Hydrodynamic Modelling for Salinity of Singapore Strait ...ipcbee.com/vol4/63-ICESD2011D30032.pdf · Prameet Rathod Environmental Science and Engineering Indian School of Mines Dhanbad,

Hydrodynamic Modelling for Salinity of Singapore Strait and Johor Strait using MIKE 3FM

Rohit Goyal Environmental Science and Engineering

Indian School of Mines Dhanbad, India

E-mail: [email protected]

Prameet Rathod Environmental Science and Engineering

Indian School of Mines Dhanbad, India

E-mail : [email protected]

Abstract—Numerical models provide quick and better understanding of the behavior of ocean waves, currents and sediment transport. Hydrodynamic modelling uses the concept of scales of motion, dimensionality of flow, physical processes and forcing mechanisms. The present study aims at modelling the flow patterns prevailed in the Singapore Strait and Johor Strait due the tidal forcing from South China Sea and Melaka Strait. The domain was forced with the measured tidal signal from the artificially truncated boundaries for a selected period from 29/03/2009 to 02/04/2009. Measurements available for the above seasons have been used for the validation of model results. Further, the hydrodynamic results were used to carry out the Water Quality Modelling to study the temperature and salinity variation. Based on the simulation results, flow off Singapore Strait was observed which was used to simulate the Temperature and Salinity of Singapore Region.

I. INTRODUCTION Coastal zone, the interface of land, ocean, and

atmosphere, is clearly of major economic and social importance. It is defined as the region from the 200m water depths at the sea to the 200m elevations on the land. Coastal landscapes consist of beaches, cliffs, dunes, sand pits, barriers, islands, tidal flats, deltas, tidal inlets etc. These are results of sediment transport generated by hydrodynamic forces. The coastal region has acquired greater attention in the recent days due to increasing utilization of its resources. But, the vulnerability of the coastlines and coastal resources due to adverse impact from natural extreme events, pollution due to industrial discharges, etc. remind us to have a wise and sustainable management of coast and coastal resources. Thus, the computational hydrodynamic study becomes more complex and tricky to bring out the real physics. Hydrodynamic (HD) modelling[1] is a prerequisite to environmental/ecological modelling, as it influences the biological and chemical processes. A good understanding of scales of motion, dimensionality of flow, physical processes and forcing mechanisms are essential for HD modelling.

Singapore has two main monsoon seasons The Northeast monsoon Season (December –March) and the Southwest monsoon season (June -September).

The objectives of the study are framed as follows: • To simulate tide-driven currents along the coast of

Singapore. • To simulate water level and flow off Singapore. • To quantify the flow velocity.

• To check the variations of Temperature and Salinity in the domain.

II. STUDY REGION:

FIG:1 Study area (locaion map)

III. COMPUTATIONAL TOOLS AND METHODS

The three-dimensional MIKE 3 FM model developed by the Danish Hydraulic Institute was used to study the hydrodynamics of the region for 3D free-surface flows. MIKE 3[2] is applicable to the simulation of hydraulic and related phenomena in lakes, estuaries, bays, coastal areas and seas where stratification or vertical circulation is important. MIKE 3 Flow Model FM is a modelling system based on a flexible mesh approach. The flexible mesh is most suitable for irregular boundaries of the water body. The modelling tool has been developed for applications within oceanographic, coastal and estuarine environments.

3.1) METHODOLOGY MIKE3 FM is capable of taking irregular domains

(unlike in case of MIKE 3) [3]. The process starts from the generation of mesh i.e., defining model domain and grid size.

295295

2011 2nd International Conference on Environmental Science and Development IPCBEE vol.4 (2011) © (2011) IACSIT Press, Singapore

Once the domain was formed (after defining the open and closed boundaries) thereafter mesh was generated .Mesh can be made smooth by increasing the number of iterations .Scatter data of depth or the bathymetry values was imported on the mesh and was interpolated by either using method of Natural

Neighbour or Linear method. After Interpolation it was exported as either .mesh file or .dfsu file .The exported file contains the actual bathymetry of the region.

Once the field work for the model was done the generated bathymetry was given as an input to the model. According to the needs module was selected and simulation time and date was given. After that various conditions like Bed resistance, Flood[4] and Dry, density, Eddy Viscosity, Solution Technique etc can be given according to the quality of required output data. Thereafter boundary conditions were specified and model was made to run and outputs were specified as required.

3.1.1) DATA USED Data on currents, sea level, and winds were collected off

Singapore. The data was collected for the period March 29, 2009 to April 2nd, 2009.

3.1.2) GENERATION OF MESH AND BATHYMETRY Setting up a mesh includes appropriate selection of the

area to be modelled, adequate resolution of the bathymetry, flow, wind and wave fields under consideration and definition of codes for open and land boundaries. Mesh file was generated with the MIKE Zero Mesh Generator [5]. The mesh file was an ASCII file (.mesh extension) that includes information of the geographical position and water depth at each node point in the mesh. The file also includes information about the node connectivity of the triangular and quadrangular elements.

Fig 2: Mesh file

Specifications required generating the domain and thereafter meshing involved: • Maximum element area: 0.0003 ^2 • Smallest allowable angle: 26 • Maximum number of nodes:100000 • Interpolation Method: It has been divided into three parts;. a. The Natural Neighbour method. b. The linear method. c. The inverse distance weighted method. • Size of Bounding: 1000% beyond convex hull • Extrapolation: 00 • Time taken: 47 sec

Fig 3: Interpolated Bathymetry used for computation

3.1.3) BOUNDARIES The water levels at the west, south and east open

boundaries of the model were specified based on the tidal elevations predicted using MIKE3 FM tidal prediction module. The rest of the domain was interpolated by means of MIKE3 FM interpolation tool. The Northern boundary was closed by land therefore it was considered as a no flow boundary. A constant water level and zero velocity were used as initial conditions at all grid points [6]. The tidal level at the open boundaries was used as the boundary condition and the flow direction at the open boundary was considered to be perpendicular to the boundary.

3.1.4) CALIBRATION FACTORS

Model calibration is defined as fine tuning of parameters until the numerical model results and the field measurements were within an acceptable tolerance by modifying the boundary conditions and improving the hydro-meteorological forcing input[7].Two parameters adjusted during calibration were:

(i) Roughness coefficient – used in the bottom friction formulation

296296

(ii) Eddy viscosity – parameterized horizontal mixing of momentum

3.2) MODEL VALIDATION BY WATER LEVEL The validation process gives an indication of the model's

sensitivity and confidence that the results it produced were consistent with measurements [8]. The zonal (u) and meridional (v) components of the currents derived from the measured elevation of March, 2009 and April, 2009 were used for the validation of model results. Water level measurements were available for the period 29th March to 2nd April, and for that period simulated values were validated . Comparison between measured and simulated water levels is presented in Fig 4 for all the stations . The results indicated that modelled water levels matched very well with measured water levels throughout region as shown:

Fig:4 Comparison of water level at different stations

3.3)TEMPERATURE SALINITY MODULE The temperature/salinity (TS) module was invoked from

the specification of the density since baroclinic density (density depends on temperature and/or salinity) was selected. The TS module sets up additional transport equations for temperature and salinity[9]. Additionally the calculated temperature and salinity are feed-back to the hydrodynamic equations through buoyancy forcing induced by density gradients. Various parameters considered were: • Temperature Range: 20 – 30 degrees • Salinity Range: 25psu – 35psu • Solution Techniques: Low order, fast algorithm • Scaling Factor: 1 • Drying depth: 0.005 m • Flooding Depth: 0.05m and • Wetting depth: 0.1 m • Reference Temperature: 29 degrees • Reference Salinity: 30 psu

3.4) DOMAIN AND MESH GENERATED Having validated the model , we redefined the whole

domain to a comparatively smaller one so as to increase the model speed to a good extent .For doing this in the mesh generator module points, lines and polygons were deleted collectively and a smaller domain was generated . Thereafter with the new domain a mesh was created taking the parameters as follows: • Interpolation Method: Natural Neighbour • Time Taken: 40 sec • Simulation time: 1.5 days • Number of scatter points: 80976 • Number of Elements: 6453 • Mesh Nodes: 4728

297297

Fig 5: Modified domain used to calculate salinity module

3.5) BOUNDARIES (NEW DOMAIN)

The New domain consisted of 7 open boundaries .The values for the boundaries were obtained from the simulated outputs of the earlier larger domain which was used for the validation of the model also. The data at these node point were extracted using Data Extraction FM tool of MIKE Zero[10].

Fig 6: Shows location of open boundaries of the new domain

IV. RESULTS AND DISCUSSION The model was run without wind for the period

29/03/2009 to 30/03/2009

4.1) WATER LEVEL The surface elevations had been simulated for every 15

minutes. The simulated water level at Singapore showed that the surface elevation was primarily contributed by tides. It was assumed that wind contribution to water level variation was negligible in the context. The presence of higher water level at the east end of the domain compared to the south and the west showed that water level along the Singapore Strait was increasing eastwards. That increase may be due to the funnelling effect and/or the bathymetric effects [11]. The currents were varied with the water level variation. Higher the water level, stronger was the current speed.

Fig 7.Time series plot for surface elevation without using the wind.

4.2) CURRENTS Tidal currents were significant in this region. The current

speed during ebb tide was almost similar as compared to flood tide since the currents were considered without wind. During ebb tide, the predominant flow was towards south east and during the flood tide flow was towards North West Tidal currents oscillate mainly in the longshore direction with little net cross-shore current. The onshore and offshore currents (u-component) were meagre compared to the alongshore currents (v-component), irrespective of the period of simulation. The current slows down during the tidal slack that was, just before current reversal takes place. There was no contribution by the fresh water discharges as such, but some contribution can be expected later, but that would be very minimal.

298298

Fig.8: Velocity vectors depicting the instantaneous flow during flood

tide

The current meter value and the simulated output showed that, the alongshore component of velocity goes on decreasing towards the coast. That may be due to some difference in bed friction. The magnitude of southward flow was similar to that of the northward flow. That indicated no role of wind, which could not force the surface water. At times, the u-component of simulated current showed some variation with the measured current. The residual currents were estimated after removing the tides from the measured data. The upwelling phenomenon [12] was absent during this period. The result was in agreement with the results of modelling that the coastal current would influence the tidally driven flow, though marginally. The simulated current showed that the currents were having distinct variability corresponding to the tide. It may be noted that in the pole ward flow, density driven flow was the existent force that drives the surface water northwards.

Fig 9: Velocity vectors depicting the instantaneous flow during ebb tide

4.3) SALINITY Point 2 was located at the South West corner of the

domain where boundary value of west side was set to 33 psu and south was also set to 33 psu. So we see an expected increase in the value from 32 psu to 33 psu .While point 3 was on the South east side of the domain where boundary value of east side was set to 32.7 psu and south as said was set to 33 psu. So here also there was expected increase of value from 32ap approaching towards 32.7 psu

Fig 10: Shows variation in salinity at point 2 and point 3

V. CONCLUSION The study examines the hydrodynamics of Singapore

strait and Johor strait using MIKE 3 FM HD model. Results obtained from model simulation matched very well with the measurements. Hence, the model was further used for the simulation of hydrodynamics for other days and points also. The simulated hydrodynamics reasonably agreed with most of the earlier studies. The tidal flows were modelled accurately on coarse grids since they were large-scale processes. The water level showed very marginal variation. But, it was the current that have marked particular pattern of flow. The circulation along the nearshore region was purely south west during NE monsoon period. The model has provided a general understanding of the surface flows off the Singapore and Johor Strait. The reason for variability of currents from the actual measurements may be due to the usage of lower resolution bathymetry[13] and limitations of the model to generate precise HD of the region. The Hydrodynamic results can be further used to study a wide range of phenomena related to hydrodynamics, such as water quality, heat and salt transport and sediment transport processes[14]. The reason for initial fluctuations in the result of water level was due to the warm up period of the model which was around 6 time steps i.e., 1 hrs 30 minutes.

The Model also provided specifications of the Temperature and salinity off the Singapore region and Johor Strait. Few variations were observed in the data. An almost continuous increase of Salinity was observed from the North to South West (or south east) direction during the first 6 hrs of flow (The period for which salinity was calculated).Salinity results were in according to the measured value).

299299

REFERENCES [1] Numerical simulation of flow off Ratnagiri, west coast of India by Dr

P Vethamony and Jaffer Shariff. [2] MIKE 3 FM, user guide and reference manual, (2001). Reference

manual. Denmark: 70 pp. [3] MIKE 3 FM, user guide and reference manual, (2003). Reference

manual. Denmark: 40 pp. [4] Pond, S., Picard GL. (1983). Introductory dynamical oceanography,

Pergamon Press. [5] Shetye, S. R., Govea AD, Shenoi SSC, Micheal GS, Sundar D,

Almeida AM, Santanam K. (1991). “The coastal current off western India during the northeast monsoon.” Deep-Sea Research 38: 1517-1529.

[6] Shetye, S. R., Shenoi SSC, Antony MK, Krishna Kumar V. (1985). “Monthly-mean wind stress over along the coast of north Indian Ocean.” Proc. Indian Acad. Sci. (Earth Planet. Sci 94: 129-137).

[7] Shankar, D., Vinayachandran PN, Unnikrishnan AS. (2001). “The monsoon currents in the north Indian Ocean.”.

[8] Shankar, D. (2000). “Seasonal cycle of sea level and currents along the coast of India.” CURRENT SCIENCE 78.

[9] MIKE 3 FM tidal analysis and prediction module environment (2007). Danish hydraulic institute.

[10] HD, U. G. a. R. M. f. M. (2007). Reference Manual. DHI Software 2007. Denmark: 90 pp.

[11] David Huntley, A., Eduardo Siegle, Mark Davidson A. (2002). “Modelling water surface topography at a complex Inlet System – Teignmouth, UK.” Journal of Coastal Research 36: 675-685.

[12] Environment and Pollution Law Manual by S.K.Mohanty. [13] Waves, Tides and Shallow-Water Processes by the Open University. [14] Hydrodynamics and Transport of Water Quality Modelling by James

L.Martin.

300300