Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Michael T. W. TAY

Michael T.W. Tay

Thesis

Submitted in partial fulfilment of the requirements for the award of the degree of Doctor of Philosophy of the University of Portsmouth.

2018

Abstract The engineering management of river channel entrances has often been regarded as a vast and complex subject matter. Understanding the morphodynamic nature of a river channel entrance is the key in solving water-related disasters, which is a common problem in all seasonally influenced tropical countries. As a result of the abrupt increase in population centred along coastal areas, human interventions have affected the natural flows of the river in many ways, mainly as a consequence of urbanisation altering the changes in the supply of sediments. In tropical countries River channel entrances are usually seasonally driven. This has caused a pronounced impact on the topography of river channel entrances especially in the East Coast of Borneo which is influenced by both North East (NE) and South West (SW) monsoons on top of related parameters such as waves and tides, sedimentation as well as climate change. Advanced numerical modelling techniques are frequently used as a leading approach to investigate the complicated nature of river channel entrances to represent the actual conditions of a designated area. Furthermore, the conventional ‘command and control’ approach of river channel entrance management has largely failed in this region due to the lack of process- based understanding of river channel entrances.

This research focuses on creating a mesoscale numerical modelling method by integrating a reductionist and reduced complexity model approach, along with local seasonal conditions to represent the Petagas River channel entrance in demonstrating an actual river channel entrance. In assisting the modelling method, data and data- driven analysis is emphasised because the available resources and the subsequent adopted management strategies in solving these problems varies from country to country. The Petagas River channel entrance model is established with a set of numerical equilibrium equations and calibrated parameters to reflect the features of regional conditions with the combination of local validated modelling parameters. The model accurately represents the effects of both NE and SW monsoons in terms of hydraulics and sediment dynamics at the Petagas River channel entrance, as verified by site data and satellite imagery.

The validated model serves as a predictive tool to evaluate various management solutions, which eventually helped in proposing the most viable solution. The proposed management solution includes a training channel and a breakwater resulting in an improved flushing capacity, thus preventing sedimentation and upstream flooding, allowing proper navigation through the river. This adopted solution is in accordance with local regulations and management plans that the local government adopt. In conclusion, the accurate investigational result from this numerical model successfully demonstrates its scientific role in solving problems related to a dynamic river channel entrance, and can serve as a numerical modelling approach for solving similar river channel entrance problems especially in tropical countries with predominant inter-annual seasonal variations.

Declaration Whilst registered as a candidate for the above degree, I have not been registered for any other research award. The results and conclusions embodied in this thesis are the work of the named candidate and have not been submitted for any other academic award.

List of Tables

TABLE 1.1 CHARACTERISTICS OF DELTAIC DEPOSITIONAL SYSTEMS (GALLOWAY 1975) ...... 3 TABLE 5.1 CONSTITUENT ANALYSIS FOR KOTA KINABALU PORT BASE IN MALAYSIA TIDE TABLE 2010 (DEPARTMENT OF SURVEY AND MAPPING MALAYSIA, 2010) ...... 78 TABLE 5.2 TIDAL PLANES FOR KOTA KINABALU PORT (DEPARTMENT OF SURVEY AND MAPPING MALAYSIA, 2010) ...... 78 TABLE 5.3 RMSE WATER LEVEL COMPARISON TO KOTA KINABALU TIDAL STATION ...... 97 TABLE 5.4 RMSE WATER LEVEL COMPARISON TO KOTA KINABALU TIDAL STATION ...... 97 TABLE 6.1 ROOT MEAN SQUARE ERROR (RMSE) OF SIMULATED AND MEASURED CURRENT SPEED, DIRECTION AND WATER LEVEL...... 118 TABLE 6.2 EXTREME STORM SURGE LEVEL (ABOVE MSL) CORRESPONDING TO 2, 5, 10, 50 AND 100 YEARS RETURN PERIODS FOR EXISTING CONDITIONS AT SELECTED LOCATIONS...... 134 TABLE 6.3 OFFSHORE DESIGN WAVE CONDITIONS BASED ON GROW WAVE DATA ...... 136 TABLE 6.4 EXTREME WAVE SETUP (ABOVE MSL) CORRESPONDING TO 2, 5, 10, 50 AND 100 YEARS RETURN PERIODS FOR EXISTING CONDITIONS AT SELECTED LOCATIONS...... 138 TABLE 6.5 RECOMMENDED DESIGN WATER LEVEL (ABOVE MSL) AT BREAKWATER ENTRANCE...... 139 TABLE 6.6 RECOMMENDED DESIGN WATER LEVEL (ABOVE MSL) AT END OF TRAINING WALLS...... 139 TABLE 6.7 SIGNIFICANT WAVE HEIGHTS FOR DESIGN DETERMINED AT BREAKWATER ENTRANCE AND END OF TRAINING WALLS. . 140

iii

List of Figures

FIGURE 1.1 TRIANGULAR CLASSIFICATION OF DELTAIC DEPOSITIONAL SYSTEMS (MODIFIED AFTER GALLOWAY (1975))...... 2 FIGURE 2.1 BASIC WAVE NOMENCLATURE (SAWARAGI, 1995)...... 8 FIGURE 2.2 WAVE CLASSIFICATION BY FREQUENCY (AFTER KINSMAN, 1965) ...... 9 FIGURE 2.3 FACTORS AFFECTING WIND WAVE DEVELOPMENT (BROOKS/COLE, 2002) ...... 9 FIGURE 2.4 WAVES ENTERING SHALLOW WATER (SVERDRUP & ARMBRUST, 2009) ...... 10 FIGURE 2.5 WAVE REFRACTION DIAGRAM (BROOKS/COLE, 2002) ...... 11 FIGURE 2.6 WAVE REFRACTION DIAGRAM (CEM, 2006) ...... 12 FIGURE 2.7 WAVE DIFFRACTION PATTERNS AROUND BREAKWATERS (JOHNSON, 1951) ...... 13 FIGURE 2.8 TYPES OF WAVE BREAKING (SARPKAYA & ISAACSON, 1981) ...... 14 FIGURE 2.9 APPLICATION OF VARIOUS WAVE THEORIES (LE MÉHAUTÉ, 1976) ...... 16 FIGURE 2.10 MODELLING APPROACH FRAMEWORK FOR THE SIMULATION OF A RIVER CHANNEL ENTRANCE...... 22 FIGURE 2.11 DEFINITIONS OF CONTROL VOLUMES FOR CELL CENTRED FINITE VOLUME METHOD IN QUADRANGULAR AND TRIANGULAR MESH ...... 29 FIGURE 2.12 EXAMPLE OF MODELLED PHYSICAL PROCESSES (DHI C, 2012) ...... 32 FIGURE 2.13 MANAGEMENT UNIT C5-19 EXTRACTED FROM ...... 36 FIGURE 2.14 COASTAL PROTECTION CONSTRUCTED BY LOCAL RESIDENTS MADE OUT OF USED TYRES ...... 37 FIGURE 2.15 ARMOURED ROCK REVETMENT (SNH, 2000A)...... 39 FIGURE 2.16 TETRAPOD REVETMENT (PATEL, 2015)...... 40 FIGURE 2.17 TIMBER REVETMENT -BREASTWORK (SNH, 2000B) ...... 40 FIGURE 2.18 TYPICAL COASTAL REVETMENT CROSS-SECTION (PILE BUCK, 2014) ...... 41 FIGURE 2.19 EROSION OCCURRENCE AT ADJACENT DUNE FRONT AS A RESULT OF THE CONSTRUCTION OF COASTAL STRUCTURE 42 FIGURE 2.20 REVETMENT CONSTRUCTION USING LORRY AND EXCAVATOR (VARELA, 2016)...... 44 FIGURE 2.21 REVETMENT CONSTRUCTION USING BARGE AS TRANSPORT (ARENA KISH, 2015)...... 44 FIGURE 2.22 RUBBLE MOUND BREAKWATER IN CIVITAVECCHIA (ITO, 1969) ...... 45 FIGURE 2.23 OVERVIEW OF BREAKWATER ARMOUR UNITS (MUTTRAY & REEDIJK, 2009) ...... 46 FIGURE 2.24 RUBBLE MOUND/GRAVITY BREAKWATER (ZABEZ, 2015) ...... 47 FIGURE 2.25 VERTICAL BREAKWATER (ZABEZ, 2015)...... 48 FIGURE 2.26 CURTAIN WALL BREAKWATER (EVANS, 1955)...... 48 FIGURE 2.27 SHEET PILE BREAKWATER (EVANS, 1955)...... 49 FIGURE 2.28 FLOATING BREAKWATER (EVANS, 1955)...... 49 FIGURE 2.29 PORT OF TANJUNG PELEPAS IN MALAYSIA WITH DREDGED CHANNEL ALLOWING HANDLING OF MULTIPLE CARGO VESSELS BIRTHING (PELABUHAN TANJUNG PELEPAS, 2012)...... 51 FIGURE 2.30 CUTTER HEAD SUCTION DREDGER AT WORK IN TERENGGANU RIVER CHANNEL ENTRANCE MALAYSIA (DR. NIK & ASSOCIATES, 2014) ...... 52 FIGURE 3.1 A MAP OF THE PETAGAS RIVER CHANNEL ENTRANCE LOCATION (SAT IMAGE FROM GOOGLE MAPS, 2009) ...... 55 FIGURE 3.2 OBLIQUE AERIAL PHOTO VIEW OF PETAGAS RIVER CHANNEL ENTRANCE ...... 56 FIGURE 3.3 SATELLITE IMAGERY AT THE VICINITY OF THE SITE (A) CAPTURED YR. 2002 BEFORE THE FURTHER RECLAMATION (B) CAPTURED YR.2010 AFTER THE FURTHER RECLAMATION OF YR. 2006 (GOOGLE MAPS, 2015)...... 58 FIGURE 3.4 PREDOMINANT SURFACE CURRENTS IN JANUARY FROM SOUTH CHINA SEA PLOT (GROWMODEL, 2009)...... 61 FIGURE 3.5 PREDOMINANT SURFACE CURRENTS IN JULY FROM SOUTH CHINA SEA PLOT (GROWMODEL, 2009)...... 62 FIGURE 4.1 BATHYMETRIC AND TOPOGRAPHIC SURVEY AT THE VICINITY OF PETAGAS RIVER ENTRANCE. THE INDIVIDUAL SURVEY POINTS ARE AS INDICATED IN THE FIGURE...... 67 FIGURE 4.2 LOCATION OF ADCPS ...... 69 FIGURE 4.3 RBR TGR-1050 TIDE GAUGE ...... 71 FIGURE 4.4 TIDE GAUGE INSTALLED AT PETAGAS RIVER CHANNEL ENTRANCE (VIEW FROM SEA) ...... 71 FIGURE 4.5 DIAGRAM OF CONNECTIONS BETWEEN TEMPORARY BENCHMARK, MARKING AT POLE AND SENSOR ...... 72 FIGURE 4.6 SEDIMENT GRAB SAMPLE LOCATIONS AND MEAN GRAIN SIZE AT THE VICINITY OF THE PETAGAS RIVER CHANNEL ENTRANCE ...... 74 FIGURE 5.1 FLOW DIAGRAM ON NUMERICAL MODELLING APPROACH ...... 77 FIGURE 5.2 TIME SERIES PLOT OF SIMULATED TIDAL LEVEL AT KOTA KINABALU ...... 79 FIGURE 5.3 TYPICAL WIND AND PRESSURE DISTRIBUTION IN JANUARY FROM SOUTH CHINA SEA PLOT ...... 80 iv

FIGURE 5.4 TYPICAL WIND AND PRESSURE DISTRIBUTION IN JULY FROM SOUTH CHINA SEA PLOT ...... 81 FIGURE 5.5 MONTHLY WIND ROSES (JANUARY TO JUNE) BASED ON HINDCAST DATA (1970 - 2009) ...... 82 FIGURE 5.6 MONTHLY WIND ROSES (JULY TO DECEMBER) BASED ON HINDCAST DATA (1970 - 2009) ...... 83 FIGURE 5.7 MONTHLY WAVE ROSES (JANUARY TO JUNE) BASED ON HINDCAST DATA (1970 - 2009) ...... 84 FIGURE 5.8 MONTHLY WAVE ROSES (JULY TO DECEMBER) BASED ON HINDCAST DATA (1970 - 2009) ...... 85 FIGURE 5.9 OFFSHORE WAVE ENERGY DISTRIBUTION AS A FUNCTION OF SIGNIFICANT WAVE HEIGHT (A), PEAK WAVE PERIOD (B) AND MEAN WAVE DIRECTION (C)...... 86 FIGURE 5.10 COMPARISON OF BATHYMETRIC SURVEYS FROM YEAR 2006 (TOP LEFT), 2009 (TOP RIGHT) AND 2010 (BOTTOM)...... 88 FIGURE 5.11 CHANGES IN BED LEVELS AT THE RIVER CHANNEL ENTRANCE FROM 2006 TO 2010 ...... 89 FIGURE 5.12 OVERVIEW OF MODEL MESH AREA ...... 92 FIGURE 5.13 LOCAL MODEL MESH AREA ...... 93 FIGURE 5.14 LOCATIONS OF CALIBRATED TIDAL STATIONS WITHIN REGIONAL MODEL ...... 95 FIGURE 5.15 MIKE 21 FLOW MODEL FM, HYDRODYNAMIC MODULE: BED RESISTANCE ...... 96 FIGURE 5.16 MIKE 21 FLOW MODEL FM, HYDRODYNAMIC MODULE: EDDY VISCOSITY (HORIZONTAL) CONSTANT VALUE INPUT ...... 98 FIGURE 5.17 MIKE 21 FLOW MODEL FM, HYDRODYNAMIC MODULE: WIND FRICTION ...... 98 FIGURE 5.18 DETAILED VIEW OF BATHYMETRY FOR THE EXISTING CONDITION IN THE VICINITY OF THE RIVER CHANNEL ENTRANCE...... 101 FIGURE 5.19 DETAILED VIEW OF BATHYMETRY FOR PROPOSED LAYOUT 1 FOR RIVER CHANNEL ENTRANCE IMPROVEMENT WORKS (BROWN AREA INDICATES LAND TO BE RECLAIMED)...... 102 FIGURE 5.20 DETAILED VIEW OF BATHYMETRY FOR PROPOSED LAYOUT 2 FOR RIVER CHANNEL ENTRANCE IMPROVEMENT WORKS (BROWN AREAS INDICATE LAND TO BE RECLAIMED) ...... 103 FIGURE 5.21 DETAILED VIEW OF BATHYMETRY FOR PROPOSED LAYOUT 3 FOR RIVER CHANNEL ENTRANCE IMPROVEMENT WORKS (BROWN AREAS INDICATE LAND TO BE RECLAIMED)...... 104 FIGURE 5.22 DETAILED VIEW OF BATHYMETRY FOR PROPOSED LAYOUT 4 FOR RIVER CHANNEL ENTRANCE IMPROVEMENT WORKS (BROWN AREAS INDICATE LAND TO BE RECLAIMED)...... 105 FIGURE 5.23 DETAILED VIEW OF BATHYMETRY FOR PROPOSED LAYOUT 5 FOR RIVER CHANNEL ENTRANCE IMPROVEMENT WORKS (BROWN AREAS INDICATE LAND TO BE RECLAIMED)...... 106 FIGURE 6.1 (A) OVERVIEW OF TRANSFORMATION OF WAVES FROM OFFSHORE TO NEARSHORE FOR TYPICAL NE MONSOON CONDITION (B) DETAILED VIEW OF THE TRANSFORMATION OF WAVES FROM OFFSHORE TO NEARSHORE FOR TYPICAL NE MONSOON CONDITION (C) OVERVIEW OF TRANSFORMATION OF WAVES FROM OFFSHORE TO NEARSHORE FOR TYPICAL SW MONSOON CONDITION (D) DETAILED VIEW OF TRANSFORMATION OF WAVES FROM OFFSHORE TO NEARSHORE FOR TYPICAL SW MONSOON CONDITION...... 108 FIGURE 6.2 OVERVIEW OF EXISTING WAVE PATTERN FOR HIGH (LEFT) AND LOW (RIGHT) TIDE DURING NE MONSOON...... 109 FIGURE 6.3 EXISTING WAVE PATTERN AT VICINITY SITE FOR HIGH (LEFT) AND LOW (RIGHT) TIDE DURING NE MONSOON...... 110 FIGURE 6.4 OVERVIEW OF EXISTING WAVE PATTERN FOR HIGH (LEFT) AND LOW (RIGHT) TIDE DURING SW MONSOON...... 110 FIGURE 6.5 EXISTING WAVE PATTERN AT VICINITY SITE FOR HIGH (LEFT) AND LOW (RIGHT) TIDE DURING SW MONSOON...... 111 FIGURE 6.6 COMPARISON OF PREDICTED AND SIMULATED TIMESERIES FOR THE 4 TIDAL STATIONS NAMELY PULAU MANTANI BESAR, PULAU MANGALUM, KOTA KINABALU AND KUALA PAPAR ...... 113 FIGURE 6.7 COMPARISON OF MEASURED AND SIMULATED CURRENT AT ADCP 1 (DEEP WATER) ...... 115 FIGURE 6.8 COMPARISON OF MEASURED AND SIMULATED CURRENT AT ADCP 2 (SHALLOW WATER) ...... 116 FIGURE 6.9 COMPARISON OF MEASURED AND SIMULATED CURRENT AND WATER LEVEL IN THE RIVER ...... 117 FIGURE 6.10 DETAIL VIEW OF WAVE PATTERN COMPARISON BETWEEN EXISTING CONDITION AND CONCEPTUAL LAYOUTS DURING NE MONSOON AT HIGH TIDE...... 120 FIGURE 6.11 DETAIL VIEW OF WAVE PATTERN COMPARISON BETWEEN EXISTING CONDITION AND CONCEPTUAL LAYOUTS SW MONSOON AT HIGH TIDE...... 121 FIGURE 6.12 MANAGEMENT MEASURES PROPOSED FOR PETAGAS RIVER CHANNEL ENTRANCE BY MAINTAINING THE EXISTING RIVER CHANNEL ALIGNMENT WITH BREAKWATERS (GOOGLE MAPS, 2009) ...... 122 FIGURE 6.13 CURRENT SPEED AND DIRECTION FOR NE MONSOON FOR (A) EXISTING CASE DURING FLOOD TIDE CONDITIONS (B) FOLLOWING IMPROVEMENT WORKS DURING FLOOD TIDE CONDITIONS (C) EXISTING CASE DURING EBB TIDE (D) FOLLOWING IMPROVEMENT WORKS DURING EBB TIDE...... 124

FIGURE 6.14 CURRENT SPEED AND DIRECTION FOR SW MONSOON FOR (A) EXISTING CASE DURING FLOOD TIDE CONDITIONS (B) FOLLOWING IMPROVEMENT WORKS DURING FLOOD TIDE CONDITIONS (C) EXISTING CASE DURING EBB TIDE CONDITIONS (D) FOLLOWING IMPROVEMENT WORKS DURING EBB TIDE CONDITIONS...... 125 FIGURE 6.15 COMPARISON OF MAXIMUM CONCENTRATIONS OF TOTAL SUSPENDED SEDIMENTS (TSS) DURING THE NE MONSOON PERIOD FOR (A) EXISTING CONDITION AND (B) FOLLOWING IMPROVEMENT WORKS. COMPARISON OF MAXIMUM CONCENTRATIONS OF TOTAL SUSPENDED SEDIMENTS (TSS) DURING THE SW MONSOON PERIOD FOR (C) EXISTING CONDITION AND (D) FOLLOWING IMPROVEMENT WORKS. COMPARISON OF MAXIMUM CONCENTRATIONS OF TOTAL SUSPENDED SEDIMENTS (TSS) DURING THE INTER MONSOON PERIOD FOR (E) EXISTING CONDITION AND (F) FOLLOWING IMPROVEMENT WORKS...... 129 FIGURE 6.16 COMPARISON OF BED THICKNESS CHANGES (NET DEPOSITION) DURING THE NE MONSOON PERIOD FOR (A) EXISTING CONDITION AND (B) FOLLOWING IMPROVEMENT WORKS. COMPARISON OF BED THICKNESS CHANGES (NET DEPOSITION) DURING THE SW MONSOON PERIOD FOR (C) EXISTING CONDITION AND (D) FOLLOWING IMPROVEMENT WORKS. COMPARISON OF MAXIMUM CONCENTRATIONS OF BED THICKNESS CHANGES (NET DEPOSITION) DURING THE INTER MONSOON PERIOD FOR (E) EXISTING CONDITION AND (F) FOLLOWING IMPROVEMENT WORKS. IN THIS MONSOON CONDITION, IT SHOWS THE ACCURACY OF THE MODEL BY COMPARING TO THE SATELLITE IMAGE AS SHOWN IN FIGURE 6.196.19. THIS RESULTS SHOWN IS BASED THE OUTCOME OF THE MODEL RUN OVER A PARTICULAR MONSOON SEASON...... 131 FIGURE 6.17 SATELLITE IMAGE OF CURRENT RIVER CHANNEL ENTRANCE CONDITION (GOOGLE MAPS 2009) ...... 132 FIGURE 6.18 EXAMPLE OF STORM SURGE LEVELS WITH WIND FROM WEST (270ON). LEFT: TYPHOON GREG WITH CONSTANT WIND OF 22.6M/S (MOST SEVERE). RIGHT: STORM WITH CONSTANT WIND SPEED OF 15.6M/S (NINTH MOST SEVERE)...... 135 FIGURE 6.19 WAVE SETUP DUE TO 100-YEAR DESIGN WAVE WITH DIRECTIONS CORRESPONDING TO DOMINANT NE (LEFT) AND SW (RIGHT) MONSOON DIRECTIONS...... 136 FIGURE 6.20 DETAIL VIEW OF WAVE SETUP SUNGAI PETAGAS RIVER CHANNEL ENTRANCE DUE TO 100-YEAR DESIGN WAVE FROM DOMINANT NE MONSOON DIRECTION...... 137 FIGURE 6.21 DETAIL VIEW OF WAVE SETUP SUNGAI PETAGAS RIVER CHANNEL ENTRANCE DUE TO 100-YEAR DESIGN WAVE FROM DOMINANT SW MONSOON DIRECTION ...... 137

Definitions and Abbreviations Definitions

Definitions Description Current Direction Direction in which the water is flowing. Sedimentation Deposition of sediments, i.e. sands, silts or clays Water Depth Depth relative to MSL (unless other datum specified). Wind Direction Direction from where the wind is blowing Wind Speed Given 10 m above mean sea level (unless other level specified) Predictive Model Model that forecasts the characteristics of a site based on available information.

Abbreviations

Abbreviations Description 2D Two dimensional 3D Three dimensional ADCP Acoustic Doppler Current Profiler GROW Global Reanalysis of Ocean Waves by Oceanology International KKIA Kota Kinabalu International Airport DHI DHI Water & Environment ICZM Integrated Coastal Zone Management HD Hydrodynamic ADCP Acoustic Doppler Current Profiler IM Inter-monsoon KMS Global Tide Model (National Survey and Cadastre Agency, Denmark) MT Mud Transport SW Spectral Waves NOAA National Oceanic and Atmospheric Administration, USA RMN Royal Malaysian Navy ASCII American Standard Code for Information Interchange BRSO Borneo Rectified Skew Orthomorphic – Borneo Map Projection System XLS Excel File Format TOPEX/Poseidon (T/P) Ocean Surface Topography from Space ASMA Alam Sekitar Malaysia Sdn. Bhd. RCM Reduced Complexity Model

EVA Extreme Value Analysis GIS Geographic Information System FVM Finite Volume Method CC-FVM Cell Centred Finite Volume Method WCI Wave Current Interaction ACES Automated Coastal Engineering System ADI Alternating Directional Implicit BCR Benefit Coast Ratio NCES National Coastal Erosion Study ISMP Integrated Shoreline Management Plan TSS Total Suspended Sediment

vii

Compass Directions

Abbreviations Description E East ENE East-northeast ESE East-southeast N North NE Northeast NNE North-northeast NNW North-northwest NW Northwest SE Southeast SSE South-southeast SSW South-southwest SW Southwest W West WNW West-northwest WSW West-southwest

Water and Tidal Levels

Abbreviations Description CD Chart Datum HAT Highest Astronomical Tide LAT Lowest Astronomical Tide LSD Land Survey Datum MHHW Mean Higher High Water MLLW Mean Lower Low Water MSL Mean Sea Level WL Water level

viii

For Mom and Dad

Acknowledgements This thesis would have not been possible without the support and encouragement of many. Firstly, I would like to express my sincere gratitude to my main supervisor Dr. Jiye Chen for the continuous support of my Ph.D study and related research, for his patience, motivation, and immense knowledge. His guidance helped me in all the time of research and writing of this thesis. I could not have imagined having a better supervisor and mentor for my Ph.D study. I would like to express my gratitude specifically to Dr. Steven Mitchell, a member of my supervisory team; your generous and insightful input and expertise on the subject were crucial throughout the research and in the production of this thesis. Dr John Williams for your encouragement throughout the research process.

I wish to thank all my past and present colleagues at the Department of Irrigation and Drainage Malaysia for giving me much insight to the local condition and advice to the research. Datuk Ir. Ng Chee Hing the Director of DID for allowing me to adopt Petagas River channel entrance project as a real-life example and allowing this research as a contribution to humanity. Jurukur Vitus Tan and Associates for survey and field data, to the Department of Survey and Mapping (JUPEM) for supplying the tidal datasets, and to Kota Kinabalu International Airport (KKIA) for the wind data sets. DHI Water & Environment provided technical advice and assistance with the numerical modelling.

Special thanks to my family. Words cannot express how grateful I am to my parents for all of the sacrifices that you have made on my behalf. Your prayer for me was what sustained me this far. I would also like to thank all my friends who supported me and incentivied me to strive towards my goal.

MICHAEL TAY TSHUNG WANG KOTA KINABALU, JUNE 2016

Contents

Abstract ...... i Declaration ...... ii List of Tables ...... iii List of Figures ...... iv Definitions and Abbreviations ...... vii Acknowledgements ...... x Contents ...... xi 1 Introduction ...... 1 1.1 Background ...... 1 1.2 Research Questions ...... 4 1.3 Aim and Objectives ...... 5 1.4 Thesis Structure ...... 6 2 Literature Review ...... 7 2.1 Wave Generation ...... 8 2.2 Hydrodynamics ...... 19 2.3 Modelling Approaches ...... 21 2.4 Data and Data-driven Approaches ...... 24 2.5 Mapping and meshing method of coupled model ...... 26 2.5.1 Mapping ...... 26 2.5.2 Meshing ...... 28 2.6 Reductionist Models ...... 30 2.7 Sediment Deposition ...... 32 2.8 Management Solution Consideration ...... 35 2.9 Engineering Solutions Considered ...... 39 2.9.1 Revetment ...... 39 2.9.2 Breakwater ...... 45 2.9.3 Dredging ...... 50 2.10 Knowledge Gap ...... 53 3 Description of the research area ...... 54 3.1 Geomorphology of the Site ...... 54 3.2 Metocean Condition ...... 59 3.2.1 Tides ...... 59 xi

3.2.2 Monsoon Seasons ...... 60 3.2.3 Offshore Wind and Wave Pattern ...... 60 3.2.4 Regional Current Conditions ...... 61 4 Data Utilisation ...... 63 4.1 Bathymetric and Topographic Survey ...... 67 4.2 Acoustic Doppler and Current Profile Data Collection ...... 68 4.3 Wind Conditions ...... 70 4.4 Tides ...... 71 4.5 River Discharge ...... 73 4.6 Sediment Properties ...... 74 4.7 Wave Conditions ...... 75 5 Methodology...... 76 5.1 Data and Data-driven Analysis ...... 78 5.1.1 Tidal Regime Analysis ...... 78 5.1.2 Offshore Wind Climate Analysis ...... 79 5.1.3 Offshore Wave Climate ...... 84 5.1.4 Bed-Changes Process ...... 88 5.2 Mapping and Meshing...... 90 5.3 Reduced Complexity Model ...... 94 5.4 Reductionist Model ...... 95 5.4.1 Hydrodynamic Modelling ...... 95 5.4.2 Sediment Dynamics Model ...... 100 5.4.3 Conceptual Layout Consideration ...... 101 6 Outcome and Analysis for Engineering and Management Solutions...... 107 6.1 Existing Morphodynamic Conditions ...... 107 6.1.1 Wave Conditions ...... 107 6.2 Model Calibration and Verification Outcome ...... 112 6.2.1 Model Calibration ...... 112 6.2.2 Model Verification ...... 114 6.3 Conceptual Layout Consideration Outcome ...... 119 6.3.1 Wave Condition Comparison ...... 119 6.3.2 Management Solution Chosen ...... 122 6.3.3 Design Conditions for Management Solutions ...... 133 7 Answer to Research Questions ...... 141

xii

7.1 Answer to Research Question 1 ...... 142 7.2 Answer to Research Question 2 ...... 144 7.3 Answer to Research Question 3 ...... 145 8 Conclusions ...... 148 8.1 Concluding Statements ...... 148 8.2 Research Limitation and Recommendation for Further Studies ...... 150 References ...... 152

xiii

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

1 Introduction 1.1 Background The processes that occur in and around river channel entrances comprise a vast and very complex subject that has been studied by countless researchers over a long period of time (Anthony et al., 2013; Brack et al., 2000; Pruszak et al., 2005; Samaras & Koutitas, 2014a; Tay et al., 2016; N. Wang et al., 2017; Wright et al., 1980). Changes in the morphology of river channel entrances have become more dramatic over the last century as a result of sea-level change, fluvial processes and urbanisation (Pruszak et al., 2005). The configuration and landform/shape of a river channel entrance also strongly depends on the wave energy adjacent to the shore (Wright & Coleman, 1972). Subsequently, the variation in morphology as a function of ocean waves and river discharge has also been considered in the formation of river channel entrance discharge systems (Wright & Coleman, 1973). The morphology of a river channel entrance is affected by tidal currents and waves, in addition to the sources from ocean energy, which include various types of currents, storm surge, and climate change causing eustatic rises in sea level (Bonaldo et al., 2015; Galloway, 1975; Stanley & Warne, 1994; Wong & Codignotto, 2007). Most major river channel entrances, especially in Southeast Asia, formed during the Holocene period caused by the influx of large quantities of mobile sediments (Woodroffe et al., 2006).

Today, more than three billion people live in the coastal zone around the world, and this figure is increasing each year (Hinrichsen, 1998; SOE, 2011). The coastal zone is especially prone to conflicts because it is densely populated, has culturally importance and is the locus for growth in multiple sectors including industry, agriculture, transport, trade and tourism (Harvey & Caton, 2010). With increasing influence of human activities along the coastlines including urbanisation, there has been an increase in surface water run-off, causing an increase in river flow and sediment influx towards river channel entrances. This is due to the transportation of fine sediments from the upper water catchment during low flow conditions, which are subsequently flushed seawards under high freshwater flows (Mitchell et al., 2015). About 80% of the world coastal populated area is vulnerable to flooding, of which 38% is located in Asia, mostly in South East Asia (Veerbeek, 2007).

It has been claimed that the sustainability of a river channel entrance is more vulnerable to human intervention than to climate change (Syvitski, 2008). River discharge and the nature of the dynamic changes in river channel entrance are known to have changed over rather shorter time-spans. This is because of anthropogenic and climatic factors (Milliman, Farnsworth, & Albertin, 1998; Syvitski, Vörösmarty, Kettner, & Pamela, 2005a; Milliman J. D., Farnsworth, Jones, Xu, & Smith, 2008; Meade, 1996).

Human interventions, particularly in deltas, have caused the redefinitions of existing architectural river delta classification as shown in Figure 1.1 (Syvitski & Saito, 2007). 1

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

River Dominated

Wave Tide Dominated Dominated

Figure 1.1 Triangular classification of deltaic depositional systems (modified after Galloway (1975)).

According to Galloway, Figure 1.1 illustrates the subdivision classification of deltaic discharge regimes (Galloway, 1975). Table 1.1 further compares the morphologic and stratigraphic characteristics of the subdivision classification. This classification ultimately shows the categorisation of Petagas River as compared to other deltaic systems.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

River Wave Tide Dominated Dominated Dominated

Geometry Elongate to lobate Arcuate Estuarine to irregular

Channel Type Straight to sinuous Meandering Flaring straight to distributaries distributaries sinuous distributaries

Bulk Muddy to mixed Sandy Variable composition

Framework Distributary Coastal barrier Estuary fill and facies mouth bar and and beach ridge tidal sand ridges channel fill sands, sands delta margin sand sheet.

Framework Parallels Parallels Parallels orientation depositional slope depositional strike depositional slope

Table 1.1 Characteristics of Deltaic Depositional Systems (Galloway 1975)

Sediment deposition in river channel entrances involves several active processes. Only confined fluvial flow, wave flux, and tidal currents appear to be capable of transporting relevant sediment load from the catchment downstream and ultimately to the river channel entrance. Ultimately, sedimentation occurs in distributary channels and distributary mouth bars, coastal barriers, beach ridges and tidal estuaries. On the seaward side, oceanic currents, wind drift, density currents and storm surge have not been shown to have significant sediment dynamic effects within river channel entrances. Sand facies form the stratigraphic framework of a delta, and the genetic classification of delta based on its relation to tidal domination, wave domination and sediment load are important in differentiating the skeleton of a delta as shown in Table 1.1.

The future of a river channel entrance can only be determined if we truly understand its complex nature, generally by the combination of climatic and anthropogenic intervention (Le et al., 2007). Hence, better understanding of coastal river channel entrance morphology is crucial to improve coastal resource planning and management.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

1.2 Research Questions In order to demonstrate the relevant representation of a Numerical Modelling Approach for the Management of a Seasonally Influenced River Channel Entrance, the focus will be on the following research questions for the Petagas System:

I. Do the monsoon seasons influence the topography of the river channel entrance and if so how do they influence patterns of sedimentation? II. What is the optimum management solution to the sedimentation problem? III. Can this modelling technique be used as a reference for solving similar river channel entrances problems?

All the above research questions will be answered in length in Chapter 7.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

1.3 Aim and Objectives This research aims to establish a suitable numerical modelling technique in the simulation of river mouth problems typically in Sabah to help local government in achieving management solutions. This investigation consists of some aims given below to complete the target.

• Evaluate and Obtain necessary data for the purpose of numerical modelling, mapping and data-driven analysis; • Analyse seasonally influenced characteristics on Petagas River Channel Entrance Behaviour; • Carry out mapping and meshing of the site to identify critical area of interests; • Establish seasonal influenced wave model to represent the ocean dynamics using MIKE 21 Spectral Wave model; • Utilise a reductionist multi-model approach to set up the model, including the effect of monsoon seasons for modelling the sedimentation of a river mouth; • Assess the calibration and validation of the model using measured site data to ensure the accuracy of the model; • Set-up a predictive model and assess the effectiveness of various management solutions; • Determine a viable management solution to minimise the dynamic sedimentation behaviour of the Petagas River Channel Entrance; • Ensure the management plan is in accordance with the regulations of the local government.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

1.4 Thesis Structure The thesis is structured as an approach to numerical modelling, where Chapter 2 covers the literature on the major approach adopted as well as considering various management solution in relation to hydrodynamics and sediment dynamics in relation to standard industrial practice and local policy as shown in Section 2.8. Chapter 3 describes the overall topography and met ocean condition necessary to cover the Petagas River channel entrance. Chapter 4 describes the necessary modelling data in building a mesoscale model. Chapter 5 describes the methodology which elaborates on execution of modelling approach on data utilisation, model platform and model setup. In Chapter 6, Section 6.2 demonstrates the outcome from the calibration of the hydrodynamic model and further describes the effectiveness of the coupled model in calibration and verification approach and results of the comparison between actual site data and model simulation to illustrate the reliability of the model. Section 6.3 shows the outcome of the well represented reductionist model in terms of wave condition, it then introduces a set of management solutions aimed at minimising sedimentation behaviour in relation to inter-annual climate variations in the monsoon seasons, together with a demonstration of the effectiveness of the reduced complexity model approach. Chapter 7 provides answer to research questions at the early stages of the research with reference to Section 1.2. In the final chapter of the thesis (Chapter 8), conclusions are drawn, including the importance of numerical modelling in solving future river channel entrance problems and its ability to provide suggestions for policy-makers in managing a river channel entrance subjected to monsoon influences. Further recommendations for future research development and research limitations regarding Numerical modelling approaches for the management of seasonally influenced river channel entrances is also included in this chapter.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2 Literature Review

This literature review chapter provides insight into each of the core elements of the numerical modelling approach for the management of seasonally influenced river channel entrances. Section 2.1 explains the overall wave theory adopted by the model. Section 2.2 gives the overall hydrodynamic theory as well as an explanation of the calibration and verification of the model. Section 2.3 shows the overall modelling approaches in which the Reduced Complexity Model is also explained. The core data and data-driven approaches together with the Mapping and Meshing method are then further explained in Section 2.5 and 2.6 respectively, before attending to the actual numerical modelling methodology used.

Section 2.6 gives an overall explanation of the Reductionist Model approach. Section 2.7 explains the sediment dynamics method adopted in the model. The management solution criteria are then found in Section 2.8, which also considers the local management policy at the site. With the consideration of the previous sections, Section 2.9 shows the management solution considered. Lastly, Section 2.10 identifies the knowledge gap.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.1 Wave Generation Waves are the undulatory motion of a water surface with fluctuations, accompanied by local currents, accelerations and pressure fluctuations. Their simplest form is sinusoidal as shown in Figure 2.1:

Figure 2.1 Basic Wave Nomenclature (Sawaragi, 1995).

According to Figure 2.1, the high water levels are wave crests, the low levels are known as wave troughs. The vertical distance between a crest and trough is the wave height (H). The distance over which the wave pattern repeats itself is the wavelength (L). Waves propagate with a velocity C, and the time that is required for a wave to pass a particular location is the wave period (T). The inverse of the wave period is the wave frequency (f).

It is well known that water waves can be classified into capillary waves, ultra gravity waves, gravity waves, infragravity waves, long period waves and trans-tidal waves regarding wave period (Sawaragi, 1995). Capillary waves have very short wave periods (order 0.1 sec) to long waves with wave periods expressed in minutes or hours. Waves also vary in height from a few millimetres for capillary waves to a few metres for gravity waves and 10s of metres for long-period waves. A classification of wave frequency of various types of waves is shown in Figure 2.2. Gravity or wind generated waves are in the middle range of wave frequencies as described in Figure 2.2 (Kamphuis, 2000).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.2 Wave Classification by frequency (after Kinsman, 1965)

The initial formation of waves by the wind is described in the Miles-Phillips theory of wave generation developed in the 1960s (Phillips, 1962). Wind waves are gravity waves formed by the transfer of wind energy to water. Wind forces convert capillary waves to wind waves. They have periods between 1 to 30 sec, a wave height that is seldom greater than 10 m and mostly of the order of 1 m (Kamphuis, 2000). Waves generated locally by wind are known as sea waves, which have peaked crests and broad troughs. Wind waves continue to grow until the sea is fully developed or becomes limited by fetch restriction, wind duration or saturated with energy. After the wind waves leave the generating area of the fetch, they become smoother and lose their rough appearance, being then known as swell waves with the features of longer wavelength and smaller wave height. The three parameters that dictate the growth of wind waves are fetch F, wind velocity V and its duration 푡푑 as shown in Figure 2.3. The close relationship between wind and wave characteristics is reflected in a scale devised in 1805 by Rear Admiral Francis Beaufort, an Irish Navy officer, in the Beaufort scale (Sundar, 2016). In general, wind speed and direction are vital in the study of establishing weather windows, and the design, construction and maintenance of structures in the coastal environment.

Figure 2.3 Factors Affecting Wind Wave Development (Brooks/Cole, 2002)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

As a wave propagates from deep water towards the shoreline, the wavelength decreases, but at a slower rate than the rate of decrease of water depth. From the viewpoint of the relative water depth, h/L (h; the still water depth and L; the wavelength), the water wave can be classified into deep water waves (h/L greater than 0.5), intermediate/transitional water waves (h/L less than 0.5) and shallow water waves (h/L less than 0.05) as shown in Figure 28 (Robert, 1993).

Figure 2.4 Waves Entering Shallow Water (Sverdrup & Armbrust, 2009)

The Celerity of Wave C is derived as;

퐿 퐶 = Equation 2.1 푇

where C is celerity, L is wave length and T is wave period (Robert, 1993).

As waves enter shallow water, the wave speed and length decrease, the wave period and energy remain unchanged, while the wave height increases. The easiest conservative quantity to follow is the constant energy flux per wave, using the Unchanged Energy Flux Equation as highlighted by (Robert, 1993) as below;

Equation 2.2 퐹퐸 = 퐸퐶푛

Where 퐸 is average energy; 퐹퐸 is the energy flux and 퐶푛 is the celerity / wave phase velocity in a particular space.

Equating the energy flux in deep water ( 퐻표, 퐶표) to the energy flux at any shallow water location ( 퐻푥, 퐶푥) results in the general shoaling relationship (Robert, 1993);

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

퐻 1 퐶 1/2 Equation 2.3 푥 = | 표⌉ 퐻 2푛 퐶 표 푥

Therefore, by knowing the deep water height and period (퐻표, 푇표) and the bathymetry of a coastal region, the shoaling wave characteristics (퐻푥, 퐶푥, 퐿푥) can be calculated at any point, x, prior to breaking.

Wave refraction is the bending of the wave crest impinging at an angle to the coastline and the associated bathymetric iso-lines as described in Figure 2.5. This transformation is triggered by the friction induced to the water particles close to the sea bed as well as varying water depths underneath. Water crests in shallower water move more slowly than those in deeper water. As the wave crest in deeper water catches up, the curving of initially straight wave crests results in a tendency of the trajectory of the wave rays to veer away from being normal to the wave crest, ultimately becoming perpendicular to the bathymetric contours (Dean & Dalrymple, 2004).

Figure 2.5 Wave refraction diagram (Brooks/Cole, 2002)

Application of Snell’s law as shown in Equation 2.4 to a plane parallel to the bathymetry confirms that the wave crest tends to turn to align with the bathymetric contours (Benassai, 2006).

sin 휃 /푐 = sin 휃 /퐶 1 1 2 2 Equation 2.4

Where 휃1 is the wave direction in deep water and 휃2 is the wave direction in shallow water, and ∁1 represents the celerity in deep water; ∁2 is the celerity in shallow water.

However, most offshore bathymetry is both irregular and variable along the coastline, which can only be solved approximately by various numerical model techniques; one of the earliest is by Noda (Noda, 1974) using RCPWAVE (Ebersole, 1985), which is 11

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

based on a finite difference steady-state phase resolving model using linear wave theory and development has been ongoing since then. Application of ray tracing method that uses Snell’s law with idealised bathymetry has been introduced by the US Army Corps of Engineers by setting up a suite of computational programs, the Automated Coastal Engineering System-ACES (Meadows, 2003).

Considering two or more wave rays propagating towards parallel bathymetry as shown in Figure 2.6, the rays can either converge or diverge. The energy per unit area may increase (convergence) or decrease (divergence) as a function of the perpendicular distance of separation between wave rays 푏표 and푏푥 . By using the shoaling geometric Equation 2.5 highlighted by (Madsen & Sørensen, 1992), modifications can be made to account for convergence and divergence of wave rays

퐻 1 퐶 1/2 푥 = | 표⌉ Equation 2.5 퐻 2푛 퐶 표 푥

The wave height 퐻1 variation highlighted by (Svendsen & Hansen, 1977) depends on the sea bed morphology and at a given point and can be written as:

Equation 2.6 퐻 = 퐻 퐾 퐾 1 0 푠 푟

where 퐻0 is the deep water wave height, 퐾푠 is shoaling coefficient and 퐾푟 is the refraction coefficient.

Figure 2.6 Wave refraction diagram (CEM, 2006)

Wave diffraction occurs when waves encounter surface-piercing obstacles such as a breakwaters or similar barriers, which interrupt a portion of the regular wave train. The principle of diffraction pertaining to the design of breakwaters has considerable practical importance as discussed by Dunham (Johnson, 1951). After passing the obstacles,the waves carry energy and the wave crest into the shadow zone behind 12

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

the barrier. The turning of the waves into the sheltered region is due to the changes in wave height as shown in Figure 2.7.

There are three primary types of important wave-structure diffraction: (1) diffraction at the end of a single breakwater (semi-infinite); (2) diffraction through harbour entrance (gap diffraction); and (3) diffraction around an offshore breakwater. For each type, there is a geometric shadow zone on the sheltered side of the structure, a reflected wave zone on the front or incident wave side of the structure, and an “illuminated” zone in the area of direct wave propagation as described in Figure 2.7 The US Army Corps of Engineers RCPWAVE is a PC compatible program capable of doing diffraction calculations (Meadows, 2003).

Figure 2.7 Wave diffraction patterns around breakwaters (Johnson, 1951)

It has been suggested by (Byron, 2016) that there are two circumstances in which wave breaking occurs: (1) the wave becomes steeper causing the crest to collapse as it unable to support the local wave height, this generally happens when the wave height reaches 1/7 of the wavelength; or (2) the speed of the water that circles the crest of the wave exceeds the speed of the wave, which in turn triggers the water in the crest to overtake the waveform causing it to fall over and break. A rule of thumb is that a wave will break when its height exceeds ¾ of the local water depth. The height of the wave when it breaks will be about 1.5 times larger than its height in deep water.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

In solitary wave theory, a theory applicable to shallow water troughless waves, at the break point it holds Relative Breaking Depth 퐻푏 (Weishar & Byrne, 1972).

퐻 = 0.78ℎ Equation 2.7 푏 푏 where ℎ푏 denotes breaking condition

Wave breakers can be placed into four categories as in Figure 2.8: Spilling, Plunging,

Collapsing and Surging, which are based on the surf-similarity parameter 휉0 (Weishar & Byrne, 1972) :

휉 = tanβ(퐻 |퐿 )−1/2 Equation 2.8 0 표 표

Where β is the slope of a plane beach.

Spilling 휉0 < 0.5 Plunging 0.5 < 휉0< 3.3 Surging/Collapsing 휉0 > 3.

Figure 2.8 Types of Wave Breaking (Sarpkaya & Isaacson, 1981)

Plunging waves lose their waveform after breaking, in contrast with spilling breakers that maintain their waveform. This results in larger energy release by the plunging breakers with higher turbulence level and bed load agitation.

In the aspect of bottom friction, energy dissipation occurs, which leads to wave height reduction as the water depth becomes shallower (Benassai, 2006). The dissipation is the work done by the wave orbital velocity against the bottom turbulent shear stress →, which is expressed (Weishar & Byrne, 1972) as 휏

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

→ = 𝜌 퐶푓 |→| → 휏 푢 푢 Equation 2.9

where → is the turbulent shear stress, 𝜌 is the water density, 퐶푓 is the friction 휏 coefficient and → is the velocity immediately outside the bottom boundary layer. 푢

This is based on bottom dissipation in finite-depth water waves (Weishar & Byrne, 1972). There are three dissipation mechanisms, namely friction, percolation, and bottom motion in dissipating wave energy in shallow water depending on wave properties and bottom sediment properties (Hsiao & Shemdin, 1978).

Historically, wind-generated waves have been described by several different theoretical developments. A water wave theory is developed under the assumptions that the fluid is incompressible and inviscid, the wave motion is irrational, and the wave is periodic and uniform in time and space with respect to the wave period (T) and the wave height (H) (Sawaragi, 1995).

The most common approach is Stokes’ Wave Theory, introduced by Sir George Stokes in 1847 (Stokes, 1847), as a solution for nonlinear wave problems in intermediate and deep water by means of a low-order Stokes expansion perturbation series (e.g. up to 2nd, 3rd or 5th order). In fluid dynamics, a Stokes wave is a nonlinear and periodic surface wave on an inviscid fluid layer of constant mean depth. Higher orders of approximation of Stokes theory are constantly developed with the use of recent computer models to describe the finite higher amplitude waves more precisely. In contrast, Small Amplitude Wave Theory, first discussed by (Airy, 1845), is best used for small waves in deep water. Cnoidal wave theory by (Korteweg & deVries, 1895) is commonly applied in shallower water as it accounts for distortion of wave shape from the interaction of the fluid motion with the bottom. As the waves reach very shallow water and are about to break, Solitary wave theory by both (Boussinesq, 1872) and (Munk, 1949) can be used (Kamphuis, 2000).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.9 Application of Various Wave Theories (Le Méhauté, 1976)

In Fig 2.9, ℎ is mean water depth; H is wave height; 휏 is wave period and g is gravity acceleration.

The light blue area indicates the range of validity of Cnoidal wave theory; light yellow for Airy wave theory; and the dashed blue lines demarcate between the required order in Stokes' wave theory. The light-grey shading gives the range extension by numerical approximations using fifth-order stream-function theory, for high waves (H > ¼ Hbreaking).

Waves play an important role in coastal morphology as they greatly influence the behaviour of sediment transport in a coastal zone (Gambino & Sussman, 1998). To accurately predict the behaviour of the wave at a river channel entrance, the near shore wave condition is derived using offshore hindcast wave data. Long-term wave data are important in determining the nearshore wave conditions for design and assessment of sediment transport capacities.

The MIKE software developed by DHI is used for modelling based on numerical simulation. MIKE 21 in which the “21” represents a two-dimensional one-layer model, offering different modules for hydrodynamics, sediment transport, waves and water quality studies in rivers, estuaries, coastal and oceanic zones. MIKE 21 is constantly undergoing development and improvement to enhance user experience such as speeding up processing time (Aackermann et al., 2013) and improving the user 16

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

interface. Hence it is used to investigate the dynamics of the river channel entrance of the Petagas River.

First, the wave is generated in the model using MIKE 21 SW. In this process, reliable information on the wave climate at the site is probably the single most important factor in the accurate determination of the sediment transport at the location (Soomere et al., 2008). Time series of hindcast offshore wind and wave data from the GROW global wind wave model has been transformed to the site using MIKE 21 SW, resulting in the further improvement in the definition of the near shore directional wave statistics and longshore variations in wave energy. GROW (Global Reanalysis of Ocean Waves) is a global deep water wave model with a resolution of 0.625° in latitude by 1.25° in longitude. The model is driven by assimilated marine surface wind fields and validated against offshore wave measurements. The model is developed and run by Oceanweather Inc (GROW, 2009a).

The Cartesian coordinates conservation equation (DHI a, 2014) for wave action is;

휕푁 푠 + ∇ ∙ (푣⃗푁) = Equation 2.10 휕푡 𝜎

where N(x⃗⃗, σ, θ, t) is action density, x⃗⃗( x, y ) are Cartesian co-ordinates, v⃗⃗(cx, cy, cσ, cθ) is the propagation velocity of waves group in four-dimensional phase space x⃗⃗, σ, θ, and S is the source term for the energy balance equation. ∇= 4D differential operator in the x⃗⃗, σ, θ − space.

The energy Source term equation is used to represent the superposition of source functions describing various physical phenomena 푆 (DHI a, 2014) as;

푆 = 푆푥 + 푆푦 + 푆σ + 푆θ Equation 2.11

Where 푆푥 푎푛푑 푆푦 represent the generation of the wave velocity in x and y directions,

푆σ is the wave height energy transfer, 푆θ is the dissipation of wave energy due to depth.

The four characteristic propagation speeds in 푐푥 푎푛푑 푐푦 direction (DHI a, 2014) are influenced by the following;

푑푥⃗ Equation 2.12 푣⃗(푐 , 푐 ) = = 푐⃗ + 푈⃗⃗⃗ 푥 푦 푑푡 푔

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Where,

푐𝜎 𝑖푠 Propagation Speed in the 𝜎 direction as shown in (Equation 2. 13)

푐휃 is Propagation Speed in the 휃 direction as shown in (Equation 2. 14) further elaborated as follows;

푑𝜎 휕𝜎 휕푑 휕푈⃗⃗⃗ 푐 = = = [ + 푈⃗⃗⃗ ∙ ∇ 푑] − 푐 푘⃗⃗ ∙ Equation 2.15 𝜎 푑푡 휕푑 휕푡 푥̅ 푔 휕푠

푑휃 1 휕𝜎 휕푑 휕푈⃗⃗⃗ 푐 = = − [ + 푘⃗⃗ ∙ ] Equation 2.16 휃 푑푡 푘 휕푑 휕푚 휕푚

Here, s is the space coordinate in wave direction 휃 , and 푚 is a co-ordinate perpendicular to s, ∇푥̅ is 2D differential operator in the 푥⃗-space (DHI a, 2014).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.2 Hydrodynamics Hydrodynamic model MIKE 21 HD simulates water level variations and flows in response to a variety of forcing functions in lakes, estuaries and coastal regions (DHI b, 2007). The model is based on Alternating Direction Implicit (ADI) techniques to combine the equation for mass and momentum in the space-time domain. Rapid developments in hardware mean that the model solves the equation mainly using the cell centred finite volume method (DHI b, 2007). The equations are highlighted in (DHI b, 2007) for conservation of mass and momentum integrated over the vertical plane (Equation 2.17), description of flow (Equation 2.18) and water level variations (Equation 2.19) and are written as:

휕ξ 휕p 휕q 휕d + + = Equation 2.17 휕푡 휕푥 휕푦 휕푡

휕p 휕 푝2 휕 푝푞 휕ξ 푔푝√푝2+푞2 + ( ) + ( ) + 푔ℎ + − 휕푡 휕푥 ℎ 휕푦 ℎ 휕푥 퐶2∙ℎ2 1 휕 휕 [ (ℎ휏푥푥) + (ℎ휏푥푦)] − 훺푞 − 푓푉푉푥 + 𝜌푤 휕푥 휕푦 Equation 2.18 ℎ 휕 (푝푎) = 0 𝜌푤 휕푥

휕p 휕 푝2 휕 푝푞 휕ξ 푔푝√푝2+푞2 + ( ) + ( ) + 푔ℎ + − 휕푡 휕푥 ℎ 휕푦 ℎ 휕푥 퐶2∙ℎ2 1 휕 휕 [ (ℎ휏푥푥) + (ℎ휏푥푦)] − 훺푞 − 푓푉푉푥 + Equation 2.19 𝜌푤 휕푥 휕푦 ℎ 휕 (푝푎) = 0 𝜌푤 휕푥 where ξ (x, y, t) represent the surface elevation (m), h (x, y, t) is the water depth, 푑(푥, 푦, 푡) is the time-varying water depth (m), p, q(x, y, t) are the flux densities in the x- and y-directions (푚3⁄푠 /푚) = (uh,vh); (u,v)=depth averaged velocities in x- and y-directions, C(x, y) represents Chezy resistance (to describe bed resistance 1 (푚 ⁄2/푠)) (DHI b, 2007).

Modelling is performed for pure tide and the dominant NE and SW monsoon wave conditions. Field measurements are obtained for calibration and verification of the hydrodynamic models such as tidal variations and current speed. The Key Parameters in this module are Bed Friction and Eddy Viscosity as further expounded in Section 5.4.1.1.

To assess the verification outcome of the model, the root-mean-square-error (RMSE) measure according to (Barnston, 1992) as shown in Equation 2.20 is adopted. This is used to establish a quantitative evaluation of the forecast competency of the outcome of the model. Then, comparison is made to the local guidelines in terms of tolerable limit (DHI b, 2007).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

RMSE is defined in the context of, as before, standardised variables such as the square root of the mean of the squared differences between corresponding elements of the forecasts and observations:

2 1⁄2 푁 Equation 2.21 푅푀푆퐸푓표=[∑𝑖=1(푍푓𝑖 − 푍표𝑖) /푁]

Where Σ is summation; N is sample size; 푍푓𝑖 − 푍표𝑖 is the difference between simulated and measured outcome. The tolerable limit of the outcome is discussed is Section 6.2.2.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.3 Modelling Approaches River channel entrances are amongst the most dynamically evolving and critical type of natural environment. They are also immensely complex. Therefore, a model that is created must be able to achieve the ability to make quantitative predictions of morphological change at a scale that is sufficient to cater for long-term strategic coastal management (French & Burningham, 2009). This scale is known as mesoscale that as a whole can be quantified between time horizons versus spatial dimensions in the order of 101 to 102 years and 101 to 102 km respectively. With this scale, modelling forecast should be robust enough for policy makers to make a decision (French et al., 2016), which serves the objective of this research. The dynamic nature of mesoscale coastal evolution can accurately be represented by deploying two types of models, that is reductionist and reduced complexity model – RCM (Liang et al., 2015). Reductionist model is better in including a board spectrum of scale which includes all processes that can potentially affect the system’s evolution (Woodroffe & Murray-Wallace, 2012), being well suited to the assessment of any potential coastal management solution (Walkden et al., 2015). A reduced complexity model is more suited in geomorphology terms in that it focuses more on the particular area of interest that is crucial to describe a situation (Walkden & Hall, 2005a). Reduced complexity models have also been proven to be effective in simulating well- represented models over large spatial and temporal scales as shown by Walkden & Hall (2005b).

Predicting large-scale morphodynamics is made possible by applying reduced complexity model approach, which in this case is needed to integrate the temporal pattern mainly the monsoon seasons into the model. Reduced complexity model demonstrates a suitable approach in which this type of model only focuses on processes that are of interest. This is achieved through parameterising rather than explicitly simulating the pertinent effects of a model. This involves a large-scale coverage seawards in modelling the ocean dynamics, mainly wave generation and tidal level. Although the reductionist model can be applied in this case, it is proven to be computationally intensive. Although computational power is constantly improving mainly with the advancement of multi-threading and faster cache speed, using these models to perform regional and decadal-scale simulations of morphological change is still challenging from a practical standpoint (van Maanen et al., 2016).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Apart from the core types of numerical modelling methods it is important to further explore the modelling potential of a model in generating quantitative prediction of a river channel entrance evolution and the outcome of a proposed management solution. The reduced complexity model can be further enhanced by integrating the model in a framework as shown in Figure 2.10 that is structured by overarching conceptual models, by including reductionist model approach into the local model and data-driven analysis. Using such an approach, the best of the largely separate modelling communities can be brought together to fully utilise the best that each of these diverse approaches has to offer.

Figure 2.10 Modelling approach framework for the simulation of a river channel entrance.

The techniques integrate the regional Reduced Complexity Model with Conceptual Model and Reductionist Model, all of which are either supported by data or data- driven techniques. Reduced Complexity models are the “backbone” of this model as they are the primary framework for simulating dynamic behaviour of river channel entrances. Hence it should be noted that these models are closely linked to the other approaches, and that vital output and input between integration processes is highly significant as shown in Figure 2.10. Further explanation of the main capabilities of each of the techniques is discussed in the following chapters to demonstrate how each of the fields can better inform the reduced complexity models.

Previous studies have been carried out that successfully integrate these two simplified numerical modelling approaches into one overarching framework (Nichollas, et al., 2012). Accordingly, this study adopts the approach of the reduced complexity model

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

in correctly representing the wave and hydrodynamic modelling in combination with the reductionist model in representing sediment dynamics and management solution prediction. Therefore, the open coast, the river channel entrance, the inner shelf and their interaction are included in the system as a whole, with all sediment dynamics and the response of the system when potential management measures are in place.

The overall methodology of this research is divided into a number of steps, some of which will run in parallel. This research involves the heavy use of numerical models for the various stages which is further explained in Chapter 5.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.4 Data and Data-driven Approaches The ever increasing pressure on the coastal zone due to the overwhelming population inhabiting the coastal area necessitates the implementation of more advanced and stringent coastal management and monitoring systems (Bradbury et al., 2002; Coco et al., 2013; van Koningsveld & Mulder, 2004; van Maanen et al., 2016). There is a large amount of readily available datasets available such as Global Reanalysis of Ocean Waves by Oceanology International (GROW) and National Survey and Cadastre - Denmark (KMS) with the ability to provide a major source of scientific insights. These datasets are sufficient to be utilised for a range of different purposes with its direct application of data-driven method depending on requirements. As an example, advanced statistical approaches such as extreme value analysis (EVA) can be used to determine design wave conditions for the design consideration of mitigation structures (Castillo & Sarabia, 1994). These outputs can then be used to make prediction of future wave conditions based on appropriate extrapolations. One of the most useful method of utilising the data is by revealing links between datasets of two or more different variables such as canonical correlation analysis (Horrillo- Caraballo & Reeve, 2008), which is an excellent demonstration of the effectiveness of such methods in making future predictions of an unknown variable such as beach morphology based on another variable –wave height, which also allows predictions to be made of future scenarios.

Other than using data to make future predictions based on statistical methods of behaviours and extrapolation of trends, the outcome of the data analysis can also be used as in input and to build the reduced complexity model. A few demonstrations have been shown in Jarrett (1976) for example, who demonstrates the relationship between tidal prism and the cross-sectional area of an open channel using equilibrium models for tidal inlet systems. Walton & Adams, (1976) shows a similar empirical relationship that exists between tidal prism and ebb-tidal delta volume. The Equilibrium models can also be used to demonstrate the relationship between a basin area, tidal range and flat volume (Bruun, 1954; Dean, 1977; Eysink & Biegel, 1992) demonstrating a straightforward example of equilibrium beach profile derived from empirical studies. ASMITA is a good example of a numerical modelling program that utilises these empirical relationships having the capability to simulate tidal flats and estuaries (Rossington et al., 2011).

With the advances in data-driven methods, which includes making direct predictions or being used as a form of input to reduced complexity models, the applicability of these methods is limited by the integrity of the collection and derivation of the data. Data-driven approaches also provide rather limited physical insights despite being able to formalise data interpretation from large data sets. Data-driven approaches are often seen as a method with limited capacity to enhance understanding of the dataset from which they have been derived (Cunge, 2003). Interpretation of the outcome of data-driven approach requires in-depth understanding of the mechanism

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

of the particular data-driven method. However, a variety of solutions are available to overcome these disadvantages and can be used to scrutinise the dataset and the procedure involved as shown in Gevrey et al. (2003).

This research has utilised data and datasets available from various sources as explained in Chapter 4. Figure 2.10 shows the role of data and data-driven approaches as a vital essence in the sense of observational data and exchange of information. The data-driven approach has also played a major role in the further derivation of the seasonal pattern in the region, which in this case in analysed using statistical wind pattern and wave climate in the wind rose and wave rose expression. Windroses are a class of diagrams designed to display the distribution of wind direction experienced at a given location over a period (Heidorn, 2005). This is especially useful to analyse climatic seasonal behaviour of climatic season (Bordbar et al., 2014). A wave rose shows the long-term distribution of wave height and direction. Both of these are categorised as directional roses, where each concentric circle represents a different frequency, emanating from zero at the centre to increasing frequencies at the outer circles. Each spoke is broken down into colour- coded bands with intensity ranges from wind speed and wave height, respectively. Directions follow the nautical convention ranges between 00 – 3600 (Waveclimate, 2011). Although monsoon seasonal temporal pattern research is performed in this region (Loo et al., 2015; Nurukamal & Razali, 2016) mainly focusing on the Peninsular Malaysia, monsoon seasonal temporal pattern analysis is never performed in the vicinity of the site using the directional rose method. Hence, to show a succinct view of monsoon seasons in terms of how Wind and Wave Climate behaviour at the research location directional roses method is used in this research as shown in Section 5.1.2 and Section 5.1.3. This brings a further understanding of the importance of including seasonal influence in the numerical model.

Apart from seasonal behaviours, data-driven analysis is also used to predict extreme events such as extreme wave conditions using Extreme Value Analysis (EVA). This is for the purpose of determining the height of the river management structures based on wave conditions. Overall, the integrity and the role of data and data-driven- approach represent the input of quantitative information into the overall modelling approach. The outcome of this can be seen in Section 6.3.3.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.5 Mapping and meshing method of coupled model 2.5.1 Mapping Despite the current technical advancement in modelling (van Maanen et al., 2016), conceptual frameworks for the analysis of river channel entrance systems have arguably taken a back seat and failed in demonstrating the further understanding and the challenges in terms of river channel entrance management (Nicholls et al., 2012). The need for an integrative system-based outlook becomes more necessary to cater for strategic application. Also, the evaluation of management and engineering options has progressed to address the broader time and space scales for the dynamic behaviour of a river channel entrance, and possibly the overall river channel entrance configuration, which is affected by climate change and sea-level rise (French & Burningham, 2013).

Mapping can also be utilised to identify shoreline management paradigm which has been carried out in many countries (Mulder et al., 2011; Nicholls et al., 2013). Geohazards, mainly referring to erosion in this case, typically occur as the adjacent waterbody is often not considered in a river channel entrance in which flood risks, tidal and surges-related phenomena are of greater concern. The geohazards in the open coast and the enclosed estuarine are different.

While geohazards occur commonly in open coastal areas, a different approach to their management has shown a lack of appreciation of the nature and significance of the sedimentary and morphodynamic interactions between river channel entrance and the open ocean. This is well illustrated in the UK as well as Malaysia, where the second edition of the shoreline management plan (SMP) has either failed to include river channel entrances altogether or to consider the dynamic nature of river channel entrance in a rather selective and inconsistent manner (Hunt et al., 2011).

The constant changes in the management of the dynamic nature of river channel entrance are more challenging than the short-term variability that often being considered in the study of a particular river channel entrance based on observational record (Cowell et al., 2003). It is stated that low-order coastal changes needed to be evaluated in a broader spatial scope, which includes sediment dynamics interaction within the vicinity of the river channel entrance, including both the open ocean and the river basin of the river it serves.

To properly simulate a well-represented dynamic river channel entrance behaviour, the model requires a system-level approach to demonstrate a collaborative interaction between the open coast, the river basin and the river channel entrance. The process of conceptually analysing the coast highlighting this interconnectivity is better exemplified by utilising a spatial analysis software in the form of a geographic information system (GIS). GIS is a computerised system that facilitates three crucial phases: data entry, analysis and presentation. GIS main functionality is to enable multiple spatial analysis via a variety of datasets. To perform this function, GIS can 26

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

integrate numerous data files, geomorphic e.g., site topography for both land and sea and human intervention e.g., revetment, breakwater, dredging, etc., to improve further analysis of a variety of management measures for river channel entrances.

These survey data are then integrated into the GIS software to form the overall map to model integration with a suitable overlay, especially in the vicinity of the river channel entrance, which is the most crucial part of the research. The overall basic principle of a model must take into account the specific contexts of a place where modern processes interact with antecedent geology, historical morphology, and engineering interventions, and local landform dynamics which are forced by tidal, wave and sediment supply boundary conditions at a broader scale.

In order to achieve these principles, the following steps must be adopted:

a) The familiarisation of existing knowledge on the specific place of interest b) Formulation of relevant scientific questions and management issues c) The implementation and deployment of predictive models

Hence mapping, in this case, is for the purpose of not only numerical modelling analysis and input, but is also applicable in giving proper consideration to the Shoreline Management Plan as shown in Section 2.8 and the historical coastal behaviour that is related to a river channel entrance as shown in Section 5.1.4. The overall mapping methodology is elaborated in Section 5.2. The meshing of the overall model based on this GIS mapping data is further explained in the next section.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.5.2 Meshing The river channel entrance is a dynamic zone between the ocean and the riverine system. In order to accurately represent a river channel entrance, it requires a multi- scale approach. The dynamics of a river channel entrance domain can only be represented over various seasons and space scales. To represent the Petagas River channel entrance, the model domain is separated into two; regional and local. This method has been successfully applied in several models (de Brye et al., 2010), consisting of the utilisation of models with variable mesh in which each of the models acts as a boundary condition for a finer one. The flexibility of a variable mesh provides the possibility to accurately represent the topography (Legrand et al., 2007) and is able to simulate a wider range of areas and times (Deleersnijder & Lermusiaux, 2008). This type of state-of-the-art modelling approach has been utilised as of late (Guerin et al., 2016; WANG et al., 2016) mainly in first world countries but has not been applied to Borneo region. Hence, a variable mesh model using separate model domain is being utilised for this purpose. With this method, a larger domain will be considered, and the transition will be consistent. Most importantly, all this can be achieved within a reasonable computational cost.

The finite volume method used in the model is based on integral conservation law rather than on partial differential equations. The integral conservation law is used for small control volumes that are identified by the mesh. The characteristic of the numerical formulation is demonstrated using Space Discretization. Space discretization is carried out in spectral and geographical space using the cell-centred finite volume method (CC-FVM). The integral conservation law is the rate of change of the sum of quantity with density 푢 in a fixed control volume 푣 that is equal to the sum of flux on the quantity through the boundary 휕푣 as shown in Equation 2.22 as highlighted by (DHI, 2009) below:

휕푡 ∫ 푢. 푑푣 + ∫ 푓(푢) . 푑퐴 = 0 푣 휕푣 푓 = 푣푢 − 푑∇푢 = 푓푙푢푥 푓푢푛푐푡𝑖표푛 Equation 2.23 휕푢 ∫ . 푑푣 + ∫ ∇푓 . 푑푣 = 0 푣 휕푡 푣

Where = 퐹푙푢푥 ; d = distance; 푢 = density; A=area; 푣 = control volume

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The cell-centred finite volume method is formed using the averaged variable value formed from the mesh cells within the cell centre Vi as shown in Figure 2.11.

Figure 2.11 Definitions of control volumes for cell centred finite volume method in quadrangular and triangular mesh

The types of mesh and the density of the mesh determine the meshing criteria. In this case, there are two key types of mesh as shown in Figure 2.11 - triangular and quadrangular/rectangular. The quadrangular mesh is used when there is uniform directional flow (channel/river), and the triangular mesh is used when the flow is two- directional (ocean flow). The mesh density is determined by the bathymetry of the area and the importance of the evaluated areas (DHI, 2009).

The models use a flexible mesh technique, which implies a combination of different shape and density of mesh under one model to represent the bathymetry. This allows progressive horizontal refinements of e.g. the water depths from offshore to nearshore areas. The area in this case is relatively complex due to the extensive presence of offshore shoals and fringing coral reefs that needed to be resolved accurately or at least to the accuracy of available bathymetrical charts for this area. The unstructured meshes for the models, which are projected to UTM-50 and aligned with true north, are based on bathymetric data extracted from the British Admiralty Sea Charts with reference to Mean Sea Level, MSL (+1.23mCD) as well as available survey data.

The mesh density is gradually increased as it reaches the river channel entrance to avoid sharp changes as this may result in excessive element distortion (Silva et al., 2011).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.6 Reductionist Models Reductionist Models are typically used to simulate hydraulic and sediment dynamics in high-resolution mesh density with small time steps, even down to seconds. They are capable of resolving flow circulation patterns and sediment dynamic behaviour over spatial scales that cover a large part of continental shelf (Souza et al., 2007a) down to the local scale of morphological features typically found in coastal environments such as mud flats and channel-shoal patterns (Brown et al., 2015). With the advance of numerical modelling tools, many commercial model codes are currently available such as MIKE (Kozyrakis et al., 2015), Delft3D (Rahbani, 2015), FVCOM (Wu & Tang, 2010), POLCOMS (Souza et al., 2007b), TELEMAC (Moulinec et al., 2011), and ROMS (Warner et al., 2008). Most of these numerical software model packages are capable of operating in 2D and 3D mode depending on the required outcome. Models that are produced under the same suit usually have the ability to be coupled or interact with each another, for instance, coupling shortwave generation and propagation models in MIKE SW to account for wave-current interactions by combining with MIKE HD to generate wave-current interaction (WCI) (Warner et al., 2008) as well as to assess the combined outcome effect of waves-currents on bed shear stress. Salinity intrusion in the vicinity of the coastal zone is an example of a complex phenomenon to be well-represented in a model as it includes different variables throughout a water column. Therefore, a 3D model can be utilised if needed for analysis with the consideration of temperature variability, salinity gradients in shelf seas, or fresh water interaction in estuaries. The sediment transport module in MIKE 21 ST models the pathways of sand transport based on current, tides and waves, and the simulation also includes initial rates of bed level changes to identify potential areas of erosion or deposition (DHI f, 2014). Hydrodynamic and sediment transport computations can be conducted as part of stand-alone modules, or in combination with an output that includes morphodynamic calculations by considering varying bed level changes as a result of the sediment budgets and feeding back the updated bathymetry in the respective time step (Latteux, 1995; Roelvink, 2006).

To build a well-represented model of a river channel entrance, the relationship between the open sea, river channel entrance and the river basin should be combined. Regarding model integration framework as shown in Figure 2.10, the critical role for reductionist models lies in the provision of the boundary conditions from the reduced complexity model. Input in terms of hydrodynamic forcing such as tidal behaviour and wave condition can be provided in this case. These types of boundary conditions are likely to be seasonally influenced, and a reductionist local model is therefore expected to produce a vast range of possible scenarios for both pre- and post- engineering measures. Most importantly, reductionist models can play a major role in testing detailed multiple hypothesis as applied to the reduced complexity models.

The exchange of information between models is necessary in terms of simulating the behaviour of bed changes, which involves the evaluation of sediment source and sink

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

that can be executed in the reduced complexity model, and this output can then be fed back into the reductionist local area model for a more accurate representation of the topography of the modelled area. Most importantly, the outcome of the reductionist local model can be used further to compare and validate the conceptual model used as a hypothesis, as the reductionist model by itself does not handle configurational changes that well (French et al., 2016).

The quality of a reductionist model is mostly dependent on the various parameters such as sediment properties, which include sediment diffusivities, settling velocities and cohesive processes, which require expert knowledge to quantify. Although the proper specification of the closure model for turbidity is strenuous, it is vital for producing a sediment dynamic result (Wang et al., 2012). With the rapid development of reductionist models over the last few decades, it is undeniable that the model needs to be utilised when the level of detail and accuracy is of the utmost concern.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.7 Sediment Deposition Sediment deposition is predicted using the combination of ocean and river energy. The modelling is based on the detailed modelling of waves and tides, as well as wave driven currents performed under MIKE 21 HD.

Sediment deposition occurrs when the velocity of current decreases until a stage where the horizontal current is no longer able to support the transport of the particle; then gravity takes over for the sediment settlement (Julien & Vensel, 2005). For this reason, it is crucial to represent siltation. The sediment deposition will be solved using DHI MIKE 21 Mud Transport (MT) module.

Attention is drawn towards cohesive sediment because mud distributes landwards and sand tends to be drawn seawards (Zhou, et al., 2015). The examples include cases in the Wash Bay, UK (Amos, 1995) and Jade Bay, Germany (Friedrichs, 2012). On that account, the subsequent improvement works should be subjected mainly to mud with secondary attention given to fine sand. Although MT is used predominately to analyse cohesive sediment, the model also examines non-cohesive fine suspended sediment (DHI c, 2012), which is applicable in this case.

Sediment deposition was analysed using the MIKE 21 MT (Mud Transport) model with the combined effect of waves and current. The model included the transportation of sediment, its deposition, resuspension and erosion, with consideration of the variation of hydrodynamic conditions, tides and seasons.

Figure 2.12 Example of modelled physical processes (DHI c, 2012)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

According to DHI (2012c) the governing equation for MIKE 21 MT module is based on the influence of waves introduced through the bed shear stress (Teisson et al., 1993), as shown Equation 2.24 as follows;:

휕푐𝑖 휕푢푐𝑖 휕푣푐𝑖 휕푤푐𝑖 휕푤 푐𝑖 + + + − 푠 휕푡 휕푥 휕푦 휕푧 휕푧 푖 푖 푖 휕 푣푇푥 휕푐 휕 푣푇푦 휕푐 휕 푣푇푧 휕푐 𝑖 Equation 2.25 = ( 푖 ) + ( 푖 ) + ( 푖 ) + 푆 휕푥 𝜎푇푥 휕푥 휕푦 𝜎푇푦 휕푦 휕푧 𝜎푇푧 휕푧

Where 푡 is time, 푥 , 푦 , 푧 represent cartesian co-ordinates, 푢 , 푣 , 푤 are the flow velocity components in the x y z directions respectively, 푐𝑖 is the mass concentration, 𝑖 𝑖 푤푠 is the fall velocity, 𝜎푇푥 represents turbulent schmidt number which is the ratio between the rates of turbulent transport of momentum (eddy viscosity) and the turbulent transport of mass (Tominaga & Stathopoulos, 2007), 푣푇푥 is the anistropic eddy viscosity and finally 푆𝑖 is the source term (Teisson, 1991).

Then, the deposition is solved by the equation 2.26 Transport of cohesive sediment (Krone, 1962):

푆퐷 = 푊푆 ∙ 퐶푏 ∙ 푃퐷 Equation 2.27

where 푊푆 is the settling velocity of the suspended sediment ( 푚⁄푠 ), 퐶푏 is the suspended sediment concentration near the bed, and 푃퐷 is derived in Equation 2.28 as highlighted by (DHI c, 2012), and is the expression of the probability of deposition:

휏푏 푃퐷 = 1 − 휏푐푑 Equation 2.29

Where 휏푏 is the bed shear stress and 휏푐푑 is the critical shear stress for erosion.

A coupled model incorporating the outputs from hydrodynamic and spectral waves was applied to represent sedimentation. Simulation was undertaken for different monsoon seasons. The accuracy of the sediment dynamics model depends heavily on the hydrodynamic model as elaborated in Section 2.2, which is controlled by natural processes such as waves on a large scale (Deloffre et al., 2006; Lesourd et al., 2003). Hence, the wave field, which is the key input parameter is generated from the wave simulation and the verified hydrodynamic module.

Early transport models commonly use general mean grain size to represent the whole spectrum of sediment information (Engelund & Hansen, 1967). However, this method is unable to predict the changes in grain size, in addition to underestimating the transport rate of the finer fractions (Wilcock & Kenworthy, 2002), in which the analysis is needed in this case. The grab sample method is used for its conceptually 33

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

simple method, being cost efficient and amenable to scheduling advantages. Hence, some grab samples are obtained from the vicinity of the river channel entrance during the monsoon seasons. The grab sample fractions are further elaborated in Section

5.4.2. As highlighted by (DHI c, 2012) the average settling velocity 푊푠 for each fraction is then determined using Stokes’s Law in Equation 2.30 as: (𝜌 − 𝜌)푔푑2 푊 = 푠 푠 18 ∙ 푣 Equation 2.31

In which 𝜌푠 is sediment density; 𝜌 is the density of water; 푔 is the acceleration due to gravity; 푑 represents grain size; 푣 is the viscosity of the water. Stokes’s law mainly applies to discrete particles and it is not to be applied for cohesive sediment because these particles come together in large numbers to form flocs.

The model also considers the erosion of the bed layer, which is also relevant regarding the transfer of sediment from the bed to the water column. Erosion takes place in the area where the bed shear stress is larger than the critical shear stress for erosion.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.8 Management Solution Consideration Coastal flooding is a pressing issue as a consequence of poor governance principles (Eisenman et al., 2007) as the outcome of the poor understanding of the causes and consequences of floods (Schneider, 2005). Hence the need for more detailed scientific approach is in this subject area. While many studies show the outcome of mitigation from a physical and social point of view, it is important to go further by analysing the results of various mitigation options as being both physically and socially constructed (Touili et al., 2014). The main means of assessing the key risks faced by a coastline as well as providing a coherent, integrated strategy to mitigate the socio-economic and ecological impacts of flooding on the coastal environment is a model currently adopted by the UK Government, which concentrates on Shoreline Management Plans (SMPs) and helps in describing the proper flooding and erosion management in a specific coastal area. These plans were initially drawn up in the 1990s and have now been updated to the second edition. The SMPs are based on one of four different management policies as described as follows:

1. Hold the (existing defence) line – Build or maintain artificial defences to preserve the position of the shoreline. 2. Managed realignment – Allow the shoreline to move naturally, but managing the process to direct in certain areas. This is usually done in low-lying areas, but may occasionally apply to cliffs. 3. Advance the line – New defences and built created on the seaward side of the existing coastline. In this case, a product is proposed to address not only the erosion problem but gain land area. 4. No active intervention – There is no planned action to be taken for the existing problem. Flood will not be able to be mitigated ‘do nothing scenario’. (Hardiman, 2013)

The first method is the most widely practised in the industry by utilising a combination of hard and soft engineering solution to hold to the existing condition or shoreline. However, Shoreline Management Plans prefer to use option 4, which involves no active intervention, as in ‘Do Nothing’. An intervention that causes changes to coastal planning needs to be proven to show sufficient benefits. One of the many methods used to analyse this is by evaluating the Benefit Cost Ratio (BCR) of a proposed intervention, which should be “Robustly more than 1” (Sugdon, 2004) to be considered for funding.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Similar guideline to this approach also operates in Malaysia where coastal management policy is based on Integrated Shoreline Management Plan (ISMP), describing the options of coastal area management to deal with flooding and erosion. By dividing sections of coast into ‘management units’, the ISMP of Sabah has identified the following occurrences in the vicinity of the Petagas River channel entrance:

1. Critical erosion threat (Class 1 of NCES 1985) with immediate danger for damage or loss of value. 2. Sedimentation has further increased following the reclamation for the KKIA runway extension, this is due to the northward littoral drift being trapped at the end of the reclamation. 3. The vicinity of the river channel entrance is located at management unit C5- 19: Tg Dumpil to Sg Petagas as shown in Figure 2.13, which has management objectives as follows: a. Improve river channel entrance navigability and flood conveyance b. Protect village against erosion c. Improve flushing capacity d. Manage coastline retreat (Ministry of Natural Resources and Environment Malaysia, 2012a)

Figure 2.13 Management Unit C5-19 extracted from (Ministry of Natural Resources and Environment Malaysia, 2012b)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

According to the numerical modelling results, the stated management objectives can be achieved by advancing the coastline with new defences and created on the seaward side of the existing coastline.

Currently, short-term/immediate action is taken by dredging the river channel entrance occasionally to allow sufficient discharge through the river channel entrance. The local residence in Contoh Village, which is beside the Sg. Petagas River channel entrance is trying to maintain the existing shoreline by building artificial defences using constructing provisional shoreline protection, which is made out of tyres and concrete pile stubs as shown in Figure 2.14.

Figure 2.14 Coastal Protection Constructed by Local Residents made out of Used Tyres

With major population and economic growth concentrated around the coastal area in Sabah especially in Kota Kinabalu, the proper management of the river channel entrance is of critical importance. The majority of coastal villages depends on fisheries as their main source of income with tourism and agriculture becoming more important over the years. The increasing population resulting from urbanisation has changed the environmental conditions in coastal areas (Terrados et al., 1998; Wilkinson, 2000). Deforestation for the purpose of urban development in river basin areas has resulted in increased soil erosion and overland runoff which further transport sediments into rivers and eventually deposited downstream in coastal waters. For that reason, the coastal zone is not to be seen in isolation. To maintain socio-economic growth, proper management, legislation and effective enforcement is crucial (Jakobsen et al., 2007). In this research, the focus of the model will be in the vicinity of Petagas River channel entrance, where the findings will be used to propose an optimum mitigation 37

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

measure. The methodology applied to develop a mitigation measure for the Petagas River Channel Entrance is considered an example to be followed by others for the management of river channel entrances especially in monsoon-influenced regions, and around the world, in resolving the impact caused by discharge restriction of river channel entrances on the coastal environment.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.9 Engineering Solutions Considered Engineering solutions that are common in the management of a river channel entrance in this region are revetments, seawalls, groynes and breakwaters. Usually, structural solutions have an immediate short term and occasionally long-lasting environmental impacts. A common example is the construction of a groyne as a training channel in a river channel entrance, which often results in downdrift erosion of the adjacent stretch of beach. Hence selection and application of a particular structural solution require an adequate understanding of the consequences of its impacts on and adjacent to the area. One or a combination of two or more of the mentioned structures is suitable and efficient in addressing the sedimentation problem in a river channel entrance.

The few options considered in this study are breakwater, revetments, reclamation and dredging.

2.9.1 Revetment A revetment is a sloping structure that is used to maintain existing shoreline or to protect the slope of the open channel, to provide protection from wave action or to retain in situ soil or fill as a defence against erosion (Graaff, 2009). Revetments can be constructed as both buried and exposed structures. Revetments these days are usually made out of rocks as shown in Figure 2.15 or cast concrete blocks called Tetrapods, as shown in Figure 2.16.

Figure 2.15 Armoured Rock Revetment (SNH, 2000a).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.16 Tetrapod Revetment (Patel, 2015).

In times gone by, revetments made of timber were widely used. Timber revetments are originally made out of planks laid against wooden frames. However, they are largely being replaced by armoured rocks or concrete blocks. The easy workability of timber allows them to serve various purposes in which they are used as a permeable ‘fence’ along the upper beach or a final wave protection wall when built as an impermeable vertical breastwork along the upper shoreline (Keating et al., 2012)

Figure 2.17 Timber Revetment -Breastwork (SNH, 2000b)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Rock revetments are constructed by armouring the dune face. The rugged face formed by the rocks dissipates waves energy to prevent erosion. Rock revetments are widely used in areas where the property value outweighs the construction of the revetment, or the infrastructure is fixed.

There are two types of rock/concrete revetment normally utilised in this region; rock revetment and riprap revetment. The major difference between these two is the absence of the toe and splash apron in riprap revetment as shown in Figure 2.18.

Figure 2.18 Typical Coastal Revetment Cross-section (Pile Buck, 2014)

Rock revetment should be constructed with in-depth study and analysis of a shoreline prior to the implementation. Local knowledge of processes and beach movements are usually obtained from measurements and observations over a period of time (SNH, 2000a).

The main design and construction consideration for rock revetments are rock grading (Hudson’s Equation) and construction method. As with all rock structures, the crest elevation, face slopes and crest width must be determined with strict consideration, mainly determined by the site tidal datum with other factors such as storm surges and sea level rise (Van der Meer, 1987).

The rock grading is dictated by the quality of rock, determined by hardness and size of the rock. It is largely dependent on the rock source from quarries where the availability/cost of the suitable rock is adequate. Subsequently, the design determination of rock size is based on wave surge, wave period and direction, structure slope, risk factor and cross-sectional design (Lee et al., 2016).

As a general rule, rock that weights between 1 to 3 tonnes is adequate provided it is placed in a double layer format with a slope inclination of not more than 30 degrees to the horizontal. With this in mind, the risk factor is minimum. Determination of revetment face slope is a compromise between a gentler angle that absorbs more wave energy and is therefore exposed to less toe scour (Muttray & Reedijk, 2008).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

This eventually allowed the usage of smaller rocks compared to a more extreme angle which tends to give a structure a lower footprint with less rock volume. A larger rock size with a slope of up to 45 degrees is a reasonable compromise provided the natural dune slope is maintained. The rock size is required to be larger in the area in which the structure is below High-Water Mark of Ordinary Spring Tides - HWMOST.

The revetment structure is to be constructed when rocks are placed in a shallow trench with a layer of geotextile laid underneath to prevent the migration of sand as well as to minimise settlement of structure and to promote equal settlement if it occurs. The geotextile should be wrapped all around the base layer of the revetment structure which includes the toe of the structure that is set below the lowest measured seabed level as shown in Figure 2.18.

The length of the revetment protection area must be sufficient to protect valuable backshore assets. To prevent localised scour, the structure connection ends must be buried between 5 – 10 m over the final 20-40 m length of the structure, depending on the expected rate of future erosion. The sloping angle of this final section should be decreased to enhance the wave absorption capability. In maintaining the rock revetment function, revetment length may need to be extended if erosion occurs at the adjacent dune front continue (SNH, 2000a).

Figure 2.19 Erosion Occurrence at Adjacent Dune Front as a result of the Construction of Coastal Structure

The crest level of a revetment structure is determined by considering various factors with the major one being the design water level, which ideally is based on the Highest Astronomical Tide (HAT) or storm surge of a particular site. Added to that, the wave run-up, wave overtopping, structure settlement, freeboard and sea-level rise factor need to be considered as well (US Army Corps of Engineers, 1995). During the extreme season, some overtopping damage is inevitable. The designer’s judgement 42

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

in determining the acceptable level of risk will depend on the potential damage and the probability of the occurrence of extreme events. Another key element on the give- and-take of the footprint of the structure is to strike a balance in between aesthetic and function of a particular rock revetment.

Public safety is a concern where large rock revetment is constructed in a populated area to protect valuable backshore assets. To increase the wave energy dissipation, rocks on the dune face of a rock revetment structure is placed randomly to form a rough surface with large voids. The safety of walking on the large boulders is often misjudged by the public. The boulders often move when walked on, and the voids may be large enough to fall or climb into, which cause injury or people to become trapped. As the structures extend down to the level below the mean sea water, the lower rocks will be slippery to walk on during low tide due to algae coverage. The public may fall into the water with the possibility of drowning. Hence, warning signs should be displayed highlighting the dangers of unstable dune faces such as slippery algae growth, buried defences, gaps in rocks, etc. (SNH, 2000a)

The construction method of a rock revetment is usually the next major determinant factor on the design of a bank protection based on the site location. The main method of transporting rocks is separated into by land and sea transportation. The ground transportation method is preferred in this region provided the site is accessible with large machinery to transport boulders as shown in Figure 2.20. Erosion often occurs when valuable assets have been in place without any access for construction, in this case, a marine-based method will be used where the raw materials and machinery are transported using barge as shown in Figure 2.21. Accessibility is often a major issue in this region and will affect the consequent cost of the product. Rock revetments provide a long-term solution for major bank and shoreline protection usually with a design life of 30 years with proper maintenance (Hudson et al., 2015).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.20 Revetment Construction Using Lorry and Excavator (Varela, 2016).

Figure 2.21 Revetment Construction Using Barge as Transport (Arena Kish, 2015).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.9.2 Breakwater A breakwater is a structure constructed for the purpose of forming an artificial harbour with a basin, in order to be protected from the effect of waves as well as to provide safe berthing for water vessels (Sciortino, 2010). Breakwaters intercept longshore currents and help to prevent beach erosion (Britannica, 2015).

There are many different types of breakwaters; natural rock and concrete, or a combination of the two. These are the materials which form 95 percent or more of all the constructed breakwaters. During ancient times, a breakwater was made from simple mounds, consisting of stones. Breakwaters existed as early as 2000 B.C., when a stone masonry breakwater was constructed in Alexandria, Egypt. Figure 2.22 Shows a rubble mound breakwater discovered in Civitavecchia, Italy, which was built by the Roman Emperor Trajanus (A.D. 53-117) and is recognised as being the oldest existing rubble mound breakwater (Ito, 1969).

Figure 2.22 Rubble Mound Breakwater in Civitavecchia (Ito, 1969)

Laboratoire Dauphinois d’Hydraulique introduced the Tetrapod in the 1950s, which is the first interlocking armour unit. The main advantage of the tetrapod is the various improved core function of the traditional armoured units, which includes better interlocking capacity, larger and more voids in the surface causing better wave energy dissipation and reducing wave run-up.

Concrete armour unit production was popular between 1950-1970 (US Army Corps of Engineers, 1984). However, it has only been applied in a very limited number of projects. It is usually placed uniformly in a double-layer format of random dumping to form a volume. The stability of concrete armour structure is largely dependent on the concrete block self-weight and its interlocking capability. However, the failure of Sines breakwater in Portugal in 1978 and the invention of Accropode by Société Grenobloise d’Etudes et d’Applications Hydrauliques in the year 1981 more or less ended the use of concrete armour units. The failure of the breakwater in Sines was the result of the slender dimension of armour units, which are designed for maximum interlocking but lack in terms of their overall structural stability. Apart from that, the gradual breakage of armoured units has also caused progressive failure to occur. Single layer randomly placed armour units such as Accropode I by ‘Sogreah’, and below has gained popularity since its invention in 1981 and has become the leading armour unit worldwide for the next 20 years. Since then, Sogreah have come up with various improvement to concrete armour units which includes Ecopode in the year 45

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

1995 and Accropode ii in the year 1999. Core-Loc and A-Jack were also invented as spinoffs by USACOE in 1995.

In the 1960s a different concept of single layer armoured units was also developed. This concept consists of single armour layer with hollow blocks that are placed uniformly side-by-side. The principle of this concept is not based on weight or interlocking but on friction, which has improved the very high wave dissipation capability and provides a very stable hydraulic outcome. An iconic example of the application of this concrete armour blocks is the protection of the manmade island of the Burj Al Arab Hotel in Dubai UAE (Muttray & Reedijk, 2008).The similarity of this future development and improvement of these armoured units has tended to concentrate on enhancing interlocking strength and wave dissipation capabilities, which resulted in more economical than multiple layers armoured units.

An overview of the overall evolvement of breakwater armour units is shown in Figure 2.23.

Figure 2.23 Overview of Breakwater Armour Units (Muttray & Reedijk, 2009)

The function of the breakwater is designed to control wave action and longshore drift. Depending on the design of a breakwater, its core function is to dissipate, radiate, transmit and to frequency and reflect wave energy. The direction of the wave spectrum will change as the wave approaches and in contact with the breakwaters (Iñigo J. Losada et al., 1993). The partial standing wave pattern is likely to be seen in all types of breakwater. Thus, wave transformation which includes wave energy dissipation, transmission, modulus and phase, and refraction plays an important role in defining the wave regime around all the dune face seaward, leeward and inside of the breakwater (Hughes & Fowler, 1995; Losada et al., 1997; Sutherland & O’Donoghue, 1998; Vilchez et al., 2016).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Depending on the typology, a breakwater can also be designed to encourage longshore drift. Various examples have shown that the construction of a coastal structure such as a breakwater interferes with beach volume and the natural coastline. In order to maintain drift, in-depth study of the site location is needed as the dimension and shape of the overall breakwater structure is to be determined according to the site morphology data monitoring over a sufficient period of time in recording the changes of beach volumes and profiles as well any erosion of the coastline (Barton & Brown, 2014).

There are three major types of breakwater, Rubble Mound type, which is the most common, followed by the vertical and non-gravity breakwater, which has been gaining in popularity lately (Takahashi, 2002). Rubble mound type breakwaters are recognised as the most traditional breakwater and are still the most commonly used in coastal defence structures globally (Thaha et al., 2015). The simple function of the outer face is to dissipate wave energy. The best rubble mound breakwater is the breakwater that can dissipate a large portion of incoming energy, which results in smaller transmitted and reflected waves.

Figure 2.24 Rubble Mound/Gravity Breakwater (Zabez, 2015)

The concept of the vertical breakwater is to reflect waves, in contrast to the rubble mound breakwater which functions in breaking waves (Takahashi, 2002). A Vertical breakwater is only suitable to be constructed in solid foundation such as a rock or coral reef. An unreinforced Vertical Breakwater as shown in Figure 2.24 shall not be constructed in locations where the water level is more than 2 metres and/or subject to strong wave action. The composite type of vertical breakwater is suitable for soft foundations. The vertical breakwater normally sits on the top part of the mound projecting above sea level if used on soft foundations. Special consideration should be given when deciding on the position and shape of the vertical breakwater as it does not absorb wave energy; instead it reflects everything back. This is very common in harbours that experience “choppy-sea” scenarios.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.25 Vertical Breakwater (Zabez, 2015).

Gravity breakwaters consist of a foundation which supports the structure and maintains its balance; the core structure which handles the transformation of wave energy to the foundation, and the superstructures that control the wave overtopping and if necessary provide access (Vilchez et al., 2016).

Gravity breakwater is preferred in many cases as it is constructed easily with commonly available equipment such as excavator, tipper truck, barge, etc, coupled with naturally and easily obtained resources such as boulders and concrete. Most gravity breakwaters have been proven to perform well with stable durability and minimal maintenance. However, the depletion of natural resources such as quarried boulders with its increased handling cost over recent years has led to another emerging competitor in the concrete armour unit. Quarried boulders are needed to form the outer layer of gravity breakwater in protecting structures against wave attack.

Non-gravity type breakwaters are a particular kind of breakwater, which comprise piling, floating or pneumatic types. Curtain wall breakwaters as shown in Figure 2.26 are one type of non-gravity breakwater, held in position by piles, driven into the seabed or rubble mound. Alternatively, the top side of the skirt can be attached to moorings, making it a floating type structure which is easily replaced. The skirt reflects wave energy which only allows minimal energy to move through the provided gap underneath. This kind of breakwater is useful when placed in a location with short wave period because short waves will be attenuated when hitting the structure.

Figure 2.26 Curtain Wall Breakwater (EVANS, 1955). 48

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Another type of non-gravity breakwater is the sheet pile breakwater as shown in Figure 2.27. This type of breakwater works on the same principle as a vertical wall breakwater. It is held in placed by being driven into the seabed. One of the major concerns of this type of breakwater is scouring, especially if the seabed consists of soft materials such as sand. Hence, proper toe design consideration is necessary.

Floating type breakwater in Figure 2.28 is a type of non-gravity type breakwater which works on the same principle as the curtain breakwater. The difference is that the floating barrier is extended to the level to reflect the necessary wave energy. It is held in place by mooring chains fastened to the seabed. The additional benefit of this type of breakwater is its ability to serve as a mooring to berth floating craft on the lee side of the breakwater.

Figure 2.27 Sheet Pile Breakwater (EVANS, 1955).

Figure 2.28 Floating Breakwater (EVANS, 1955).

Pneumatic breakwaters work by calming the wave by injecting bubbles underneath the waterbody. This process has been around for 100 years. Previously the outcome of pneumatic breakwater could not be predicted (EVANS, 1955). However, a numerical solution has been discovered (Zhang et al., 2010) after continuous research in this area. The major benefit of a pneumatic breakwater is its minimal structural footprint and that it can be operated when needed. However, it is only effective for short wavelengths.

Breakwater layout/orientation is a major design consideration, which includes navigational aspects, wave penetration, and environmental effects. This consideration is only possible using detailed numerical modelling study of a breakwater site and with sufficient data collection.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.9.3 Dredging Dredging is the maritime transportation of natural materials from one part of the water environment to another by specialised dredging vessels and involves collecting or clearing away any materials from the bed of a river or sea. Dredging is considered a construction process and a dredged area may be considered as a temporary structure where the dredged area may be filled in due to the dynamic sediment nature of a sea or river bed. Since the beginning of civilisation and communities’ evolution, water-borne transportation has been growing in importance, resulting in the increased requirement of easily accessible ports and harbours to the shipping industry. Therefore, one of the major reasons for dredging in nearly all the major ports in the world is to create or extend shipping access channels and turning basins, in order to provide appropriate water depths along waterside facilities to accommodate supersized vessels, known as capital dredging. Many of the channels eventually require maintenance dredging, i.e. the removal of sediments deposited in the shipping channel to ensure sufficient dimension for shipping vessels (IADC / IAPH, 2010).

Dredging can be categorised as either capital or maintenance. Where capital dredging is usually categorised as a one-off operation, maintenance dredging is used to describe dredging of a recurrent nature. Dredging is undertaken to serve a variety of purposes and is achieved using different types of dredging method and equipment. The reasons for dredging include excavating material from underwater to create more conveyance in a lake, river, or the open ocean, and to obtain construction materials for various construction purposes, which also includes material replacement to achieve a better structural foundation for the construction of breakwater structure. In Malaysia, one of the major uses of dredging is to improve the hydraulic efficiency of a river channel (Cohen et al., 2015). The Port of Tanjung Pelepas (PTP) in Johor as shown in Figure 2.29, which is the largest and most important transhipment hub in Malaysia was completed in 1998 with 2800m of waterfront, followed by the second phase in 2003 with deeper and wider dredged channel allowing additional container berths, terminals and industrial buildings (Renkena & Kinlan, 2000).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.29 Port of Tanjung Pelepas in Malaysia with dredged channel allowing handling of multiple cargo vessels birthing (Pelabuhan Tanjung Pelepas, 2012).

Dredging can be done with the utilisation of different types of equipment with the ability to dig, transport and dump a certain amount of underwater soil in a certain time (Vlasblom, 2003a). Important factors that play a part in the ultimate choice of dredger include the quality and type of material to be dredged, placement or relocation alternatives, availability of equipment or cost of mobilisation and accuracy (Rotterdam Public Works Engineering Department, 2001).

The most common type of dredging methods available locally are cutter head suction dredgers (Mechanical Dredger) as shown in Figure 2.30. These utilise hydraulic centrifugal pumps to provide the dislodging and lifting force and remove the material in a liquid slurry form. They usually work well in loose, ‘unconsolidated’ silts, sands, gravels and soft clays. In cohesive materials teeth or water jets may be needed to break up the material. The cutter-head dredger uses rotating mechanical devices, called cutters, mounted ahead of the suction head. The cutters excavate the material into suitable sized quantities, which are then sucked into the suction pipe as a solid/water slurry and pumped to the surface. They are characterised by high production rates and able to effectively dig silts, clays, sand, gravel, cobbles and fractured and sound rocks (Vlasblom, 2003b).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 2.30 Cutter Head Suction Dredger at work in Terengganu River Channel Entrance Malaysia (Dr. Nik & Associates, 2014)

Dredging has always been an invisible industry as the work is mostly done underwater, and tends to be low-key, pragmatic and highly technical. The dredging industry is constantly striving to improve its performance, both with respect to technical and managerial efficacy and in relation to environmental protection.

For the purposes of this research, the river channel is dredged to maintain channel orientation and to improve channel conveyance capacity, by ensuring a suitable depth for sufficient hydraulic efficiency and to ensure sufficient navigational manoeuvrability within it. A trained channel is considered in this case to minimise the maintenance cost and maintenance frequency of the river channel entrance.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2.10 Knowledge Gap Previous modelling studies have shown that the wind condition in the South China Sea is governed by the North East (NE) and South West (SW) Monsoon with a mean wind speed of around 10-14m/s (Abdullah, 1992; ICZM Task Force, 1998; Lokman Husain et al., 1995; Zakariya, 2012). Typically, the North East Monsoon season lasts from December to March and the South East Monsoon season lasts from June to October, with the remainder of the year known as the transition period (Malaysian Meteorological Department, 2012). A modelling study by Xue et al. (2012) showed a similar directional shifting of coastal current from North East to South West monsoon period in the South China Sea, also known as the wetting region where frequent tropical storms and typhoons occur (Santini & di Paola, 2015). Waves at the river channel entrance play a major role in sediment transport and coastal morphology (Amoudry, 2008; Chiang & Hsiao, 2011; Valsamidis et al., 2013). The overall pattern in the month-to-month wave directions are related to the dominant directions of the monsoon winds (Mangor, 2007). The current direction is parallel to the prevailing wind direction of the South China Sea.

Several good comparative examples can be found, as evidenced by a series of studies performed in the Mekong River Delta in Cambodia, which also discharges into the South China Sea (Kocar & Fendorf, 2012; Schmitt & Albers, 2014; United Nations, 1969; Xue et al., 2012). Numerical modelling in this case is particularly useful in predicting the outcome of different management solutions (Stark et al., 2016; Wang et al., 2014) or even the future outcome of any river channel entrance because it helps save time and cost compared to physical modelling. The coastal processes and geomorphic behaviour in the Petagas River channel entrance have never previously been explored as an integrated system. In addition, the way the system evolves under future conditions as an integrated approach that fully appreciates the interconnectivity between the river basin discharge, river channel entrance and the open coast will be examined further in this research.

Based on the preceding information there is an apparent gap between the correlation of temporal patterns and their effects on river channel entrance analysed as an integrated system as a whole. Also, integrating the viability of management solutions utilising numerical modelling analysis is often disregarded. Hence, this research focuses on bridging the gaps between these aspects. The overall answers to these research gaps can be found in Chapter 7.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

3 Description of the research area 3.1 Geomorphology of the Site Located in Sabah, Malaysia (Formerly known as North Borneo), the Petagas River serves as the main outlet for the Moyog river basin, covering an area of 297 km2 (DID, 2011). As the urban centre of Kota Kinabalu city continues to expand on to the Moyog floodplain, this has resulted in the transition of land use from rural development and cultivation of rice paddies to intensive urban development. This has caused immense changes in river characteristics (DID, 2011). Apart from being one of the largest and increasingly most important river basin in Sabah, the Moyog basin has recorded rainfall intensities of up to 400mm over a period of 24 hours, which is one of the highest in Malaysia (DID, 2011).

The Petagas River channel entrance is located immediately to the south of Kota Kinabalu International Airport (KKIA). This river channel entrance discharges into the South China Sea through a single south-westerly channel about 50m wide. The growing importance of river channel entrances in Sabah has been acknowledged recently as one of the leading causes of flooding (Syvitski et al., 2014). Hence, the necessity for more advanced and sophisticated techniques capable of measuring physical changes in river channel entrance morphology is crucial. Therefore, computer modelling is widely used due to its cost effectiveness and time-saving capability.

The Sungai Petagas River channel entrance features a textbook example of a dynamically evolving river channel entrance, with problems frequently encountered in this region, and is the main focus of this research. The Sungai Petagas River channel entrance is located on the West Coast of Sabah at Putatan, a few kilometres south of Kota Kinabalu, Malaysia. Figure 3.1 shows a map of the Petagas River channel entrance Location.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 3.1 A map of the Petagas River channel entrance Location (Sat Image from Google Maps, 2009) 55

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The Petagas River channel entrance is located on the west coast in the state of Sabah, Malaysia as shown in Figure 3.1 to the immediate south of Kota Kinabalu International Airport facing an overall south-westerly direction and discharging into the South China Sea, a few kilometres to the south of Kota Kinabalu city. To the north of the river channel entrance, the coastline is strongly curved due to wave sheltering caused by the islands and coral reefs in the TAR Marine Park just off the Tanjung Aru headland. The coastline to the south of the river channel entrance has a south- westerly to north-easterly orientation.

Tides plays a major role in the circulation process. The tidal regime is mixed, mainly diurnal with one tidal cycle per day with an F value of 2 (Lisitzin, 1974) based on a constituent analysis undertaken by the Kota Kinabalu Port Authority (Malaysian Meteorological Department, 2009) and further derivation is given in Section 3.2.1.

Figure 3.2 Oblique aerial photo view of Petagas River channel entrance

Sedimentation is one of the most common problems to occur in a river channel entrance, and one of the most complex processes (Jayappa & Narayana, 2009; Mashriqui, 2003; Nelson et al., 2013; Nylen & Ramel, 2011; Robakiewicz, 2010). This complex phenomenon occurs throughout Malaysia under the hydraulic conditions found when the velocities are low, causing sand and silt to be deposited on the river bed (DID, 2001). The study of sedimentation dynamics behaviour in relation to temporal patterns needs to be carried out with prolonged and reliable data collection (Roy et al., 2012). An understanding of sedimentation requires thorough research of the hydrodynamic processes of the tides, the river flow, and the morphology of the bed (DID, 2001). From hydrological data obtained for the Petagas River, it has been shown that the main wave direction during the SW monsoon period is the main driver for the sediment transport and morphology at the river channel entrance, which is the focus point of this research.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Due to the sedimentation problem in the Petagas River channel entrance, the Moyog River, which is connected to the Putatan River further upstream, also sustained serious flooding problems (Sabahkini.net, 2012). According to the Petagas local authorities, the navigational condition of the river has worsened lately. The shallow bathymetry of the Petagas River channel entrance is prone to sedimentation due to increased bed friction; causing further river flow deceleration and lateral expansion of the river channel entrance area (Wright, 1977). Apart from the ocean energy, sediment grain size is a major factor in the sediment dynamics of a river channel entrance (Orton & Reading, 1993).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

As a result of sedimentation, the bathymetry of the river channel entrance is relatively uneven as shown in Figure 3.2. Petagas River used to supply sediment to maintain the Tanjung Aru Beach to the north, but reclamations for the extension of Kota Kinabalu International Airport (KKIA) have blocked the sediment transport as shown in Figure 3.3b. Hindrance caused by the reclamations has further increased the sedimentation problem of the shallow river channel entrance.

a b

Figure 3.3 Satellite Imagery at the Vicinity of the site (a) Captured yr. 2002 before the further reclamation (b) Captured yr.2010 after the further reclamation of yr. 2006 (Google Maps, 2015).

The river channel entrance of Petagas River is connected to the South China Sea by a single channel about 50m wide. The Petagas River is one of the major discharge channels for the Moyong River basin (297km2). Recent development has caused increased discharge and sediment transport from the river basin (DID, 2011), which has been studied by many researchers (Syvitski et al., 2014). The river channel entrance continues to respond to a dynamic morphology due to interannual climate variation of Northeast (NE) and Southwest (SW) monsoon seasons. Interannual climatic events cause fluvial flooding and flushing of sediment downstream (Mitchell, 2013), as evidenced in the Petagas river channel entrance where rapid development within the upstream river basin has resulted in a great volume of loose sediment being flushed down to the Petagas River channel entrance. As a result, navigational conditions have worsened. Within the Moyong River basin lies Moyong River which is linked to Petagas River and both are prone to overspill causing frequent fluvial flooding (DID, 2011).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

3.2 Metocean Condition The following Sections feature descriptions and statistics on tidal, wind and wave patterns in the Petagas River channel entrance.

3.2.1 Tides The tidal signal at any location in the world’s oceans consists of a large number of individual waves of different frequencies caused by gravitational forcing. These waves are referred to as “constituents”, where each constituent is described by amplitude and phase. As these waves behave linearly and do not interact, a measured water level record can be decomposed into a series of constituents, which summarise the tidal characteristics of the area. Tidal predictions for the area may then be made for any time in the past or future by reconstituting a water level time series from the extracted tidal constituents. Tidal regime analysis is shown in section 5.1.1.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

3.2.2 Monsoon Seasons Monsoons occur due to the migration of large-scale sea-breezes caused by warming phenomena. The occurrence of these phenomenon is due to low pressure in the continental landmass, which heats up the air faster here than over the ocean. This causes the air pressure above the land to be lower than usual, which in turn causes the wind over the sea to blow landwards and bring the moist air from the sea, as a result of solar radiation (Huffman et al., 1997). Seasonal occurrence of the monsoon is caused by the slight inclination of the earth’s pitch at an angle of 23.5ᵒ. This causes opposing seasons between the northern and the southern hemisphere (Kripalani & Kulkarni, 1997). In Malaysia, the occurrence of the monsoon season is divided into three main periods, the North East Monsoon that occurs between early December and ends in March; the South West Monsoon that begins early June and ends in October; the remainder of the period is known as the Inter-monsoon.

In Malaysia, this movement of wind occurs more significantly on a seasonal basis every year. In fact, during the monsoon season in Malaysia, the winds that blow regularly are usually within three months between monsoon seasons and other seasons. There are four monsoon seasons in Malaysia. Two of the main seasons are known as the Southwest Monsoon season and Northeast monsoon season. The Southwest monsoon is said to have drier conditions than the Northeast monsoon.

3.2.3 Offshore Wind and Wave Pattern The weather patterns over the entire South East Asia region are mostly governed by persistent pressure systems developed over the large landmasses of the Asian and Australian continents. A high-pressure system develops over the continent in the winter hemisphere while a low pressure system develops over the continent in the summer hemisphere (Nurukamal & Razali, 2016). This is the cause of the occurrence in monsoon seasons. Further analysis is carried out on the behaviour of offshore wind pattern in Section 5.1.2. As for off-shore wave climate, the wave behaviour is largely influenced by wind as detailed in Section 2.1. This is further proven by the data driven analysis carried out in Section 5.1.3.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

3.2.4 Regional Current Conditions The monsoon winds dominate the surface current systems of the sea and hence dominate the currents on the continental shelves in the area. This is further illustrated in Figure 3.4 and Figure 3.5

Figure 3.4 Predominant surface currents in January from South China Sea Plot (Growmodel, 2009).

As shown in Figure 3.4 during the NE monsoon period water flows from the Pacific into the South China Sea through the Bashi Channel/Luzon Strait towards the China Coast. The flow is forced by the monsoon winds. The moving water masses create a strong south-westerly current along the South China coast. It continues along the Vietnam coast, where strong currents are observed. A counter-current (return flow) develops off the Borneo coast as a result of the wind-induced set-up from the persistent north-easterly winds. These current joins the water from Sulu Sea and again merges with the main drift off the China Coast. This leads to a weak north- easterly current at the site, even during the NE monsoon.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 3.5 Predominant surface currents in July from South China Sea Plot (Growmodel, 2009).

During the SW monsoon period shown in Figure 3.5, the current direction develops parallel to the prevailing wind direction develops. A strong inflow of water from the south occurs through the Karimata Strait and runs northwards along Peninsula Malaysia towards the Vietnam coast where the strongest currents are generated. One branch generates a weak circulation pattern in the Gulf of Thailand, while the main drift flows along the southern Vietnam coast and continues towards Hainan Island and the coast of South China. The wide and uniform drift in the northern part of the sea exhibits a deflection to the right of the wind. The monsoon drift pushes the water of the South China Sea through the Bashi Channel and Luzon Strait. Due to the moderate supply of water from the Java Sea, a counter current offshore of Sabah develops and flows towards the southwest, providing additional water to the main monsoon drift. This return flow may at times affect the site as its separation point along the coastline is not stationary.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4 Data Utilisation

To carry out the research assessment, data have been collected from various sources. These include:

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

I. Bathymetric and Topographic Survey Data • Field Survey • BRSO Map II. Acoustic Doppler and Current Profile Data Collection (Transects Measurement) for model sensitivity tests. • Flow Velocity (Current Speed) • Flow Direction

Water Level

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

III. Wind Conditions • From Global Forecast System

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

IV. Tides • KMS Tide Model • Local Tidal Stations V. River Discharge • Gauging Station VI. Sediments • Grab Samples • Gauging Station VII. Wave Conditions • GROW Ocean Wind Wave Model

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.1 Bathymetric and Topographic Survey The detailed bathymetric survey in the vicinity of the river channel entrance was produced using an echo sounder. Bathymetrical survey data are presented in ASCII format using BRSO map projection with vertical datum being equal to chart datum (CD). Chart Datum (CD) in this project site is based on the Kota Kinabalu Port Benchmark which is 1.23m below mean sea level. Track lines of the bathymetric survey are shown in Figure 4.1 where the interval of the survey line is 100m, by the recommendation in DID Manual (DID, 2009). Measured Bathymetry is used in the river channel and the vicinity of the river channel entrance except for surrounding sea, where BRSO map projection is used. The bathymetry in the river channel entrance is shallow with depth variation between +0.5m and +1.25 m, as a result of sediment deposition causing very shallow conditions along the banks and in the middle of the channel entrance. The seabed becomes gradually deeper seawards, whereas the river has an average depth of -2.5m msl in the middle of the channel.

Survey Points

Figure 4.1 Bathymetric and topographic survey at the vicinity of Petagas River Entrance. The individual survey points are as indicated in the figure.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.2 Acoustic Doppler and Current Profile Data Collection An Acoustic Doppler Current Profiler (ADCP) is an instrument to measure current speed and direction over the entire water column. An ADCP measures data non- intrusively with unmatched efficiency compared to other equipment (Kim & Muste, 2012). The main component ADCP measures are the flow velocity, flow direction and the water level overflow depths and transect. A comprehensive range of scientific and practical analyses related to ocean and river hydrodynamics are supported by the data obtained from the instrument (Dinehart & Burau, 2005; Guerrero et al., 2014). ADCP data investigation and collection is conducted by validating non- homogeneous numerical flow solutions (Harding et al., 2015). For model validation, comparison with simulated results was carried out with the data obtained from an ADCP at the site in this study as shown in Figure 4.2. The ADCP was deployed in three locations (river channel, deep water and shallow water) between 7th May 2010 at 11:00 to 21st May 2010 at 13:30 for each location. The location of the ADCP deployment is determined by considering the river channel entrance as a whole including the river, the open sea, and the vicinity of the river channel entrance, for the purpose of model sensitivity tests. As for the measurement duration, this was decided according to the Guidelines for Preparation of Coastal Engineering Hydraulic Study and Impact Evaluation published by (DID, 2001). While it would have been better to have more deployment and prolonged duration, cost and time constraints are the determining factor. In this case, the deployment locations served the purpose of determining current speed, direction and water level in three different locations. As for the measurement duration, 14 days is sufficient to cover a full neap-spring tidal cycle, with other effects due to river flow and storm surge being considered unnecessary for the purposes of calibrating the model.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 4.2 Location of ADCPs

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.3 Wind Conditions To include the regional effect of wind on the local current conditions, nowcast wind data were obtained from the Global Forecast System (GFS). The Global Forecast System (GFS) is a global numerical weather forecast model run by NOAA (National Oceanic and Atmospheric Administration, USA). Data (pressure and wind fields) were nowcast using a spatial resolution of 1°×1° at 3-hour intervals.

Wind speed and direction measured at KKIA during the survey period were purchased from the Malaysian Meteorological Department to assess the effect of local wind conditions on current conditions. Time series of hindcast offshore wind data (speed and direction) were purchased from the GROW wind wave model as shown in Section 3.2.2.

In a nutshell, the GROW data forms an input to the boundary condition to generate offshore waves in the Wave model, while GFS data are input to the hydrodynamic model to include the wind effects in the hydrodynamic model. As for the KKIA wind data (airport adjacent to the site), these are used to enhance the accuracy of the local model mainly in the vicinity of the river channel entrance. Both GROW and GFS data are results from global numerical weather forecast models and are well suited for numerical model input purposes.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.4 Tides For tidal forcing and verification of simulated water levels, data from the KMS global tide model and nearby tidal stations were used.

The KMS global tide model is a database representing the major diurnal (K1, O1, P1 and Q1) and semidiurnal tidal constituents (M2, S2, N2 and K2) with a spatial resolution of 0.25°×0.25° based on TOPEX/POSEIDON altimetry data. The data are mainly applicable in relatively deep water, i.e., in depths greater than 20m.

Information on tidal variations at nearby stations was obtained from tide tables given by the UK Admiralty and the Royal Malaysian Navy.

The main tidal observations were obtained using one submersible tide gauge (RBR TGR-1050) positioned at BRSO coordinates Lat: N 654467.526 Long: E 704401.545 (Figure 4.4). The tide gauge was configured for one month of self-recording with 10- minute intervals between observations.

Figure 4.3 RBR TGR- 1050 Tide Gauge

Figure 4.4 Tide gauge installed at Petagas River Channel Entrance (View from Sea)

The location of the RBR TGR-1050 tide gauge was approximately 500m away from the river channel entrance. The tide gauge was tied to the Land Survey Datum and Chart Datum so that the observed data could be analysed to derive the possible harmonic constants and various tidal variations (HAT, MHHW, MLHW, MSL, MHLW, MLLW and LAT)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 4.5 Diagram of connections between temporary benchmark, marking at pole and Sensor

The tide gauge was deployed by attaching it to a pole inserted vertically into the seabed. The tide gauge was submerged approximately 1.794m below the surface, and pressure sensor was fixed at an offset of 0.3m from the seabed to prevent the sensor from being blocked by mud or silt. The end of the pole was levelled and tied into the survey control station; the purpose of doing this was to relate the tide gauge data to the Land Survey Datum (LSD). The reduced level (RL) of the transducer or pressure sensor related to LSD could be computed through direct reduction from the end of the pole to the sensor position.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.5 River Discharge The industrial norm for collecting river discharge data makes use of discharge rating curves established upstream using a river gauge (Herschy, 2008). An alternative method is based on acoustic velocity measurements and is only installed in areas of interest and generally in developed countries. Whichever method used, the collection point of flow data are usually located upstream of the tidal limit to overcome methodological difficulties caused by tidal influence (Herschy, 2008)(Moftakhari et al., 2013). A major shortcoming of this monitoring strategy lies in the uncertainties caused by possible freshwater additions and losses between the river channel entrance and the gauging station, which is usually located way upstream and away from the tidal limit. Hence to address this issue, correlation of discharge data derived from ADCP transects (directly measured at high or low tide or otherwise corrected for tidal flow) at Sungai Petagas and discharge data from Sungai Moyog is attempted.

A river transect analysis approach for discharge as shown by a case study (Zhou et al., 2014) features an excellent representation of discharge flow analysis, where the spatial pattern of the neighbourhood riverscapes changes longitudinally from upstream to the river channel entrance. Therefore, the assessment of the maximum river discharge is based on the river cross-sections closer to the Petagas River channel entrance. As for the determination of the seasonal average discharge value, it is based on water level and discharge data obtained from DID at Sungai Moyog gauging stations upstream. The normal discharge value is taken as 41 푚3⁄푠 for NE monsoon; 40 푚3⁄푠 for SE monsoon and 20 푚3⁄푠 for Inter monsoon. The maximum discharge is 250 푚3⁄푠. These values are based on the gauge data upstream as well as the channel capacity of the Petagas River.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

4.6 Sediment Properties Concentrations of suspended sediments measured in Sungai Petagas are found to vary significantly within a tidal cycle. During the flood tide, sediments brought into suspension at the shallow river channel entrance are transported into the river and sediments from upstream are temporarily stored due to the inflow from the sea. Sediments are flushed out during the ebb tide where higher concentrations in the sea are measured (sediment concentrations in the sea and river are in opposite phase).

To assess sediment loads in the river (not affected by the tides) data were obtained from Alam Sekitar Malaysia (ASMA) at the Sungai Moyog station. The measured sediment concentrations vary with the seasons but show a surprisingly higher average concentration outside the monsoon periods where run-off from the hinterland is supposedly lower. The suspended sediment has irrespective of season been determined as having an average load of 25mg/l. This may be considered relatively low and does not include any contributions downstream of Sungai Moyog, but there are at present no data available to refine this figure.

Figure 4.6 Sediment grab sample locations and mean grain size at the vicinity of the Petagas River Channel entrance

Sediment sampler is collected using Van Veen Grab Sampler. The weight of the Van Veen grab is about 18 kg and suited to take large samples in soft river or sea beds. The long lever arms allow it to cut deep into softer bottoms. Grab sampling is sufficient in this case as opposed to a core sampler because only top soil sediment characteristic is needed for evaluation.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Hence, sediment grab samples have been collected and analysed for grain size distribution at different locations. These samples are collected and ultimately distributed into multiple fractions to better reflect the sediment properties in the vicinity of the river channel entrance.

South of the river mouth, the mean grain size increases significantly towards the shore reflecting the effect of wave radiation stresses increasing inwards through the surf zone. The measured mean grain sizes and cross-shore profiles correlate well with the dynamic equilibrium beach/shoreface profile being a common guiding principle for shoreline change.

4.7 Wave Conditions To demonstrate a well-represented model, a wave condition is required at the boundary. This will vary for different monsoon seasons. For determination of nearshore wave conditions for design and assessment of sediment transport capacities, long-term wave records are required. Since no long-term measurements are available near the site, nearshore wave conditions were derived through transformation of hindcast offshore wave data. Wave and wind data were obtained from the GROW wind wave model covering a period of forty (40) years.

The GROW wave model is a global wind wave model based on reanalysis of wind field and assimilation to wave measurements. The model has a spatial resolution of 0.625°×1.25° with wave and wind information given at three-hour intervals. The model is run by Oceanweather Inc, which is also a provider of metocean design data for deep water oil and gas fields in the South China Sea. The model does not include shallow water effects and is only applicable in relatively deep water, i.e. depths greater than 200m.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5 Methodology

Figure 5.1 which is sub-categorised into Mapping and Conceptual Model, Reduced Complexity Model, Reductionist Model in which all of the three models is backed by Data and Data-driven Analysis.

The reductionist model is mainly applied in the nearshore local model to describe hydraulics and to simulate sediment dynamics and the behaviour of proposed management works of the river channel entrance. The reductionist model is produced in a high-resolution grid with short time steps in order to accurately resolve flow patterns and sediment dynamics over spatial scales ranging from large parts of the local model down to the vicinity of the coastal environment of interest.

The reason for these simplifications of models is to gain efficiency in computational time and resources with minimal sacrifice in model representation. These simplified models are to strike a balance in the model complexity spectrum where descriptive construction and intuitive simplifications is included over the hierarchy of natural process as opposed to high complexity model which constantly involves the topographic characteristics of a location.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Data and Data Driven Analysis • Raw Data • Seasonal Justification • Historical Bed Changes

Mapping and Conceptual Model • GIS Mapping • Survey Mapping • Meshing • Conceptual Management Measures

Reduced Complexity Model

Regional Model (Open Sea) - Wave Generation at larger scale. MIKE 21 SW Model

Reductionist Model • Local Model - Hydrodynamic conditions at the immediate vicinity of the river channel entrance. • Model Sensitivity Tests. • Management Measures Model (Predictive Model)

Figure 5.1 Flow Diagram on Numerical Modelling Approach The overall methodology of this research is executed in 4 separate fields as shown between Section 5.1 - 5.4.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.1 Data and Data-driven Analysis Various data are gathered for the purpose of further assessment for Petagas River channel entrance research - the details of the primary data collected are elaborated in Chapter 4. The data gathered not only serve as model input but also function as a source for data-driven analysis which is further elaborated in this subsection. Which includes tidal regime analysis, seasonal wind pattern and offshore wave climate analysis and bed changes process.

5.1.1 Tidal Regime Analysis Constant constituent and tidal planes for Kota Kinabalu Port are based on Malaysia Tide Table 2010 and are presented in Table 5.1 and Table 5.2. It is noted that the tidal planes defined in the tide tables have been computed and determined from thirteen (13) years of tidal observations. The tidal regime is mixed, mainly diurnal with one tidal cycle per day and F=2 as derived using in Equation 5.1 by (Lisitzin, 1974). Diurnal Tides have one high and one low tide each day (Lisitzin, 1974). The term ‘mixed’ implies that there are two high tides and two low tides daily, but of unequal shape. This is further proven in Figure 5.2 by the generated time series of tidal levels in Kota Kinabalu, where there are five peaks and five lows in high latitude over a period of 5 days which balance out to be one tidal cycle per day, indicating mainly diurnal tides. The time series also indicates mixed tidal behaviour mainly occurs during neap tidal periods. This further enhances the tidal accuracy of the model. Constituents Amplitude M2 0.24 S2 0.10 K1 0.37 O1 0.31 Table 5.1 Constituent analysis for Kota Kinabalu Port base in Malaysia Tide Table 2010 (Department of Survey and Mapping Malaysia, 2010)

퐾1+ 푂1 퐹 = = 2 푀2+푆2 Equation 5.1

Name of Tidal Plane Abbreviation Tidal level (ref. to RMN Datum) Highest Astronomical Tide HAT 2.40 Mean Higher High Water MHHW 2.13 Mean Sea Level MSL 1.23 Mean Lower Low Water MLLW 0.32 Chart Datum CD 0.00 Table 5.2 Tidal planes for Kota Kinabalu Port (Department of Survey and Mapping Malaysia, 2010)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.2 Time series plot of simulated tidal level at Kota Kinabalu

5.1.2 Offshore Wind Climate Analysis During the NE monsoon, offshore winds are generally from the NNE to NE sectors, whereas the winds during the SW monsoon are from the S to SSW sectors with slightly lower intensity. The monsoon winds are not very strong compared to ‘storm’ systems, but during the height of the monsoon and especially over the areas of open sea, the winds are relatively consistent for extended periods of time. This is reflected in the wave climate, the surface currents and variations in the mean sea level due to very large-scale wind set-ups. Outside the two monsoonal periods, there is little wind directionality or strength. Typical wind and pressure distributions over the South China Sea for the NE and SW monsoons are shown in Figure 5.3 and Figure 5.4; respectively.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Wind data used for the present study has been obtained from GROW at a location offshore Kota Kinabalu covering a period of 40 years from 1970 to 2009 (see Section 4.3).

Figure 5.3 Typical wind and pressure distribution in January from South China Sea Plot (GROW, 2009b).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.4 Typical wind and pressure distribution in July from South China Sea Plot (GROW, 2009b).

The offshore hindcast wind data reveal that the wind conditions in the vicinity of the study area (and offshore) are strongly influenced by the Northeast (NE) and Southwest (SW) monsoon seasons. Wind conditions offshore of Kota Kinabalu clearly show the monsoon patterns. The offshore wind conditions are governed by the Northeast (NE) and Southwest (SW) monsoon seasons with a magnitude of 10-14m/s. Typically, the Northeast (NE) Monsoon season is from December to March and the Southwest (SW) monsoon season is from June to October. April to May and November are known as transition periods. Figure 5.5 and Figure 5.6 shows the monthly wind rose in terms of wind intensity (speed given in m/s) and direction (coming from) based on 40 years of hindcast data. 81

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.5 Monthly wind roses (January to June) based on hindcast data (1970 - 2009)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.6 Monthly wind roses (July to December) based on hindcast data (1970 - 2009)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.1.3 Offshore Wave Climate The offshore waves in the study area are primarily generated by the South China Sea monsoon winds as discussed in Section 6.1.1. The directional distributions of the offshore waves are therefore quite similar to the distributions of the offshore winds.

Figure 5.7 Monthly wave roses (January to June) based on hindcast data (1970 - 2009)

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.8 Monthly wave roses (July to December) based on hindcast data (1970 - 2009)

Wave roses for the whole 40-year data set for each month from January to June and from July to December are shown in Figure 5.7 and Figure 5.8; respectively. The overall pattern in the month-to-month wave directions are related to the dominant directions of the monsoon winds.

It can be seen in Figure 5.7 and Figure 5.8 that larger waves approach from north- easterly directions (November to March) and are associated with the Northeast monsoon. Weaker waves, approaching from south-westerly directions, can also be seen (June to October) corresponding to the Southwest monsoon.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.9 Offshore wave energy distribution as a function of significant wave height (a), peak wave period (b) and mean wave direction (c).

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.9 shows the wave energy distribution as a function of significant wave height, peak wave period, and mean wave direction. The values represent the contribution to the time-averaged wave energy calculated over the entire period covered by the data. It shows that 75% of the wave energy carried by wave height varying from 0.2m to 1.4m. The peak wave period from 5s to 8s carries around 50% of the energy. Mean wave direction is governed mainly by the NE monsoon, with directions from 15oN to 60oN and during the SW monsoon directions from 215oN to 270oN.

From this data driven analysis, the influence of the SW monsoon in the vicinity of the river channel entrance is felt between June to October, and the NE monsoon occurs between November to March. This is somewhat at odds with the general statement by (Malaysian Meteorological Department, 2008) in which the start and end are a month earlier during the SW monsoon.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.1.4 Bed-Changes Process The construction and extension of the runway for KKIA and related construction activities, as demonstrated in previous chapters have significant impacts on both the coastline and the Sungai Petagas River channel entrance.

Using the data-driven analysis method, comparison of bathymetric surveys for the river channel entrance carried out in year 2006 and 2009 with the bathymetric and topographic survey from year 2010 shows that the river channel entrance has changed significantly since the works for extension of the runway commenced as described in Figure 5.10.

Figure 5.10 Comparison of bathymetric surveys from year 2006 (top left), 2009 (top right) and 2010 (bottom).

As shown in Figure 5.10, the significant changes to the depth contours in the red box area are mainly due to the changes in current conditions caused by reclamation for the extension of the runway. It is worth noting that comparing the year 2009 and

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

2010, the changes are notable and cannot be explained by natural causes and must therefore be ascribed to dredging activities. As a result of the runway extension, the river channel entrance has become very shallow with a sand bar stretching approximately 200m southwards from the runway reclamation. Significant accretion has also occurred along the southern river bank close to the tip of the headland.

Figure 5.11 Changes in bed levels at the river channel entrance from 2006 to 2010

The surveys have further been used to estimate the bed level changes and transported sediment volumes within the vicinity of the river channel entrance. The estimated bed level changes are shown in Figure 5.11 with areas denoted A to F used for estimation of sediment volumes. The volume extracted at the sand spit (Area E) is estimated at 63,000m3, not including natural backfilling since the dredge. The changes to the depth contours (Area B) have resulted in a deepening corresponding to a volume of approximately 17,000m3, probably deposited within the river channel entrance. The estimated deposition within the river channel entrance (Areas A, C, D and F) amounts to a total of 280,000m3 of which probably 80-100,000m3 was deposited along the southern river bank (Kuala Petagas) since the sand spit was removed as a result of changes in flow orientation from the river discharge. The accretion is belived to have occurred in zones A & D, which are sheltered from the wave action. This is also further proven by the model in Figure 6.16c and satellite image in Figure 6.17 as described in Section 6.3.2.2.2.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.2 Mapping and Meshing. Mapping, in this case, is done using GIS where a further understanding of the topographic nature of the river channel entrance can be understood as previously described in Section 2.5.1. This enables a graphical representation of the geomorphological nature of the research area. Mapping also functions as a valuable input (Bathymetry Study) before meshing is carried out. Mapping is also necessary to propose and analyse the potential mitigation measures suggested as shown in Section 5.4.3 to enable a graphical presentation.

The survey area covers approximately 1.4km along the length of the Petagas River. It begins from 400m upstream after the bridge, crossing towards the downstream to the Petagas River channel entrance and then extends seawards on the left towards Contoh Village (approx. 6.5km x0.5km). The echo sounder was calibrated at the start and end of each day's work when sounding was required. The calibration method known as 'Bar Check' was carried out using a circular bar plate attached to a measured length of chain. The bar-check calibration was to account for the daily variation in acoustic velocity propagation through the water column. The bar was also used to check the echo sounder's transducer draft to make sure the echo sounder accuracy is in good condition throughout the duration of the hydrographic survey work.

The spacing of the survey lines along the Petagas River is at 200m intervals. The survey lines connecting the profiles were not parallel to the riverbanks but approached the riverbanks at an oblique angle. The spacing between the survey lines at the offshore are approx. 100m-300m intervals, depending on the location of the survey line. The area that covered in the survey with the length is approximately 3km to Tanjung Dumpil and 3km towards the edge of the airport runway. The sounding data are fixed at an interval of 10m length. The sounding lines were designed perpendicular to the beach and river banks.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

As for the river topographic survey, detailed survey work was carried out 1.4km inland from the river channel entrance. From the river channel entrance, shoreline profiling survey work has been started towards the Tanjung Dumpil on the left, and from river channel entrance until Tanjung Aru on the right. The survey work was started based on a good and proper planimetric traverse datum and height control at the site. The planimetric datum relies on the proper old mark, and the height datum is based on the BM SA 2130 (RL 2.815m). The accuracy of planimetric control is better than 1:8000 and the height not more than 1:0.02√푘 where k is total distance as specified in the survey regulations. Height datum from BM SA 2130 is transferred to the control point and TBM in the study using the levelling method. Meanwhile during the topography and shoreline profiling survey work, height datum is transferred from one station to another by tachymetry method together with survey traversing.

Meshing is carried out by achieving a sufficiently high spatial resolution for nearshore sediment transport modelling at a reasonable computational run; the numerical wave transformation is performed in two stages. The GIS approach for this coupled system is an intricate process that requires bathymetry data from official sea charts using C- MAP and detailed site survey data for nearshore site for the generation of denser mesh for better model accuracy. The bathymetry and mesh used for the transformation of waves from offshore are shown in Figure 5.12 and Figure 5.13. For nearshore wave modelling a higher spatial resolution is required and a local and more detailed model is used to transform the waves from deeper water (40m MSL) to near shore. Figure 5.12 shows an overview of the bathymetry and unstructured mesh applied for the finer local model, with detailed views for the near shore of Sungai Petagas River channel entrance shown in Figure 5.13.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Depth

Figure 5.12 Overview of Model Mesh Area

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Depth

Figure 5.13 Local Model Mesh Area

Figure 5.12 and Figure 5.13 shows the meshing condition from the ocean boundary to the river. In general, there are five types of mesh density as categorised by area 1, 2, 3, 4 and 5. Mesh in area 1 accepts the input from the boundary and feeds into area 2 and gradually transfers the computing input parameter to area 3 and 4 which is also considered as the transition zone. Mesh in area 1, 2, and 3 are considered coarse compared to the rest. Area 5 is considered to be the most critical area as it is the river channel entrance where the ocean flow meets the river flow, hence a fine mesh is used to provide accuracy. Apart from that, a refined mesh can also be found in area 4 or in those areas where any variability in bathymetry is critical to the outcome.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.3 Reduced Complexity Model The regional model wave generation is carried out for the purpose of wave transformation to well represent wave characteristics into the reductionist model in the later stage. The model simulates the wave propagation from deep water to nearshore areas. The wave condition for this research is generated using the MIKE 21 SW spectral wave model as elaborated in Section 2.1.

In a nutshell, the spectral wave model is built using mapping and meshing of the model domain data from Bathymetric Survey Data elaborated in Section 4.1 and using the mapping and meshing method as mentioned in the previous chapters. Then, the GROW data are input as the boundary condition to generate offshore waves in the Wave model and transferred to the model as elaborated in Section 4.7.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.4 Reductionist Model

5.4.1 Hydrodynamic Modelling The prevailing hydrodynamic conditions in the vicinity of the river channel entrance are of importance for understanding the subsequent hydrodynamic and morphological impacts for pre-existing and post management measures. The hydrodynamic conditions for this research are obtained using the MIKE21 HD flexible mesh modelling system as mentioned in Section 2.2.

5.4.1.1 Model Calibration Parameters To ensure the tidal variation and current conditions are predicted accurately within the model domain, the model is calibrated against tidal stations along the north-west coast of Sabah.

Figure 5.14 shows the locations of relevant published tidal stations within the regional model. The outcome of the calibration is shown in Section 6.2.1.

Figure 5.14 Locations of calibrated tidal stations within regional model

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

During the calibration process, model parameters considered are as follows:

Bed Friction

The MIKE21 hydrodynamic model is governed by the depth-integrated Navier-Stokes equations subject to the shallow water assumptions and Reynolds-averaging of sub- grid scale turbulence. In the governing equations, a Manning formulation of bed friction is adapted and the drag coefficient highlighted by (DHI b, 2007) 퐶푑 is expressed in Equation 5.2 below as;

푔 퐶푑 = 1 2 ⁄3 푀 ℎ Equation 5.3

Where g is gravitational acceleration; h is the depth of the waterbody; M is the Manning Number (the Manning Number is also seen in the literature as n=1/M). The Manning Number is well known in both traditional as well as numerical hydraulics. For the MIKE21 hydrodynamic model, the relationship with Nikuradse bed roughness length can be estimated from the Manning number using Equation 5.4 (DHI b, 2007).

25.4 푀 = 푘1⁄6 푠 Equation 5.5

Where 푘푠 is the surface roughness height, Nikuradse’s equivalent sand roughness.

Due to the large water depth, the bed resistance is of relatively little importance and a constant Roughness height value of 0.05 (mm) has been applied throughout the model area. This equates to a constant value of Manning’s M of 41.8, which is input into the model as shown in Figure 5.15.

Figure 5.15 MIKE 21 Flow Model FM, Hydrodynamic Module: Bed Resistance

Attempts have been carried out to further prove the relevance of the Manning number input (bed friction) compared to the tidal level outcome and it is proven to be irrelevant as far as tidal level is concerned (Table 5.3):

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Manning Number input RMSE in Tidal Level 35 7.8 % 40 7.8 % 41.8 7.8 % 45 7.8 % Table 5.3 RMSE Water Level Comparison to Kota Kinabalu Tidal Station

Eddy Viscosity

In most hydrodynamic models the turbulence closure problem, resulting from the Reynolds averaging of the Navier-Stokes equations over the water column to yield the depth integrated St. Venaint equations, is solved through the Boussinesq eddy viscosity concept relating the Reynolds’ stresses to the mean velocity field. In this manner, the problems of describing the turbulent fluctuations are transformed into a description of an eddy viscosity.

It is essential to understand that in a numerical model classical turbulence is only one of several processes with similar behaviour and characteristics. In the discrete world, processes that are not resolved by the adopted grid are typically called sub-grid scale processes. In analogy similar to that for turbulent fluctuations, these processes lead to terms similar to the eddy viscosity. Often all these terms are lumped together into one (eddy) formulation.

The eddy viscosity, E, varies in time and space and depends on the changing flow features. It is calculated by the Smagorinsky formulation in each grid point and at each time-step by (Smagorinsky, 1963):

After a few attempts the constant Cs was obtained as per Table 5.4:

Cs Value RMSE in Tidal Level 0.2 8.4 % 0.24 8.0 % 0.28 7.8 % 0.3 8.2 % Table 5.4 RMSE Water Level Comparison to Kota Kinabalu Tidal Station

Therefore, the constant used in the Smagorinsky formulation is set to be Cs = 0.28, as this resulted in the lowest variation. This figure has been used extensively in many well-calibrated MIKE 21 hydrodynamic models. This constant apart from being recommended by DHI Water and Environment (DHI b, 2007) is also used widely in various numerical modelling endeavours such as (NINGSIH et al., 2011) on modelling of flooding due to tidal influence along the Coast and on nearshore modelling for tsunami runup modelling (LESCHKA et al., 2009); all which were applied to modelling approaches in this region.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

This input is applied in the model as Horizontal Eddy Viscosity as shown in Figure 5.16 :

Figure 5.16 MIKE 21 Flow Model FM, Hydrodynamic Module: Eddy Viscosity (Horizontal) Constant Value Input

This has become the basis for hydrodynamic modelling and is validated with two different sets of current data measured around the vicinity of the river channel entrance using ADCP. The integrity of the outcome is then assessed using the RMSE method as further expound in Section 6.2.2

Wind Friction

The wind friction in relation to wind speed is input as linear variation between 0.001255 at 7 m/s and 0.002425 at 25 m/s which is the recommended value according to the model (DHI e, 2012). Although detailed studies have been carried out concerning wind friction, none has been proven to be relevant to this region as there are no proven wind friction data sources (Peng & Li, 2015; Port & Engineers, 2009; Sokolov & Chubarenko, 2012; Soon, 2009; Urlburt, 2007).

The wind friction value is kept as the recommended value as shown in Figure 5.17 as there is limited availability of wind data in this region.

Figure 5.17 MIKE 21 Flow Model FM, Hydrodynamic Module: Wind Friction

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.4.1.2 Model Verification Model Verification is the process where model predictions are then compared to measurements of separate and independent periods with different conditions. When the processes are justified, the model is able to reproduce a reasonable prediction of water levels and currents in different seasons and to predict various mitigation measures.

For verification purposes, water levels and currents have been measured near Sungai Petagas River channel entrance (seaward of the reclamation for the runway extension of Kota Kinabalu International Airport) and upstream of the river channel entrance location as shown in Figure 4.2 previously. This configuration allows a comprehensive coverage of location coverage to be achieved. In this case the elements are the deeper sea, the dynamic river channel entrance and the river serving as a discharge conveyance for the upper catchment. The outcome of the calibration is shown in Section 6.2.

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.4.2 Sediment Dynamics Model The main sediment material of concern here is fine sediment, because the surveys have shown this to be the prevailing sediment class. The behaviour of fine sediment brought into suspension and deposition by mitigation measured have been simulated using MIKE21 MT (Mud Transport), which is a multifraction cohesive sediment transport model that describes the processes of erosion, transport and deposition of mud or sand/mud mixtures under the influence of currents and waves. This is further elaborated in Section 2.7. The model operates interactively with the aforementioned developed hydrodynamic model and includes the temporal and spatial effects of wave conditions determined by the developed wave model. This modelling method adopted is known as coupling.

The key input for MIKE 21 MT is the distribution of sediment into three fractions. These include coarse (fines), medium (fines) and fine (fines) with mean sediment size of 0.08mm, 0.04mm and 0.007mm respectively, according to the distribution curve. This is according to grab sample tests results; the sediment is categorised as loam with almost equal distribution between 3 fractions. The other important input is the sediment load, which has been determined by the average of 25mg/l based on data obtained from the Alam Sekitar Malaysia (ASMA), a local authority administrating the Moyong river station upstream.

Sedimentation was represented over a full neap-spring tidal cycle for NE and SW monsoon periods in the evaluation of the existing condition and the selected management measures, which is further described in Section 6.3.2.2.

100

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

5.4.3 Conceptual Layout Consideration Various conceptual layouts for river channel entrance management works are assessed based on local management policies prior to a more detailed investigation of a preferred layout. These were proposed in the interests of improving the local environment and the livelihoods of local people.

Figure 5.18 Detailed view of bathymetry for the existing condition in the vicinity of the river channel entrance.

A detailed view of the existing river channel entrance conditions for Sungai Petagas is shown in Figure 5.18, and five (5) conceptual layouts for improvement work are shown in Figure 5.19 to Figure 5.23. A dredged channel without any training structures to improve the flushing capacity through the river channel entrance has proven to be inefficient in the long-run due to the historical high deposition rates and not been considered as a viable solution. The presented layouts are conceptual in the sense that they do not consider depth requirements, length of training structures and breakwaters, orientation of artificial beaches and terminal structures. Hence, the channel depth is determined by the natural depth at the seaward end of the training structures. However, the conceptual layouts allow for an assessment of the different impacts of the sediment transport and sediment deposition, which has to be considered for a workable long-term solution to improve flushing capacity and ultimately reduce the risk of upstream flooding and to improve the navigational conditions. The five conceptual layouts assessed are described below:

101

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.19 Detailed view of bathymetry for proposed layout 1 for river channel entrance improvement works (brown area indicates land to be reclaimed).

I. Layout 1 as shown in Figure 5.19. Diversion of existing river channel entrance with training utilising the existing revetment at the end of the runway reclamation. The newly diverted channel is dredged to -2mCD. The area south of the trained channel to be reclaimed forming a natural northward extension of the beach.

102

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.20 Detailed view of bathymetry for proposed layout 2 for river channel entrance improvement works (brown areas indicate land to be reclaimed)

II. Layout 2 as shown in Figure 5.20. Training of the river channel entrance maintaining the natural orientation of the river channel entrance with channel dredged to -1mCD. The training requires revetments to be constructed along both river banks. Reclamations are to be constructed towards the airport and at the western part of Kg Contoh.

103

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.21 Detailed view of bathymetry for proposed layout 3 for river channel entrance improvement works (brown areas indicate land to be reclaimed).

III. Layout 3 as shown in Figure 5.21. Partial training of river channel entrance to enhance sediment by-pass toward the north. River channel entrance maintaining the natural orientation of the river with channel dredged to -1mCD. Involves protected reclamations immediately south of the runway reclamation (round shaped) and at the western part of Kg Contoh.

104

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.22 Detailed view of bathymetry for proposed layout 4 for river channel entrance improvement works (brown areas indicate land to be reclaimed).

IV. Layout 4 as shown in Figure 5.22. Training of river channel entrance maintaining the existing river channel alignment with protective breakwaters and dredged to -2mCD. Involves training of river channel entrance similar to Layout 2, but with the construction of protective breakwaters and training walls.

105

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 5.23 Detailed view of bathymetry for proposed layout 5 for river channel entrance improvement works (brown areas indicate land to be reclaimed).

V. Layout 5 as shown in Figure 5.23. Training of river channel entrance with revetments and training wall extended to -3mCD to avoid sediments discharged from the river to deposit at the entrance. Involves training of river channel entrance similar to Layout 2, but with an orientation perpendicular to the depth contours.

The detailed outcomes for the waves, currents, and sedimentation in the existing and all the proposed layouts are given in Section 6.3.3.

106

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6 Outcome and Analysis for Engineering and Management Solutions.

This chapter will focus on the results of the verified model and the interpretation of various results as shown in each of the following sections.

There are three sections in this chapter; Section 6.1 deals with the existing morphodynamic condition of the Petagas River channel entrance which includes wave changes and the data-driven analysis to show the historical bathymetric bed changes. Section 6.2 shows the outcome of the model sensitivity test where calibration and verification is done according to the correlation between two different sets of measured data, namely tidal stations and ADCP data from the vicinity of the site. With the verified model, various management solutions are then assessed and determined based on the modelling outcome. Section 6.3.3 contains a determination of the crest level of the relevant structures based on data-driven analysis.

6.1 Existing Morphodynamic Conditions 6.1.1 Wave Conditions Wave climate is one of the important factors determining the complex behaviour of sandbar and coastal morphology (Hsu et al., 2006; Yin et al., 2012). Wave transformations for typical NE and SW monsoon conditions generated from MIKE21 SW are shown in Figure 6.1a and Figure 6.1c, with detailed wave fields near shore shown in Figure 6.1b and Figure 6.1d.

From the figures, most of the offshore waves are from the north-north-east (NNE) and the west-south-west (WSW), (associated with the monsoon seasons). In general, the largest waves approach from a north-easterly direction. Wave refraction causes the waves to become increasingly perpendicular to the depth contours in shallower water as the waves propagate towards the shoreline. The wave heights reduce near the coastline due to the effects of wave breaking and bottom friction. Figure 6.1a clearly illustrates the effects of diffraction, refraction and sheltering for offshore waves from northerly and north-easterly directions, resulting in reduced wave penetration into nearshore areas. Also, during the NE monsoon, refraction and sheltering change the predominant wave direction from NNE to westerly towards the shoreline, resulting in a more direct exposure to severe waves. A minor shoaling effect can be seen during both monsoons.

107

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

a B

c d

Figure 6.1 (a) Overview of transformation of waves from offshore to nearshore for typical NE monsoon condition (b) Detailed view of the transformation of waves from offshore to nearshore for typical NE monsoon condition (c) Overview of transformation of waves from offshore to nearshore for typical SW monsoon condition (d) Detailed view of transformation of waves from offshore to nearshore for typical SW monsoon condition.

108

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The shallow water in the vicinity of the river channel entrance affects the propagation of waves and the wave conditions at the river channel entrance vary with the water levels over a tidal cycle. These effects are included in the simulations by modelling the transformation of waves from offshore as shown in the first half of this chapter with the water level varying according to the tide. Results of the wave simulations are presented in Figure 6.2 to Figure 6.5 showing local wave fields for high water levels and low water levels during both NE and SW monsoon conditions.

Figure 6.2 Overview of existing wave pattern for high (left) and low (right) tide during NE monsoon.

109

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.3 Existing wave pattern at vicinity site for high (left) and low (right) tide during NE monsoon.

Figure 6.4 Overview of existing wave pattern for high (left) and low (right) tide during SW monsoon.

110

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.5 Existing wave pattern at vicinity site for high (left) and low (right) tide during SW monsoon.

In the wave simulation outcome as demonstrated in Figure 6.2, from the colour scale it can be seen that the largest offshore waves approach from the NE direction. Large- scale diffraction can be seen around the island to the north of the shore, as a result of the change of wave direction from the NNE to NW landwards of the island. The significantly lower wave exposure landwards and the vicinity of the river channel entrance during the NE monsoon period is due to the Sheltering of a string of islands (Pulau Gaya and Tunku Abdul Rahman Marine Park) as shown in Figure 6.2. The main reason to include the tide in this wave comparison is for the demonstration of the effect of water level, which has a marked influence on the wave conditions with more wave energy penetrating to the nearshore area during high tide conditions. The significant wave height at the river channel entrance is approximately 0.7m during high tide NE monsoon period, while higher significant wave heights of 0.9m are observed during high tide SW monsoon period as compared to low tide periods for both monsoon periods.

111

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.2 Model Calibration and Verification Outcome 6.2.1 Model Calibration Figure 6.6 shows a time series comparison between the simulated and predicted water levels for four tidal stations around the study area as shown in Figure 5.14 in Section 5.4.1.1. The simulations reproduce the predicted tides well along the modelled coastlines and are capable of capturing well both the tidal amplitudes and phases at the tidal stations.

112

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.6 Comparison of predicted and simulated timeseries for the 4 tidal stations namely Pulau Mantani Besar, Pulau Mangalum, Kota Kinabalu and Kuala Papar

113

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.2.2 Model Verification During the period where current measurements were performed, strong winds and heavy rain were reported, which affected the current conditions at the site.

Measured and simulated water levels, current speeds and directions at locations seaward of the reclamation and in the river are compared in Figure 6.7 to Figure 6.9. For the seaward locations, the water levels are reproduced well during both neap and spring tides. The measured current speed is seen to be low at both locations as shown in Figure 6.7 (ADCP 1) and Figure 6.8 (ADCP 2) with a strong influence of wind and little directionality due to the tide. Therefore, the accuracy in reproducing the current conditions with a forcing by tide and river discharge alone will be affected. Hence, local wind measurements from the airport given by hourly mean wind speed and direction have been applied, unfortunately with little success. The hourly mean wind speed is very low (for longer periods around 2 m/s), and little correlation is found between wind and current directions. In the vicinity of the river channel entrance, The current velocity is influenced by the discharge of the river and is governed by a narrow channel that directs the discharge over the river channel entrance tidal flat.

A comparison between the current speed measured in the river and the simulated results is shown in Figure 6.9. From the comparison, it is noted that the model is able to accurately simulate the water level and current speed, as well as to capture the effect of high discharge following a period of heavy rainfall on 02 August - 21:35 at peak velocity. The measured current directions are not bi-directional (as expected in a river) and significant cross currents are observed. The cross currents are ascribed to helical flow conditions due to river bends situated near to the measuring station and the uneven bathymetry of the shallow river.

114

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.7 Comparison of measured and simulated current at ADCP 1 (deep water)

115

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.8 Comparison of measured and simulated current at ADCP 2 (shallow water)

116

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.9 Comparison of measured and simulated current and water level in the river

117

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

ADCP ID RMSE Criteria RMSE Criteria RMSE Criteria (Current (Current (Water Speed) Direction) Level) ADCP 1 15% 30% 49.06° 45° 9.4% 10% ADCP 2 14.8% 30% 56.43° 45° 8.7% 10% ADCP 3 7.8% 30% 8.3° 45° 8.9% 10% Table 6.1 Root Mean Square Error (RMSE) of simulated and measured current speed, direction and water level.

Table 6.1 shows the statistical analysis outcome of the difference between measured and simulated results. All the criteria fall well within the tolerable limits according to the local hydraulic study guidelines (DID, 2001), except for Wind direction (not shown in the table) which did not fully meet the range criteria. This may be due to the site- specific challenges, which from a modelling perspective include:

• With the current field being wind-dominated, the current model is susceptible to any errors or resolution issues in the wind and pressure fields. • The currents are dominated by non-tidal effects driven by a combination of regional wind and pressure systems driving ocean circulation in the South China Sea, often opposing local winds. • The complex bathymetry at the edge of the continental shelf with large depth gradients combined with the local rugosity of the coral reef itself, surrounded by deeper channels. This leads to complex flows where currents are magnified by the transition from the deep ocean up over the continental shelf, combined with local channelling and sheltering effects by the reef.

To capture these effects, a 3D modelling approach covering a larger scale should be adopted, but this is not described here.

For a site with the present complexity especially with a rapidly changing bathymetry, the model is considered to perform well as it generally captures both the magnitudes, directions and variations over the depth well. In terms of the application of the model to assist the mitigation measures in decision making, the key is that the model can capture the correct range of current magnitudes and overall directional patterns, which will determine the location of sedimentation and areas at risk. This is seen to be the case, and by considering climatic scenarios for the modelling that includes net current flows in all the key directions and a range of magnitudes, it is ensured that the model predicted sedimentation and impacts are conservative. On this basis, the modelling tool is considered fit for purpose.

It is stressed that this calibration is done based on hydrodynamics aspect only and it is not within the scope of the work to carry out a full-on calibration and verification works with regards to the morphodynamics and wave climate.

118

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3 Conceptual Layout Consideration Outcome Various conceptual layouts for river channel entrance management works as elaborated in Section 5.4.2 are assessed and verified with the local regulation as elaborated in Section 2.8 before more detailed investigations of a preferred layout is carried out.

As highlighted before, the verified model in Section 6.2.2 is then used as a predictive model to assess the various conceptual layouts. The model platform used for modelling of the conceptual layouts is a coupled flexible mesh (FM) model that helps to simulate the mutual interaction between waves and currents using a dynamic coupling between the hydrodynamic (HD) and spectral wave (SW) modules. The model has direct input to the mud transport (MT) module with a complex formulation for combined wave and current actions used to derive the bed shear stresses and the amount of sediment in suspension and transported as bed load. To run the model information on wave heights, periods, mean directions and directional spreading are required for derivation of the so-called radiation stresses, which are included in the hydrodynamic model together with the “normal” driving forces for tides and winds to include the wave driven currents and water levels. Thus the model simulated the combined wind, wave and tidal driven water level and current fields. The model has a very detailed local mesh that properly resolves the surf zone and the processes in the vicinity of the river channel entrance.

6.3.1 Wave Condition Comparison Outputs from the detailed wave model primarily consist of radiation stresses that are used as inputs to the current model. The radiation stresses are responsible for driving the longshore (wave-driven) currents in the surf zone in the near shore reductionist model.

Figure 6.10 shows wave patterns for the existing condition and the five conceptual layouts for high tide during a typical NE monsoon period, while Figure 6.11 shows the wave patterns during a typical SW monsoon period.

According to Figure 6.10 and Figure 6.11, the reclamations and revetments proposed in the layouts will protect Contoh Village from direct wave exposure. The relatively wide entrance in Layout 3 results in a wave pattern that is not much different from the existing conditions. With the presence of breakwater in layout 4, wave penetration into the river is significantly lower compared to other layouts, and this is beneficial to avoid sedimentation resulting an improved river discharge. Some kind of positive ‘feedback’ loop may then be possible due to the resulting increased scour caused by the higher discharge.

119

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.10 Detail view of wave pattern comparison between existing condition and conceptual layouts during NE monsoon at high tide.

120

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.11 Detail view of wave pattern comparison between existing condition and conceptual layouts SW monsoon at high tide.

121

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.2 Management Solution Chosen As previously discussed, the main objective of this research is to validate a numerical model of sedimentation in the Petagas River channel entrance by assessing the effectiveness of introducing a set of proposed improvement works, aimed at minimising the dynamic sedimentation behaviour caused by inter-annual climate variations in the monsoon seasons.

For the purpose of proposing a solution that is practical and adaptable various layouts are assessed, and layout 4 is selected as the preferred solution as it is a solution that will comply with the local policy as discussed in Section 2.8. The major requirement for local policy makers is a management solution that allows a sufficient discharge through the entrance throughout the two monsoon seasons; this will then solve the upstream flooding issue. The proposed management solution should allow continuous access through the river channel entrance for both large and small scales of fishing craft, thereby helping to sustain the local economy. This layout is advantageous in terms of solving the current issues, and it will in the long term provide opportunities for local development.

Coastal structures commonly lead to detrimental coastal impacts in the vicinity. Designated improvement works with minimum impacts on the coastal environment are thus preferred in this setting

Airport Runway Extension

Contoh

Village

Figure 6.12 Management measures proposed for Petagas river channel entrance by maintaining the existing river channel alignment with breakwaters (Google Maps, 2009)

122

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The layout shown in Figure 6.12 involves the training of the river channel entrance to follow the natural alignment of the access channel. The training requires revetments to be constructed along both river banks as well as around the existing sand spit and seawards. Reclamations are being carried out to extend the airport to the south and also along the western side of Contoh Village. Breakwaters are to be constructed, and the main channel will be dredged.

123

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.2.1 Hydrodynamic conditions The instantaneous flood and ebb tide currents under different seasonal conditions are shown in Figure 6.13 and Figure 6.14, for the existing case and after the proposed improvement works. Comparing these figures allows an assessment to be made to improve the effectiveness of the works for the different modelled conditions.

a b

c d

Figure 6.13 Current speed and direction for NE monsoon for (a) existing case during flood tide conditions (b) following improvement works during flood tide conditions (c) existing case during ebb tide (d) following improvement works during ebb tide.

124

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

a b

c d

Figure 6.14 Current speed and direction for SW monsoon for (a) existing case during flood tide conditions (b) following improvement works during flood tide conditions (c) existing case during ebb tide conditions (d) following improvement works during ebb tide conditions.

125

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The current direction is seawards, reflecting the direction of the monsoon, as shown in Figure 6.13 and Figure 6.14. The river current is high during ebb tide conditions (Figure 6.13c, d and Figure 6.14c, d) with a difference of about 50% in terms of flow velocity compared with flood tide conditions (Figure 6.13a, b and Figure 6.14a, b). The flow patterns in the vicinity of the proposed improvement works are generally smooth, with minimal differences between the two sets of figures in the immediate vicinity of the works, where eddies are generated downstream of the structures. The flow patterns in the trained channel, the breakwater basin and the channel entrance are found to be smooth, with limited transverse currents and merely weak eddy formations (Figure 6.13b, d and Figure 6.14b, d).

Eddies are easily formed in the basin during flood tide conditions; however, the eddies are weak and not significant for navigation. The improvement works, therefore, provide an easily navigable access channel into the river.

126

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.2.2 Sediment Dynamics The main sediment source is from the Moyong catchment upstream. The changes in land use due to development have caused an increase in sediment load in the channel, which is ultimately flushed downstream. This process tends to be more intense during monsoon seasons where rainfall intensity is high within the catchment. Sediments discharged from Petagas River have historically fed the adjacent coastlines, although the recent reclamation for the extension of the runway for KKIA has created a river channel entrance area with weak currents, limited wave exposure and increased sedimentation rates.

To assess the impacts of the river channel entrance improvement works on the discharge of sediments, a sediment plume model has been setup (MIKE 21 MT). The model includes the transport of fine sediments (mud), deposition, resuspension and erosion considering the variation of hydrodynamic conditions with the tide and seasons. The suspended sediments have been modelled using three (3) fractions of sediments varying from fine silt to very fine sand, with the latter being present at the river channel entrance. Since deposition and erosion will vary over a tidal cycle with the possibility of different processes governing during neap and spring tides, the net deposition is given for a full neap-spring tidal cycle.

The simulations cover a full neap-spring tidal cycle and have been carried out for NE, SW and inter-monsoon periods. They are represented in terms of sediment plumes, in which a conservative approach is shown by demonstrating maximum concentrations of total suspended sediments (TSS) to show sediment behaviour in the worst-case scenario. This demonstration is for the purpose of showing sediment behaviour in suspension.

127

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.2.2.1 Sediment Plumes Maximum sediment plumes for the existing condition and with the prediction of the improvement works are shown in Figure 6.15. As shown in Figure 6.15a, c, e, it demonstrates that the sediment plume is very much contained within the river channel entrance area with a limited excursion along the adjacent coastlines throughout all seasons. It is useful to identify the fate of the river-borne sediment under different monsoon conditions because this will help us to understand the likely patterns of deposition.

After the management measures are established, the sediments are flushed into the sea. During the inter-monsoon period, the sediments may be lost to deeper water as shown in Figure 6.15f, whereas the sediments tend to be transported towards the adjacent coastlines during the NE and SW monsoon periods as shown in Figure 6.15b, d.

128

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

a b

c d

e f

Figure 6.15 Comparison of maximum concentrations of total suspended sediments (TSS) during the NE monsoon period for (a) existing condition and (b) following improvement works. Comparison of maximum concentrations of total suspended sediments (TSS) during the SW monsoon period for (c) existing condition and (d) following improvement works. Comparison of maximum concentrations of total suspended sediments (TSS) during the Inter monsoon period for (e) existing condition and (f) following improvement works.

129

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.2.2.2 Sedimentation As previously stated the main sediment source of sediment is the Moyong catchment upstream. The changes in land use due to development have caused an increase in sediment to be washed into the channel and ultimately flushed downstream. This process tends to be worsened during monsoon seasons where rainfall intensity is high within the catchment (Mitchell, 2013).

Numerical Two-dimensional mud transport simulations were carried out for the typical NE and SW monsoon periods in order to provide information on sedimentation, as shown in Figure 6.16.

NE monsoon conditions lead to a concentration of net deposition to the south of the river channel entrance, while SW monsoon conditions caused net deposition to the north. The configuration of the breakwaters demonstrates an effective means of directing the sediment into deeper waters. The beach is predicted to accrete to the south of the breakwaters, while erosion is likely to occur downdrift, along with the small stretch of coastline between the breakwater and the KKIA reclamation. Sediments are deposited inside the harbour basin as well.

Figure 6.16 delineates the changes of bed thickness for the NE, SW and inter- monsoon conditions respectively, as generated by the model. As illustrated in the figure, it demonstrates minimal or the absence of net sediment deposition in the trained part of the river channel or at the channel exit, which appears to be well flushed. The sediments are deposited outside the training walls and breakwaters, resulting in the formation of a new flat.

Following the management measures, sedimentation is predicted to take place in the lee zones within the harbour basin. The flow eddies generated on both sides of the access channel within the breakwater will flush away the sediment that enters through the breakwaters from the open sea and the river system towards the more protected parts of the basin, in which the transport capacity is low, and the sediments can settle.

The areas of sedimentation are mainly determined by the Northeast (NE) and Southwest (SW) monsoon. The sedimentation is strongly related to the current speed as shown in Section 0. The net deposition appears to cover a wider area during the SW monsoon (Figure 6.16c, d) as the current and wind directions are uniform during this period.

130

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

a b

c d

e f

Figure 6.16 Comparison of bed thickness changes (net deposition) during the NE monsoon period for (a) existing condition and (b) following improvement works. Comparison of bed thickness changes (net deposition) during the SW monsoon period for (c) existing condition and (d) following improvement works. Comparison of maximum concentrations of bed thickness changes (net deposition) during the Inter monsoon period for (e) existing condition and (f) following improvement works. In this monsoon condition, it shows the accuracy of the model by comparing to the satellite image as shown in Figure 6.176.19. This results shown is based the outcome of the model run over a particular monsoon season. 131

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Land Reclamation (Airport Runway Extension)

Deposition Zone

Existing Sand Bars

Contoh Village

Figure 6.17 Satellite image of current river channel entrance condition (Google Maps 2009)

Figure 6.17 shows the existing river channel entrance during the South West Monsoon period; the existing sand bars have formed since the land reclamation for the KKIA runway which had extended to the north of the river channel entrance in the indicated location. By comparing the model results as shown in Figure 6.16 with the satellite image, the bed changes in the deposition zone reflect the bed thickness changes shown in the model in Figure 6.16c.

132

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

6.3.3 Design Conditions for Management Solutions Design conditions of structural measures for preferred management solution of training wall and breakwater of the river channel entrance are based on the data- driven analysis approach. Works information on design water levels and wave heights is required to determine the crest level of the structures and to ensure the stability of armour layers.

The water level for the design of coastal structures is determined by a number of components including the tide, seasonal variations, general sea level rise and storm conditions resulting in surge and wave setup in the coastal zone. Some of these factors and their joint probability are associated with some uncertainty and the design water level will be varying with the level of conservatism required. As explained in the methodology chapter, the design water level is determined by the sum of mean higher high water (MHHW), seasonal variations, global sea level rise, storm surge and wave setup with the level given by the two last components depending on the return period for design.

The water level fluctuation of greatest concern in coastal design is storm surge or the meteorological tide, which is an increase in water level resulting from shear stress by onshore wind over the water surface. Storm activity can cause both a rise (setup) and fall (set down) of the water level. The temporary increase in water level coupled with major wave action is the cause of the world’s most disastrous flooding and coastal damage. For instance, the shoreline along the southern borders of the North Sea was flooded in 1953 as a storm surge caused dyke breaches, which resulted in extensive property damage and 1835 lives were lost in Netherlands. The threat of severe storm surge from Hurricane Floyd in 1999 resulted in the evacuation of 3 million people and 50 deaths in the United States (Kamphius, 2000).

During a storm surge, the water level at a downwind shore will be elevated until gravity (the slope of the water surface) counteracts the shear stress from the wind. Although the prediction of storm surge heights is dependent on many factors, an estimate of sea surface slope at the coast, caused by wind stress, 휏푠 effects, as highlighted by Chen & Liew, (2003) as;

푑푧 휏푠 = λ Equation 6.1 푑푥 훾푑

푑푧 where is the sea surface slope; 훾 is water depth; λ is an experimentally 푑푥 푑 −푚2 determined variable, ranges from 0.7-1.8 and 휏푠 = wind stress in 푘푔 as shown in Equation 6.2 (Chen & Liew, 2003) as ;

2.22 휏푠 = 0.58 푈10 Equation 6.3

133

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

where 푈10 = the wind speed in m/s measured at 10m above the sea level.

Storm surge can occasionally be determined directly from long-term measurements if the parameters are available at the site. However, with a lack of measurements, the storm surge has to be determined on the basis of hydrodynamic modelling of water levels and currents driven by extreme wind conditions causing water to be “trapped” and pushed against the shoreline.

The modelling is based on a large-scale hydrodynamic model (MIKE 21 HD) with wind forcing based on extreme offshore wind conditions extracted from the GROW wind wave data. The wind speed for the twenty-six worst offshore storm events has been used to estimate the storm surge. It is noted that the storm surge is not only determined by the wind speed, but also by the wind direction relative to the orientation of the coastline. Based on a few directional tests with the wind from 15, 255, 270, 285, 315 and 345o N, it has been found that a wind direction 270oN results in the largest surge. Examples of simulated storm surge are shown in Table 6.2.

Based on results from the simulations, surge levels have been extracted at the locations for the breakwater entrance and the end of the training walls. Based on extreme value analysis the surge levels are given and determined in Table 6.2

Return Period (Years) 2 5 10 50 100 Breakwater, 2m CD 0.12 0.17 0.20 0.27 0.30 Training Wall, 1m CD 0.12 0.17 0.20 0.28 0.31 Table 6.2 Extreme storm surge level (above MSL) corresponding to 2, 5, 10, 50 and 100 years return periods for existing conditions at selected locations.

134

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.18 Example of storm surge levels with wind from west (270oN). Left: Typhoon Greg with constant wind of 22.6m/s (most severe). Right: Storm with constant wind speed of 15.6m/s (ninth most severe).

Breaking waves on a beach give rise to a setup of the mean water level. The magnitude of this wave driven setup of the water level, called wave setup, depends primarily on the magnitude of wave breaking and the angle at which they approach the shore. The wave setup is based on a reductionist model coupling the wave and hydrodynamic processes driven by extreme offshore wave conditions (MIKE 21 HD and SW).

135

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Return Period Design Design Peak Design Offshore Design Water (Years) Offshore Wave Wave Periods Mean Wave Levels Heights Tp (s) Directions WL Hs (m) MWD (°N) (mCD) 2 2. 7 8.2

MHHW 5 3.3 9.1 270°N 10 3.7 9.7 (2.13m) 50 4.7 10.9 100 5.1 11.4 Table 6.3 Offshore design wave conditions based on GROW wave data

The simulations have been performed with a wave direction of 270oN (perpendicular to the beach) resulting in the maximum wave setup and a water level equal to MHHW.

Examples of generated wave setup for a 100-year event for wave directions corresponding to NE and SW monsoon directions are shown in Figure 6.19 is described in Figure 6.21.

Figure 6.19 Wave setup due to 100-year design wave with directions corresponding to dominant NE (left) and SW (right) monsoon directions.

136

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Figure 6.20 Detail view of wave setup Sungai Petagas River channel entrance due to 100-year design wave from dominant NE monsoon direction.

Figure 6.21 Detail view of wave setup Sungai Petagas River channel entrance due to 100-year design wave from dominant SW monsoon direction

137

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

The wave setup is seen to increase throughout the breaking zone, and it varies slightly from the location of the breakwater entrance to the end of the training walls (see Table 6.4).

Location Return Period (Years)

2 5 10 50 100 Breakwater at 2m CD 0.06 0.11 0.16 0.27 0.32 Training Wall at 1m CD 0.08 0.14 0.18 0.30 0.35 Table 6.4 Extreme wave setup (above MSL) corresponding to 2, 5, 10, 50 and 100 years return periods for existing conditions at selected locations.

The design of coastal structures usually requires the determination of a maximum and minimum design water level (D.W.L.) (Meadows & Wood, 2003). Design water level is computed from the addition of various components of water level fluctuation shown in Equation 6.4 as ;

D.W.L = MHHW + 푆푉+ 푆푤 + 푆푠 + SLR Equation 6.5

where MHHW is the mean high high water level, 푆푣 is seasonal variation level, 푆푤 is wave setup and 푆푠 is the Storm Surge and SLR is a provision for sea level rise.

The recommended design water levels are based on the principle of Equation 6.5 accounting for return periods at the breakwater entrance and at the end of the training walls as given in Table 6.5 and Table 6.6 respectively. Seasonal variation and values for MHHW have been obtained from tide tables of Kota Kinabalu Port. The provision for global sea level rise is based on the most recent report from the Intergovernmental Panel on Climate Change (IPPC), with the rise determined as the central estimate for all climatic scenarios.

138

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Return Period (Years) Parameter 2 5 10 50 100 MHHW (m) 0.9 0.9 0.9 0.9 0.9 Seasonal Variation (m) 0.1 0.1 0.1 0.1 0.1 Provision for Global Sea Level Rise (m) 0.4 0.4 0.4 0.4 0.4 Storm Surge (m) 0.12 0.17 0.20 0.27 0.30 Wave Setup (m) 0.06 0.11 0.16 0.27 0.32 Design Water Level (mMSL) 1.58 1.68 1.76 1.94 2.02 Table 6.5 Recommended design water level (above MSL) at breakwater entrance.

Return Period (Years) Parameter 2 5 10 50 100 MHHW (m) 0.9 0.9 0.9 0.9 0.9 Seasonal Variation (m) 0.1 0.1 0.1 0.1 0.1 Provision for Global Sea Level Rise (m) 0.4 0.4 0.4 0.4 0.4 Storm Surge (m) 0.12 0.17 0.20 0.28 0.31 Wave Setup (m) 0.08 0.14 0.18 0.30 0.35 Design Water Level (mMSL) 1.60 1.71 1.78 1.98 2.06 Table 6.6 Recommended design water level (above MSL) at end of training walls.

139

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Descriptions of the estimation of extreme waves include those given by Sarpkaya and Isaacson (1981), Isaacson and MacKenzie (1981), Muir and El-Shaarawi (1986), and Gran (1991). In developing design wave conditions, hindcast or recorded data of one or more sea state parameters (e.g. significant wave height), are fitted by a suitable probability distribution and further extrapolated to estimate an appropriate design sea state with a corresponding specified return period or annual risk. An estimation of the maximum individual wave height 퐻푚 as highlighted by (Isaacson & Foschi, 2000) as shown in Equation 6.4 within the design sea state with significant height 퐻푠 can be made as well and often taken as ;. Equation 6.4 퐻푚 ≈ 1.8 ~ 2.0 퐻푠

Design wave conditions at the breakwater heads and at the end of the training walls for preliminary design have been determined from 40 years of GROW wave data, transformed to the river channel entrance using a spectral wind wave model (MIKE 21 SW). For the wave transformation, the design water levels determined in section 6.3.1have been considered. This is shown in Table 6.7.

Significant wave height Return Period (Years) Hs (m) 2 5 10 50 100 Mean Sea Level Training wall 1.50 1.65 1.75 1.90 2.00 Breakwater 1.70 1.90 2.00 2.25 2.30 Design Water Level Training wall 1.90 2.15 2.35 2.70 2.85 Breakwater 1.95 2.30 2.50 2.90 3.10 Table 6.7 Significant wave heights for design determined at breakwater entrance and end of training walls.

140

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

7 Answer to Research Questions

The overall research objectives and aims as shown in Section 1.3 have been achieved through integrating each of the approaches as shown the previous chapters 2 through 6. Each section has contributed to the progress in the fields of river channel entrance geomorphology, data used, modelling methodology and outcome, and each were written with a disciplinary focus. In this chapter, the main results will be synthesised, followed by placing the results into an integrative perspective. Research questions posed in Section 1.2 will be answered here.

141

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

7.1 Answer to Research Question 1 Do the monsoon seasons influence the topography of the river channel entrance and if so how do they influence patterns of sedimentation?

The monsoon seasons affect the topography of the river channel entrance in various ways. The monsoon winds will dominate over the surface current systems of the open sea in which the predominant driving force is the ocean current as described in Chapter 2. From the view of water catchment, the river flow varies over the monsoon season due to different rainfall pattern, with tropical storms occurring during monsoon season.

The offshore wave analysis generated in this analysis is shown in Section 3.2 in which wind is the main driving force of the wave direction. Wave roses have been generated by utilising 40 years of wave data to support this statement, and the overall pattern of the wave direction is closely related to the dominant monsoon wind direction as shown in Section 6.1, which illustrates the detailed generated wave field near the vicinity of the river channel entrance. The topography of the site has caused most of the offshore wave derived from the north-north-east (NNE) and west-south-west (WSW) as a result of the influence of monsoon seasons towards the geomorphology of the vicinity of the river channel entrance. Also, the NE monsoon refraction and sheltering have changed the predominant wave direction from NNE to westerly towards the shoreline, resulting in more direct exposure to severe waves. A minor shoaling effect can be seen during both monsoons. The river and the catchment will generate more runoff and cause a higher river discharge as a result of monsoon rainfall with an average of 40 m3/s during Northeast and Southwest Monsoon and 20 m3/s during inter-monsoon periods. However, during extreme weather occurrences (Typhoons) which occur during the monsoon period, extreme discharges can reach up to 230 m3/s causing significant volume of loose sediments flushing down to the Petagas River channel entrance (Mitchell, 2013).

The dynamic topography of the river channel entrance is constantly varying throughout the year as a result of monsoon seasons as shown in Section 6.3.2. During the NE monsoon a net southerly direction of sediment transport results in minimal bed changes. Contrarily, the SW monsoon causes more bed changes due to a higher northerly ocean current resulting in higher sediment transport. During the inter- monsoon, the hydrodynamic behaviour of both land and sea is the calmest and will produce higher sedimentation in the vicinity of the river channel entrance within the shallow water as well as over the tidal flat created to the south of the airport runway reclamation extension. Monsoon season will provide a form of flushing effect to the river channel entrance resulting in higher natural conveyance in the formed channel. However, this sudden episodic event may cause fluvial flood upstream if the channel entrance is under capacity as a result of sedimentation causing a backwater effect (Mitchell et al., 2012).

142

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

With the preferred management solution in place as shown in Figure 6.12 consisting of a training channel following the natural alignment of the existing access channel as well as reclamation to form the channel and the natural coastal alignment, a constantly adequate open channel is available throughout the monsoon seasons with sufficient discharge capacity.

143

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

7.2 Answer to Research Question 2 What is the optimum management solution to the sedimentation problem?

Sedimentation at the river channel entrance is the root cause of many social and environmental problems, which include flooding as a result of poor management due to the poor understanding of the consequences of devastation flooding (Hoggart et al., 2014). Hence the optimum management solution must be able to solve sedimentation problem with a solution that will be able to mitigate socio-economic and ecological impacts of a particular area. This is shown in Section 2.8 where further description on Shoreline Management Plans (SMPs) is evaluated.

In this research several engineering options are considered, amongst them are breakwaters, revetments, reclamation and dredging. Due to the dynamic nature of the river channel entrance, a combination of these solutions is to be adopted and tested with different conceptual layouts as shown in Section 6.2 using the verified numerical model to ensure the feasibility of the management measures to Petagas River Channel Entrance and its performance throughout all seasons.

The optimum engineering solution selected is as shown in Section 6.3.2. It consists of a training channel that requires revetments to be constructed along both river banks and around the existing sand spit and seawards, with training wall further extending seawards to discharge sediment into the deeper sea, thus avoiding sediment being discharged from the River and deposited at the entrance, leading to sedimentation. With this training channel, sediments are flushed to the deeper sea and transported by ocean current towards the adjacent coastline during the NE and SW monsoon periods.

Reclamations are being carried out to extend the airport to the south and also along the western side of Contoh Village to maintain the original orientation of the channel and to create more land use. The breakwater is then constructed to minimise wave penetration into the vicinity of the river channel entrance as shown in Section 6.3.2.

The breakwater option is selected as opposed to other options considered, mainly because of its ability to provide a shelter for small fishing vessels and thus contribute to the local economy. The analysis stated that river channel entrance improvement works are able to maintain a constant open channel flow throughout all seasons that will be able to mitigate flooding problem upstream and navigation through the river channel entrance by small vessels mainly to sustain the livelihood of local villagers.

144

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

7.3 Answer to Research Question 3 Can this modelling technique be used as a reference for solving similar river channel entrances problems?

Coastal morphology includes constant evolvement as the combined consequences of both natural and human-induced factors that cover a broad range of spatial and temporal scales of effects. Therefore, there is no one size fits all solution when it comes to solving river channel entrances problems.

Much of the research work carried out over the years is very much subject specific and has resulted in a lack of application-based research. This can seen in recent research (Anthony, 2015) which focuses mainly on wave influences, but even so it supports the present findings on interactions between fluvial and marine processes; Long et al. (2016) and Wei et al. (2017) describe comprehensive research on the effects of human intervention and management measures but is lacking in terms of considering the seaward influence of the East China Sea. Similarly, Garel et al. (2015) also studied the effect of human intervention on deltas but their findings lack any sugnificant attempt to integrate the effects seawards and landwards. Towards application-based research, Pianca et al. (2014) cover a good data-driven mesoscale based approach but this is restricted to backwards analysis of the existing condition and predictive analysis is limited as compared to numerical model approach. The work of Garel et al. (2014) covers some good research on littoral drift of sediment but again it would be beneficial if a larger scale were covered both seawards and landwards. The consideration of different environmental forcing taking place at different time scales, and the effect of these on the river channel entrance especially during the monsoon is still incomplete especially in this region as it is the determining factor as shown throughout chapter 6. In the vicinity of river channel entrances, special interest is drawn due to the vast chain of effects brought about by the land- sea continuum in terms of the flow of water and sediment. Although the impact of changes in coastal sediment dynamics and river basin is well established, the integration of these two fields to address river channel entrance problems is rare (Samaras & Koutitas, 2014b). Hence, this application-based research approach will further add into the current body of literature in this area. Most importantly, this modelling technique is created as a reference for solving similar river channel entrance problems.

The main themes developed throughout this research, related to human intervention, sedimentation, and the effect of monsoons have clearly been addressed by other authors as described above. The present findings reinforce the importance of planning and coordination of different coastal activities and a central realisation that in engineering the coast one must ensure that the key economic, environmental, and social drivers are discussed and understood. This research thus highlights the importance of understanding simultaneous multiple stressors as referred to by Mitchell et al (2015) and the references therein. In light of ongoing research on the 145

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

response of all coastal areas to sea level rise (Anthony, 2015), such understandings rely heavily on the proper consideration of sedimentation in river channel entrances, which must be informed by properly validated numerical models, as in the present case.

In general, the overall approach of the modelling technique spatially consists of the integration between the open sea, coastal and the River watershed in relation to the river channel entrance. Hence, a comprehensive understanding of the geomorphology of the specific river channel entrance is crucial to analyse, identify and solve its problems. For the purpose of enhancing the model predictability and representation for management purposes, the time horizon is included, generated and analysed based on hindcast data in terms of temporal pattern as further elaborated in Section 3.2 for a prolonged period of 40 years in this research. The type of model involving both period and space is known as mesoscale model as further elaborated by French et al. (2016). To build a well-represented model, the integration of data-data driven approach is shown in Section 2.4.1. Apart from data used as a direct input to the model, other measurements such as bathymetry, tidal constituents, boundary conditions, wind forcing and more are shown in Chapter 4. Data-driven analysis is utilised to predict the extreme weather condition in which extreme wave height is determined in this research for structural design consideration. Section 5.2 shows mapping and meshing techniques in which mapping is predominantly used to build the geomorphic characteristics of the model and to produce an enhanced understanding of the river channel entrance vicinity, which is important in the conceptual state of modelling. Also, mapping can be used as an end product representation of a modelling outcome. Meshing is then built on top of mapping in which the mesh density is determined with regards to the priority of the “area of interest” in the model. The reductionist and reduced complexity models are used to good effect as shown in Section 6.2. The reduced complexity model is used as a backbone input in correctly representing the overall hydrodynamics in the model in combination with the reductionist model in representing sediment dynamics and management solution prediction. Therefore, these models are able to incorporate accurately the open coast, river channel entrance, watershed with their interaction in a system as a whole, including all sediment dynamics and the response of the system when potential management measures are in place, all into one state of the art overarching framework as illustrated in Figure 2.10.

Using this modelling approach, a successful representation of the Petagas River Channel Entrance has been achieved in combination with the effect of monsoon seasons, predominately on the hydraulics and sediment dynamics aspects at the river channel entrance. The simulation is verified by site data collected and is well represented when compared to the satellite image taken as shown in Section 6.1. The good accuracy of the model then serves a predictive model to evaluate various conceptual management strategies with different engineering measures in place with the aim to solve coastal flooding problem and to improve the livelihood of the 146

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

residents in the vicinity of the river channel entrance. Finally, an optimum management solution is selected to provide a constant opening of river channel entrance that is sufficient to cater for flood discharge throughout all seasons and provide a well-defined river channel that will allow the passage of small crafts regardless of the state of the tide; and with an entrance located in deeper water, safety measures due to wave breaking along the shoreline are reassured as shown in Section 6.3.2 . Most importantly with the utilisation of this modelling technique, the proposed management measures successfully solve the sedimentation problems in the vicinity of the river channel entrance by increasing channel currents, thereby enabling flushing of sediment into the deeper water. Thus, the well-represented model will serve as an example for future establishment of numerical model of river channel entrances, which can then be utilised as a numerical tool for solving similar river channel entrance problems.

147

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

8 Conclusions The Petagas River Channel demonstrates a real-world example of a seasonally influenced dynamic river channel entrance. Without any intervention, the river channel entrance is predicted to be silted up and block altogether in the near future, causing flooding upstream and thus affecting the livelihood of the populations living in the rapidly developing Moyong river basin.

8.1 Concluding Statements Some conclusions can be drawn as per the aims set out in Section 1.3 of this research, as follows:

• Sufficient data have been evaluated and obtained for the purpose of utilisation of various numerical models, mapping and data-driven analysis. • Mapping has enabled the analysis of critical area of the site and the presentation of the viability of the proposed management solution. The meshing process is dependent on mapping as an overlay to identify the critical study area. • The reductionist multi-model set up is successfully implemented using the demonstrated numerical modelling technique. • The model accurately investigated the effect of monsoon seasons on the hydraulics and sediment dynamics aspects at the river channel entrance. The simulation is verified by site data collected and is further well represented when compared to the satellite image taken. • Good accuracy of the model enables further assessment of the Petagas River entrance problem from obtaining the impacts of the proposed management works. Through the predictive model, the proposed improvement works, which include a training channel and a breakwater, results in increased current speeds, along with a more predictable and smoother flow and limited cross current, easing the problems of navigation through the river channel entrance.

The model simulated an improved conveyance during high discharge events; this could, in turn, mitigate flooding upstream due to the effects of increased scouring of the channel. The proposed improvement option will solve the sedimentation problems by increasing channel currents, thereby enabling flushing of sediment into the deeper water.

In terms of contribution, similar research has been carried out previously on a subject specific basis as explained in the last part of Chapter 7 but there is a distinct lack of integrative approach-based research. Hence, on a larger scale this research will be able to contribute to a wider discussion on the subject of managing sedimentation in river channel entrances around the world and to bridge the gap between them. With the technique of effectively combining model components and assembling them to represent as a system sufficiently is crucial, as demonstrated in this modelling approach in which the coupled reductionist model has successfully represented a sufficient range of real-world situations - waves and tidal variants are accurately 148

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

simulated over the monsoon seasons. This is shown by using a variable mesh resolution throughout the model domain for efficient utilisation of computational cost. The result of the wave generated model as shown in Section 0 has successfully generated waves in offshore regional model, which is further transferred to a nearshore local model between the monsoon seasons. The wave generated local model is then coupled with hydrodynamic model to produce tide and current patterns for the existing condition. The calibration and verification of the model is also done after this process to assess the accuracy of the model with the actual site condition as shown in Section 6.2.

Together with the approach of integrated reduced complexity model, the model has successfully proven to be effective in simulating realistic emergent behaviours and topographic change over sufficient spatial and temporal scales. A sediment source is input into the river basin and coupled with the hydrodynamic model to simulate sedimentation in the river channel entrance between the monsoon seasons. Following that, the prediction of the river channel entrance with regards to the management measure is then assessed by coupling the proposed management scheme in Section 6.3.2. As a result of the accuracy of the modelling prediction, the proposed management solution is accepted by the local government as it is in line with the local shoreline management plan as shown in Section 2.8.

The numerical modelling approach used in accurately predicting this management scheme demonstrates a solution to meet the requirement of policy makers and can be adopted as long-term solution to similar local areas especially in the vicinity of river channel entrance discharging into the South China Sea. Most importantly this research demonstrates the sound development of seasonal influenced river channel entrance management strategy which requires sufficient prediction of coastal change at the mesoscale.

This approach also demonstrates the major importance of simulating the coastal environment as an overall system and to appreciate the crucial interactions that occur between the open coast, river basin and the shallow vicinity of the river channel entrance. With the strength of each currently available model types, a coupled multi- model framework is the key in serving as an example for future establishment of numerical model of river channel entrance, which can be used as a numerical modelling approach for solving similar dynamic river channel entrance problems especially on the east coast of Borneo.

Solving river channel entrance problems is very much geographically dependent in terms of the social-economic impact it brings, so places such as the Mekong Delta in Vietnam and Yangtze River Delta in China tend to receive a lot of research attention because of their economic and social importance. While the Petagas river has received less attention to date, with its strategic location now and into the future it

149

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

envisaged that this research will act as a start point for much more detailed research, especially in view of climate change and the clear strategic importance of the airport.

8.2 Research Limitation and Recommendation for Further Studies While this thesis contains promising results in terms of integrating multiple modelling approaches, there are many challenges for further research. For example, detailed study of the river basin is crucial as future urbanisation is predicted to cause increased discharge. As a consequence of urbanisation, the sediment properties may change leading to different sedimentary behaviour. Hence, a proper catchment model may be carried out in the future especially in terms of land use. Sediment load varies especially over land use. Hence long-term sediment data collection should be undertaken in key locations in the catchment for the better representation of sediment load. However, a proper prediction for the future of a constantly evolving river due to ongoing development remains a challenging subject for researchers.

An enhanced regional model can be carried out in the future; this can be achieved by a model focusing on the South China Sea region instead of relying on GROW global model. This may lead to a model with enhanced accuracy which will result in better confidence and reliability of the model. Also, wind friction parameter in South East Asia region, especially in terms of proven wind friction input value in this humid region remains a debate among researchers and is a topic for further research. All these aforementioned enhancements will in turn lead to a better model overall. I would believe the best model sensitivity test would be to carry out a post modelling site survey should the mitigation measured every materialised comparing to the generated model. This requires long-term data gathering especially on sediment accretion can be performed to verify the model accuracy further.

Integrated flood model/catchment model will be ideal and bring this research to a whole new level. This will result in expertise from 2 major hydrodynamic fields (Flood + Coastal). Due to the limitation at the time of the research, the regional model is derived based on the data purchased from GROW (Global Reanalysis of Ocean Waves by Oceanology International) – Global model. However, if given more funding and time a complete regional model with enhanced accuracy may be produced covering the whole South China Sea region with a part of the Pacific and Indian ocean as the boundary condition. With many factors to be included in the numerical modelling analysis, it is impossible to quantify each of the parameters input accurately. Sometimes we use the default input suggested. However, 2 of the major factor that may affect the outcome of the model are Bed Friction and Wind Friction. While Channel bed friction is relatively easy to quantify, oceanic bed friction may be difficult as to represent accurately. Different samples from various micro-regions need to be obtained from the ocean floor in order to well represent this input. As for wind friction, this is a parameter that is highly regionally dependent as it is climatically influenced and to date remains a research field to be further explored. However, with the limited availability of wind datasets especially in this region, to accurately quantify this value 150

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

remain a challenge. Sediment properties that are relative to discharge flow and are seasonally dependent can be considered in the future. However, collecting sediment data using specialist equipment in this region can be socio-economically challenging. More ADCP measurements can be obtained as well as consideration for salinity data for enhanced calibration and verification modelling. There is not enough data on temperature and relative humidity to determine its relevancy to Cd (Drag Coefficient) value. This may be a subject for further research, as the given funding, time and scope of this research is deemed overwhelming to be included. Consideration should also be given to changes in the adjacent coastline, advancement and retraction due to human or natural causes. The effect of littoral drift could be considered further in view of its importance in terms of sediment budgets.

151

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

References

Aackermann, P. E., Pedersen, P. J. D., Engsig-Karup, A. P., Clausen, T., & Grooss, J. (2013). Development of a GPU-accelerated mike 21 solver for water wave dynamics. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 7686 LNCS, 129– 130. https://doi.org/10.1007/978-3-642-35893-7-15

Abdullah, S. M. S. (1992). The Coastal Zone in Malaysia: Processes, Issues and Management Plan.

Amos, C. L. (1995). Siliciclastic tidal flats. Geomorphology and Sedimentology of Estuaries. Developments in Sedimentology, 53, 273–306. https://doi.org/10.1016/S0070-4571(05)80030-5

Amoudry, L. (2008). A review on coastal sediment transport modelling. Proudman Oceanographic Laboratory NERC, (189), 46. Retrieved from http://nora.nerc.ac.uk/8360/

Anthony, E. J. (2015). Wave influence in the construction, shaping and destruction of river deltas: A review. Marine Geology, 361, 53–78. https://doi.org/10.1016/j.margeo.2014.12.004

Anthony, E. J., Gardel, A., Proisy, C., Fromard, F., Gensac, E., Peron, C., … Lesourd, S. (2013). The role of fluvial sediment supply and river-mouth hydrology in the dynamics of the muddy, Amazon-dominated Amap??-Guianas coast, South America: A three-point research agenda. Journal of South American Earth Sciences, 44, 18–24. https://doi.org/10.1016/j.jsames.2012.06.005

Arena Kish. (2015). Updating & Reconstruction. Retrieved from http://arenakish.com/2015/10/08/5/

Barnston, A. G. (1992). Correspondence among the Correlation, RMSE, and Heidke Forecast Verification Measures; Refinement of the Heidke Score. Weather and Forecasting. https://doi.org/10.1175/1520- 0434(1992)007<0699:CATCRA>2.0.CO;2

Benassai, G. (2006). Introduction to coastal dynamics and shoreline protection. Southampton [u.a.]: WIT Press.

Bonaldo, D., Benetazzo, A., Sclavo, M., & Carniel, S. (2015). Modelling wave-driven sediment transport in a changing climate: a case study for northern Adriatic Sea (Italy). Regional Environmental Change, 15(1), 45–55. https://doi.org/10.1007/s10113-014-0619-7

Bordbar, M. H., Pedram, M., & Hassanzadeh, S. (2014). Behaviour of surface atmospheric flow passing over the northeast of the Persian Gulf. Meteorological Applications, 21(2), 271–277. https://doi.org/10.1002/met.1325

Brack, K., Stevens, R. L., & Paetzel, M. (2000). Holocene environmental change and

152

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

the accumulation ± erosion È teborg , SW Sweden balance of sheltered river- mouth sediments , Go. Marine Geology, 170, 347–362.

Bradbury, A., McFarland, S., Horne, J., & Eastick, C. (2002). Development of a Strategic Coastal Monitoring Programme for Southeast England. Coastal Engineering 2002. ASCE. https://doi.org/doi:10.1142/9789812791306_0269

Britannica, E. (2015). Breakwater. Retrieved from http://www.britannica.com/technology/breakwater

Brooks/Cole. (2002). Factors Affecting Wind Wave Development slide. Retrieved April 30, 2017, from http://www.coa.edu/stodd/oceanweb/oceanography/Oceanlectures02/Lecture7 /sld015.htm

Brown, J. M., Amoudry, L. O., Souza, A. J., & Rees, J. (2015). Fate and pathways of dredged estuarine sediment spoil in response to variable sediment size and baroclinic coastal circulation. Journal of Environmental Management, 149, 209– 221. https://doi.org/10.1016/j.jenvman.2014.10.017

Bruun, P. (1954). Coast erosion and the development of beach profiles. U.S. Beach Erosion Board.

Byron, I. (2016). Wave Energy and Wave Changes with Depth. Retrieved April 15, 2017, from https://manoa.hawaii.edu/exploringourfluidearth/physical/waves/wave-energy- and-wave-changes-depth

Castillo, E., & Sarabia, J. M. (1994). Extreme Value Analysis of Wave Heights. Journal of Research of the National Institute of Standards and Technology, 99(4), 445– 454.

Chen, W.-F., & Liew, J. Y. R. (2003). The civil engineering handbook. CRC Press.

Chiang, Y.-C., & Hsiao, S.-S. (2011). Coastal Morphological Modeling. In Sediment Transport in Aquatic Environments. InTech. https://doi.org/10.5772/22076

Coco, G., Zhou, Z., van Maanen, B., Olabarrieta, M., Tinoco, R., & Townend, I. (2013). Morphodynamics of tidal networks: Advances and challenges. Marine Geology, 346, 1–16. https://doi.org/10.1016/j.margeo.2013.08.005

Cohen, M., Dholander, J., Kolman, R., & Taal, M. (2015). Beyond Sand and Sea.

Cowell, P. J., Stive, M. J. F., Niedoroda, A. W., Vriend, H. J. De, Swift, D. J. P., Kaminsky, G. M., & Capobianco, M. (2003). The Coastal-Tract ( Part 1 ): A Conceptual Approach to Aggregated Modeling of Low-Order Coastal Change. Journal of Coastal Research, 19(4), 812–827. https://doi.org/10.2307/4299222

Cunge, J. A. (2003). Of data and models. Journal of Hydroinformatics, 75–98. Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.610.265&rep=rep1& type=pdf 153

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

da Silva, L. F. M., Öchsner, A., & Adams, R. D. (2011). Handbook of Adhesion Technology. … of Adhesive Technology, 1–8. https://doi.org/10.1007/978-3- 642-01169-6 de Brye, B., de Brauwere, A., Gourgue, O., Kärnä, T., Lambrechts, J., Comblen, R., & Deleersnijder, E. (2010). A finite-element, multi-scale model of the Scheldt tributaries, river, estuary and ROFI. Coastal Engineering, 57(9), 850–863. https://doi.org/10.1016/j.coastaleng.2010.04.001

Dean, R. (1977). Equilibrium beach profiles : U.S. Atlantic and Gulf coasts. Newark Del.: Dept. of Civil Engineering and College of Marine Studies University of Delaware.

Dean, R. G., & Dalrymple, R. A. (2004). Water Wave Mechanics for Engineers and Scientists.

Deleersnijder, E., & Lermusiaux, P. (2008). Special issue:multi-scalemodeling: nestedgrid. Ocean Dynamics, 335–498.

Deloffre, J., Lafite, R., Lesueur, P., Verney, R., Lesourd, S., Cuvilliez, A., & Taylor, J. (2006). Controlling factors of rhythmic sedimentation processes on an intertidal estuarine mudflat - Role of the turbidity maximum in the macrotidal Seine estuary, France. Marine Geology, 235(1–4 SPEC. ISS.), 151–164. https://doi.org/10.1016/j.margeo.2006.10.011

DHI. (2009). MIKE ZERO : Creating 2D Bathymetries Bathymetry Editor & Mesh Generator.

DHI a. (2014). MIKE 21 SW - Spectral Wave Model: Scientific Documentation. DHI Water and Environment.

DHI b. (2007). MIKE 21 FLOW MODEL HD Hydrodynamic.

DHI c. (2012). MIKE 21 & MIKE 3 FLOW MODEL FM, Mud Transport Module Scientific Documentation, 46.

DHI e. (2012). MIKE 21 {&} MIKE 3 FLOW MODEL FM Hydrodynamic Module Step- by-step training guide.

DHI f. (2014). MIKE 21 ST: Non-Cohesive Sediment Transport Module: User Guide.

DID. (2001). Guidelines for Preparation of Coastal Engineering Hydraulic Study And Impact Evaluation (for hydraulic studies using numerical models), (December), 1–29.

DID. (2009). Coastal Management. Kuala Lumpur: Department of Irrgation and Drainage Malaysia.

DID. (2011). Master Plan Study for Moyong River Basin.

Dinehart, R. L., & Burau, J. R. (2005). Repeated surveys by acoustic Doppler current profiler for flow and sediment dynamics in a tidal river. Journal of Hydrology, 154

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

314(1–4), 1–21. https://doi.org/10.1016/j.jhydrol.2005.03.019

Dr. Nik & Associates. (2014). Services - Dr. Nik & Associates Sdn. Bhd. Retrieved July 15, 2017, from http://www.drnik.com.my/svc_pmc_p1.php

Ebersole, B. A. (1985). Refraction‐Diffraction Model for Linear Water Waves. Journal of Waterway, Port, Coastal, and Ocean Engineering, 111(6), 939–953. https://doi.org/10.1061/(ASCE)0733-950X(1985)111:6(939)

Eisenman, D. P., Cordasco, K. M., Asch, S., Golden, J. F., & Glik, D. (2007). Disaster Planning and Risk Communication With\r\nVulnerable Communities: Lessons From Hurricane Katrina. American Journal of Public Health, 97(Supplement 1), S109–S115. https://doi.org/10.2105/AJPH.2005.084335

Engelund, F., & Hansen, E. (1967). A monograph on sediment transport in alluvial streams. Teknisk Forlag, 65. https://doi.org/10.1007/s13398-014-0173-7.2

EVANS, J. T. (1955). Pneumatic and Similar Breakwaters. Proceedings of Royal Society of London, 231(1187), 457–466. https://doi.org/10.1680/ipeds.1956.11769

Eysink, W. D., & Biegel, E. J. (1992). Impact of sea level rise on the morphology of the Wadden Sea in the scope of its ecological function (phase 2): Investigations on empirical morphological relations. H1300. Retrieved from http://repository.tudelft.nl/islandora/object/uuid:3a321491-a1c9-4eb2-b536- d35cb81302d8?collection=research

French, J., Burningham, H., Thornhill, G., Whitehouse, R., & Nicholls, R. J. (2016). Conceptualising and mapping coupled estuary, coast and inner shelf sediment systems. Geomorphology, 256, 17–35. https://doi.org/10.1016/j.geomorph.2015.10.006

French, J., Payo, A., Murray, B., Orford, J., Eliot, M., & Cowell, P. (2016). Appropriate complexity for the prediction of coastal and estuarine geomorphic behaviour at decadal to centennial scales. Geomorphology, 256, 3–16. https://doi.org/10.1016/j.geomorph.2015.10.005

French, J. R., & Burningham, H. (2009). Coastal geomorphology: trends and challenges. Progress in Physical Geography, 33(1), 117–129. https://doi.org/10.1177/0309133309105036

French, J. R., & Burningham, H. (2013). Coasts and climate: Insights from geomorphology. Progress in Physical Geography, 37(4), 550–561. https://doi.org/10.1177/0309133313494962

Friedrichs, C. T. (2012). Tidal Flat Morphodynamics: A Synthesis. Treatise on Estuarine and Coastal Science, 3, 137–170. https://doi.org/10.1016/B978-0-12- 374711-2.00307-7

Galloway, W. E. (1975). Process framework for describing the morphological and stratigraphic evolution of deltaic depositional systems. Deltas: Models for

155

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Exploration, (September), 87–98.

Gambino, J. P., & Sussman, J. M. (1998). A Comparison of Sediment Transport Models for Combined Current and Wave Flows by A Comparison of Sediment Transport Models for Combined Current and Wave Flows.

Garel, E., Sousa, C., & Ferreira. (2015). Sand bypass and updrift beach evolution after jetty construction at an ebb-tidal delta. Estuarine, Coastal and Shelf Science, 167, 4–13. https://doi.org/10.1016/j.ecss.2015.05.044

Garel, E., Sousa, C., Ferreira, Ó., & Morales, J. A. (2014). Decadal morphological response of an ebb-tidal delta and down-drift beach to artificial breaching and inlet stabilisation. Geomorphology, 216, 13–25. https://doi.org/10.1016/j.geomorph.2014.03.031

Gevrey, M., Dimopoulos, I., & Lek, S. (2003). Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecological Modelling, 160(3), 249–264. https://doi.org/10.1016/S0304-3800(02)00257-0

Graaff, J. van de. (2009). Coastal Wiki. Retrieved from http://www.coastalwiki.org/wiki/Seawalls_and_revetments

GROW. (2009a). GROW. Retrieved September 9, 2009, from http://www.oceanweather.com/metocean/grow/index.html

GROW. (2009b). GROW.

Guerin, T., Bertin, X., & Dodet, G. (2016). A numerical scheme for coastal morphodynamic modelling on unstructured grids. Ocean Modelling, 104, 45–53. https://doi.org/10.1016/j.ocemod.2016.04.009

Guerrero, M., Rüther, N., & Archetti, R. (2014). Comparison under controlled conditions between multi-frequency ADCPs and LISST-SL for investigating suspended sand in rivers. Flow Measurement and Instrumentation, 37, 73–82. https://doi.org/10.1016/j.flowmeasinst.2014.03.007

Hardiman, N. (2013). Shoreline Management Plan policies – what do they mean? Retrieved from http://apps.environment-agency.gov.uk/wiyby/134834.aspx

Harding, S. F., Richmond, M. C., Romero-Gomez, P., & Serkowski, J. A. (2015). Effects of non-homogeneous flow on ADCP data processing in a hydroturbine forebay. Flow Measurement and Instrumentation. https://doi.org/10.1016/j.flowmeasinst.2015.12.010

Harvey, N., & Caton, B. (2010). Coastal Management in Australia. Coastal Management in Australia, 1–342. https://doi.org/10.1017/UPO9780980723038

Heidorn, K. C. (2005). The Weather Doctor Almanac 2005. Retrieved April 6, 2017, from http://www.islandnet.com/~see/weather/almanac/arc2005/alm05mar.htm

Herschy, R. W. (2008). Streamflow Measurement, Third Edition. Retrieved from https://books.google.de/books?id=N2hYxTzyzzAC 156

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Hinrichsen, D. (1998). Coastal population and altered land cover in coastal zones. Washington D.C.

Hoggart, S. P. G., Hanley, M. E., Parker, D. J., Simmonds, D. J., Bilton, D. T., Filipova- Marinova, M., … Thompson, R. C. (2014). The consequences of doing nothing: The effects of seawater flooding on coastal zones. Coastal Engineering, 87, 169– 182. https://doi.org/10.1016/j.coastaleng.2013.12.001

Horrillo-Caraballo, J. M., & Reeve, D. E. (2008). An investigation of the link between beach morphology and wave climate at Duck, NC, USA. Journal of Flood Risk Management, 1(2), 110–122. https://doi.org/10.1111/j.1753- 318X.2008.00013.x

Hsiao, S. V., & Shemdin, O. H. (1978). Bottom Dissipation in Finite-Depth Water Waves. Coastal Engineering 1978, 434–448. https://doi.org/10.1061/9780872621909.026

Hsu, T. J., Elgar, S., & Guza, R. T. (2006). Wave-induced sediment transport and onshore sandbar migration. Coastal Engineering, 53(10), 817–824. https://doi.org/10.1016/j.coastaleng.2006.04.003

Hudson, T., Consulting, J., Keating, K., Chatterton, J., & Williams, A. (2015). Cost estimation for coastal protection – summary of evidence, 1–46. Retrieved from www.environment-agency.gov.uk%0Awww.environment-agency.gov.uk

Huffman, G. J., Adler, R. F., Arkin, P., Chang, A., Ferraro, R., Gruber, A., … Schneider, U. (1997). The Global Precipitation Climatology Project (GPCP) Combined Precipitation Dataset. Bulletin of the American Meteorological Society, 78(1), 5– 20. https://doi.org/10.1175/1520-0477(1997)078<0005:TGPCPG>2.0.CO;2

Hughes, S., & Fowler, J. (1995). Estimating wave-induced kinematics at sloping structures. Journal of Waterway, Port, Coastal, and …, 1995(August), 209–215. https://doi.org/10.1061/(ASCE)0733-950X(1995)121:4(209)

Hunt, S., Guthrie, G., Cooper, N., & Roberts, H. (2011). Estuaries and Shoreline Management Plans–lessons learned from Round 2. In Littoral 2010–Adapting to Global Change at the Coast: Leadership, Innovation, and Investment (p. 6003). Retrieved from http://coastnet- littoral2010.edpsciences.org/articles/litt/abs/2011/01/litt-06003/litt-06003.html

IADC / IAPH. (2010). Dredging for Development, 80. https://doi.org/978-90-75254- 16-7

ICZM Task Force, S. G. (1998). Climate. Kota Kinabalu.

Isaacson, M., & Foschi, R. (2000). On the selection of design wave conditions. International Journal of Offshore and Polar Engineering, 10(2), 99–106.

Ito, Y. (1969). A treatise on historical development of breakwater design.

Jakobsen, F., Hartstein, N., Frachisse, J., & Golingi, T. (2007). Sabah shoreline management plan (Borneo, Malaysia): Ecosystems and pollution. Ocean and 157

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Coastal Management, 50(1–2), 84–102. https://doi.org/10.1016/j.ocecoaman.2006.03.013

Jarrett, J. T. (1976). Tidal Prism-Inlet Area Relationships. Fort Belvoir.

Jayappa, K. S., & Narayana, A. C. (2009). Coastal environments : problems and perspectives. I.K. International Pub. House.

Johnson, J. W. (1951). Generalized Wave Diffraction Diagrams. Coastal Engineering Proceedings, 1(2), 2. https://doi.org/10.9753/ICCE.V2.2

Julien, P. Y., & Vensel, C. W. (2005). Review of Sedimentation Issues on the Mississippi River, (November), 62.

Kamphius, W. J. (2000). Introduction to Coastal Engineering and Management. https://doi.org/10.1142/7021

Kamphuis, J. W. (2000). Water Waves. In Introduction to Coastal Engineering and Management.

Keating, K., Pettit, A., & Wood, A. (2012). Cost estimation for coastal protection – summary of evidence.

Kim, D., & Muste, M. (2012). Multi-dimensional representation of river hydrodynamics using ADCP data processing software. Environmental Modelling and Software, 38, 158–166. https://doi.org/10.1016/j.envsoft.2012.05.011

Kocar, B., & Fendorf, S. (2012). Arsenic Release and Transport in Sediments of the Mekong Delta. Interdisciplinary Studies on Environmental Chemistry— Environmental Pollution and Ecotoxicology, 6, 117–124.

Kozyrakis, G. V., Delis, A. I., Alexandrakis, G., & Kampanis, N. A. (2015). Numerical modeling of sediment transport applied to coastal morphodynamics. Applied Numerical Mathematics. https://doi.org/10.1016/j.apnum.2014.09.007

Kripalani, R. H., & Kulkarni, A. (1997). Rainfall variability over South‒east Asia— connections with Indian monsoon and ENSO extremes: new perspectives. International Journal of Climatology, 17(11), 1155–1168. https://doi.org/10.1002/(SICI)1097-0088(199709)17:11<1155::AID- JOC188>3.0.CO;2-B

Krone, R. B. (1962). Flume Studies of the Transport of Sediment in Estuarine Shoaling Processes. Washington D.C.

Latteux, B. (1995). Techniques for long-term morphological simulation under tidal action. Marine Geology, 126(1–4), 129–141. https://doi.org/10.1016/0025- 3227(95)00069-B

Le, T. V. H., Nguyen, H. N., Wolanski, E., Tran, T. C., & Haruyama, S. (2007). The combined impact on the flooding in Vietnam’s Mekong River delta of local manmade structures, sea level rise, and dams upstream in the river catchment. Estuarine, Coastal and Shelf Science, 71(1–2), 110–116. 158

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.1016/j.ecss.2006.08.021

Lee, A., Kim, S. E., & Suh, K.-D. (2016). An easy way to use artificial neural network model for calculating stability number of rock armors. Ocean Engineering, 127(October), 349–356. https://doi.org/10.1016/j.oceaneng.2016.10.013

Legrand, S., Deleersnijder, E., Delhez, E., & Legat, V. (2007). Unstructured, anisotropic mesh generation for the Northwestern European continental shelf, the continental slope and the neighbouring ocean. Continental Shelf Research, 27(9), 1344–1356. https://doi.org/10.1016/j.csr.2007.01.009

LESCHKA, S., KONGKO, W., & LARSEN, O. (2009). ON THE INFLUENCE OF NEARSHORE BATHYMETRY DATA QUALITY ON TSUNAMI RUNUP MODELLING, PART II: MODELLING. In Asian and Pacific Coasts 2009 (pp. 157–163). World Scientific Publishing Company. https://doi.org/10.1142/9789814287951_0017

Lesourd, S., Lesueur, P., Brun-Cottan, J. C., Garnaud, S., & Poupinet, N. (2003). Seasonal variations in the characteristics of superficial sediments in a macrotidal estuary (the Seine inlet, France). Estuarine, Coastal and Shelf Science, 58(1), 3– 16. https://doi.org/10.1016/S0272-7714(02)00340-2

Liang, M., Voller, V. R., & Paola, C. (2015). A reduced-complexity model for river delta formation - Part 1: Modeling deltas with channel dynamics. Earth Surface Dynamics, 3(1), 67–86. https://doi.org/10.5194/esurf-3-67-2015

Lisitzin, E. (1974). Sea-level changes. Elsevier Scientific Pub. Co.

Lokman Husain, M., Yaakob, R., & Saad, S. (1995). Beach Erosion Variabiltiy during a Northeast Monsoon: The Kuala Setiu Coastline, Terengganu, Malaysia. Pertanika J. Sci. & Technol, 3(2), 337–348. https://doi.org/ISSN,0128-7680

Long, H., Xing, P., Bing, Z., Zhong, J., & Lun, S. (2016). Geomorphology Decadal morphological evolution of the Yangtze Estuary in response to river input changes and estuarine engineering projects. Geomorphology, 265, 12–23. https://doi.org/10.1016/j.geomorph.2016.04.022

Loo, Y. Y., Billa, L., & Singh, A. (2015). Effect of climate change on seasonal monsoon in Asia and its impact on the variability of monsoon rainfall in Southeast Asia. Geoscience Frontiers, 6(6), 817–823. https://doi.org/10.1016/j.gsf.2014.02.009

Losada, I. J., Losada, M. A., & Baquerizo, A. (1993). An analytical method to evaluate the efficiency of porous screens as wave dampers. Applied Ocean Research, 15(4), 207–215. https://doi.org/10.1016/0141-1187(93)90009-M

Losada, I. J., Silva, R., & Losada, M. A. (1997). Effects of reflective vertical structures permeability on random wave kinematics. Journal of Waterway, Port, Coastal and Ocean Engineering, 123(December), 347–353.

Madsen, P. A., & Sørensen, O. R. (1992). A new form of the Boussinesq equations with improved linear dispersion characteristics. Part 2. A slowly-varying bathymetry. Coastal Engineering, 18(3–4), 183–204.

159

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.1016/0378-3839(92)90019-Q

Malaysian Meteorological Department. (2008). Weather Phenomena. Retrieved July 31, 2008, from https://web.archive.org/web/20080320105546/http://www.kjc.gov.my/english/ education/weather/monsoon01.html

Malaysian Meteorological Department. (2009). Malaysian Tide Table.

Malaysian Meteorological Department. (2012). General Climate in Malaysia. Retrieved May 1, 2017, from http://www.met.gov.my/en/web/metmalaysia/climate/generalinformation/mala ysia

Mangor, K. (2007). Wave climate classification according to wind climate. Retrieved May 1, 2017, from http://www.vliz.be/wiki/Wave_climate_classification_according_to_wind_climat e

Mashriqui, H. S. (2003). Hydrodynamic and Sediment Transport Modeling of Deltaic Sediment Processes. Journal of Chemical Information and Modeling. https://doi.org/10.1017/CBO9781107415324.004

Max Barton, M. E., & Brown, S. (2014). Littoral drift barriers and the problem of proving accelerated recession. Engineering Geology for Society and Territory - Volume 4: Marine and Coastal Processes, 9–12. https://doi.org/10.1007/978-3- 319-08660-6_2

Meadows, G. A. (2003). Wave mechanics. In Coastal Engineering (Vol. 5, p. 205).

Meadows, G. A., & Wood, W. L. (2003). Water Level Fluctuations Study. Coastal Engineering.

Ministry of Natural Resources and Environment Malaysia. (2012a). Integrated Shoreline Management Plan For Negeri Sabah. Retrieved from http://ww2.sabah.gov.my/phb/wp-content/uploads/2013/08/Exec-Summary- Integrated-Shoreline-Management-Plan-Final-Report-Nov-2012.pdf

Ministry of Natural Resources and Environment Malaysia. (2012b). Integrated Shoreline Management Plan For Negeri Sabah.

Mitchell, S., Akesson, L., & Uncles, R. (2012). Observations of turbidity in the Thames Estuary, United Kingdom. Water and Environment Journal, 26(4), 511–520. https://doi.org/10.1111/j.1747-6593.2012.00311.x

Mitchell, S. B. (2013). Turbidity maxima in four macrotidal estuaries. Ocean and Coastal Management, 79(2013), 62–69. https://doi.org/10.1016/j.ocecoaman.2012.05.030

Mitchell, S., Green, M. O., MacDonald, I. T., & Pritchard, M. (2015). Field studies of estuarine turbidity under different freshwater flow conditions, Kaipara River, New Zealand. Estuarine, Coastal and Shelf Science. 160

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.1016/j.ecss.2016.06.009

Moftakhari, H. R., Jay, D. A., Talke, S. A., Kukulka, T., & Bromirski, P. D. (2013). A novel approach to flow estimation in tidal rivers. Water Resources Research, 49(8), 4817–4832. https://doi.org/10.1002/wrcr.20363

Moulinec, C., Denis, C., Pham, C. T., Rougé, D., Hervouet, J. M., Razafindrakoto, E., … Gu, X. J. (2011). TELEMAC: An efficient hydrodynamics suite for massively parallel architectures. Computers and Fluids, 51(1), 30–34. https://doi.org/10.1016/j.compfluid.2011.07.003

Mulder, J. P. M., Hommes, S., & Horstman, E. M. (2011). Implementation of coastal erosion management in the Netherlands. Ocean and Coastal Management, 54(12), 888–897. https://doi.org/10.1016/j.ocecoaman.2011.06.009

Muttray, M., & Reedijk, B. (2008). Design of Concrete Armour Layers. Ocean and Coastal Management.

Muttray, M., & Reedijk, B. (2009). Design of Concrete Armour Layers. Hansa International Maritime Journal, 6(October), 111–118.

Nelson, S. M., Zamora-Arroyo, F., Ramírez-Hernández, J., & Santiago-Serrano, E. (2013). Geomorphology of a Recurring Tidal Sandbar in the estuary of the Colorado River, Mexico: Implications for restoration. Ecological Engineering, 59, 121–133. https://doi.org/10.1016/j.ecoleng.2012.12.095

Nicholls, R. J., Bradbury, A., Burningham, H., Dix, J., Ellis, M., French, J., … Whitehouse, R. (2012). iCOASST - INTEGRATING COASTAL SEDIMENT SYSTEMS. Coastal Engineering Proceedings; No 33 (2012): Proceedings of 33rd Conference on Coastal Engineering, Santander, Spain, 2012, (DECEMBER). https://doi.org/10.9753/icce.v33.sediment.100

Nicholls, R. J., Townend, I. H., Bradbury, A. P., Ramsbottom, D., & Day, S. A. (2013). Planning for long-term coastal change: Experiences from England and Wales. Ocean Engineering, 71, 3–16. https://doi.org/10.1016/j.oceaneng.2013.01.025

NINGSIH, N. S., HADI, S., UTAMI, M. D., & RUDIAWAN, A. P. (2011). MODELLING OF STORM TIDE FLOODING ALONG THE SOUTHERN COAST OF JAVA, INDONESIA. In Advances in Geosciences (pp. 87–103). https://doi.org/10.1142/9789814355353_0006

Nurukamal, A. M. M., & Razali. (2016). Modeling the wind direction behaviors during the monsoon seasons in Peninsular Malaysia. Renewable and Sustainable Energy Reviews, 56, 1419–1430. https://doi.org/10.1016/j.rser.2008.08.009

Nylen, L., & Ramel, E. (2011). The effects of inlet sedimentation on water exchange in Maha Oya Estuary, Sri Lanka.

Orton, G. J., & Reading, H. G. (1993). Variability of deltaic processes in terms of sediment supply, with particular emphasis on grain size. Sedimentology, 40(3), 475–512. https://doi.org/10.1111/j.1365-3091.1993.tb01347.x

161

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Patel, P. (2015). Ever Thought, Why Mumbai’s Marine Drive Has These Weird Shaped Rocks At The Seashore? Retrieved from http://www.clubpimble.com/fun- pimble/ever-thought-why-mumbais-marine-drive-has-these-weird-shaped- rocks-at-the-seashore

Pelabuhan Tanjung Pelepas. (2012). Pelabuhan Tanjung Pelepas Sdn Bhd. - Home. Retrieved July 15, 2017, from http://www.ptp.com.my/

Peng, S., & Li, Y. (2015). OPEN A parabolic model of drag coefficient for storm surge simulation in the South China Sea. Nature Publishing Group, 1–6. https://doi.org/10.1038/srep15496

Phillips, O. M. (1962). Recent developments in the theory of wave generation by wind. Journal of Geophysical Research, 67(8), 3135–3141. https://doi.org/10.1029/JZ067i008p03135

Pianca, C., Holman, R., & Siegle, E. (2014). Mobility of meso-scale morphology on a microtidal ebb delta measured using remote sensing. Marine Geology, 357, 334– 343. https://doi.org/10.1016/j.margeo.2014.09.045

Pile Buck. (2014). Chapter 6 - Coastal Revetments for Wave Attack - Pile Buck Magazine. Retrieved December 30, 2017, from http://www.pilebuck.com/highways-coastal-environment-second- edition/chapter-6-coastal-revetments-wave-attack/

Port, C., & Engineers, C. (2009). NUCLEAR SITES SITE SAFETY REPORTS NUMERICAL MODELLING OF COASTAL PROCESSES Report No . 1010 / 3 / 101 SITE SAFETY REPORTS NUMERICAL MODELLING OF COASTAL PROCESSES.

Pruszak, Z., van Ninh, P., Szmytkiewicz, M., Hung, N. M., & Ostrowski, R. (2005). Hydrology and morphology of two river mouth regions (temperate Vistula Delta and subtropical Red River Delta). Oceanologia, 47(3), 365–385.

Rahbani, M. (2015). A comparison between the suspended sediment concentrations derived from DELFT3D model and collected using transmissometer - A case study in tidally dominated area of Dithmarschen Bight. Oceanologia, 57(1), 44–49. https://doi.org/10.1016/j.oceano.2014.06.002

Renkena, A., & Kinlan, D. (2000). Tanjung Pelepas Port Malaysia’s newest container port. World Dredging, Mining and Construction, 36(11), 6–20.

Rob Varela. (2016). Powerful surf causing Ventura’s beaches to shrink. Retrieved from http://archive.vcstar.com/news/local/ventura/powerful-surf-causing- venturas-beaches-to-shrink---2bfbbf60-6f84-4a49-e053-0100007f88fa- 369158251.html

Robakiewicz, M. (2010). Vistula River Mouth – History and Recent Problems, 57(2), 155–166.

Robert, M. S. (1993). Small Amplitudes Wave Characteristics. In Basic Wave Mechanics: For Coastal and Ocean Engineers (pp. 13–14).

162

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Roelvink, J. A. (2006). Coastal morphodynamic evolution techniques. Coastal Engineering, 53(2–3), 277–287. https://doi.org/10.1016/j.coastaleng.2005.10.015

Rossington, S. K., Nicholls, R. J., Stive, M. J. F., & Wang, Z. B. (2011). Estuary schematisation in behaviour-oriented modelling. Marine Geology, 281(1–4), 27– 34. https://doi.org/10.1016/j.margeo.2011.01.005

Rotterdam Public Works Engineering Department. (2001). Construction and Survey Accuracies for the Execution of Dredging and Stone Dumping Works. Rotterdam.

Roy, P. D., Caballero, M., Lozano, S., Morton, O., Lozano, R., Jonathan, M. P., … Macías, M. C. (2012). Provenance of sediments deposited at paleolake San Felipe, western Sonora Desert: Implications to regimes of summer and winter precipitation during last 50 cal kyr BP. Journal of Arid Environments, 81, 47–58. https://doi.org/10.1016/j.jaridenv.2012.01.008

Sabahkini.net. (2012). PETAGAS RIVER MOUTH TOO SHALLOW. Retrieved May 1, 2017, from http://sabahkinimirror.blogspot.my/2012/06/petagas-river-mouth- too-shallow.html

Samaras, A. G., & Koutitas, C. G. (2014a). Comparison of three longshore sediment transport rate formulae in shoreline evolution modeling near stream mouths. Ocean Engineering, 92, 255–266. https://doi.org/10.1016/j.oceaneng.2014.10.005

Samaras, A. G., & Koutitas, C. G. (2014b). The impact of watershed management on coastal morphology: A case study using an integrated approach and numerical modeling. Geomorphology, 211, 52–63. https://doi.org/10.1016/j.geomorph.2013.12.029

Santini, M., & di Paola, A. (2015). Changes in the world rivers’ discharge projected from an updated high resolution dataset of current and future climate zones. Journal of Hydrology, 531, 768–780. https://doi.org/10.1016/j.jhydrol.2015.10.050

Sarpkaya, T., & Isaacson, M. (1981). Mechanics of wave forces on offshore structures. Van Nostrand Reinhold Co. Retrieved from https://www.amazon.com/Mechanics-Wave-Forces-Offshore- Structures/dp/0442254024

Sawaragi, T. (1995). Basic Formulation of Sea Waves. In Coastal Engineering- Waves, Beaches, Wave- Structure Interactions (pp. 1–2). Elsevier.

Schmitt, K., & Albers, T. (2014). Area Coastal Protection and the Use of Bamboo Breakwaters in the Mekong Delta. In Coastal Disasters and Climate Change in Vietnam: Engineering and Planning Perspectives (pp. 107–132). https://doi.org/10.1016/B978-0-12-800007-6.00005-8

Schneider, S. K. (2005). Administrative breakdowns in the governmental response to hurricane Katrina. Public Administration Review, 65(5), 515–516.

163

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.1111/j.1540-6210.2005.00478.x

Sciortino, J. (2010). Fishing harbour planning, construction and management. In FAO Fisheries and Aquaculture Technical Paper (p. 337).

Smagorinsky, J. (1963). GENERAL CIRCULATION EXPERIMENTS WITH THE PRIMITIVE EQUATIONS\rI. THE BASIC EXPERIMEN. Monthly Weather Review, 91(99–164), 99–164. https://doi.org/10.1175/1520- 0493(1963)091<0099:GCEWTP>2.3.CO;2

SNH. (2000a). Summary 14: ROCK REVETMENTS. Retrieved from Pooja Patel

SNH. (2000b). Summary 15: TIMBER REVETMENTS. Retrieved from http://www.snh.org.uk/publications/on- line/heritagemanagement/erosion/appendix_1.15.shtml

SOE. (2011). Australia State of the Environment 2011: Chapter 3 Atmosphere. State of the Environment 2011 Committee, Canberra, ACT. Canberra.

Sokolov, A., & Chubarenko, B. (2012). Wind Influence on the Formation of Nearshore Currents in the Southern Baltic : Numerical Modelling Results, 59(1), 37–48. https://doi.org/10.2478/v10203-012-0003-3

Soomere, T., Kask, A., Kask, J., & Healy, T. (2008). Modelling of wave climate and sediment transport patterns at a tideless embayed beach, Pirita Beach, Estonia. Journal of Marine Systems, 74(SUPPL.). https://doi.org/10.1016/j.jmarsys.2008.03.024

Soon, E. (2009). Wind wave effects on hydrodynamic modeling of ocean circulation in the South China Sea Author Wind wave effects on hydrodynamic modeling of ocean circulation in the South China Sea.

Souza, A. J., Holt, J. T., & Proctor, R. (2007a). Modelling SPM on the NW European shelf seas. Geological Society, London, Special Publications, 274(1), 147–158. https://doi.org/10.1144/GSL.SP.2007.274.01.14

Souza, A. J., Holt, J. T., & Proctor, R. (2007b). Modelling SPM on the NW European shelf seas. Geological Society, London, Special Publications, 274(1), 147–158. https://doi.org/10.1144/GSL.SP.2007.274.01.14

Stanley, D. J., & Warne, A. G. (1994). Worldwide initiation of Holocene marine deltas by deceleration of sea-level rise. Science, 265, 228–231. https://doi.org/10.1126/science.265.5169.228

Stark, J., Plancke, Y., Ides, S., Meire, P., & Temmerman, S. (2016). Coastal flood protection by a combined nature-based and engineering approach: Modeling the effects of marsh geometry and surrounding dikes. Estuarine, Coastal and Shelf Science, 175, 34–45. https://doi.org/10.1016/j.ecss.2016.03.027

Stokes, G. G. (1847). On the Theory of Oscillatory Waves. Cambridge Philosophical Society, 8, 197–237. https://doi.org/10.1017/CBO9780511702242.016

164

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Sugdon, R. (2004). Developing the Cost-Benefit Framework for the Appraisal of Flood and Coastal Erosion Risk Management Projects, 37. Retrieved from http://cdn.environment-agency.gov.uk/geho0609bqdf-e-e.pdf

Sundar, V. (2016). Ocean Wave Mechanics: Applications in Coastal and Offshore Structures . Retrieved from http://vt.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwdV1LawIxEB6kX gQPLSraB90_oGQnj909l9r2UHoRPEqe4GUP7ervbybu-mjxGAJDAplX- L5vADgu2PxPTDBBGUot6IJVuoiFv5FBWGaDlqgNXpLgwXXUmIRoPmI6qLtv4TX_ NROTsMTxp45cKgV6qlVbNACSnGz07dhy5EpiIp9L8o3Fx1v3BpHzSp6lPOQxYLMT

Sutherland, J., & O’Donoghue, T. (1998). Wave phase shift at coastal structures. Journal of Waterway, Port, Coastal and Ocean Engineering, 124(2), 90–98.

Svendsen, I. A., & Hansen, J. B. (1977). The wave height variation for regular waves in shoaling water. Coastal Engineering, 1, 261–284. https://doi.org/10.1016/0378-3839(77)90018-7

Sverdrup, K. A., & Armbrust, E. V. (Elisabeth V. (2009). An introduction to the world’s oceans. McGraw-Hill.

Syvitski, J. P. M. (2008). Deltas at risk. Sustainability Science, 3(1), 23–32. https://doi.org/10.1007/s11625-008-0043-3

Syvitski, J. P. M., Cohen, S., Kettner, A. J., & Brakenridge, G. R. (2014). How important and different are tropical rivers? - An overview. Geomorphology, 227, 5–17. https://doi.org/10.1016/j.geomorph.2014.02.029

Syvitski, J. P. M., & Saito, Y. (2007). Morphodynamics of deltas under the influence of humans. Global and Planetary Change, 57(3–4), 261–282. https://doi.org/10.1016/j.gloplacha.2006.12.001

Takahashi, S. (2002). Design of vertical breakwaters. Port and Airport Research Institute - Japan, 1996(34), 126. https://doi.org/10.1142/9789814282413_0004

Tay, M. T. W., Mitchell, S. B., Chen, J., & Williams, J. (2016). Numerical modelling approach for the management of seasonal influenced river channel entrance. Ocean and Coastal Management, 130, 79–94. https://doi.org/10.1016/j.ocecoaman.2016.06.004

Teisson, C., Ockenden, M., Le Hir, P., Kranenburg, C., & Hamm, L. (1993). Cohesive sediment transport processes. Coastal Engineering, 21(1–3), 129–162. https://doi.org/10.1016/0378-3839(93)90048-D

Terrados, J., Duarte, C. M., Fortes, M. D., Borum, J., Agawin, N. S. R., Bach, S., … Vermaat, J. (1998). Changes in Community Structure and Biomass of Seagrass Communities along Gradients of Siltation in SE Asia. Estuarine, Coastal and Shelf Science, 46(5), 757–768. https://doi.org/10.1006/ecss.1997.0304

Thaha, A., Maricar, F., Aboe, A. F., & Dwipuspita, A. I. (2015). The breakwater, from wave breaker to wave catcher. Procedia Engineering, 116(1), 691–698.

165

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.1016/j.proeng.2015.08.352

Tominaga, Y., & Stathopoulos, T. (2007). Turbulent Schmidt numbers for CFD analysis with various types of flowfield. Atmospheric Environment, 41(37), 8091– 8099. https://doi.org/10.1016/j.atmosenv.2007.06.054

Touili, N., Baztan, J., Vanderlinden, J. P., Kane, I. O., Diaz-Simal, P., & Pietrantoni, L. (2014). Public perception of engineering-based coastal flooding and erosion risk mitigation options: Lessons from three European coastal settings. Coastal Engineering, 87, 205–209. https://doi.org/10.1016/j.coastaleng.2014.01.004

United Nations. (1969). 083097Eb.pdf. In Mekong River Delta Model Study. Paris. Retrieved from http://unesdoc.unesco.org/images/0008/000830/083097Eb.pdf

Urlburt, H. A. E. H. (2007). Wind Stress Drag Coefficient over the Global Ocean *, 5856–5864. https://doi.org/10.1175/2007JCLI1825.1

US Army Corps of Engineers. (1984). Shore protection manual. https://doi.org/10.5962/bhl.title.47829

US Army Corps of Engineers. (1995). Design of Coastal Revetments , Seawalls , and Bulkheads. US Army Corps of Engineers. Washington DC.

Valsamidis, A., Różyński, G., & Reeve, D. E. (2013). Data-driven analysis of medium- term wave–seabed interactions at a non-tidal beach with multiple bars. Marine Geology, 344, 144–154. https://doi.org/10.1016/j.margeo.2013.08.003

Van der Meer, J. W. (1987). Stability of breakwater armour layers - design formulae. Coastal Engineering. https://doi.org/10.1016/0378-3839(87)90013-5 van Koningsveld, M., & Mulder, J. P. M. (2004). Sustainable Coastal Policy Developments in The Netherlands. A Systematic Approach Revealed. Journal of Coastal Research, 202, 375–385. https://doi.org/10.2112/1551- 5036(2004)020[0375:SCPDIT]2.0.CO;2 van Maanen, B., Nicholls, R. J., French, J. R., Barkwith, A., Bonaldo, D., Burningham, H., … Walkden, M. J. A. (2016). Simulating mesoscale coastal evolution for decadal coastal management: A new framework integrating multiple, complementary modelling approaches. Geomorphology, 256, 68–80. https://doi.org/10.1016/j.geomorph.2015.10.026

Veerbeek, W. (2007). Ranking of the world’s cities most exposed to coastal flooding today and in the future.

Vilchez, M., Clavero, M., & Losada, M. A. (2016). Hydraulic performance of different non-overtopped breakwater types under 2D wave attack. Coastal Engineering, 107, 34–52. https://doi.org/10.1016/j.coastaleng.2015.10.002

Vlasblom, W. J. (2003a). Introduction to Dredging Equipment. Designing Dredging Equipment, (May), 1–27.

Vlasblom, W. J. (2003b). Plain Suction Dredgers, 1–35. 166

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

Walkden, M., Dickson, M., Thomas, J., & Hall, J. W. (2015). Simulating the Shore and Cliffs of North Norfolk (pp. 187–211). Springer Netherlands. https://doi.org/10.1007/978-94-007-5258-0_7

Walkden, M., & Hall, J. W. (2005a). A Mesoscale Predictive Model of the Evolution and Management of a Soft-Rock Coast. Journal of Coastal Research, 52, 535– 563. https://doi.org/10.2112/JCOASTRES-D-10-00099.1

Walkden, M., & Hall, J. W. (2005b). A Mesoscale Predictive Model of the Evolution and Management of a Soft-Rock Coast. Journal of Coastal Research, 52, 535– 563. https://doi.org/10.2112/JCOASTRES-D-10-00099.1

Walton, T. L., & Adams, W. D. (1976). Capacity of Inlet Outer Bars to Store Sand. In 15th Coastal Engineering Conference (pp. 1919–1937). https://doi.org/10.9753/icce.v15

Wang, N., Li, G., Qiao, L., Shi, J., Dong, P., Xu, J., & Ma, Y. (2017). Long-term evolution in the location, propagation, and magnitude of the tidal shear front off the Yellow River Mouth. Continental Shelf Research. Elsevier. https://doi.org/10.1016/j.csr.2017.01.020

WANG, Q., KANG, H., & WANG, K. (2016). Three-dimensional unstructured modelling of wave-induced circulation over a plane and irregular beach. Journal of Hydrodynamics, Ser. B, 28(2), 219–226. https://doi.org/10.1016/S1001- 6058(16)60623-6

Wang, Tang, L. Q., Wang, C. H., Liu, C. J., & Dong, Z. Di. (2014). Combined effects of channel dredging, land reclamation and long-range jetties upon the long-term evolution of channel-shoal system in Qinzhou bay, SW China. Ocean Engineering, 91, 340–349. https://doi.org/10.1016/j.oceaneng.2014.09.024

Wang, Z. B., Hoekstra, P., Burchard, H., Ridderinkhof, H., De Swart, H. E., & Stive, M. J. F. (2012). Morphodynamics of the Wadden Sea and its barrier island system. Ocean and Coastal Management, 68, 39–57. https://doi.org/10.1016/j.ocecoaman.2011.12.022

Warner, J. C., Sherwood, C. R., Signell, R. P., Harris, C. K., & Arango, H. G. (2008). Development of a three-dimensional, regional, coupled wave, current, and sediment-transport model. Computers and Geosciences, 34(10), 1284–1306. https://doi.org/10.1016/j.cageo.2008.02.012

Waveclimate. (2011). WAVECLIMATE.COM: Help pages. Retrieved April 6, 2017, from http://www.waveclimate.com/clams/redesign/help/app/rose_output.html

Wei, W., Dai, Z., Mei, X., Liu, J. P., Gao, S., & Li, S. (2017). Shoal morphodynamics of the Changjiang (Yangtze) estuary: Influences from river damming, estuarine hydraulic engineering and reclamation projects. Marine Geology, 386, 32–43. https://doi.org/10.1016/j.margeo.2017.02.013

Weishar, L. L., & Byrne, R. J. (1972). FIELD STUDY OF BREAKING WAVE CHARACTERISTICS. Retrieved from https://icce-ojs-

167

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

tamu.tdl.org/icce/index.php/icce/article/viewFile/3282/2946

Wilcock, P. R., & Kenworthy, S. T. (2002). A two-fraction model for the transport of sand/gravel mixtures. Water Resources Research, 38(10), 1–12. https://doi.org/10.1029/2001WR000684

Wilkinson, C. (2000). Status of Coral Reefs of the World : Australian Institute of Marine Science, 1–363.

Wong, P. P., & Codignotto, J. (2007). Coastal Systems and Low-Lying Areas. Climate Change 2014 Impacts, Adaptation, and Vulnerability, 361–410. https://doi.org/10.1017/CBO9781107415379.010

Woodroffe, C. D., & Murray-Wallace, C. V. (2012). Sea-level rise and coastal change: the past as a guide to the future. Quaternary Science Reviews (Vol. 54). https://doi.org/10.1016/j.quascirev.2012.05.009

Woodroffe, C. D., Nicholls, R. J., Saito, Y., Chen, Z., & Goodbred, S. L. (2006). Landscape variability and the response of Asian megadeltas to environmental change. Global Change and Integrated Coastal Management (Vol. 5). https://doi.org/10.1007/1-4020-3628-0

Wright, L. D. (1977). Sediment transport and deposition at river mouths: A synthesis. Bulletin of the Geological Society of America, 88(6), 857–868. https://doi.org/10.1130/0016-7606(1977)88<857:STADAR>2.0.CO;2

Wright, L. D., & Coleman, J. M. (1972). River delta morphology: wave climate and the role of the subaqueous proffile. Science (New York, N.Y.), 176(4032), 282– 4. https://doi.org/10.1126/science.176.4032.282

Wright, L. D., & Coleman, J. M. (1973). Variations in morphology of major river deltas as functions of ocean wave and river discharge regimes. American Association of Petroleum Geologists Bulletin, 57(2), 370–398. https://doi.org/10.1306/819A4274-16C5-11D7-8645000102C1865D

Wright, L. D., Thom, B. G., & Higgins, R. J. (1980). Wave influences on river-mouth depositional process: Examples from Australia and Papua New Guinea. Estuarine and Coastal Marine Science, 11(3), 263–277. https://doi.org/10.1016/S0302- 3524(80)80083-1

Wu, X. G., & Tang, H. S. (2010). Coupling of CFD model and FVCOM to predict small- scale coastal flows. Journal of Hydrodynamics, 22(5 SUPPL. 1), 284–289. https://doi.org/10.1016/S1001-6058(09)60208-0

Xue, Z., He, R., Liu, J. P., & Warner, J. C. (2012). Modeling transport and deposition of the Mekong River sediment. Continental Shelf Research, 37, 66–78. https://doi.org/10.1016/j.csr.2012.02.010

Yin, J., Zou, Z.-L., Dong, P., Zhang, H.-F., Wu, G.-Q., & Pan, Y.-N. (2012). Experimental research on unstable movements of sandbars under wave actions. Water Science and Engineering, 5(2), 175–190.

168

Numerical Modelling Approach for the Management of Seasonally Influenced River Channel Entrance

https://doi.org/10.3882/j.issn.1674-2370.2012.02.006

Zabez. (2015). Maj 2015 – zabezpieczeniahydrotechniczne. Retrieved July 15, 2017, from https://zabezpieczeniahydrotechniczne.wordpress.com/2015/05/

Zakariya, R. Y. Z. A. S. S. R. (2012). 2-Dimensional Hydrodynamic Patterns of Pahang River Estuary, Pahang, Malaysia during Northeast and Southwest Monsoon. International Annual Symposium on Sustainability Science and Management, (July), 1357–1371.

Zhang, C. X., Wang, Y. X., Wang, G. Y., & Yu, L. M. (2010). Wave dissipating performance of air bubble breakwaters with different layouts. Journal of Hydrodynamics, 22(5), 671–680. https://doi.org/10.1016/S1001- 6058(09)60102-5

Zhou, T., Ren, W., Peng, S., Liang, L., Ren, S., & Wu, J. (2014). A riverscape transect approach to studying and restoring river systems: A case study from southern China. Ecological Engineering, 65, 147–158.

169

END.