltation in the North−South Passage is shown in Fig. 11. The amount of siltation and erosion in the North Passage both increase gradually in the first 30 years, but the growth rate of erosion is slightly larger than that of siltation. Therefore, the overall erosion and siltation volume of the North Passage is negative in the first 30 years. After 30 years, erosion and deposition amount tends to be stable, and the total amount is about −2.95×10 8 m 3 after stabilization.
According to the simulation period, the deposition amount in the South Passage is larger than the erosion amount. Therefore, the net erosion and deposition volume is positive, which means that the South Passage as a whole is in a state of siltation. The changing trend of the South Pas-sage is the same as that of the North Passage, and the erosion/siltation amount of the South Passage increases obviously in the first 30 years. There is almost no change in the amount of erosion and siltation during the period of 30−40 years. It indicates that the South Passage reached a relatively stable state within 30 to 40 years. According to the simulation results, the deposition volume of the South Passage in the future 40 years is 3.80×10 8 m 3 .

Discussion
5.1 Causes for the erosion amount larger than the measured value of the North Passage in the validation model The simulated North Passage scouring amount in the verification model is larger than the measured value. After analysis, it is mainly caused by two reasons.
First of all, because the mathematical model cannot perform real-time simulation of the project under construction, the calculated amount of erosion in the main channel of the North Passage is greater than the measured value. According to the previous research and analysis of the measured data, during the first and second phases (1998−2004), the volume change of the North Passage was not obvious and decreased slightly. During 2004 to 2010, i.e. the period from the end of the second phase to the third phase of the project, the main channel volume of the North Passage decreased significantly, which was resulted from the large amount of siltation between the groins but just mild erosion in the main channel. After the completion of the DNCP in 2010, the sediment volume of the North Passage increased. This is mainly due to the fact that during the construction of the DNCP of the North Passage, the double dikes and groins have enhanced the sediment transport in the main channel of the North Passage and the entrance of the South Passage, resulting in enhanced erosion. At the same time, during the construction and maintenance dredging, the dredged soils were thrown into the area between the groins, which also caused the increase of the silt amount in the area between the groins. The simulation time of the numerical model is from 2002 to 2011, which corresponds to the second and third phases of the DNCP construction. During this time, the construction of double dike and the spur dike group cannot be reflected in the model.
Since the DNCP project has been added at the beginning of the simulation, erosion was thereby intensified, making the calculated value of the erosion amount of the North Passage larger than the measured value.
Another reason is the use of generalized upstream runoff. Although the annual average flow is consistent with the actual situation, the morphological evolution under extreme runoff conditions, such as flood, cannot be accurately simulated. In the mouth sand area where the North Passage is located, siltation occurs in the flood season and erosion occurs in the dry season. Without considering the effects of a large flood in the simulation, the simulated erosion could be severer than the measured value.

Reasons for future morphological evolution to stabilize
It is controversial whether the interaction and feedback between hydrodynamics and morphology in the process of long-term morphodynamic process will last forever, or there will be a stable state. However, according to the current theoretical studies and observations, the landform will eventually reach a stable state under certain hydrodynamics and sediment transport conditions.
After the construction of the Yangtze Estuary DNCP, the morphological balance state and balance time of the South−North Passage of the Yangtze Estuary attract worldwide concentration and research. The morphological equilibrium can be judged by the split ratio and the amount of erosion and deposition in a certain area. It can be considered that the morphology has reached equilibrium when the split ratio and the erosion and siltation amount in the study area show a stable fluctuation state or non-significant change.

Split and sediment diversion ratio prediction
The time-varying ebb tide split ratio of the North−South Passage is shown in Fig. 12. Under the initial bathymetry (2016), the simulated ebb tide split ratio of the North Passage is about 42.5%. With the increase of simulation time, the split ratio of the North Passage increases slightly. The ebb tide split ratio in the North Passage climbs to about 48% by the thirtieth year. During the period of futhure 30−40 years, there is no obvious change of the split ratio. After 40 years, the ebb tide split ratio in the South Passage will decrease from 57.5% to 51.9%.
The variation of ebb tide sediment diversion ratio with time in the North−South Passage is shown in Fig. 13. At present, the sediment diversion ratio of ebb tide in the North Passage is 35.3%. It rises to 42.3% after 40 years in simulation.
Therefore, in the future, the ebb tide split ratio and sediment diversion ratio of the North Passage will continue to be smaller than those of the South Passage and tend to be stable.

Tidal asymmetry changes in the North Passage
The tidal asymmetry and the non-linear effects of runoff and tide directly affect the residual sediment transport,  which determines the long-term morphological evolution in estuaries. Therefore, the tidal asymmetries of the current situation and 40-year-morphological evolution are compared.
The M2 tide is the main tidal component in the Yangtze Estuary. The asymmetric tide wave method proposed by van de Kreeke and Robaczewska (1993) is used to analyze the tidal asymmetry of the North Passage.
Taking the entrance of the North Passage as the starting point of observation and the offshore area downstream of the training dike as the end point, a total of 12 observation points are set up, which are arranged according to the starting distance. The change of tidal wave characteristics can be seen in Fig. 13. It can be seen that under the current bathymetry, the M2 tidal level amplitude a M2 and the M2 tidal current amplitude v M2 decrease from the open sea to the upstream in the project section. a M4 and v M4 increase from the open sea to the upstream in the project section. After 40year morphodynamics simulation, a M2 and v M2 both increase. This is because the main channel of the North Passage is eroded during simulation, which reduces the resistance along the channel and facilitates the spread of the M2 tidal component to the upstream. a M4 and v M4 will decrease after 40 years, because less M2 tidal components are converted to M4 ones due to non-linear effects such as friction.
The tidal wave deformation characteristics are reflected by a M4 /a M2 and v M4 /v M2 . It can be seen that a M2 and v M2 increase, and a M4 and v M4 decrease, so the deformation of the tide level and current decreases. At the same time, it can be seen from the phase difference between the M2 and M4 tidal components that the interaction of the M2 and M4 tidal components causes the dominant tide flow, and this state has not changed after 40 years of morphological change. Therefore, the residual sediment transport direction caused by the M2 and M4 tidal effects has not changed, but the transport rate has decreased after 40-year morphological evolution.

Conclusions
Based on the measured data, the morphological evolution of the North−South Passage before and after the construction of the Yangtze Estuary Deep Navigation Channel Project has been analyzed. A validation was carried out on the morphological evolution from 2002 to 2011, and the model results show that the numerical model can basically reflect the evolution trend in the North−South Passage. And the morphological evolution in the next 40 years is simulated. The main conclusions are as follows.
(1) After the DNCP, the North Passage has experienced rapid adjustment stages and self-adaptive stage. At present, it shows the characteristics of slow erosion and volume increase in the main channel, but slow siltation and volume reduction between the groins. The DNCP has a significant influence on the morphological evolution of the main chan-nel and the surrounding area of the South Passage. Right after the construction of the project, the upper section of the South Passage was eroded and the down section silted. At present, it is in a mild erosion state.
(2) A two-dimensional morphodynamic numerical model was established by Delft3D. The time-varying morphological acceleration factor was applied and representative boundary conditions of the tide and runoff were determined by the cross-section residual sediment transport method, making the model suitable for the North−South Passage area of the Yangtze Estuary under the influence of the projects. Through the hydrodynamics and morphological change validation, the mathematical model can basically reflect the erosion and deposition characteristics in the North−South Passage.
(3) The morphological evolution of the North−South Passage of the Yangtze Estuary in the next 40 years after 2016 is predicted. The results show that the erosion and deposition rate of the North Passage will slow down in the future. The present state of erosion in the main channel and siltation in the area between the groins will continue in the future. The South Passage will show a pattern of upper section erosion and lower section deposition in the future. In particular, the erosion and deposition pattern in the North Jiangya Channel and Jiuduan Shoal area will change obviously, which is unfavorable to the stability of the South Passage.
(4) Due to the erosion of the main channel of the North Passage and siltation in the main channel of the South Passage, the split ratio and sediment diversion ratio of the North Passage will increase in the future, but are still smaller than those of the South Passage. The direction of the residual sediment transport caused by the M2 and M4 tidal components in the North Passage does not change, but the transport rate will decrease after 40 years of morphological evolution.
(5) According to the prediction, the North−South Passage could reach a relatively stable state in the next 40 years under the influence of the DNCP.
(6) Due to the limitation of the grid size, the model did not generalize the width of the 12.5 m deep water channel well. At the same time, the effects of extreme floods, storm surges, waves and channel dredging on the long-term evolution of the North−South Passage were not considered. The research on the impact of these factors should be carried on in the future study.