International Journal of Remote Sensing and Earth Sciences Vol. 6: 65-76 (2009) © IReSES
NUMERICAL SIMULATION OF COHESIVE SEDIMENT TRANSPORT IN JAKARTA BAY
I. M. Radjawane1 And F. Riandini2
Abstract. The 3D-numerical model has been applied to simulate the current circulation and cohesive sediment transport in the Jakarta Bay, Indonesia. Sediment load comes from 3 river mouths i.e. Angke River, Karang River, and Ancol River. The model was simulated to analyze the effect of tidal current and river discharge. A constant westerly and easterly wind was used as input of the model to see the influence of monsoonal season. The numerical results showed that the tidal current flows from east to western part of the bay during ebb tide and vice versa during flood tide. The surface current circulation was dominantly influenced by the tidal current compared with the wind and river discharge effects. High turbidity level was found near the river mouths with the range of 50 to 100 mg/l. This high sediment concentration was caused by the effect of sediment load from the river upstream. In the offshore area of the bay the sediment concentration decreases up to 10 mg/l. The movement of sediments followed the current circulations. During the flood tide, the sediment concentration from the mouth of Angke River moved to the western part of the bay. Model simulated for increasing the river discharge into two times showed that the sediment distributed to the offshore direction two times longer compare with the normal debit. The transport of sediment from the Angke and Karang Rivers to the offshore area reached > 6 km, while it just reached + 2,5 km from the Ancol River . Keywords: 3D-numerical model, Cohesive sediment, Effect of tidal current and river discharge
1. Introduction trough Java Sea and resulting the co- Jakarta Bay is located at western side of oscillation tides in the central part of Java northern part of Java Island, a shallow bay Sea. The monsoon wind system prevails with an average depth of about 15 m, an over the Java Sea that consists of the area of 514 km2, and a shoreline about 72 northwest monsoon during December to km long. The bay receives highly polluted February, southeast monsoon during June water from 19 rivers that run through the to August, and two transitional monsoons. Jakarta Metropolitan Area (JMA) and There area several rivers, such as the poured out into the Java Sea (UNESCO, Citarum, Ciliwung, Angke, Ancol, and 2000). Karang Rivers. The physical processes affected the dynamics of sedimentation and The current circulation in this bay is erosion in this area. Bathymetry and controlled mainly by the three physical coastline change is one issue in JMA. Wind processes, i.e. tides, wind and river direction and river discharges are seasonal discharges (Koropitan, et. al., 2009 and based data. Rachmayani, 2004). Tides are dominated by the diurnal components, especially the K1 component coming from Flores Sea and ______1 Oceanography Department, Faculty of Earth Sciences and Technology, Bandung Institute of Technology. E-mail: [email protected] 2 Research Institute of Water Resources and Development, Bandung, Indonesia.
65
Radjawane and Riandini
Further, this area becomes worse and simulate the cohesive sediment transport, causes the sea floor and coastline the ECOMSED model was improved by surrounding the Jakarta Bay changed due applying flocculation and consolidation to the anthropogenic effect, i.e. human processes, resulting in the following three activities as reported from UNESCO specific module developments. (2000). Large scale sand extraction started Suspended Sediment Module with harbour dredging in the Jakarta Bay The transport of suspended sediment is area. The dredged material was generally described by the following advection- dumped elsewhere in the bay and diffusion equation: continuation of sand extraction for building ∂cccc∂∂∂∂∂∂∂ started on a small scale and was carried out +−()()uwcAiisδ 3 = H + A H + K H (1) ∂∂txi ∂ x112233 ∂ x ∂ x ∂ x ∂ x ∂ x manually in the 1970s and has intensified yearly in order to provide construction where, c: concentration of the suspended materials. Mangrove area has been sediment, ui: velocity components. AH: the transformed into reclamation area for horizontal diffusivity and KH: the vertical luxury residences, commercial activities, eddy diffusivity. and industrial zones. Based on comparison At the water surface, zs, the net satellite image between 1971 and 2004, sediment flux is zero. At the sediment- there are 80% of land use changes from water interface, zb, the flux is estimated by vegetative area into urban area and almost the rates of erosion and deposition, Ferosion at the coastal area (Koropitan, 2009). and Fdeposition. The bottom boundary A comprehensive study to understand condition can therefore be formulated as the spatial and temporal distribution of follows: ∂c cohesive sediments in this bay due to the ()uwcKis− −= H 0 and ∂x3 main physical processes is needed. In this xZ3 = s study, the monsoon effect, tides, and river ∂c uwcK−− = E discharges are simulated into a ()is H bc, (2) ∂x3 hydrodynamic model to get a more clear xZ3 = b understanding concerning with current with EFb, c=+ erosion F deposition . circulation in this bay. The dynamic of The erosion rate is represented by sedimentation and erosion processes is Partheneiades’s formulation (Partheniades, simulated with a sediment transport model. 1965) as: Several scenarios based on generating τ b forces of the current circulation are FMerosion = −1 (3) τ conducted to understand the variation and e dynamics of seasonal distribution. where, M is a positive empirical erosion parameter,τ is the bed shear stress, τ is the 2. Model Descriptions b e critical shear stress for erosion, is a A three-dimensional finite difference function of the concentration of the top bed model system for hydrodynamic and layer, which itself is given by the state of cohesive sediment transport, ECOMSED consolidation. The erosion coefficient M (HydroQual, 2002), is used in this may also be function of the concentration. simulation. This model solves the Navier- The deposition rate is calculated according Stokes equations with free surface to Krone’s formula (Krone, 1962): boundary conditions and the advection- F = Pwc (4) diffusion equations of the temperature, deposition d s salinity, and any other variable. To
66 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009
Numerical Simulation Of Cohesive Sediment Transport
where Pd is the probability for deposition 2.2.2. A Model for Turbulence-Induced described by, Flocculation τ b A complete flocculation model should Pd =−1 (5) τ d include both aggregation and floc breakup where τd is the critical shear stress for processes. By assuming a simultaneous deposition. acting of aggregation and breakup, the differential equation for the flocculation of Flocculation Module cohesive sediment under the influence of 2.2.1. Relation between Settling Velocity turbulent shear, is described with respect to and Floc Size the number concentration, N, as: ∂∂NN()11−−φφ* p ∂ ∂ ( ) Winterwerp (1998) developed three- + +−uwNDiiδ 3, sr −() s +Γ= T ∂∂tx()12.5 +φ ∂ x ∂ x dimensional Eulerian model of the iii(8) p −−kGDNkGDDDN'1()φ 32 +qq+ 1 − 2 evolution of the settling velocity of fine- ABp* () grained cohesive sediment in turbulent where, Ds : molecular diffusivity, open channel flow: ΓT :turbulent diffusivity, G : the square root nf −1 α ()ρρsw− g 3−nf D of turbulent dissipation rate divided by wDsr= p 0.687 (6) 18βµ 1+ 0.15Re p molecular dynamic viscosity, δi3 : Kronecker’s delta. The parameters k’A and which D is the actual floc size, Dp is the diameter of the primary particle, and n is kB are defined as follows (e.g. Winterwerp, f 1998): fractal dimension for sediment particles. α q 3 − p µ and β are coefficients depending on the kee'Acd= π kaeD= 2 and Bbp spherity of the particles, and Re is the Fy particle Reynolds number. where ec, ed, and eb are efficiency Mud flocs seldom settle as individual parameters for collision, diffusion, and particles. When their concentration breakup respectively. µ is the dynamic becomes high enough, the settling flocs viscosity of the suspension, Fy is the start to hinder each other in their strength of the mud flocs, and numbers of movement, generally known as hindered power, p and q are to be established settling. The effective settling velocity ws empirically. The definition and relation of in suspension of cohesive sediment the number concentration, N, and the mass affected by the process of hindered settling concentration, c, and the volumetric defined as: concentration φ are: 3−n f 3 c D ()11−−φφ* ()p φ = fs ND = (9) ww= ρs Dp ssr12.5+ φ (7) where f s is the shape factor. where, φ the volumetric concentration of The aggregation and floc breakup terms are set to zero at the water surface the flocs (φ=c/cgel), cgel is the gelling concentration at which a space-filling and at the horizontal water-bed interface. network forms, and c is the sediment Hence the boundary conditions lead: concentration by mass. The factor (1-φ*) ∂N ()()uwNDis− −+Γ sT =0 and accounts for the return-flow effect. c/c ∂x gel 3 xZ= can exceed unity in a consolidating fluid 3 s ∂N mud layer. ()()uwNDis−−+Γ= sT E bN, (10) ∂x3 xZ3 = b
International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009 67
Radjawane and Riandini
The source-sink term Eb,N represents the Data used for this research are input for exchange with the bed and is modeled with hydrodynamic and sediment transport classical formula of Partheniades and models. The data are: bathymetry, water Krone, so that the water-bed exchange level, salinity, and temperature, and formulation is consistent with the one for concentration of total suspended solid. the mass balance. Bathymetry Depth data of Jakarta Bay extracted from 3. Application to Jakarta Bay the World Digital Chart (GEBCO, 2009) as 3.1. Data Collection seen in Figure 1.
Figure 1. Bathymetry Map of Jakarta Bay. *Source: GEBCO (2009). Sea level Salinity and Temperature Mean sea level data is obtained from Temperature and salinity data used for prediction of ORI Tide at the northern of initial condition of Jakarta Bay are shown Jakarta Bay, at points 106° E - 6° S and in Table 1. 106.967° E - 6° S. o o Table 1. Observed data of average temperature ( C) and salinity ( /oo) in the Jakarta Bay. Parameter Surface Bottom Temperature 33.3 31.2 Salinity 31.6 32.5 *Source: Report of Monitoring Water Quality of Jakarta Bay BPLHD, 2005. o o Table 2. Observed data of average temperature ( C) and salinity ( /oo) in river mouths. River Temperature Salinity mouth Spring Neap Spring Neap Angke 32.55 31.3 5 5 Karang 33.1 31.1 18 13 Ancol 32 30,9 20 6 *Source: Report of Monitoring Water Quality of Jakarta Bay BPLHD, 2005.
68 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009
Numerical Simulation Of Cohesive Sediment Transport
3.2. Boundary Conditions constituents of the Jakarta Bay. River Boundary conditions for the model were boundary discharges of each river are used at open and river boundaries. The shown in Table 3, while Table 4 shows for open boundaries are water level, temperature, salinity, and sediment temperature, salinity, and sediment concentration. concentration. Water level boundary condition was predicted from tidal Table 4. River discharge. Velocity (m/s) Discharge (m3/s) River mouth Spring Neap Spring Neap M. Angke -0.060 0.030 -21.6 10.8 M. Karang 0.010 -0.010 3.6 -3.6 M. Ancol -0.008 0.000 4.6 0.0
Table 5. River boundary conditions.
0 0 Temperature ( C) Salinity ( /00) Sediment River mouth Concentration Spring Neap Spring Neap (mg/l)
M. Angke 32.5 31.3 5 5 200
M. Karang 33.1 31.1 18 13 200
M. Ancol 32.0 30.9 20 6 200
3.3 Discretisation and Model Desain Simulation of sediment tranport was Model domain shown in Figure 2, is done to identify distribution of sediment divided horizontaly to 144 x 54 grids and around the Jakarta Bay during spring and 10 vertical layers. The various parameters neap tide. Source of sediment was assumed used in the numerical simulations are listed come from upstream of the three rivers, in Table 6. as follows : namely the Muara Angke (A’), Muara Karang (B’), and Muara Ancol (C’).
Table 6. Parameter setting. ∆ X 250 m ∆ Y 250 m External interval time 1 s Internal interval time 10 s Simulation time 15 days Initial concentration of TSS 10 mg/l
International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009 69
Radjawane and Riandini
Figure 2. Model domain.
4. Simulation Results of Sediment 4.1.1.2. Flood Tide Transport Figure 4 shows distribution of sediment Results of sediment transport model are concentration during flood condition. Weak shown in the form of distribution surface current velocity cause deposition of concentration during spring tide and neap sediment around the river mouths therefore tide. surface sediment concentration is decrease. 4.1. No Wind 4.1.1.3 Flood to Ebb 4.1.1. Spring Tide Strong current during flood to ebb 4.1.1.1. Ebb to Flood condition increases surface sediment Distribution of sediment concentration concentration around the river mouths as for ebb to flood condition is shown in shown in Figure 5. This strong current is Figure 3. The figure shows that high capable to push high concentration of concentration of sediment occured around sediment seaward. river mouths and coastal water near the 4.1.1.4. Ebb Tide river mouths. This high concentration was Figure 6 shows distribution of sediment caused by high sediment load from concentration during ebb condition. upstream of the rivers. Sediment Distribution of sediment is similar to the concentration seaward of the river mouths previous tidal condition with spreading is small, less than 10 mg/l. From the figure distance + 6,5 from the river mouths. we can see that due to tidal current, sediment from Muara Angke River move eastward.
Figure 3. Sediment distribution during ebb to Figure 4. Sediment distribution during flood flood condition (spring tide). condition (spring tide).
70 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009
Numerical Simulation Of Cohesive Sediment Transport
Figure 5. Sediment distribution during flood to Figure 6. Sediment distribution during ebb ebb condition (spring tide). condition (spring tide).
4.1.2 Neap Tide spring tide. Distributions of sediment The simulation results during the neap concentration for neap tide condition are tide condition show similar pattern with the shown in Figures 7-10. Maximum distance spring tide condition. However, weak of sediment spreading reaches + 5 km from current velocity during neap tide causing the river mouths. spreading of sediment is narrower than the
Figure 7. Sediment distribution during ebb to Figure 8. Sediment distribution during flood flood condition (neap tide). condition (neap tide).
Figure 9. Sediment distribution during flood Figure 10. Sediment distribution during ebb to ebb condition (neap tide). condition (neap tide).
4.2. Simulation with Wind 4.2.1 West Wind For scenario of simulation with wind Simulation results of west wind are data, there are 2 scenarios representing east shown in Figures 11-14. In this simulation, and west monsoons. Wind speed was sediment concentration seems to spread - assumed to be constant 2 m/s for both seaward compared with no wind scenarios. simulation. Influence of wind seen from direction of sediment spreading to East.
International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009 71
Radjawane and Riandini
Figure 11. Sediment distribution during flood Figure 12. Sediment distribution during flood condition (neap tide). to ebb condition (neap tide).
Figure 13. Sediment distribution during ebb Figure 14. Sediment distribution during ebb to condition (neap tide). flood condition (neap tide).
4.2.1 East Wind simulation. Influence of wind can be seen Simulation results of East Wind are from westward spreading of sediment shown in Figures 15-18. In this simulation, concentration, as shown in Figures 16 and sediment concentration g seems to spread 17. seaward compared with no wind
Figure 15. Sediment distribution during flood Figure 16. Sediment distribution during flood to condition (neap tide). ebb condition (neap tide).
Figure 17. Sediment distribution during ebb Figure 18. Sediment distribution during ebb to condition (neap tide). flood condition (neap tide).
72 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009
Numerical Simulation Of Cohesive Sediment Transport
4.3 Simulation with Big Discharge the figures, we can see that sediments The purpose of this scenario of move seaward farther than the actual ones. simulation is to investigate influence of At the Muara Angke and Muara Karang discharge changing to sediment waters (marked A’ and B’ in Figure 2, distribution. River discharges used in this respectively), distance of sediment simulation were two times of the actual spreading reaches more than 6 km seaward, discharges. Simulation results of this while in the Muara Ancol waters (marked scenario are shown in Figures 19-22. From C’ in Figure 2) it is only + 2.5 km.
Figure 19. Sediment distribution during flood Figure 20. Sediment distribution during flood to ebb condition (neap tide). condition (neap tide).
Figure 21. Sediment distribution during ebb Figure 22. Sediment distribution during ebb to flood condition (neap tide) condition (neap tide)
5. Distribution of Tidal Current influences to sediment distribution. It Generally, all simulation results of seems similar influence of discharge to current circulation show same pattern. current pattern and sediment distribution. Currents flow from east to west during ebb Large Simulation results of wind scenario condition, and flow eastward during flood. are shown in Figures 31–34 for West wind Simulation results of current pattern for no and in Figures 35–38 for East wind. wind scenario are shown in Figures 23–26 River discharges only gave small for spring tide and in Figures 27–30 for influences to current pattern. However, neap tide. they influence distribution of sediment Influence of winds to current pattern spreading. The results of simulation are were small, but they had rather great shown in Figures 39–42.
International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009 73
Radjawane and Riandini
5.1. No Wind (Spring tide)
Figure 23. Current pattern during flood condition Figure 24. Current pattern during flood to ebb (spring tide). condition (spring tide).
Figure 25. Current pattern during ebb condition Figure 26. Current pattern during ebb to flood (spring tide). condition (spring tide).
5.2. Wind (Neap Tide)
Figure 27. Current pattern during flood condition Figure 28. Current pattern during flood to ebb (neap tide) condition (neap tide)
Figure 29. Current pattern during ebb condition Figure 30. Current pattern during ebb to flood (neap tide) condition (neap tide)
74 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009
Numerical Simulation Of Cohesive Sediment Transport
5.3. West Wind
Figure 31. Current pattern during flood condition Figure 32. Current pattern during flood to ebb (neap tide). condition (neap tide).
Figure 33 Current pattern during ebb condition Figure 34. Current pattern during ebb to flood (neap tide). condition (neap tide). 5.4. East Wind
Figure 35. Current pattern during flood condition Figure 36. Current pattern during flood to ebb (neap tide). condition (neap tide).
Figure 37. Current pattern during ebb condition Figure 38. Current pattern during ebb to flood (neap tide). condition (neap tide). 5.5. Large Discharge
Figure 39. Current pattern during flood condition Figure 40. Current pattern during flood to ebb (neap tide). condition (neap tide).
International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009 75
Radjawane and Riandini
Figure 41. Current pattern during ebb condition Figure 42. Current pattern during ebb to flood (neap tide). condition (neap tide).
6. Conclusion physical process on the ecosystem of In this study, the three-dimensional Jakarta Bay: a coupled physical- hydrodynamics and sediment transport ecosystem model experiment, ICES, model, ECOMSED, have been modified Journal of Marine Science, 66:336-348. by introducing some method from Rachmayani, R., 2004, 2-D Numerical COSINUS project. The modified model Model of Horizontal Distribution of was used for predicting tidal flow and BOD-DO by using Quickest Method suspended sediment transport in the (manuscript in Bahasa and abstract in Jakarta Bay. From the model results we English), Final Project, Department of concluded that the models were capable of Oceanography, Institut Teknologi reproducing the hydrodynamics and Bandung. cohesive sediment transport processes in UNESCO, 2000, Reducing megacity the Jakarta Bay. It was found that higher impacts on the coastal environment– concentration in shallow area occurred due Alternative livelihoods and waste tidal condition and river discharge. management in Jakarta and the 7. References Seribu Islands, Coastal Region and Berlamont, J. E. and E. Toorman (eds), Small Island Papers 6, UNESCO, 2000, COSINUS Final Scientific Paris, 59 pp. Report, Hydraulics Laboratory, K.U. Winterwerp, H., 1999, On the Leuven. dynamics of high-concentrated mud Bruens, A. W., 2003, Entraining mud suspension, Report 99-3 suspension, PhD-thesis, Delft Communication on Hydraulic University of Technology, Department of Civil Engineering and Geosciences. Engineering, Department of Civil Engineering and Geosciences, Delft Dyer, K. R., 1997, Estuaries A Physical – University of Technology. Introduction, Wiley. Winterwerp, J. C., 2002, On the HydroQual, 2002, A Primer for flocculation and settling velocity of ECOMSED Version 1.3, User Manual, estuarine mud, Continental Shelf HydroQual Inc., New Jersey. Research, 22:1339-1360. Koropitan, A. D., Ikeda, M., Damar, A, and Yamanaka, Y., 2009, Influences of
76 International Journal of Remote Sensing and Earth Sciences Vol. 6, 2009