TWO-DIMENSIONAL CHARACTERISTICS OF HYDRODYNAMICS AND SEDIMENT TRANSPORT IN THE PEARL RIVER ESTUARY

  

C.H. Wang1, Onyx W.H. Wai and Y.S. Li

Department of Civil and Structural Engineering

The Hong Kong Polytechnic University

Hung Hom, Kowloon, Hong Kong, China

1Corresponding author; Email: chonghao.wang@polyu.edu.hk;

Tel: (852) 2766 4472; Fax: (852) 2766 6389

 

Abstract: A two-dimensional depth-integrated model was used to investigate the tidal-induced characteristics of hydrodynamics and sediment transport in the Pearl River Estuary (PRE). A computation domain with far enough open-sea boundaries was adopted which aimed to alleviate the errors induced by inaccuracy of open-sea boundary conditions. The computed hydrodynamic results were compared with measurements. In general, the computed results were in good agreement with the observed data. Based upon the computed results, some hydrodynamic processes such as the tidal process, current and residual flows in the PRE were analyzed. Computed results showed that the Eulerian residual flows are mainly due to runoff from the Pearl Rivers and the Stokes drifts are due to tidal asymmetry. Furthermore, spatial distribution of vertically averaged suspended sediment concentration was studied and compared with satellite imagery. The phenomenon that the sediment-laden flow in the main channels appears to be clearer than that in shoals was explained through the analysis of sediment-carrying capacity of flow.  

Keywords: Pearl River estuary, two-dimensional model, hydrodynamics and sediment transport, residual flow, sediment-carrying capacity

1    INTRODUCTION

The Pearl River is the fourth longest river in China and a major river in southern China. The delta network discharges into the South China Sea through eight outlets, which can be divided into two groups (see Fig.1). The four east outlets, comprising Humen, Jiaomen, Hongqimen and Hengmen, discharge into Lingding Sea. The four west outlets are Modaomen, Jitimen, Hutiaomen and Yamen, which all discharge into South China Sea directly. With the flourish of economic development in the PRE, many water-related problems accompanied. Many hydraulic researchers pay more and more attentions to the hydrodynamic and environmental issues of the PRE in recent years (Xu et al. 1985, Tian 1986, Wang et al. 1992, Ying et al. 1993). Field investigations have also been performed, which provide the useful information for further research.

Traditional methods for investigating tidal current and mass transport in an estuary can be concluded as: field data analysis, physical modeling, and mathematical modeling. However, the data analysis method should be based on numerous measurements, and the model scale and huge financial support limit physical modeling. With the development of numerical schemes and computer capacity, numerical model simulation is becoming more and more attractive. Although physical properties of a real estuary are three-dimensional, 2D models are useful in gaining a general understanding of the overall mass distribution pattern in an estuary, especially in a large computation domain. 2D models (van Rijn 1986, Celik and Rodi 1988, Hu and Kot 1997) are available in the literature. In the following sections, a 2D model is briefly described. Then simulation results and analysis are shown; at last some conclusions are drawn. Because Lingding Sea and Hong Kong Waters are of interest in this project, most of the analysis below is focusing on the Lingding Sea and Hong Kong waters.

2    MODEL DESCRIPTION

In the present project, a vertically integrated 2D hydrodynamic and mass transport model, which was modified from 3D model developed by Wai et al. (1996), was employed. In this model, computation domain is divided with four-node isoparameter finite elements in the horizontal domain for easy adaptation to complex boundary configurations. The two-step Lax Wendroff scheme followed by the Kawahara et al. (1978)’s time marching scheme. Details of this model can be seen in Wai et al. (1996).

Fig. 1 shows the computation domain, which covers all the eight Pearl River outlets and entire Hong Kong waters. The purpose of setting open sea boundary far from the area of interest is to reduce errors due to inaccuracy of open sea boundary conditions. The total area covered by the computation domain is about 20,400km2. In this model, the area to be simulated was divided into 4799 four-node elements. Four main harmonic tidal constituents (M2, S2, K1, and O1) were used as the forcing to drive the flow at the open sea boundaries and the known fresh water discharges were the riverine momentum at the eight Pearl River outlets. Because there is not tidal gauge located at the open sea boundaries (see Fig. 1), the amplitudes and phases of the four tidal constitutes were extrapolated from the nearest known gauge stations. Based on the consideration that the boundaries were set far from the eight outlets and the salt water hardly get the outlets during wet seasons (Xu et al. 1985, Ying et al. 1993), it was decided to treat the river outlets as net inflow systems of fresh turbid water at the upstream river boundaries. The measurement data from 3rd to 7th Aug., 1992 during neap tides and data from 10th to 14th Aug., 1992 during spring tides was used to validate the model results. The Hong Kong Civil Engineering Department provided the field data and bathymetric information.

3    RESULTS AND ANALYSIS

3.1    Hydrodynamics

Fig. 2a shows the comparison of computed and predicted tidal elevations at two tide gauges. It could be seen that both the magnitude and phase of calculated tidal level time series were simulated well. The maximum absolute error was less than 0.2 m. Tide in PRE is of irregular semi-diurnal type. Because of the serious restriction of Ma Wan strait at Hong Kong Waters to the tidal propagation, tides from NE delay about 2-3hrs at Chiwan compared with Waglan Island (see Fig. 2b). Duration in ebbing is longer than that in flooding, that may be explained by the strong influence of fresh water from the eight outlets (Xu J.L. et al. 1985, Ying Z.F. et al. 1993) and bathymetry and particularly the coastline (Wang J.M. et al. 1992). Fig. 3 shows that the computed profiles of velocity at WF5 and WF7 were in good agreement with the measurements during spring tides. Also, it can be seen that flow velocity in ebbing is slightly larger than that in flooding, particularly during spring tides.

The flow pattern in flooding and ebbing during a spring tide is complex in PRE. During flooding, tidal flow coming from offshore and Hong Kong Waters, pass though the Lantau channel and Umston channel into Lingding Sea. With the effect of Coriolis force and the transverse surface slope which generally incline to the SE direction (Wang et al. 1992, Ying et al. 1993), the salt water flows riverward along the channels, the direction is corresponding with the coastline in the east region. In the West shoal, because of the higher seabed roughness, tidal energy dissipates quickly; the flow moving riverward is relatively smaller. During ebbing, flow velocity in the West shoal is relatively larger, this is due to the transverse surface slope and strong freshwater from Jiaomen, Hongqimen and Hengmen. Flows in the West shoal, mostly go into along the West channel and partly pass though the Middle shoal into the East channel, then go into South China Sea through Lantau and Umston channels.

Based on the computed velocity field, the residual flow in the PRE was also analyzed. Fig. 4 shows the Eulerian residual flow from non-tidal drift and Stokes drifts due to tidal pumping (Sylaios and Boxall 1998) during spring tides (from 16:00 11th to 17:00 12th, Aug. 1992). It can be seen that Eulerian residual flow is dominant compared with Stokes drifts in the PRE. The maximum Eulerian residual is about 0.3m/s in the PRE during the computation period. Eulerian residual flow at the West shoal is in SE to SEE direction consistent with the ebbing flow direction. It is larger in the West channel than that in the East channel due to the strong effect of freshwater input. The Stokes drift velocity is mostly less than 0.05m/s. The magnitude of Stokes drift velocity is about one fold larger in spring tide than that in neap tide.

In general, the residual flow in the PRE is from freshwater from the Pearl River, the Stokes drifts from the tidal asymmetry is also significant in transporting mass from open sea, especially during spring tides in dry seasons (Xu et al. 1985, Ying et al. 1993).

3.2    Suspended sediment

Sediment transport in the PRE is more complex in the PRE. It is not only affected by the input from the Pearl River, but also by the tidal pumping, local resuspension, and flocculation and so forth. Fig. 5 shows the suspended sediment concentration distribution in ebbing in neap and spring tides at which suspended sediment reached equilibrium after 20 cycles. It could be seen that suspended sediment concentration in channels is less than that in shoals. This computed suspended sediment pattern in PRE resembles a satellite imagery presented by Au and Lulla (1997) which shows that the suspended sediment concentration in the main channels is lower than that in shoals. Because the sediment mainly comes from the Pearl River, hyper-concentration region is located at the outlets, the 0.1kg/m3 iso-concentration contour basically coincides with the 5m iso-depth contour in the west region during neap tides. During spring tides, because of the higher ebbing flow velocities, the 0.1kg/m3 iso-concentration shifts to SE near the West channel. Other hyper-concentration regions exit in the East shoal, south of the Middle shoal around inner Lingding island and Tonggu shoal located at NW of Lantau island, and some shallow part around the Lantau island. Sediments coming from the Pearl River will deposit quickly due to the flow dispersion, especially with the reaction of flocculation due to intrusion salt wedges (Xu et al. 1985, Ying et al. 1993). Local hyper-concentration can be seen as a result of local resuspension. According to the theory of sediment-carry capacity (Qian and Wan 1983), the sediment-carrying capacity is inversely proportional to the water depth. In tidal ebbing, the velocity in the West shoal is nearly the same magnitude as that in the main channels. However the water depth is small, thus the sediment-carrying capacity is large enough to resuspend the fine bed materials. The resuspension mechanism of suspended sediment in the PRE has been confirmed by analyzing the longitudinal near-bed shear stresses by Tian (1986). In Hong Kong Waters, sediment concentration is small. Comparing the sediment concentration patterns between neap tide and spring tide, it is found that concentration in the East channel and Umston channel varies insignificantly, because the sediments in the East channel come mainly from Humen. However sediments in the West channel come mainly from Jiaomen, Hongqili and Hengmen, partly from Humen, this feature is corresponding to the residual flow in PRE as mentioned above.

In general, sediments in PRE come mostly from the Pearl River. Sediments from sea can be neglected in wet seasons. The suspended sediment concentration in the West shoal, which is directly supplied from the riverine sediments, is higher. It is the result of the combination of the transport from river and local resuspension. However some isolated hyper-concentration situations can be explained as the result of local resuspension, because of the surplus sediment-carrying capacity in these shallow regions in flooding and ebbing.

4    CONCLUSIONS

The hydrodynamics and suspended sediment patterns in the PRE are complicated as seen from the computation results. In this study, a depth-integrated two-dimensional hydrodynamic and mass transport model, modified from a multi-layer three-dimensional model was applied. Tidal levels were calibrated using four main tidal constitutes. The simulated results were in very well agreement with the predicted ones. Model results show that tidal flow can penetrate more riverward along the East and West channel. However, the west shallow regions make the tidal wave energy dissipating quickly, thus flooding current is relatively smaller and ebbing current is larger due to the strong riverine runoff. Analysis of residual flow shows that the Eulerian residual from non-tidal drift is the dominant force in the PRE, maximum residual velocity is about 0.3m/s, and the direction of the residual flow is consistent with the direction of tidal flooding and ebbing. The Stokes drift velocity is less that 0.05m/s, but it is an important source of power to drive the mass from open sea.

Characteristic of suspended sediment concentration was also investigated. Model results show that sediment concentration in the West shoal is high due to the inputs from the Pearl River and local resuspension. Resuspension plays an important role on some isolated hyper-concentration regions in the shallow regions because of the surplus sediment-carrying capacity. Sediment concentration in deep channels is smaller than that in the nearby shoals. This suspended sediment concentration pattern is consistent with a satellite picture.

In this study, the effects of stratification due to summer runoff have not been taken into account. However, these are important for a partially mixed estuary. PRE is a partially mixed estuary with density stratification in wet seasons, intrusion of salt water and turbidity maximum (Xu et al. 1985, Tian 1986). To investigate the effects of density stratification due to salinity gradient in both horizontal and vertical directions on flow pattern and sediment transport, a 3-D model is necessary to reveal the actual situation.  

Acknowledgements

The project was supported by an internal HKPolyU grant: G-V783 and a RGC grant: PolyU 5038/98E. We are thankful for the field measurements provided by the Hong Kong Civil Engineering Department.

References

Au, K.N. and Lulla, K. (1997). Hong Kong and the Pearl River Delta as seen from space imagies. Geocarto International Centre, Hong Kong.

Celik, I. and Rodi, W. (1988). “Modeling suspended sediment transport in Noneqilibrium situations”. J. Hydr. Eng., ASCE, 114(10), 1,157-1,191.

Hu, S.L. and Kot, S.C. (1997). “Numerical model of tides in Pearl River Estuary with moving boundary”. J. Hydr. Eng., ASCE, 123(1), 20-29.

Kawahara, M., Nakazawa, S., Ohmori, S. and Hasegawa, K. (1978). “Tsunami wave propagation analysis by the finite element method”, Finite Elements in Water Resources, eds. C.A.Brebbia et al. Pentech Press, London, 2,131-2,150.

Qian, N. and Wan, Z.H. (1983). “ Mechanics of sediment transport”. Science press, Beijing.

Rijn, L.C. van (1986). “Mathematical modeling of suspended sediment in nonuniform flow”. J. Hydr. Eng., ASCE, 112(6), 433-455.

Sylaios, G. and Boxall S.R., (1998). “Residual current an flux estimates in a partially-mixed estuary”. Estuarine, Coastal and Shelf Science, 46, 671-682.

Tian, X.P. (1986). “A study on turbidity maximum in Lingding sea estuary of the Pearl River”. Tropic Oceanology, 5(2), 27-35. (in Chinese).

Wai, W.H. Onyx, Lu, Q.M. and Li, Y.S.(1996). “Multi-layer modeling of three-dimensional hydrodynamic transport processes”. J. Hydr. Res., 34(5), 677-693.

Wang, J.M., Yu, G.Y., Chen, Z.Y (1992). “Numerical Modeling of Tidal Current in Lingding Sea PRE”. ACTA Ocea. Sini., 14(2), 26-34. (in Chinese).

Xu, J.L., Li, Y.X., Can, F. X and Chen, Q.D (1985). “Evolution of Channels and Shoals in Lingding Sea PRE”, Ocean Press, Beijing. (in Chinese).

Ying, Z.F, Luo, Z.R, Wu, C.Y., Chen, Z.X. and Wang H.S.(1993). “Dynamic deposition environment and navigation channel selection in Lingding Sea”. Study on coastal dynamic physiography and its application in harbor infrastructure in South China, 1995, edited by Yang G.R et al., Zhongshan University Press, China. (in Chinese).

 

 

 

 

 

 

 

                                                  

 

                                   Fig. 1  Sketch of PRE            Fig. 2   Time series of tidal and comparison                                


Fig. 3    Comparison of observed and computed velocity

 

(a) Eulerian residual                                (b) Stokes drift

Fig. 4    Residual flow during a spring tide in PRE

(a) Neap tide                                (b) Spring tide

Fig. 5    Computed suspended sediment pattern in PRE