MORPHOLOGICAL STUDY OF OLD BRAHMAPUTRA OFFTAKE USING TWO-DIMENSIONAL MATHEMATICAL MODEL FAHMIDA NOOR. Student No: P

Transcription

1 MORPHOLOGICAL STUDY OF OLD BRAHMAPUTRA OFFTAKE USING TWO-DIMENSIONAL MATHEMATICAL MODEL FAHMIDA NOOR Student No: P DEPARTMENT OF WATER RESOURCES ENGINEERING BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY DHAKA, BANGLADESH JUNE 2013 i

2 MORPHOLOGICAL STUDY OF OLD BRAHMAPUTRA OFFTAKE USING TWO-DIMENSIONAL MATHEMATICAL MODEL Fahmida Noor Student No: P A thesis Submitted to The Department of Water Resources Engineering of Bangladesh University of Engineering and Technology in partial fulfillment of the requirement for the degree of MASTER OF SCIENCE IN WATER RESOURCES ENGINEERING Department of Water Resources Engineering Bangladesh University of Engineering and Technology Dhaka, Bangladesh June 2013 ii

3 CERTIFICATION OF APPROVAL The thesis titled Morphological study of Old Brahmaputra offtake using two-dimensional mathematical model, submitted by Fahmida Noor, Roll No P, Session April 2009, has been accepted as satisfactory in partial fulfillment of the requirement for the degree of Master of Science in Water Resources Engineering on 29 June, Dr. M. Abdul Matin Professor, WRE BUET, Dhaka-1000, Bangladesh Chairman (Supervisor) Dr. Sabbir Mostafa Khan Professor and Head, WRE BUET, Dhaka-1000, Bangladesh Member Dr. Md. Ataur Rahman Professor, WRE BUET, Dhaka-1000, Bangladesh Member Abu Saleh Khan, PEng. Deputy Executive Director Institute of Water Modelling (IWM), Dhaka-1206, Bangladesh Member (External) iii

4 DECLARATION It is hereby declared that this thesis work or any part of it has not been submitted elsewhere for the award of any degree or diploma. Fahmida Noor Signature of the Candidate iv

5 TABLE OF CONTENTS Page No Certification of Approval iii Declaration iv Table of Contents v List of Figures ix List of Tables xii List of Symbols xiv List of Abbreviation xv Acknowledgement xvi Abstract xvii CHAPTER 1: INTRODUCTION General Background of the Study Objectives of the Study Structure of the thesis 3 CHAPTER 2: LITERATURE REVIEW General Previous Studies and Researches on Offtake and River Morphology Previous Studies and Researches on Dredging Previous Studies and Researches on Bifurcation Previous Studies and Researches on Old Brahmaputra The Brahmaputra - Jamuna River System Avulsion of the Brahmaputra River The Old Brahmaputra River Study Area 21 CHAPTER 3: THEORY AND METHODOLOGY General Offtake Offtake Management Dredging River Training Structures Morphological Variables Independent Variables v

6 3.6.2 Dependent Variables Sediment Transport Sediment Transport Equations Engelund-Hansen Equation (1967) Yang's Equation (1976) Van Rijn's Equation (1984) Mathematical Modelling MIKE 21C Governing Flow Equation of MIKE 21C Morphological Model of MIKE 21C Methodology Collection of data from various sources Hydrological and morphological analysis of the collected data A 2D morphological model setup of the study area Selection of different flood events by frequency analysis Flood Routing Simulation of different options to observe the sustainability of the Old Brahmaputra offtake CHAPTER 4: DATA COLLECTION AND ANALYSIS General Data Collection Water Level Discharge Cross Sectional Bed Level / Bathymetry Data Satellite Image Bankline Data Sediment Data Data Analysis Hydrological Analysis Water Level Discharge vi

7 4.3.2 Morphological Analysis Planform Analysis Bankline Shifting Shifting of Mouth of Offtake Sediment Transport Analysis Cross Section and Longitudinal Bed Profile Analysis CHAPTER 5: APPLICATION OF MATHEMATICAL MODEL General Model Setup Grid Generation Bathymetry Generation Boundary Condition Model Calibration Model Validation Model Stability Various Option used in Model Simulation 75 CHAPTER 6: RESULTS AND DISCUSSIONS General Results and Discussions Planform/Bed level changes Cross section analysis Bed Erosion/Deposition Sediment Transport Discharge Percentage of flow diversion Water depth Sustainability of Option Summary CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS General Conclusions 107 vii

8 7.3 Recommendations 108 References APPENDIX-A APPENDIX-B viii

9 LIST OF FIGURES Page No. Figure 2.1 Catchment of the Ganges and Brahmaputra rivers 13 Figure 2.2 Development of the main rivers in Bangladesh over time 15 Figure 2.3 Details of the Brahmaputra avulsion site redrawn from maps by A) 16 Rennell (1776), (B) Wilcox (1828), (C) Allen (1843) and (D) LANDSAT satellite image (1978). Figure 2.4 Study area along the Jamuna River 22 Figure 3.1 a) Dredging along the Gorai River on GRRP-II project and b) River 24 training structures at the mouth of offtake on Gorai Offtake Management Project Figure 3.2 Proposed regulator and guide bunds at the mouth of Dhaleswari 25 Figure 3.3 Flow chart of methodology applied in the study 36 Figure 3.4 Probability plots for Gumbel distribution 39 Figure 3.5 Probability plots for Log Pearson three parameters distribution 39 Figure 3.6 Probability plots for Log Normal distribution 40 Figure 4.1 Location of discharge and water level station near the study area 45 Figure 4.2 Transect line of bathymetric survey for the study area in Jamuna 46 River Figure 4.3 Maximum, minimum and average water level at Bahadurabad station 48 of Jamuna River Figure 4.4 Water level hydrograph at Kholabarichar station 48 Figure 4.5 Maximum and minimum water level of Old Brahmaputra River at 49 Jamalpur station Figure 4.6 Maximum and minimum discharge of Jamuna River at Bahadurabad 50 station Figure 4.7 Yearly maximum discharge histogram of Old Brahmaputra River at 50 Mymensingh station Figure 4.8 Percentage of Maximum flow through Old Brahmaputra River 51 Figure 4.9 Percentage of flow diversion during dry season in different period 51 Figure 4.10 Satellite images in study area showing the off-take of the Old 53 Brahmaputra River in different period. Figure 4.11 Location of the off-take of the Old Brahmaputra River at different 54 time period. Figure 4.12 Changes of the offtake of Old Brahmaputra River at different time 55 period. Figure 4.13 Sediment rating curve of Jamuna River at Bahadurabad station 56 during Figure 4.14 Sediment rating curve of Old Brahmaputra at Mymensingh station 57 during Figure 4.15 Sediment transport of Jamuna River during Figure 4.16 Sediment transport of Brahmaputra River during Figure 4.17 An observed cross section of Old Brahmaputra River during Figure 4.18 The observed maximum, average and minimum bed level from the offtake to the d/s of Old Brahmaputra River. 60 ix

10 Page No. Figure 5.1 Computational Grid cell of Jamuna River in the study area 62 Figure 5.2 Model bathymetry of Jamuna River in the study area for postmonsoon Figure 5.3 Discharge hydrograph of Jamuna River at Bahadurabad for the years , 2001 and Figure 5.4 Discharge and water level boundary for the high flood event (1998). 65 Figure 5.5 Discharge and water level boundary for the medium flood event 65 (2005). Figure 5.6 Discharge and water level boundary for the low flood event (2001). 66 Figure 5.7 Boundary for simulation of 2011 monsoon 67 Figure 5.8 Water level calibration at Bahadurabad station in Jamuna River. 69 Figure 5.9 Water level calibration at Horichandi in Old Brahmaputra River. 69 Figure 5.10 Sediment transport calibration at Bahadurabad station in Jamuna 70 River. Figure 5.11 Comparison of water level at Bahadurabad station for Figure 5.12 Comparison of discharge at Bahadurabad station for Figure 5.13 Discrepancy ratio for water level comparison 72 Figure 5.14 Discrepancy ratio for discharge comparison 73 Figure 5.15 Comparison of cross-section at Bahadurabad location 74 Figure 5.16 Comparison of sediment transport in Jamuna River for validation 74 Figure 5.17 Model stability check for the study area 75 Figure 5.18 Schematic diagram of Option 1 76 Figure 5.19 Schematic diagram of Option 2 77 Figure 5.20 Schematic diagram of Option 3 74 Figure 5.21 Schematic diagram of Option 4 78 Figure 5.22 Design dredged section at the mouth of offtake 78 Figure 5.23 Different simulated options superimposed on base bathymetry 79 Figure 6.1 Surveyed and simulated bathymetry for the base condition for 82 different flood events Figure 6.2 Surveyed and simulated bathymetry for Option 1 for different flood 83 events Figure 6.3 Surveyed and simulated bathymetry for Option 2 for different flood 84 events Figure 6.4 Surveyed and simulated bathymetry for Option 3 for different flood 85 events Figure 6.5 Surveyed and simulated bathymetry for Option 4for different flood 86 events Figure 6.6 Simulated cross sections immediate downstream of offtake for high 87 flood event for different options Figure 6.7 Simulated cross sections immediate downstream of offtake for 88 medium flood event for different options Figure 6.8 Simulated cross sections immediate downstream of offtake for low 89 flood event for different options Figure 6.9 Simulated bed erosion and deposition at the end of monsoon in the Old Brahmaputra reach for high flood event for different options 90 x

11 Page No. Figure 6.10 Simulated bed erosion and deposition at the end of monsoon in the 90 Old Brahmaputra reach for medium flood event for different options Figure 6.11 Simulated bed erosion and deposition at the end of monsoon in the 91 Old Brahmaputra reach for low flood event for different options Figure 6.12 Volume of deposition and erosion for high flood event for different 92 options from the mouth of offtake to 20km reach of Old Brahmaputra Figure 6.13 Volume of deposition and erosion for medium flood event for 92 different options from the mouth of offtake to 20km reach of Old Brahmaputra Figure 6.14 Volume of deposition and erosion for low flood event for different 93 options from the mouth of offtake to 20km reach of Old Brahmaputra Figure 6.15 Sediment transport for different options a) at the immediate 94 downstream of offtake and b) at downstream of Old Brahmaputra River for high flood event Figure 6.16 Sediment transport for different options a) at the immediate 94 downstream of offtake and b) at downstream of Old Brahmaputra River for medium flood event Figure 6.17 Sediment transport for different options a) at the immediate 95 downstream of offtake and b) at downstream of Old Brahmaputra River for low flood event Figure 6.18 Velocity distribution along the cross section for medium flood event. 95 Figure 6.19 Discharge of Old Brahmaputra River for high flood event for 96 different options Figure 6.20 Discharge of Old Brahmaputra River for low flood event for 97 different options Figure 6.21 Discharge of Old Brahmaputra River for medium flood event for 97 different options Figure 6.22 Percentage of flow diversion through Old Brahmaputra River for 98 high flood event for different options Figure 6.23 Percentage of flow diversion through Old Brahmaputra River for low 99 flood event for different options Figure 6.24 Percentage of flow diversion through Old Brahmaputra River for medium flood event for different options 99 Figure 6.25 Percentage of flow diversion at the end of monsoon through Old 100 Brahmaputra River Figure 6.26 Simulated water depth at offtake during dry period for medium flood 101 event Figure 6.27 Boundary for three successive monsoon periods from 2005 to Figure 6.28 Simulated water depth for three successive monsoon periods 103 Figure 6.29 Percentage of flow diversion for the three successive monsoon periods 104 xi

12 LIST OF TABLES Page No. Table 2.1 The Characteristics of the Old Brahmaputra 21 Table 3.1 Peak flows of Jamuna River at Bahadurabad for selected return 38 periods using different distribution Table 3.2 Goodness of fit test among three distributions 40 Table 3.3 Probability of exceedence of Chi square 41 Table 3.4 Peak flow according to Log Pearson three parameters distribution 41 Table 4.1 Available water level data for the study area 44 Table 4.2 Available discharge data for the study area 44 Table 4.3 Available sediment data for the study area 47 Table 5.1 Summary of the parameters used for the model domain 68 Table 6.1 Simulated discharge (in m 3 /s) of Old Brahmaputra at the end of 96 monsoon for different options Table 6.2 Summary results for different options at the end of monsoon 106 xii

13 LIST OF SYMBOLS ω B b B t C C o C t D D 50 D * E g g s h H h b i b L N n p p,q Q Q s Rs,Rn s,n S x S y t U * Fall velocity average width of channel total width Chezy roughness coefficient alluvial resistance coefficient Total sediment concentration in parts per million (ppm) by weight Median sieve diameter Median diameter of bed material dimensionless particle parameter co-efficient of eddy viscosity Acceleration due to gravity Sediment transport per unit time per unit width water depth Water Level bankfull depth slope of the river typical wave length of the curved channels Bed porosity alluvial resistance exponent sinuosity Mass fluxes in the s and n direction discharge sediment discharge Radius of curvature of s and n line Coordinates in the curvilinear coordinate system Total sediment transport in x-direction Total sediment transport in y-direction Time Shear velocity xiii

14 V V cr x,y Z γ γ s ΔS e ν τ o Average flow velocity Critical average flow velocity Cartesian coordinated system Bed level Specific weight of water Specific weight of sediment particles Lateral sediment supply from bank erosion Kinematic viscosity Bed shear stress xiv

15 LIST OF ABBREVIATION 2D 3D BRIC BWDB BTM CEGIS CPP Delft 3D DHI DPP DMI EGIS FAP FMG FPCO GRRP Ha-m HEC-RAS HYMOS IWM ISPAN MIKE 11 MIKE 21C PWD RSP SWMC USA Two-Dimensional Three-Dimensional Bangladesh River Information and Conservation Project Bangladesh Water Development Board Bangladesh Transverse Mercator Center for Environment and Geographic Information Services Compartmentalization Pilot Project Three-dimensional model developed by Delft hydraulics Danish Hydraulic Institute Development Project Performa/Proposal Danish Maritime Institute Environment and GIS Support Project Flood Action Plan Flood Management Division Flood Plan Coordination Organization Gorai River Restoration Project Hectare-meter Hydraulic Engineering Center- River Analysis System Frequency Analysis tool developed by Delft Hydraulics Institute of Water Modelling Irrigation Support Project for Asia and the Near East One-dimensional One Layer Model Develop by MIKE ABBOTT Two-dimensional One Layer Curvilinear Model Develop by MIKE ABBOTT Public Works Datum River Survey Project Surface Water Modelling Centre United States of America xv

16 ACKNOWLEDGEMENT First of all, the author would like to express her gratefulness to the almighty Allah the creator of universe. Without the rahmat of Allah the author couldn t have finished the degree. The author acknowledges her deepest gratitude to her supervisor Dr. Md. Abdul Matin, Professor, Department of Water Resources Engineering, BUET for providing an interesting idea for the thesis and encouraging to work with it. His cordial supervision, valuable suggestion and expertise contributed greatly to this thesis. The author is also grateful to Dr. Sabbir Mostafa Khan, Professor and Head, Department of Water Resources Engineering, BUET, Dr. Ataur Rahman, Professor, Department of Water Resources Engineering, BUET and Mr. Abu Saleh Khan, Deputy Executive Director, IWM, who were the members of the board of Examiners. Their valuable comments on this thesis are duly acknowledged. The author would like to thank her parents for their encouragement and inspiration. The author is grateful for their guidance and blessings. Without their help and support she wouldn t have finished her M.Sc. The author is grateful for their limitless sacrifice and love. The author is truly grateful to her husband for his continuous support and inspiration. She appreciates and admires his patience and encouragement throughout her study. The author would like to express her sincere gratitude and appreciation to her all colleagues and friends for their help, support and inspiration. Author is in debt to Institute of Water Modelling (IWM), for providing necessary data, information and modelling tools to carry out this research work. Fahmida Noor June 2013 xvi

17 ABSTRACT Most of the distributaries of major river system of Bangladesh have been silted up. The Old Brahmaputra River is one of the distributaries of the Jamuna River. Nowadays, the mouth of Old Brahmaputra River becomes dying due to serious siltation of sediments in the vicinity of the offtake. The offtake channel conveys a fraction of flow from its original discharge of the main river Jamuna. The maximum diversion ratio also shows a decreasing trend over the years from 1964 to During lean period, this flow diversion is drastically reduced to below 1% in the last few decades. In order to augment the dry season flow of Old Brahmaputra, a better offtake management is required. In this thesis, a mathematical morphological model has been setup for the study area and simulated for various options. The study area covers 50 km reach of Jamuna River starting from the 25km upstream of the mouth of Old Brahmaputra offtake to 25km downstream of the offtake. It also covers the 20 km reach of Old Brahmaputra River. To setup these morphological model, MIKE21C, an advanced two-dimensional mathematical modelling software developed by DHI, has been applied. The model has been successfully calibrated and validated against the year of 2011 and 2012, respectively. Finally, assessment of morpho-dynamics at high, medium and low flood event of Jamuna for existing and design river bed condition of offtake has been conducted. Four different options have been chosen considering the dredging and river training works for the simulation. The validated model has been run with different types of options. Results of model outputs have been interpreted to understand the outcomes of the options used. The findings from the result suggest that dredging along the left anabranch of Jamuna River (Option 1) does not improve the flow condition of Old Brahmaputra rather than a huge siltation is taken place in front of offtake which impede the flow through the river. Placing of two guide bunds in the mouth of offtake (Option 2) or only dredging in the vicinity of offtake (Option 3) improves the flow condition relatively. About 200 m 3 /s discharge is available during the month of November for Option 3 whereas it was almost null for base condition. Hence, the percentage of flow for Option 3 varies from 2% to 2.4% during November, while it is less than 2% for Option 2. Again, the connection between two rivers is closed at the end of November for Option 2 whereas it sustains another one month for Option 3. The last option which is the combination of guide bund and dredging (Option 4) provide good results in terms of flow condition. Dredging make the river deeper and confined and guide bunds help to generate less siltation near the mouth of offtake. It is seen that xvii

18 approximately 250 m 3 /s discharge is available at the end of November if Option 4 is implemented for average flood event condition. In addition, about 32m 3 /s of flow can be available in the Old Brahmaputra during the lean period for this option. Therefore, it can be suggest that combination of guide bund at the mouth of offtake and the dredging (Option 4) is the most favorable and sustainable engineering solution of the offtake management. xviii

19 CHAPTER 1 INTRODUCTION 1.1 General The offtakes are important links between the main rivers and the distributaries. The distribution of discharge and sediment transport at river offtake is a key factor for the long term morphological development of the main rivers (FAP 24, 1996a). The dynamics of such offtakes have undoubtedly been of great significance during channel avulsion, such as: the occupation of the current Jamuna channel and abandonment of the Old Brahmaputra. The mouth of the river is not stable and changes when the left bank anabranch of the Jamuna River has a changing discharge. At present, serious deposition has taken place at the mouth, induced by intense char movement. For analyzing the siltation problem of this river, two-dimensional morphological model will be a useful tool to investigate the present problem to propose a better solution for the offtake management. The detail study on offtake morphology will be helpful for engineers, planners involved in offtake management. To some extent the study will suggest the possible restoration measures to be undertaken of Old Brahmaputra offtake region. 1.2 Background of the Study The major rivers, including the Ganges, Brahmaputra-Jamuna, Padma and Meghna and their numerous tributaries and distributaries make Bangladesh a land of rivers. These rivers are very dynamic in nature as most part of Bangladesh has been formed by recent sediments and rivers that are loaded with huge sediment. The catchment area of these rivers is about 1.72 million km 2 of which only 7.5% lies within the borders of Bangladesh. This catchment annually generates 120 million ha-m of runoff, only 10% of which is generated within Bangladesh. In addition to vast quantities of water, these rivers carry about 1.6 to 2.0 billion tons of sediment load every year. The large discharges and heavy sediment loads carried by these rivers results in highly variable and dynamic channel morphologies characterized by rapid adjustments to the cross-sectional geometry, bankline positions and planform attributes (Coleman, 1969). 1

20 The distribution of discharge and sediment transport at river offtakes is a key factor for the long term morphological development of the main rivers. The sediment transport distribution, in addition, may be influenced by the local planform and bathymetry at the offtake, which on the other hand are dependent on the planform and sediment transport conditions of the main river (FAP 24, 1996a). The presence of systems of moving bars further complicates conditions at offtake channels. A bar in front of the offtake may partly block the inflow to the distributary. For the development of the distributaries the offtakes are the crucial elements. At an offtake the flow and the sediment transport entering into the distributary are determined by the downstream conveyances, the local geometry of the offtake becomes less important. If more sediment is entering into the offtake than can be transported over a longer period, the distrbutary will start to aggrade. As a consequence less water is entering and hence this may start a self-accelerating process. It has been found that most offtakes are unstable: either the distributary start to die or gradually the distributary takes the function of the main river. They are start to function only during higher stages. Any change in the stages in the main river will affect the flow (and sediment transport) entering into the distributary. In particular a decrease of stages will result in less water entering into the distributary and depending on the quantities of sediment entering the distributary, it will silt up. 1.3 Objectives of the Study Objectives of the study are as follows: 1. To assess the morphological characteristics at Old Brahmaputra offtake of Jamuna River. 2. To assess the siltation pattern around the offtake by using mathematical model. 3. To suggest better offtake management option after simulating various options. 2

21 1.4 Structure of the thesis The thesis has been organized under seven chapters. Chapter 1 describes the background and objectives of the study. Chapter 2 describes different definition of relevant topics, literatures, previous studies related to this study. Theories regarding this thesis work and a brief description of the methodology have been described in Chapter 3. Chapter 4 describes the data collection from various sources and its processing. The hydrological and morphological analysis is also represented in this chapter. Chapter 5 illustrates the model set up for the study area, its hydrodynamic and morphological calibration as well as validation. Chapter 6 demonstrates the results in terms of bed level, discharge, siltation etc obtained for different option simulation. Conclusion and recommendations for further study are outlined in Chapter 7. 3

22 CHAPTER 2 LITERATURE REVIEW 2.1 General In Bangladesh, studies have been done on the morphological aspects such as planform, hydraulic geometry, erosion /deposition and bed level variation in many rivers. Most of these studies were carried out for the major rivers like Ganges, Jamuna, Padma and Meghna. Few studies have been done on the Old Brahmaputra offtake in Bangladesh. The present study will be directed towards an assessment of siltation problem around the Old Brahmaputra offtake by using two-dimensional mathematical modelling approach. In addition, previous studies and researches relevant to current study is discussed in the following articles. 2.2 Previous Studies and Researches on Offtake and River Morphology FAP 24 (1996b) studied the Gorai offtake in detail on the Special Report No 10. The distributary was selected because of the importance of the Gorai for the fresh water inflow to the Southwest region, but also because the Gorai is a well defined single channel without any tidal influence and therefore is less complex. They mainly focused on three elements: the analysis of historical data in order to understand how the hydraulic conditions at Gorai offtake have changed over the last 30 years, the collection and analysis of detail data in special surveys of the RSP, the development of two-dimensional morphological modeling to improve the understanding of which processes are important for the development of the dry season flow to the Gorai. Dey, K.C et. al. (1998) have taken attempts to derive the exponent b analytically for different total and bed load predictors with some assumptions. Seven total load and three bed load prediction formula were considered in this analysis. The relation between the exponent b and Shields parameter (Ө) also analysed. The analysis shows that the exponent b varies between 3.1 to 8.5 for total load and 3.3 to 10.5 for bed load predictors. The sensitivity of the exponent b for variation with the grain sizes (d 50 ) is found to be significant for Ackers and White and VanRijn formula. The water level slope has a small influence on the exponent b as observed at low values of the Shields parameter (Ө). 4

23 Imteaz et. Al. (2001) developed a mathematical model for the Old Dhaleswari River as well as other offtakes from Jamuna River. Due to the construction and associated river bank protection works of Jamuna Multipurpose Bridge on Jamuna River at Bangladesh, water flow through the Old Dhaleswari River was reduced significantly. As a result effectiveness and usefulness of irrigation project named Compartmentalization Pilot Project (CPP) area at Tangail located at the downstream end of Old Dhaleswari River became at a stake. A link canal was proposed with mouth at the Jamuna River connecting the Old Dhaleswari River. The model was simulated to check the effectiveness of the proposed link canal in terms of water availability to the downstream users. It is found that proposed link canal could augment water supply for the irrigators significantly. Obasi et. al (2008) constructed a physical model with meandering features and used to investigate the effect of off-take angles on the flow distribution at a concave channel bifurcation. Seven different off-take angles with varied main channel flow rates were used for the study. Predicting equations for the off-take discharge dependent on the off-take angles, main channel discharges, dispersion coefficients and Reynolds numbers were developed and calibrated statistically. Results of the study and predicting equations showed that the offtake discharge increased positively with increases in off-take angles as well as main channel discharges. The developed empirical predicting equations for the offtake discharge gave correlation coefficient values of x10-1 for both model equations with corresponding standard errors of 9.754x10-5 and 9.42 x10-5, respectively. It was observed that the predicted off-take discharge values from the model equations compared closely with those of the study suggesting that off-take discharges for concave channel bifurcations could be fairly predicted with the established model equations. Agunwamba et. al. (2009) showed how different offtake angles influence velocity distribution around the canal entrance which will in turn influence the quantity of sediments deposited along the canal bed. The problem of excessive siltation in canals (navigation, irrigation, water supply, etc) was tackled by the Schwarz-Christoffel transformation, neglecting gravity and assuming a constant depth of flow. This implies that large off take angles will encourage more intake of sediments by the canal. In addition, it was also observed that large off take angles exhibit higher and lower (wider range) velocities. That is, near the stagnation point, a large off take angle will posses lower velocities than small off take angles thus encouraging siltation, while near the point of infinite velocity, a large off take angle will posses higher velocities thereby increasing 5

24 sediment intake by canal. It is therefore recommended that canals off take angles should be as small as possible but not too small. If the off take angle is too small, the bank between the branch canal and the main canal will be eroded gradually leading to flooding and eventual destruction of the canal. The results obtained can be applied to navigation, irrigation and water supply canals. The results obtained show that the larger the off take angle, the higher will be the off take discharge as well as the off take entrance velocity distribution. The results were found to agree with both laboratory data obtained using a model and field data, giving correlation coefficients of 0.76, 0.77 and Haskoning (2012) in collaboration with IWM carried out the Bangladesh River Information and Conservation Project (BRIC). The main objective of this project is to develop a new two-dimensional hydraulic and morphological model for the Gorai off-take area and its surroundings (including data collection and re-calibration), and use this model to assess the effectiveness of the engineering interventions of the selected option that could potentially help improve water flows into the Gorai sub-basin, and reduce salinity in the downstream areas. Based on the changes over the last ten years such as planform changes and river training works constructed in the Gorai and / or the Ganges, the structural interventions of the option might need to be modified. The model is used to analyse the effectiveness of the modified option. As a result suitable modifications will be made, for which the engineering designs will be updated. Obasi et. al (2012) examined the effect of offtake angles on spatial distribution of silt material at concave bifurcation. For this purpose, a meandering physical channel was constructed. Four different off-take angles of 30 0, 45 0, 60 0 and 90 0 with varied main channel flow rates were used for this study. Predicting statistical equations dependent on the off-take angles and main channel discharges for the evaluation of the tributary channel sediment intake were developed. Results of the studies showed that even with constant main channel discharge, the tributary channel sediment intake increased significantly with higher off-take angles. It was observed that the predicting equation under estimated the tributary channel sediment yield for off-take angles between 30 o 70 o and for those between 70 o - 90 o the sediment values were over estimated for all the main channel flow rates considered. The predicting tributary sediment values equaled the experimental values at the off-take angles of 50 o 70 o but varied differently for each of the main channel flow rates. It could be seen from the various off-take angles considered that the divergence in results obtained from the experimental works and predicting equation is in the range of 6

25 2.9% % for minimum main channel flow rate and 12% - 36% for maximum main channel flow rate suggesting that the predicting equation could be useful in the evaluation of sediment yield at concave channel bifurcation. Mamun et. al. (2012) has studied the hydro-morphological analysis of the offtake of the Arial Khan River of Bangladesh to predict its sustainability. Analysis of historical hydrometric data and satellite images near the offtake and selected stations for both the parent Padma River and the bifurcated Arial Khan River had been carried out. It shows that water levels and discharges are in rising trend in both the rivers. In 1975, 2% peak flows of the Padma River is diverted to the Arial khan River which has been increased to about 3% in The offtake reach of the Arial Khan Upper River which is concerned for the study is in the trend of losing conveyance due to aggradations resulted from longterm sedimentation. It reveals that heavy sedimentation has been occurring which leads to formation of sand bars. This eventually has an impact over the dynamics of the offtake of the Arial Khan River. Again, bed topography generated from bathymetric data of 2005 at the junction with the Padma clearly demonstrate that the main channel of the Padma is aligned far north which is opposite to the offtake and there is a char ( sand bar) formed at the mouth of the Arial Khan hindering the flow and the channel has become narrowed. It has been observed from the satellite images of 1974 to 2006 that char area in the vicinity of Arial Khan offtake was 60 sq km in 1974, 72 sq km in 1985, 82 sq km in 1997 and 92 sq km in The yearly development rate of char area is 2%. The Arial Khan Upper offtake is in the trend of abandoning stage and there must be a shifting offtake which has been dynamically under development to keep the river morphologically active. Recent bed topography of the offtake, conveyance, trend analysis of recent inflows and stages in the Arial Khan River support that the river is going to be the cause of sufferings in terms of flooding in the monsoon and navigability losses and other water resources activities will be hampered due to inadequate inflows in the river. 2.3 Previous Studies and Researches on Dredging Lagasse (1986) examined the impact of dredging on the Mississippi River system. He reveals that the dredge provides the river engineer with a means of rapidly altering channel configuration and accelerating morphologic processes. In this respect, dredging constitutes a morphologic agent responsive to engineering requirements. This application is overshadowed by the volume of material moved and the number of reaches involved in 7

26 dredging operations for navigation channel maintenance. Dredging and disposal of dredged material in support of channel maintenance implies the repeated moving of alluvial sediments from the main channel region toward the periphery of the channel. The combined use of dredging, contraction dikes, and disposal of dredged material in the dike fields can induce major changes in the cross-sectional characteristics of a river. This direct physical displacement of bed material and the resulting change in channel shape can retard the movement of bed-load sediments through a river system. In both the Columbia and Mississippi River systems this lateral redistribution of sediment by dredging, when combined with contraction works, has constituted an agent for long-term morphologic change Van Rijn (1986) developed a detailed mathematical model for sedimentation of dredged channels, based on a detailed representation of all relevant transport processes such as convection, mixing and settling. This is an important advantage compared with the traditional prediction formulas, which are based on a rather strong schematization of the transport processes. He presented a sensitivity analysis showing the influence of the streamline refraction effect and the wave shoaling effect in the channel on the sedimentation process. Two applications of the proposed mathematical model are given and show reasonable agreement between measured and computed concentrations and sedimentation rates. Finally, a set of graphs was presented which can be used to get a rough estimate of the trapping efficiency of dredged channels. Mead (1999) presented two-dimensional-horizontal and two-dimensional-vertical mathematical models of flow and suspended sand transport used in typical engineering predictions of deposition in a dredged trench across an estuary. The results of the models were compared with experimental data, and it is found that the sedimentation predictions are fundamentally dependent on the specification of mobile bed sand, and that the predictions of the two types of model are qualitatively different. He provided an insight into the benefits of both modelling techniques, and indicated the need for further model development. BWDB (2010) has taken up steps to carry out dredging from the offtake, for a length of 30 km of the river. The Government approved a DPP under the name Gorai River Restoration Project Phase-II (GRRP-II) for this purpose. The objectives of the mathematical modelling study are to support the Gorai River Restoration Project Phase-II 8

27 (GRRP-II), during the capital dredging operation, for monitoring through application of the state-of-the-art tool, mathematical modelling, alongside other monitoring mechanism (which BWDB had put in place through Bathymetric Survey), and during operation of maintenance dredging activities and to study hydrodynamic, morphological and salinity intrusion phenomena of the river systems, situated at the downstream of the Gorai river, namely the Modhumoti-Kaliganga-Baleswar-Haringhata and Noboganga-Atai-Rupsha- Kazibacha-Pussur systems under base (without dredging of the Gorai) and project condition (with dredging implemented in the Gorai). 2.4 Previous Studies and Researches on Bifurcation Roosjen et.al. (1995) is used experimental model used to do experiments which lead to a better insight in the behaviour of the bifurcation. Several experiments have been designed and carried out with the test rig. The measurement errors, made during the experiments, are described, which gives a good view of the quality of the experiments. Since it was time-consuming to obtain data from the experiments, a thorough statistical analysis of the found data was carried out. Several statistical techniques were used to obtain as much information as possible from the data. With the use of this statistical analysis, nodal point relations were found for the specific types of bifurcations in the test rig, for three different upstream discharges and two different shapes of the bifurcation. It appeared that the general nodal-point relation proposed by Wang et al (1993) was appropriate. The unknown parameters of this relation were found for the circumstances of the experiments; three relations are found for the first shape of the bifurcation, with a respective upstream discharge of 20 1/s, 30 1/s and 40 1/s. Two relations are found for the second shape of the bifurcation, with a respective upstream discharge of 30 1/s and 40 1/s. It is statistically proven that some of the coefficients in these relations are comparable for the different circumstances, but others are not. In order to see whether it is possible to carry out numerical simulations of a bifurcating river, the configuration of the experimental model and the found nodal point relations were used as input for simulations with WENDY. Thus simulations were carried out of some of the experiments. It appeared that the results of these simulations were comparable to the measured data from the experiments. Therefore it is proven that if a good nodal-point relation is known, a good simulation of a bifurcating river can be carried out. 9

28 Mosselman, E. (2004) employs different categories of methods: theoretical analysis, field measurements, laboratory experiments (elementary process experiments as well as physical scale modelling) and mathematical modeling to define the morphology of river bifurcation. As a consequence, different categories of methods have been used in research on sediment transport and morphology at river bifurcations. A theoretical analysis revealed that the morphological development of bifurcated channels depends sensitively on the way in which sediment transport rates are divided over the two branches. A combination of field measurements and mathematical modelling provided an insight in the effects of grain sorting and alluvial roughness that was not given by previous physical modelling. Kleinhans, M., et. al. (2006) studied about the effect of upstream meanders on bifurcation stability and sediment division in 1D, 2D and 3D models. According to them, at river bifurcations, water and sediment are divided over two branches. The dynamics of the division determine the long-term evolution of the downstream branches, which can be studied by 1D models. For such models, a relation describing the sediment division is necessary, but this is poorly understood. They studied the division of sediment and the morpho-dynamics on a time scale of decades by idealized 2D and 3D modelling of bifurcations with upstream meanders and dominantly bed load transport. Initially, migrating alternating bars in the models caused damped quasi-periodic fluctuations in bed levels, water and sediment division until the bars are near equilibrium. Varying the radius of upstream meanders and the slope of one of the downstream channels led to subtle changes in the sediment transport direction and the location of bars and pools. This caused large differences in which branch becomes dominant and in the rate of change in discharge asymmetry. The effects of wider downstream branches or of an overall narrower or wider river are dramatic, again demonstrating the importance of bars. The resulting division of sediment, on the other hand, is similar to the division of flow discharge in all runs after the initial fluctuations have damped out and until the discharge division becomes very asymmetrical. They concluded that bifurcations are extremely sensitive to local conditions affecting the secondary currents and the sediment transport direction, and to the downstream boundary conditions. Although most combinations of parameters lead to the development of an asymmetrical discharge division, some combinations lead to a quasistable symmetrical division. Finally they discussed the limitations of the models and the applicability to natural meandering rivers. 10

29 2.5 Previous Studies and Researches on Old Brahmaputra SWMC and DHI (1999) jointly studied for High Point Rendell/ Mitsui to support the barge transportation in the Old Brahmaputra. The project area covers about 161 km stretch of the Jamuna and Old Brahmaputra River starting from Bahadurabad Ghat upto Mymensingh. The study comprises a details assessment of the dredging requirement for transportation of the barge. Two sets of cross sectional data have been used for this purpose. A one dimensional mathematical model is set up for the Jamuna and Old Brahmaputra rivers with the purpose to analyze the possible variation of water levels and velocities along the river. A two-dimensional model is also setup for a critical reach of the Old Brahmaputra River. The main purpose of the two-dimensional modeling of river is to provide the Danish Maritime Institute (DMI) with hydrodynamic model results to be applied in ship navigation simulations. These simulations had been used to give advice on navigational matters in relation to the barge transportation. Boskali (2000) investigate the technical feasibility of a standalone dredging intervention in the Old Brahmaputra in order to augment dry season flows of the river. Moreover, the study should indicate environmental and socio-economic impacts of the intervention and assess costs and benefits. The findings from this study indicate that augmenting dry season flows of the Old Brahmaputra by dredging is technically feasible. The Old Brahmaputra can be kept flowing by dredging through the offtake area, by making a connecting channel in the Jamuna (between the offtake and a main Jamuna channel) and removing the bottlenecks in the upper reach of the Old Brahmaputra. EGIS (2001) undertook the planform analysis of the Old Brahmaputra River in connection with the Khurshid Mohal Bridge project under contract from Surface Water Modelling Centre (SWMC) using satellite images. The objectives of the study were to assess the changes of average width of the different reaches of the Old Brahmaputra River with time, meandering bend migration rate of the river, bank erosion/accretion rate at different reaches of the river, sinuosity and meandering wave length and amplitudes in order to support review of the selection of the location for bridge and approach roads and estimating the different design parameters for construction bridge over the Old Brahmaputra River. Ali (2010) investigated the siltation at the intake reach of the Old Brahmaputra River. For this purpose, satellite images, cross-sectional data, water level, and discharge data of Old 11

30 Brahmaputra River covering the period from 1973 to 2005 were collected. The overall trend line of Northing is increasing and Easting is decreasing, which indicates that the mouth of the Old Brahmaputra River is shifting towards north-west ward. It is also visualized that the river has become comparatively narrower. From 1973 to 2007 both narrowing and widening of the channel have been observed in all sections but ultimately channel narrowing process appears to be a dominant process in the area. The average widening of the Intake reach for the last 34 years is m/yr and narrowing is m/yr. Considering variation in cross-sectional area it can be found that for Intake reach sedimentation is portrayed in the period 1996 to 2005 ( m 2 ) and for the remaining cross sections of the river, sedimentation is portrayed in the period 1996 to 2005 ( m 2 ). Sedimentation dominates over a period of nine years ( ) by raising of mean bed level of about 2.10 m. The movement of thalweg in the vertical direction is very dynamic over the study years. From 1988 the water level of Old Brahmaputra River at Kholabarichar station stops falling at a certain water level of Jamuna River at Bahadurabad station, which may change from year to year. This is the indication of closure of Old Brahmaputra mouth. In 1973, percentage of Jamuna River flow to the Old Brahmaputra River was 6.69% and in 2007 this percentage of flow is reduced to 0.70%. This reduction of flow reveals the siltation problem at the Intake reach of the Old Brahmaputra River. 2.6 The Brahmaputra - Jamuna River System The Brahmaputra flows through a narrow valley, which is known as the Brahmaputra valley in about east west direction for 640 km with a very low gradient. In this valley it is joined by several tributaries from both sides. On the west, the valley is open and beyond Assam it widens into a broad low lying deltaic plain of Bangladesh. The Brahmaputra, after traversing the spurs of the Meghalaya plateau, turns south and enters Bangladesh with the name of Jamuna. The total length of the river from its source in south-western Tibet to the mouth in the Bay of Bengal is about 2,850 km (including Padma and Meghna up to the mouth). Within Bangladesh territory, Jamuna is 240 km long (upto Aricha). The Jamuna enters Bangladesh east of Bhabanipur (India) and northeast of Kurigram district. Originally, the Jamuna (Brahmaputra) flowed southeast across Mymensingh district where it received the Surma River and united with the Meghna, as shown in Rennell s Atlas (1785). By the beginning of the 19th century its bed had risen due to tectonic movement of the Madhupur Tract and it found an outlet farther 12

31 west along its present course (Coleman, 1969). It has four major tributaries: the Dudhkumar, Dharla, Teesta and the Baral-Gumani-Hurasagar system. The first three rivers are flashy in nature, rising from the steep catchment on the southern side of the Himalayas. The main distributaries of the Jamuna River are the Old Brahmaputra River, which leaves the left bank of the Brahmaputra River 20 km north of Bahadurabad, and the New Dhaleswari River just south of the Bangabandhu Bridge. The Brahmaputra-Jamuna drains the northern and eastern slopes of the Himalayas, and has a catchment area of 5, 83,000 sq.km percent of which lie in China, 33.6 percent in India, 8.1 percent in Bangladesh and 7.8 percent in Bhutan (Figure 2.1). Figure 2.1: Catchment of the Ganges and Brahmaputra rivers (Source: CEGIS, 2007) The catchment area of Jamuna River in Bangladesh is about 47,000 sq. km. The average annual discharge is about 19,600 m 3 /sec (ISPAN, 1995), which is nearly twice that of the Ganges. The Brahmaputra River is characterized by high intensity flood flows during the monsoon season, June through September. There is considerable variation in the spatiotemporal distribution of rainfall with marked seasonality. Precipitation varies from as low as 120 cm in parts of Nagaland to above 600 cm in the southern slopes of the Himalayas. In Bangladesh territory rainfall varies from 280 cm at Kurigram to 180 cm at Ganges- 13

32 Brahmaputra confluence (FAP 2). Monsoon rains from June to September accounts for 60-70% of the annual rainfall. These rains that contribute a large portion of the runoff in the Brahmaputra and its tributaries are primarily controlled by the position of a belt of depressions called the monsoon trough extending from northwest India to the head of the Bay of Bengal. Deforestation in the Jamuna watershed has resulted in increased siltation levels, flash floods, and soil erosion. Occasionally, massive flooding causes huge losses to crops, life and property. Periodic flooding is a natural phenomenon which is ecologically important because it helps maintain the lowland grasslands and associated wildlife. Periodic floods also deposit fresh alluvium replenishing the fertile soil of the Jamuna River Valley. 2.7 Avulsion of the Brahmaputra River Sometime between 1776 and 1830, the course of the Brahmaputra River shifted from the east of Madhupur block to the west, and river in its new course took the name Jamuna. The term avulsion may be used to describe this shift as the change in channel alignment was achieved by an abandonment of one course and adoption of a new one some distance away rather than by progressive shifting (Bristow, 1999). There is no agreement among the scientists concerning exactly when, how and why the avulsion occurred (Morgan and McIntire, 1959; Coleman, 1969, Monsur, 1995; and Bristow 1999). In this context, use of this term should not then be taken to indicate that the shift occurred during a single event or even within a few years. For example, in 1916, Hirst suggested that the avulsion took place gradually over a sixty year period (Morgan and McIntire, 1959). What is known is that shifting did not begin in earnest earlier than 1776 because Major Rennell s map of that date clearly shows the Brahmaputra flowing east of the Madhupur, along the present course of the Old Brahmaputra River. It is also clear that shifting of the main course had been accomplished by 1830, as Colonel Wilcox s map of that date indicates that the main flow had been diverted to form the Jamuna, to the west of the Madhupur block (Figure 2.2). Regarding how the shift occurred, writing in 1810, Buchanan Hamilton noted that the Brahmaputra was at that time threatening to shift westwards along the course of Konni (or Jennai) River (Fergusson, 1863) and for many years the common opinion was that this threat was realised in the late 18 th century when, the Brahmaputra started to divert the flow through the Jennai River (Morgan and McIntire, 1959; Coleman, 1969; and Monsur, 14

33 1995). However, this opinion has recently been challenged (Bristow, 1999) on the basis that both on Rennell s map of 1776 and modern maps, the Jennai River is located to the east of the town of Dewanganj while, following its avulsion, the Brahmaputra occupied a channel west of Dewanganj. Figure 2.2: Development of the main rivers in Bangladesh over time (Source: CEGIS, 2007) There are also different opinions on the cause of the avulsion, including tectonic influence, tributary switching, and the occurrence of a very large flood (Bristow, 1999). For many years scientists believed that tectonics were the main causes of switching the Brahmaputra from the east of Madhupur to the west, with Fergusson (1863) being one of the first to attribute the avulsion specifically to uplift of Madhupur block. In 1916, Hirst advanced an explanation based on the hypothesis that the avulsion was triggered by ground sinking along the present course of the Jamuna together with compensatory uplift of the Barind and Madhupur blocks (Morgan and McIntire, 1959). Coleman (1969) suggested that geologic events such as earthquakes, which caused faulting and tilting of the Madhupur block, were liable for the avulsion. He also mentioned that the present course of Jamuna River is one of the former courses of the Teesta River, implicating tributary shifting as a supplementary cause of the avulsion. More recently, the map of sediment deposition thickness during the Holocene produced by Goodbred and Kuehl (2000b), which shows a narrow subsiding strip along the present course of the Jamuna River, supports Hirst s concept. 15

34 Morgan and McIntire (1959) concluded that the avulsion occurred gradually in response partly to tectonic tilting of the Madhupur block and partly due to the addition of the flow from the Teesta River due to its sudden avulsion from west to east, to confluence with the Brahmaputra just upstream of the point of avulsion. They hypothesize that this sudden increase in high season discharges might have led to higher stages, increased spill flows and channel scouring along the course of the Jennai River that in time promoted an avulsion by the main flow of the Brahmaputra River. Chilmari Chilmari Dewanganj Dewanganj A Madhupur B Madhupur Chilmari Chilmari Dewanganj Dewanganj Jamuna River Madhupur Madhupur C Sirajganj D Sirajganj Figure 2.3: Details of the Brahmaputra avulsion site redrawn from maps by (A) Rennell (1776), (B) Wilcox (1828), (C) Allen (1843) and (D) LANDSAT satellite image (1978). (Source: Bristow 1999) 16

35 As explained in the last section, Goodbred and Kuehl (2000a) suggest that the most recent avulsion was not a unique event, but was just the latest in a series of periodic switches of the Brahmaputra from the east to west of the Madhupur block. According to their hypothesis, the period for switching is only about 2,000 to 3,000 years. While tectonic uplift and tilting of the Madhupur block and associated subsidence in the Sylhet Basin and along the zone of weakness between the Madhupur and Barind Tracts is the underlying cause of switching, different events may act as triggers for particular avulsion events. Specifically, trigger events might include earthquakes, tributary diversions, and major floods (Makaske, 2001). To these might be added the crossing of geomorphic thresholds intrinsic to the fluvial system (Schumm, 1977) associated with channel slope adjustments driven by sediment accumulation that is alternately centered in the Sylhet Basin and the trough between the Madhupur and Barind Tracts. The earliest study of the Brahmaputra was made by Rennell (1781) who mapped the Brahmaputra (Rennell, 1776) as a braided river flowing into the Meghna River and following a course now occupied by a much smaller river known as the Old Brahmaputra. Since then the river has changed course and avulsed into its present course known as the Jamuna River. The exact timing and cause of the avulsion remains uncertain. Theories include tectonic tilting, an earthquake trigger, river capture, flooding or increased discharge resulting from an upstream tributary avulsion. The main theories are reviewed here and discussed with respect to the cartographic evidence (Figure 2.3), although it is possible that the maps themselves may not be completely accurate. Tectonic influence Morgan & McIntire (1959) assert some tectonic influence; Probably the diversion of the Brahmaputra was gradual, as most river diversions are, and was caused in part by gradual tilting of the Madhupur block. This tilting caused the Old Brahmaputra River to become antecedent in places, necessitating river scour into slowly or periodically rising, comparatively resistant Pleistocene sediments. (Morgan & McIntire, 1959). Coleman (1969, p. 336) says that... faulting was probably the major cause of the recent shift of the Brahmaputra River from its course east of the Madhupur Jungle to its present position. Winkley (1994) state that tectonic activity in 1772 altered the course of the Brahmaputra, which seems unlikely because this pre-dates Rennell s map (1776). Bangladesh is tectonically active and the rivers appear to follow structural trends (Morgan 17

36 & McIntire, 1959); tectonic subsidence has almost certainly affected the river. There is no clear link, however, between the avulsion and a specific earthquake. Tributary switching It has been suggested (Morgan & McIntire, 1959; Monsur, 1995) that the avulsion of the Brahmaputra River followed avulsion of the Teesta River, a major tributary of the Brahmaputra, which has its confluence just upstream from the avulsion node of the Brahmaputra. The sudden change of course by the Teesta River with resulting addition of waters to the Brahmaputra River may well have been a contributing factor towards diversion (Morgan & McIntire, 1959). Using evidence from Rennell s map of 1776, Morgan & McIntire (1959) suggest that the Teesta flowed into the Purnabhaba and Atrai rivers, which are tributaries of the River Ganges. They suggest that during the 1770s, when Rennell made his map, the Atrai was the major channel. Nowadays, the Teesta flows directly into the Brahmaputra. Neither Wilcox s map of 1828 nor Allen s map of 1843 show the Teesta as a significant tributary of the Brahmaputra, however, which raises a question over the assertion that avulsion of the Teesta caused the avulsion of the Brahmaputra. Allen (1843) shows that the Teesta River was still largely flowing into the Purnabhaba and Atrai Rivers, whereas the Brahmaputra was starting to change course (Figure 2.3 C). The Teesta appears in its present position on a map of 1872, based on a survey of 1868 (Thuillier, 1872). The cartographic evidence indicates that the Teesta avulsion occurred after the avulsion of the Brahmaputra. Flood event According to LaTouche (1910) and Morgan & McIntire (1959), the Teesta River changed course suddenly in 1787 during a single flood. This report is not supported by the maps of Wilcox (1828) or Allen (1843), neither of which shows the Teesta as a significant tributary of the Brahmaputra (Figure 2.3 B & C). This does not rule out a later flood event on the Brahmaputra causing the avulsion. River Capture On Rennell s map there is a distributary of the Brahmaputra called the Jenni River, which exits the Brahmaputra downstream from the avulsion node. Other authors (Morgan & McIntire, 1959; Monsur, 1995) refer to the Jennai as a precursor of the Jamuna, which opens up the possibility of river capture as a cause of the avulsion. Rennell s and Wilcox s 18

37 maps, however, show the towns of Chilmari and Dewanganj on the west (right) bank of the Brahmaputra, upstream from the Jenni (Figure 2.3 A & B). At the present day the town of Dewanganj is on the east (left) bank of the Jamuna and downstream from the Jamuna/Old Brahmaputra junction. Thus the avulsion node is upstream from the Jenni. This interpretation is supported by Allen s map of 1843, which shows the Jenni River to the east of the Jamuna, indicating that the Jenni was not the site of the Brahmaputra avulsion. This observation raises questions over Coleman s (1969) interpretation of westward migration of the Jamuna River. The large meander loops (B-6 of Coleman, 1969) may not have been formed by the Jamuna; they may relate to an earlier river, or an earlier phase of the Jenni with a higher discharge. The Jenni appears to have been decreasing in importance following the Brahmaputra avulsion (Figure 2.3D) and might have been a more significant distributary channel with larger meander loops before the avulsion. Today, Chilmari is still on the west (right) bank, but Dewanganj is on the east (left) bank. As such, it seems most likely that the avulsion node was between these two towns, with the Brahmaputra switching into an unnamed tributary of the Monaash and Joobnee Rivers, which are shown on Rennell s map (Rennell, 1776), but do not appear on modern maps. These rivers probably have been eroded by the Jamuna. A Gradual Avulsion The lack of agreement over a flood or tectonic trigger mechanism for the avulsion of the Brahmaputra may indicate that there was no significant trigger. An alternative explanation for the avulsion may be inferred from the maps of Allen (1843), Wilcox (1828) and Rennell (1776) (Figure 2.3), all of which show large mid-channel islands within the Brahmaputra River downstream from Chilmari, close to the avulsion site. It is suggested here that flow divergence around the mid-channel islands, with one channel directed towards the west (right) bank, caused bank erosion, which led to flow diversion into an existing floodplain channel, and which the Brahmaputra exploited and enlarged to form the Jamuna. Allen s map (Allen, 1843) appears to show the early stages of the avulsion, with Chilmari on the west (right) bank of the river, upstream from the junction and Dewanganj on the east (left) bank, downstream from the junction, as they are today (Figure 2.3 C & D). Allen s map also shows the original Brahmaputra River as the larger river with a braided pattern, whereas the new, as yet unnamed, Jamuna channel has a broadly sinuous course flowing past Sirajganj. The Jamuna, which was initially sinuous, still flows along this course, but has now become braided (Figure 2.3 D). The change in 19

38 channel pattern is interpreted to indicate a gradual transfer of flow from the Old Brahmaputra to the Jamuna, with the Jamuna changing from sinuous to braided as discharge increased. This supports the report of Hirst, 1916 that the changes took place gradually between 1720 and As the Jamuna has expanded, it has removed the evidence of the earlier channel system. Borehole evidence from the Jamuna floodplain, however, indicates that this is not the first time that the Brahmaputra has flowed down this valley (Umitsu, 1993). 2.8 The Old Brahmaputra River The Old Brahmaputra is one of the main distributaries of the Jamuna (Brahmaputra) that distributes part of Jamuna discharge over a large area of North Central region of Bangladesh. The old course of the Brahmaputra River, presently known as the Old Brahmaputra, takes off at Kholabarichar, approximately 10km upstream from Bahadurabad, and follows a south-easterly course via Mymensingh and Toke up to Bhairab Bazar - at the confluence with the Upper Meghna River. The total river length between the off-take and outfall is approximately 225 km, in accordance with the estimate by River Morphology and Research Circle, Bangladesh Water Development Board. The Old Brahmaputra River is at present reduced to a left bank spill channel of the Brahmaputra River and only active during the high stage of the Brahmaputra River. The discharge and sediment transport through the river is dependent on opening of the off-take with the Brahmaputra River. The mouth of the river is not stable (FAP 24,1996c). Moreover, in many occasions it was found that the location of the off-take was not well defined or there might present more than one off-take of the Old Brahmaputra River (EGIS 2001). Gradual silting of the mouth of Old Brahmaputra is creating obstruction of smooth flow from the Brahmaputra to downstream reaches. At present, serious deposition has taken place at the mouth, induced by intense char movement (FAP 3, 1993). The 225km long Old Brahmaputra River is meandering in nature and has limited capacity for passing flood discharges. In the lower areas of the basin flooding situation aggravates due to the tendency of channels to overflow towards the floodplains during the flood period. The seasonal water level variation of the river is 5.3m (at Kholabarichar, 1993). The annual flow of the river varies substantially as found from the data of 1964 to 2008 where the highest recorded discharge of the river at Mymensingh is 4890m 3 /s during monsoon (September 1, 1988) and the lowest recoded discharge at the same station was 20

39 12m 3 /s during the dry period. Rivers that experience such large fluctuation of discharge tend to be unstable and the river may change its morphology in medium to long-term perspective. At Mymensingh, maximum monthly flow is about 2900 m 3 /s. Over the years the annual flows gradually reduce from about 800 m 3 /s in 1965 to m 3 /s in the early nineties, which is equivalent to respectively 4% and 3% of the average Jamuna river discharge. Recorded highest flood level at Mymensingh station is mpwd occurred in 1998.The average sinuosity of the channel is With average bed slope 8.4cm/km near Jamalpur to 5.8cm/km near Toke (FAP 3, 1993). The average grain size of the river varies between 0.005mm to 0.348mm. The tributary, Jinjirum, coming from across the Indo-Bangladesh border joins the Old- Brahmaputra River very close to its off-take and act as the main source of dry season flow of the river Old Brahmaputra. Table 2.1: The Characteristics of the Old Brahmaputra (Source: FAP24 (1996) and IWM Database) Parameters Length from offtake to outfall at Bhairab Bazar Value 225km Discharge (Q at Mymensingh) 12~4,890m 3 /s Seasonal water level variation (at Kholabarichar, 1993) 5.3m Channel sinuosity(avg.) 1.24 Average Bed slope (upto Jamalpur) Average grain diameter (d 50 ) 8.4 cm/km 0.005mm to 0.348mm 2.9 Study Area The study area covers the 50 km reach of Jamuna River starting from the 25km upstream of the mouth of Old Brahmaputra offtake to 25km downstream of offtake and upto 20 km reach of Old Brahmaputra river. Figure 2.4 shows the study area along the Jamuna River. 21

40 Study area Figure 2.4: Study area along the Jamuna River 22

41 CHAPTER 3 THEORY AND METHODOLOGY 3.1 General Offtakes are the distributaries usually bifurcated from the major rivers. For instance, Gorai, Chandana, Hisna, Mathabhanga are some of the main distributaries generated from Ganges River; Dhaleswari, Old Brahmaputra are some of the main distributaries of Jamuna River. It is the most uncertain part of the river. The purpose of this chapter is to give a brief description of the related theories regarding offtake and offtake management through dredging or incorporating different river training structures. Moreover, various mathematical equations and formulas of MIKE21C are also explained in the following section. 3.2 Offtake Off-take of the river is the most dynamic part of the river. Often the location of the offtake shifted double from one place to the other. At the side of offtake, river course should be stable and high. The river should also flow with an average velocity and should have normal width. The variation of water level should also be low. In case the offtake has to be located on the outer side of the curve (i.e. the concave side) so that there is no problem of silting in head reaches. 3.3 Offtake Management The river-bifurcation point namely the Offtake is vital for the sustainability of a river. The sustainability also depends on the conveyance characteristics of the river-reach as a whole. Channel offtakes, which are more permanent divisions of flow, are key elements of the river morphology in Bangladesh. The dynamics of such offtakes have undoubtedly been of great significance during channel avulsions, such as: 1) the occupation of the current Jamuna channel and abandonment of the Old Brahmaputra 2) the division of flow from the Jamuna into the Dhaleshwari that is near the Bangabandhu Bridge and 3) the current concern regarding the flow reduction and infilling of the Gorai channel which is an offtake from the Ganges just upstream of the Jamuna-Ganges confluence. 23

42 Among those offtakes, extensive study has been carried out on the Gorai River such as, Gorai River Restoration Project (Phase-I and II) and Gorai Offtake Management. Those projects were conducted in the year of 1999 to 2001 and 2010 to 2012, respectively. Offtake management can be feasible through dredging along the offtake or by implementing certain river training structures on the mouth of offtake or might be a combination of both dredging and river training structures. Gorai River Restoration Project, GRRP-I and GRRP-II have been conducted by executing the dredging along the river from its mouth to 30 km downstream of the river (Figure 3.1a). On the other hand, river training structures have been proposed in the Gorai Offtake Management Project, shown in Figure 3.1b to increase the flow through Gorai. Similarly, to augment the flow of Dhaleswari and Buriganga River, a regulator has been proposed at the mouth of Dhaleswari River, shown in Figure 3.2. Dredging along the channel Revetment Flow Divider a b Figure 3.1: a) Dredging along the Gorai River on GRRP-II project and b) River training structures at the mouth of offtake on Gorai Offtake Management Project (Source: IWM, 2011a, DHV-Haskoning). 24

43 Figure 3.2: Proposed regulator and guide bunds at the mouth of Dhaleswari offtake (Source: IWM, 2011c) 3.4 Dredging Dredging is an excavation activity or operation usually carried out at least partly underwater, in shallow seas or fresh water areas with the purpose of gathering up bottom sediments and disposing of them at a different location, mostly to keep waterways navigable as well as to create harbors, channels, locks, docks and berths etc. The material removed during dredging can vary greatly and can be any combination of rocks, clays, silts or sands (Bray, 1979). A dredge is a device for scraping or sucking the seabed, used for dredging. A dredger is a ship or boat equipped with a dredge. Dredging takes place to: maintain the depth in existing ports, harbors and channels to provide ready and safe passage for commercial and recreational vessels. create new or deeper access or berths for vessels. This may mean the deepening and widening of channels and anchorages as well as the excavation of basins and marinas from areas of previously dry land. 25

44 provide material for specific purposes, e.g. beaches in coastal areas subject to erosion are sometimes re-nourished with sand dredged from other areas. bypass an artificial structure, such as a breakwater, that is an obstacle to the normal pattern of sediment movement along the coast. Hence, dredging is mainly of two types Capital dredging and Maintenance dredging. Capital Dredging The formation of a new bed configuration by dredging, whether it is stable or not, is known as capital dredging. It is carried out to create a new harbor, berth or waterway, or to deepen existing facilities in order to allow larger ships access. Maintenance Dredging The dredging work associated with recurrent or periodic removal of sediments to maintain a desired bed configuration, is known as maintenance dredging. It is carried out to deepen or maintain navigable waterways or channels which are threatened to become silted with the passage of time, due to deposited sand and mud, possibly making them too shallow for navigation. It only involves the removal of recent unconsolidated sediments. 3.5 River Training Structures River training structures are manmade structures designed and constructed in a river reach to modify the hydraulic flow and sediment response of a river. Some examples of manmade river training structures are guide bund, groyne, bandalling, hardpoint, bank revetments etc. This structures work by using the river s energy to move sediment out of the navigation channel, reducing the need for dredging. 26

45 3.6 Morphological Variables Rivers being a very dynamic systems are one of the most sensitive and dynamic components of the physical landscape and are influenced by a variety of interconnected factors. Each river responds to a number of independent inputs, the river characteristics being the resulting dependent outputs. Lane (1957) established the major factors affecting alluvial stream channel forms. The independent variables of the river catchment are determined by the climate and the geology of the basin, but not directly; also the vegetation, the weathering process and human interference play a role. The dependent variables in a particular river reach are the river characteristics which comprises of both river morphology and hydraulics Independent Variables Factors result in variables like discharge, sediment volumes, sediment characteristics and valley slope. These factors are: Climate: precipitation, temperature, evaporation humidity Hydrology: water levels and discharge versus major river shifts Topography: relief versus sediment concentrations Geology: tectonics, faults, soil types and sedimentary yields Base-level changes: sea level rise versus sedimentary structure Human interference: deforestation, sediment transport The climatic factor is an important geo morphological control of a river system. It governs the amount of precipitation and consequently the surface runoff and river discharge to determine the extent and intensity of basin denudation and indirectly the sediment yield. Hydrological factors like water level and discharge important parameters which may change the characteristics of river behavior. River basin areas of various topographies such as hilly, mountainous and coastal relief offer different sediment yield in the form of sediment amount, sediment size and shape etc. Geology and climate do not determine the discharges and the sediment transport rates in a river system directly. There are the role of the vegetation and the weathering process. The geology in combination with the climate determines the rate of weathering and this again has a significant influence on the growth of the vegetative cover. However, each independent variable is different for different river 27

46 reaches; e.g. a more downstream reach usually has to carry more sediment. Hence for all independent variables at the river reach level it holds that they are a function of the longitudinal coordinate. Otherwise, the sediments transported in a river are distinguished as far as their origin and their mode of transport is concerned. In line with this the total volume of sediment that has to be transported is divided in bed material and wash load Dependent Variables The dependent variables are classified into primary dependent variables and secondary dependent variables as follows: Primary Dependent Variables The primary dependent variables are the combination of morphological characteristics and hydraulic characteristics of a river reach. The offtakes and confluences are nodes of a river system and considered to belong to the primary dependent characteristics. The primary dependent variables are: Dependent morphological characteristics: - number of channels, n - average width of channel(s) B b, where the index b stands for bankfull conditions - sinuosity p of the river channel(s) - typical wave length of the curved channels L - total width B t, occupied by the river channel(s) - bankfull depth h b of the river channel(s) - slope of the river i b. Dependent hydraulic characteristics: - average water depth in the main channel; - wet width either in the main channel or including the floodplain; - inundation depth of the floodplain; - the roughness coefficient of the main channel and of the floodplain; - the average velocity. 28

47 Secondary Dependent Variables The secondary dependent variables in a particular river reach are more detailed, local and dynamic in nature. These are: bars bifurcations scour holes bank erosion cutoffs The dependent variables are linked to the independent variables through relationships of river mechanics. There is no single-valued relationship between the imposed variables and the dependent characteristics. The excess flow strength leads to bed degradation when banks are fixed. This bed degradation increases the water depth and reduces the longitudinal slope until the flow strength reaches a value which matches the transport of the sediment supplied from upstream. When the banks are erodible there are three additional modes of adjustment. Firstly, the river can become wider. Secondly, the longitudinal slope can be reduced by increasing the sinuosity through the expansion of meander loops. Thirdly, new channels can be carved in the floodplain at discharges above bank full, which leads to an anabranches system. The vertical responses of rivers with fixed banks are rather unambiguous, but the distribution of horizontal response over the three modes of adjustment depends on the initial state of the river and the time series of discharge and sediment supply during the transition to new equilibrium. 3.7 Sediment Transport Sediment transport aspects have been studied for the major rivers of Bangladesh. Studies were conducted by Bari (1978), Alam and Hossain (1988), Sultana (1989) and Hossain (1989). These studies were mainly concentrated to assess the applicability of few well known sediment transport equations against the data of Ganges and Jamuna. Some of these investigators however, while adjudging the efficiency of various transport equations also computed the total annual sediment flow. 29

48 3.8 Sediment Transport Equations Mainly stage, discharge, velocity, water surface and bed slope, shear stress, mean particle diameter and stream power control the sediment movement of any channel. Only a single equation cannot incorporate all these variables and predicts the sediment load. For this reason, different equations have been put forward on the basis of different independent variables. Out of many available empirical formulas the following three well known equations have been selected for the present study Engelund-Hansen Equation (1967) Engelund-Hansen s (1967) equation is based on the shear stress approach. In developing the equation Engelund-Hansen's relied on data from experiment in a specific series of test in a large flume. The sediment used in this flume had mean diameter of 0.19mm, 0.27 mm, 0.45 mm and 0.93 mm. The equation can be written as: g s = 0.05γ s V 2 D 50 /g γ s γ 1 τ o γ s γ D 50 3/2 (3.1) where, g s = Sediment transport per unit time per unit width γ s = Specific weight of sediment particles γ = Specific weight of water τ o = Bed shear stress D 50 = Median diameter of bed material V = Average flow velocity g = Acceleration due to gravity This equation is dimensionally homogeneous and any consistent set of units can be used. 30

49 3.8.2 Yang's Equation (1976) Yang (1976) proposed a sediment transport formula based on the concept of unit stream power, which can be utilized for the prediction of total bed material concentration transported in sand bed flumes and rivers. The formula (Equation.3.2) is as follows: log C, = log ωd 50 /υ log u /ω log ωd 50 /υ log u /ω log VS ω V cr S ω (3.2) The value V cr ω is given by: V cr ω = 2.5 u D log u , 0 < D υ 0.06 υ < 70 V cr ω = 2.05, 70 < u D υ where C t = Total sediment concentration in parts per million (ppm) by weight D 50 = Median diameter of bed material V = Average flow velocity D = Median sieve diameter U * = Shear velocity ν = Kinematic viscosity V cr = Critical average flow velocity ω = Fall velocity This equation is dimensionally homogeneous and any consistent set of units can be used Van Rijn's Equation (1984) A simplified method was given by Van Rijn for calculating suspended sediment transport. This method is based on computer computations in combination with a roughness 31

50 predictor. Using regression analysis, the computational results for a depth range of 1 to 20 m, a velocity range of 0.5 to 2.5 m/s and a particle range of 100 to 2000 µm were represented by a simple power function, as follows: q s,c uh = u u cr d s 1 g d 50 h 1 D 0.6 (3.3) in which q s,c = volumetric suspended load transport (m 2 /s) u cr = critical depth-averaged velocity according to Shields h = water depth u = depth-averaged velocity Equation 3.3 only requires u, h and d 50 as input data and can be used to get a first estimate of the suspended load transport. It is assumed that the instantaneous bed-load transport rate is related to the instantaneous T parameter, as follows q b = 0.1 s g 0.5 d D T m (3.4) In which T m = τ b, τ b,cr /τ b,cr = instantaneous shear stress parameter τ b, = μτ b = instantaneous effective bed shear stress τ b,cr = instantaneous critical bed shear stress D = dimensionless particle parameter When the bed load transport and the suspended load transport are known, the total load transport of bed material can be determined by summation of Equation 3.3 and 3.4. q t = q b + q s (3.5) 3.9 Mathematical Modelling Many modelling packages are available which can simulate hydrodynamic as well as morphological characteristics of a river. Moreover, modelling packages are so advances that it can now simulate intervention phenomenon on river through control structure. 32

51 There are few softwares available that can simulate the flow through control structures and sediment transport with bed level changes in river systems. There are mainly HEC-RAS, MIKE 11, MIKE 21C, Delft 3D, SMS etc MIKE 21C MIKE 21C is a special module of the MIKE 21 software package based on a curvilinear (boundary-fitted) grid, which makes it suitable for detailed hydrodynamic and morphological simulations of rivers and channels, where an accurate description of banklines is required. The numerical grid is created by means of a user-friendly grid generator. Areas of special interest can be resolved using a higher density of grid lines at these locations. The MIKE21C is particular suited for river morphological studies and includes modules to describe: Flow hydrodynamics, i.e. water levels and flow velocities are computed over a curvilinear or a rectangular computational grid covering the study area by solving the vertically integrated St. Venant equations of continuity and conservation of momentum. Helical flow (secondary currents) which is developing in channel bends due to curved streamlines is included. The time and space lag in the development of the helical flow is also described. Sediment transport, based on various model types (e.g. Van Rijn, Meyer-Peter & Müller, Engelund-Hansen, Engelund-Fredsoe, Yang, or user-defined empirical formulas). The effect of helical flow, gravity on a sloping river bed, shapes of velocity and concentration profiles are taken into account in separate bed load and suspended load sub-models based on the theories by Galappatti. Graded sediment descriptions can be applied as well by defining a number of different sediment fractions, which are treated separately by the sediment transport module Alluvial resistance due to bed material and bed forms (calculation of skin friction from the grain sizes and form drag friction from the bed forms). Scour and Deposition: Large-scale movement of bed material is computed. The continuity equation for sediment is solved for determining changes in bed elevations at each grid cell at every time-step. The effect of supply limited sediment layers can be incorporated as well. This is used for simulating for 33

52 instance downstream migration of fine material on a coarse riverbed, or for representing non-erodible (protected) riverbed areas in the modelling domain Governing Flow Equation of MIKE 21C By introducing the simplification of three dimensional Navier Stokes theorems the three dimensional flow pattern of river can be reduced to two-dimensional equations of conservations of mass and conservation of momentum in the two horizontal directions. Three directional (secondary flow) effects are maintained in the depth averaged model by introducing helical flow component in the flow equation. Flow model is valid for the shallow, gently varying topography and mildly curved wide river channels with small Froude numbers. The flow equations solved in the curvilinear hydrodynamic model of MIKE 21C are: p t 2 p pq pq s h 2 n h hrn 2 p q hr H g p gh 2 s C 2 s h 2 p 2 q 2 RHS (3.6) q t 2 pq q pq s h n h 2 hrs 2 q p hr H g q gh 2 n C 2 n h 2 p 2 q 2 RHS (3.7) H t p q s n q R p s R n 0 (3.8) where s,n = Coordinates in the curvilinear coordinate system p,q = Mass fluxes in the s and n direction H = Water Level H = Water Depth G = Acceleration of Gravity C = Chezy roughness coefficient Rs,Rn = Radius of curvature of s and n line RHS = the right hand side in the force balance, which contains Reynolds stresses, Coriolis force and atmospheric pressure. 34

53 These equations are solved by an implicit finite differential technique with variables (water flux density p and Q in two horizontal directions and water depth H) defined on a space staggered computational grid Morphological Model of MIKE 21C A morphological model of MIKE 21C is a combined sediment transport and hydrodynamic model (DHI 2009). The hydrodynamic flow field is updated continuously according to changes in bed bathymetry. Morphological model is a uncoupled model. In uncoupled model, the solution of hydrodynamics is solved at a certain time step prior to solution of the sediment transport equations. Subsequently, a new bed level is computed and the hydrodynamic model proceeds with the next time step. Sediment Continuity Equation Following calculation of sediment transport of bed material (bed load and suspended load), the bed level change can be computed from the equation: z S S 1 (3.9) t x y x y n. Se Where S x = Total sediment transport in x-direction S y = Total sediment transport in y-direction N = Bed porosity Z = Bed level t = Time x,y = Cartesian coordinated system ΔS e = Lateral sediment supply from bank erosion MIKE 21C morphological model solves the sediment continuity equation implicitly Methodology Modeling of any physical phenomenon is an iterative development of a process. Model refinements are based on the availability and quality of data, hydrological understanding and scopes of the project. The general approach that has been followed in the current study 35

54 can be summarized in the flowchart given in Figure 3.3. A brief description of the methodology and approaches are provided in this section to achieve the study objectives. It includes - Data Collection Data Checking & Analysis Input Data Preparation Model Setup Model Calibration & Validation No Yes Different Option Simulation Data Extraction Comparison of Different Outputs Figure 3.3: Flow chart of methodology applied in the study 36

55 Collection of data from various sources Historical water level and discharge data of Jamuna and Old Brahmaputra River have been collected from IWM and BWDB. Recent data on river cross section, sediment transport within the study area and satellite images have also been collected from Institute of Water Modelling (IWM). All the collected data were then compiled, processed and analyzed to the required extent for use in the model development and application. The detail description of data collection is discussed in next chapter Hydrological and morphological analysis of the collected data Based on the historical water level and discharge data, an extensive hydrological analysis have been done to understand the hydrological condition of the study area. Morphological analysis such as planform analysis, bankline shifting, shifting of mouth of offtake etc have been carried out by using different satellite images (see Chapter 4). Moreover, cross section analysis at different reaches of Old Brahmaputra River and longitudinal bed profile has been also conducted to observe the bed level condition and thalweg of the river A 2D morphological model setup of the study area Morphological assessment has been carried out with the aid of two-dimensional (2D) morphological model to resolve the governing physical processes to a sufficient degree of detail. A 2D model has been setup for 50km reaches of Jamuna including 20km reach of Old Brahmaputra River from the mouth of offtake towards downstream. The model is based on the bathymetric data of post-monsoon The model is calibrated for both hydrodynamic and morphological conditions for the hydrological monsoon of The model is also validated for hydrological monsoon of 2012 using the same calibrated parameters Selection of different flood events by frequency analysis After calibration and validation, this model has been finally used for different option simulations for different flood event i.e. high, medium and low flood events. These hydrological flood events have been selected by frequency analysis using HYMOS software tool. For frequency analysis, historical discharge data of Jamuna River were used and finally three different hydrological years have been selected. 37

56 The analysis has been made with the tool HYMOS, an advanced tool for frequency analysis for hydrological events developed by Delft Hydraulics of the Netherlands. Three extreme value methods namely Gumbel, Log Pearson Three Parameters and Log Normal Three Parameters have been used for the analysis. Table 3.1 shows the discharge of different return periods for different distribution methods. Table 3.1: Peak flows of Jamuna River at Bahadurabad for selected return periods using different distribution Probability Distribution Return Period in year Gumbel (EV1) Log Pearson Log Normal Comparison among the three distribution methods used for frequency analysis has been made based on the probability plots and goodness of fit tests (Chi square test and K-S test). The following probability plots in Figure 3.4 to Figure 3.6 shows the probability values using Gumbel, Log Pearson three parameters and Log Normal distribution respectively. 38

57 Figure 3.4: Probability plots for Gumbel distribution Figure 3.5: Probability plots for Log Pearson three parameters distribution 39

58 Figure 3.6: Probability plots for Log Normal distribution To select the best fit among the three distributions Chi square test and Kolmogprov- Smirnov (K-S) test has been made. Table 3.2: Goodness of fit test among three distributions Gumbel Log-Pearson Log-Normal Chi Square Variate K-S Test From Table 3.2 it is seen that chi square is minimum for Log Normal distribution. Again It is know that if Chi square is equal to zero the theoretical and observed frequencies agree exactly; while if chi square is greater than zero they do not agree exactly and the larger the value of Chi square the greater is discrepancy between the observed and expected frequencies. 40

59 Again, Table 3.3 presents the probability of exceedence of Chi square for the mentioned three distributions. It is found here that the probability of exceedence of Chi square value is found minimum for Log-Normal distribution. For Log Pearson distribution it is slightly higher and for Gumbel it is much higher compared to Log Normal distribution. Table 3.3: Probability of exceedence of Chi square Distribution Item Gumbel Log-Normal Log-Pearson Prob. of exceedance of Chi Square Among the Gumbel, Log Pearson three parameters and Log Normal distribution the chi square value is found minimum for the Log Normal distribution. The probability of exceedence of Chi square value is found minimum for Log-Normal distribution and it is slightly higher for Log Pearson distribution. Considering the two goodness of fit test the Log Normal distribution has been selected to estimate the flood discharges for the selected return periods. Table 3.4 gives the peak flow according to Log Normal distribution for two return periods and corresponding year that experienced equivalent flood discharges. Table 3.4: Peak flow according to Log Normal distribution Return period Peak flow according to Log Normal, m 3 /s Observed peak flow, m 3 /s ,136 68, ,03,059 1,05, Corresponding year The hydrological year of 1998 and 2005 represents the high and medium flood events respectively. The low flood event has been selected by analyzing the historical minimum discharge for the past 55 years. The corresponding year of low flood event is Flood Routing The discharge data is collected from the station named Bahadurabad which is situated at 30 km downstream from the u/s boundary of the model domain. So it is required to flood rout for getting discharge at u/s boundary of model domain. On this regard help was taken 41

60 from the Flood Management Division (FMG) of Institute of Water Modelling. They provide the discharge boundary by using one-dimensional network model, MIKE Simulation of different options to observe the sustainability of the Old Brahmaputra offtake The options are selected in such a way that the maximum amount of flow should be passed through the Old Brahmaputra River. In this regard, four different options have been chosen for the simulation. Assessment of flow (at high, medium and low flow of Jamuna) for existing and design river bed condition of offtake river has been conducted for the following management options. a) river training works b) dredging works at the vicinity of offtake c) dredging and river training at the vicinity of offtake 42

61 CHAPTER 4 DATA COLLECTION AND ANALYSIS 4.1 General In order to develop the mathematical morphological model, various kinds of data of recent and previous years have been collected and compiled. These data also form the basis for further analysis and interpretation of the model results leading to accurate assessment of hydro-morphological condition of the study area. According to the modelling requirements, a significant amount of data includes water level, discharge, cross-section, sediment, satellite images and other relevant information like bank characteristics etc have been collected. This chapter describes a brief discussion about the collected data. 4.2 Data Collection Quality data are prerequisite for reliable model setup, model results and to have understanding on the existing physical processes. To determine the present hydraulic and morphological conditions and to develop a mathematical model of Jamuna incorporating the Old Brahmaputra offtake, various data have been collected from different sources. A brief description of data is given below: Water Level Water level data at different locations are required for calculating the river slope, defining water level of different flood events as well as providing boundary of two-dimensional model and to calibrate the model. Therefore, historical water level data at Bahadurabad, Kholabarirchar station of Jamuna River and Jamalpur, Mymensingh station of Old Brahmaputra River have been collected and analyzed to get an idea about the amount of water is flowing at this location. The duration of collected data are listed in Table It is worth mentioning here that, there is no discharge or water level gauge station of Old Brahmaputra within the study area. So available water level data of the gauge stations situated at the downstream of this river have been collected and presented. For calibration purpose, water level of Horichandi at Old Brahmaputra River during 2011 is used. 43

62 Table 4.1: Available water level data for the study area Station Name Bahadurabad Station ID SW46.9L Kholabarichar SW46.7L Jamalpur Mymensingh SW225 SW228.5 Horichandi - River Name Easting (BTM) Northing (BTM) Duration Source BWDB BWDB BWDB BWDB IWM Discharge Discharge data is needed to investigate the hydrological characteristics of the river and to provide boundary for the two-dimensional morphological model. Available discharge data for the Jamuna and Old Brahmaputra River is listed below (Table 4.2): Table 4.2: Available discharge data for the study area Station Name Station ID Bahadurabad SW46.9L Mymensingh SW228.5 River Name Brahmaputra- Jamuna Brahmaputra- Jamuna Old Brahmaputra Old Brahmaputra Old Brahmaputra Brahmaputra- Jamuna Old Brahmaputra Easting (BTM) Northing (BTM) Duration Source BWDB BWDB 44

63 Figure 4.1: Location of discharge and water level station near the study area Cross Sectional Bed Level / Bathymetry Data The post-monsoon 2011 bathymetry data of Jamuna River and Old Brahmaputra River have been collected from IWM. These surveyed bathymetry data covers the reach of Jamuna River from 25 km upstream and downstream of the mouth of Old Brahmaputra offtake and 20km of the Old Brahmaputra River from its mouth. Figure 4.2 shows the transect lines of bathymetry data within this reach. The spacing between the transect line is about 1km for Jamuna River and 500 m for Old Brahmaputra River. 45

64 Figure 4.2: Transect line of bathymetric survey for the study area in Jamuna River 46

65 4.2.4 Satellite Image The satellite images of Jamuna River of 1973, 1984, 1997 and 2011 have been collected from IWM for planform analysis and to delineate the bank lines for assessment of bankline shifting Bankline Data Bankline data of Jamuna and Old Brahmaputra River have been collected for September 2011 (Figure 4.2) to generate the computational grid of the model Sediment Data Sediment and other pertinent data were used in the model as input. The collected sediment data of the Jamuna includes suspended sediment and bed sample data. Table 4.3: Available sediment data for the study area Suspended Sediment Data Bed Sample Data Station Data Period Source Bahadurabad , , & BWDB and FAP 24 Year of Sampling Source FAP 24 Mymensingh BWDB FAP Data Analysis Hydrological Analysis Water Level Yearly maximum, minimum and average water level data of Jamuna River at Bahadurabad station has been plotted and shown in Figure 4.3. The available water level hydrograph at Kholabarichar station is shown in Figure 4.4. The yearly maximum and minimum water level at Jamalpur station of the Old Brahmaputra River are also shown in Figure 4.5. By analyzing the collected historical water level data it is found that the water level is in declining trend in Old Brahmaputra River since 1950 but in Jamuna the water level did not show any trend. 47

66 Water level, mpwd Water Level, mpwd Maximum Average Minimum Figure 4.3: Maximum, minimum and average water level at Bahadurabad station of Jamuna River Oct-95 Jul-98 Apr-01 Jan-04 Oct-06 Jul-09 Apr-12 Dec-14 Figure 4.4: Water level hydrograph at Kholabarichar station 48

67 Water Level, mpwd Maximum WL Minimum WL Year Figure 4.5: Maximum and minimum water level of Old Brahmaputra River at Jamalpur station Discharge Historical discharge data of Jamuna River at Bahadurabad station that have been collected from BWDB are analysed and plotted in Figure 4.6. From this figure, the maximum and minimum discharges have been detected for the year 1998 and 2001 respectively. The historical discharge data of Old Brahmaputra are available at Mymensingh station since Figure 4.7 presents the yearly maximum discharge histogram of the Old Brahmaputra at Mymensingh station. It is observed that peak discharge of Old Brahmaputra River is decreasing from 1964 to Peak discharge was highest in 1988 and lowest in The percentage of maximum flow through Old Brahmaputra also shows a decreasing trend, shown in Figure 4.8. The percentage of maximum flow through the Old Brahmaputra River has reduced from 4.48 % to 1.74 % during 1964 to This reduction of flow reveals the siltation problem at the mouth. From Figure 4.9, it is seen that percentage of flow was decreasing year to year during dry period. In , the percentage of flow was about 5% in October, whereas it was less than 1% during

68 Yearly maximum discharge, m 3 /s Maximum discharge, m 3 /s Minimum discharge, m 3 /s Maximum Minimum Year Figure 4.6: Maximum and minimum discharge of Jamuna River at Bahadurabad station Figure 4.7: Yearly maximum discharge histogram of Old Brahmaputra River at Mymensingh station 50

69 % of flow diversion % of maximum flow through Old Brahmaputra Year Figure 4.8: Percentage of maximum flow through Old Brahmaputra River Oct 31-Oct 30-Nov 30-Dec 29-Jan 28-Feb 30-Mar Figure 4.9: Percentage of flow diversion during dry season in different period 51

70 4.3.2 Morphological Analysis Morphological analyses at the study area have been carried out using available satellite images and cross-sectional data. The planform analysis, bankline shifting, dynamics of offtake of Old Brahmaputra River, sediment transport analysis, cross section and longitudinal bed profile as well as historical changes of cross sections at different reaches of Old Brahmaputra River have been discussed below. Planform Analysis The planform analysis illustrated the rapid morphological changes around the offtake area. Hence, a representative satellite images from every decade has been shown in Figure It is seen from the figure that in 1973 the main channel of Jamuna River near Old Brahmaputra reach touched along the left bank and at that time the mouth of the offtake was apparent. Moreover, Jinjiram River was also contributing some flow to the Old Brahmaputra River during that period. In 1984, the main channel of Jamuna near offtake area was getting anabranched, though the left anabranch touched the mouth of offtake. Satellite image showed that Jinjiram River still contribute certain amount of flow to the Old Brahmaputra River. In 1997, the char near the offtake (shown in 1984 image) became attached with the left bank and a very narrow channel shared some flow to the Old Brahmaputra River from Jamuna. The major portion of flow was passed from the Jinjiram River, shown in Figure The recent image of 2011 shows that the attached char along the offtake separated from the left bank and a narrow channel is flowing near the offtake which distribute certain amount of flow to the offtake. 52

71 Figure 4.10: Satellite images in study area showing the offtake of the Old Brahmaputra River in different period. Bankline Shifting A number of dry seasons Landsat images of 1973, 1984, 1997 and 2011 were used to delineate the banklines of the Jamuna and Old Brahmaputra River. The bankline is defined as a line, which separates the floodplain from the river. Water and sand are considered as river and bare land and vegetated land are considered as floodplain. Figure 4.11 shows the bankline of different years. It is seen from the figure that the bankline of Old Brahmaputra River is shifting drastically from year to year and it didn t follow any trend. 53

72 Figure 4.11: Location of the offtake of the Old Brahmaputra River at different time period. Shifting of Mouth of Offtake Offtake of the river is the most dynamic and uncertain part of the river. Often the location of the offtake shifted double from one place to the other. Moreover, in many occasions it was found that the location of the off-take was not well defined or there might present more than one opening of a single river. The location of offtake shifted along the left bank of the Brahmaputra-Jamuna River within a 15 km stretches of the river ( ), shown in Figure

73 Figure 4.12: Changes of the offtake of Old Brahmaputra River at different time period. 55

74 Sediment discharge (ton/day) log scale Sediment Transport Analysis Sediment rating curve of Jamuna and Old Brahmaputra have been prepared from the available historical sediment data of Bahadurabad station and Mymensingh station. The sediment discharge and corresponding flow discharge of Jamuna and Old Brahmaputra have been plotted on log-log paper, shown in Figure 4.13 and Figure 4.14 respectively. From trend line analysis the following relationships have been found. For Jamuna River the equation is as follows: Q s =0.133 Q (4.1) and for Old Brahmaputra River the equation is as follows: Q s =0.052 Q (4.2) Where, Q s and Q is the sediment discharge and water discharge Data from the period No of data= y = 0.133x R² = Flow Discharge (m 3 /s) log scale Figure 4.13: Sediment rating curve of Jamuna River at Bahadurabad station during

75 Sediment discharge (ton/day) log scale Data from period No of Data = y = 0.052x R² = Discharge (m 3 /s) log scale Figure 4.14: Sediment rating curve of Old Brahmaputra at Mymensingh station during Sediment transport analysis has been done for Jamuna River by using the suspended sediment data from the period 1968 to For the Jamuna River, the combined yearly sediment transport is found 1200 million tons per year on the average (FAP 24). Figure 4.15 shows that the average sediment transport in Jamuna River is nearly 35000kg/s (~3 million tons per day). The maximum sediment transport, however, may be as high as kg/s (~10 million tons per day). Similarly for Old Brahmaputra River, collected sediment data from 1993 to 1996 have been used for sediment transport analysis. Figure 4.16 represents the yearly sediment transport hydrograph for Old Brahmaputra River. 57

76 Sediment transport, kg/s Sediment transport, (kg/s) /09/ /07/ /05/ /03/ /01/ /12/1999 Date Figure 4.15: Sediment transport of Jamuna River during Oct-92 Aug-93 Jun-94 Apr-95 Feb-96 Dec-96 Figure 4.16: Sediment transport of Old Brahmaputra River during

77 Bed level, mpwd Cross Section and Longitudinal Bed Profile Analysis To observe the bed level condition and thalweg of the river, available cross section data at different reaches of Old Brahmaputra River are also plotted and shown in APPENDIX - A. Figure A-1 shows the transect lines of cross section at Old Brahmaputra and Jamuna River. The historical cross sections are shown in APPENDIX A. The data covers the year from 1967 to A single representative cross section of 2011 is shown in Figure Left Bank Right Bank Distance from left bank, m Figure 4.17: An observed cross section of Old Brahmaputra River during The longitudinal profile of mean bed level of Old Brahmaputra River during 1967 to 1998 is shown in Figure It is seen from the figure that in the first 20km reach of the river, the mean bed level was found 10 mpwd, but at present the mean bed level is increased due to huge sedimentation. After 20 km reach, the bed level rise again and follow the bed slope around 4.5cm/km. The bed level is approximately 13 to 14mPWD in 2011 along the offtake. 59

78 Bed level, mpwd Average Minimum Maximum Distance from the offtake, km Figure 4.18: The observed maximum, average and minimum bed level from the offtake to the d/s of Old Brahmaputra River during

79 CHAPTER 5 APPLICATION OF MATHEMATICAL MODEL 5.1 General Mathematical modeling is an advance technology in engineering practice for predicting variables which is tedious to find in hand calculation. To evaluate such variables, a twodimensional morphological model, MIKE 21C has been used. Using this modeling software a mathematical model of 50km reach of Jamuna River including 20km reach of Old Brahmaputra River has been set-up. The various key steps during processing the model are described below. 5.2 Model Setup The model setup includes the generation of computational grids, the preparation of the bathymetry, boundary conditions and selection of parameters. Followed by the set up, the model is calibrated with the tuning of hydrodynamic parameters (like Chezy s bed roughness, eddy viscosity etc.), the parameters in the sediment transport magnitude formula etc. With all these tasks, the model gets ready for different sensitivity and applications runs. Detail methodology for model set-up, calibration and option simulations are presented in the following sections Grid Generation The computational grid has been generated using the bankline of Jamuna and Old Brahmaputra River extracted from satellite image of The grid is curvilinear and therefore represents the curved banklines more accurately than a rectilinear one and can follow the flow lines along the river. The resolution of the grid is 350 in the flow direction and 155 in the transverse direction. The computational grid is shown in Figure 5.1. The grid has land boundaries as well. The model simulates the hydrodynamic and morphological parameters in every computational grid point. 61

80 Mouth of offtake Figure 5.1: Computational grid cell in the study area Bathymetry Generation The bathymetry of the model has been prepared based on data from the IWM bathymetric survey carried out during post-monsoon The surveyed bathymetry of the model is shown in Figure 5.2. To simulate the morphological model with different hydrological flood events, the bathymetry data has been superimposed on the curvilinear grid (Figure 5.1). It means every grid cell contains river cross-section data and after model simulation 62

81 every grid cell produces hydrological and hydraulic parameters like, water level, discharge, water depth, velocity, bed scour and others. Flow direction Mouth of Old Brahmaputra River Bed level, mpwd Figure 5.2: Model bathymetry of the study area for post-monsoon

82 Discharge, cumec Boundary Condition To simulate the morphological model, it is necessary to use the following hydrological data at the model boundaries. Discharge at the upstream boundary Water level at the downstream boundary The application run of the model has been done for high, medium and low flood events. The 1 in 100 year return period and 1 in 2.33 year return period flood events represents the high and medium flood events. The corresponding year of high and medium flood events are 1998 and 2005, which have been found by doing frequency analysis (Figure 5.3). The low flood event has been selected by analyzing the historical minimum discharge for the past 55 years. The corresponding year of low flood event is 2001 (Figure 5.3). The upstream and downstream boundaries of different flood events have been shown in Figure 5.4 to Figure Nov 5-Jan 24-Feb 15-Apr 4-Jun 24-Jul 12-Sep 1-Nov 21-Dec 9-Feb Figure 5.3: Discharge hydrograph of Jamuna River at Bahadurabad for the years 1998, 2001 and

83 Discharge, cumec WL, mpwd Discharge, cumec WL, mpwd Q Bahadurabad d/s WL boundary of Jamuna d/s WL boundary of Old Brahmaputra May-98 Jun-98 Jul-98 Aug-98 Sep-98 Oct-98 Figure 5.4: Discharge and water level boundary for the high flood event (1998) Q Bahadurabad d/s WL boundary of Jamuna d/s WL boundary of Old Brahmaputra May-05 Jun-05 Jul-05 Aug-05 Sep-05 Oct-05 Nov-05 Figure 5.5: Discharge and water level boundary for the medium flood event (2005). 65

84 Discharge, cumec WL, mpwd Q Bahadurabad d/s WL boundary of Jamuna d/s WL boundary of Old Brahmaputra May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Figure 5.6: Discharge and water level boundary for the low flood event (2001) Model Calibration It is always customary to calibrate and validate the model before making any predictions using the model, so that the reliability of the predictions can be made as much as perfect. In this study, model has been calibrated by using the boundary of 2011 (Figure 5.7). 66

85 Q, cumec WL, mpwd u/s discharge boundary at Jamuna d/s WL boundary at Jamuna d/s WL boundary at Old Brahmaputra Mar 27-Apr 27-May 26-Jun 26-Jul 25-Aug 24-Sep 24-Oct 12 Figure 5.7: Boundary for simulation of 2011 monsoon This section provides brief description of the hydrodynamic and the sediment transport calibration of the morphological model. Hydrodynamic Calibration Hydrodynamic calibration is done with the fixed bed model. Water level calibration has been done at Bahadurabad station in Jamuna River and Horichandi location in Old Brahmaputra River. Due to non-availability of the observed water level data upto 20km downstream of Old Brahmaputra, the 2D model calibration was done by comparing the MIKE 21C simulated data with the MIKE 11 simulated data at Horichandi. The parameters which contribute in adjusting the hydrodynamic misbalance or fine tuning the numerical models in the hydrodynamic field in MIKE21C are mainly roughness (Chezy s C), co-efficient of eddy viscosity (E) where Chezy s C influence in adjusting the water level and co-efficient of eddy viscosity used for the distribution of flow by exchanging lateral momentum of flow. The Chezy value can vary spatially and is also a function of depth. The functional relation can be updated after each computational time 67

86 step. In this model, Chezy value is updated at every time step according to the updated depth from the following formula: n C C o H (m 1/2 /s), where H represents the water depth. The value C o and n are alluvial resistance coefficient and exponent respectively. Numbers of trial simulations were carried to verify the selection of the C and finally C = 40 has been fixed as initial Chezy value and for the selection of the value for alluvial resistance coefficient and the exponent, C o as 15 and the exponent, n as 0.5 were used. So by taking these values, the equation stands in the form of C = 15h 0.5. The maximum limit for Chezy s factor C is considered as 100 and the minimum as 5. The reason for using the limiting value of the Chezy s factor is to control the extreme roughness of the model domain. In all the simulations, the value of eddy viscosity is 3 for Jamuna River and 1 for Old Brahmaputra River. The summary of the parameters used in the model domain are shown in Table 5.1. Table: 5.1 Summary of the parameters used for the model domain General Momemtum Pressure Sediment Chezy C= C o h n Eddy Viscosity Time Time space Time space transport Alluvial m 2 /s Space (s) (s) (s) Equation Resistance C o n *Englund- Hansen Jamuna =3 Old Brahmaputra=1 The model is simulated for a specified period. Since the morphological changes take place primarily during the high discharges, the simulation period is selected so that it can include the main contributing part of the monsoon. On this basis the period has been chosen. So, for representing the major peak of the hydrological year total 60 days were considered starting from 1 st July to the 31 th August for calibration purpose. The hydrodynamic time-step has been set to 10 second for the simulation. Figure 5.8 and Figure 5.9 shows the calibration results at Bahadurabad station in Jamuna River and Horichandi in Old Brahmaputra River. The comparison shows more or less good agreement between the simulations. 68

87 Water level, mpwd Water level, mpwd WL calibration at Bahadurabad Bahadurabad obs WL Simulated WL Jul 19-Jul 3-Aug 18-Aug 2-Sep Figure 5.8: Water level calibration at Bahadurabad station in Jamuna River WL calibration at Horichandi Horichandi obs WL Simulated WL Jul 19-Jul 3-Aug 18-Aug 2-Sep Figure 5.9: Water level calibration at Horichandi in Old Brahmaputra River. 69

88 Sediment Transport Calibration of the Morphological Model The calibration of the morphological model is basically meant in MIKE21C to deal with the sediment transport and it s associated parameters, such as the grain size diameter, the selection of sediment transport formula, the helical flow strength etc. In this study, calibration was done on the basis of the sediment transport rating curve at Bahadurabad station for Jamuna River. For the application and for calibration of the model, twice of sediment transport found from Englund-Hansen formula is applied for better result, shown in Figure The value of grain size diameter is 0.18mm at the mouth of Old Brahmaputra River (FAP 24d). But in this case, the value of d 50 for the whole model domain is given by interpolating the value from 0.19mm to 0.17mm. Figure 5.10: Sediment transport calibration at Bahadurabad station in Jamuna River Model Validation The calibrated model is validated by running for one or more simulations for which measurements are available without changing any tuning parameters. Here, model has been hydro-dynamically and morphologically validated by using the boundary of

89 Discharge, m 3 /s Water level, mpwd Hydro-Dynamic Validation Figure 5.11 and Figure 5.12 show the comparison of observed and simulated water level and discharge at Bahadurabad station. It is seen that the simulated water level is almost matched with the observed one. The simulated discharge is also more or less similar with the observed discharge Observed WL Simulated WL Jun 10-Jul 20-Jul 30-Jul 9-Aug 19-Aug 29-Aug 8-Sep 18-Sep 28-Sep 8-Oct Figure 5.11: Comparison of water level at Bahadurabad station for Observed discharge Simulated discharge Jun 10-Jul 30-Jul 19-Aug 8-Sep 28-Sep 18-Oct Figure 5.12: Comparison of discharge at Bahadurabad station for

90 Simulated Water Level, mpwd To assessing the deviation of the model results from the observation is the presentation of discrepancy ratio. Discrepancy ratio is defined as the ratio between measured and model generated water level or discharge. Ratio from the models along with the observed one have been determined and plotted in Figure 5.13 and Figure Regression equation and values of R 2 have been given in the plot. R 2 is a statistic that will give some information about the goodness of fit of a model. In regression, the R 2 coefficient of determination is a statistical measure of how well the regression line approximates the real data points. An R 2 of 1.0 indicates that the regression line perfectly fits the data. The values of R 2 for the model with observed water level is and it is found in case of discharge y = 1.055x R² = Observed Water Level, mpwd Figure 5.13: Discrepancy ratio for water level comparison 72

91 Simulated discharge, m 3 /s y = 0.951x R² = Observed discharge, m 3 /s Figure 5.14: Discrepancy ratio for discharge comparison Morphological Validation The model has been morphologically validated against the year of The observed cross-section at Bahaduarbad location is compared with the simulated cross-section at the same location. The cross-section is match quite well at the very left channel; the simulated channel depth is well matched with the observed one. There are some variation is also found in the other locations, shown in Figure The sediment transport for 2012 is also compared with the observed sediment data of 2012 and presented in Figure The simulated value is present within the range of observed sediment transport value. 73

92 Sediment tranport, ton/day Bed level, mpwd Simulated Observed Distance from left bank, m Figure 5.15: Comparison of cross-section at Bahadurabad location Obsreved Simulated 2012 Observed 2012 Power (Obsreved ) Discharge, m 3 /s Figure 5.16: Comparison of sediment transport in Jamuna River for validation Model Stability Initially, the model has made stable by simulating hydro-dynamic condition only for a while. The simulation period was about 15 days. After some time the model become stable 74

93 Water depth, m as shown in Figure All data have been extracted after the model become stable. Morphological simulation also carried out after the stability has been reached. Date Figure 5.17: Model stability check for the study area 5.3 Various Option used in Model Simulation As well as with the base condition simulations have been done for various options. Four different options have been employed in this study to suggest the better offtake management for Old Brahmaputra River. The details of options have been described below: Option 1 (Dredging along the left channel of Jamuna River) From recent bathymetry it is seen that the left channel of Jamuna River near the offtake of Old Brahmaputra River has been almost dried up. Therefore, dredging alignment has been fixed along the left channel of Jamuna River by following the gradient slope at that location. Figure 5.18 shows the schematic diagram of Option 1. 75

94 Figure 5.18: Schematic diagram of Option 1 Option 2 (Placing of guide bunds at the mouth of offtake) In Option 2, two guide bunds are placed at the mouth of offtake in the existing bathymetry of river. This option has been selected to observe the response of river due to placing a river training structure at the mouth of offtake. The schematic diagram of Option 2 is shown in Figure Figure 5.19: Schematic diagram of option 2 76

95 Option 3 (Dredging in the vicinity of offtake) In Option 3, dredging has been done in the vicinity of offtake to the certain reach (5.5km) downstream of Old Brahmaputra River. Option 1 is also superimposed in this option to having continuation of flow in the vicinity of offtake and finally connected them by maintaining a gradual slope. The schematic diagram of this option with dredging alignment is shown in Figure Figure 5.20: Schematic diagram of Option 3 Option 4 (Placing two guide bunds and dredging in the vicinity of offtake of Old Brahmaputra River) Option 4 is similar with Option 3 except two guide bunds are placed at the mouth of offtake like in Option 2. The schematic diagram of this option is shown in Figure The design dredge section for these options is also shown in Figure The minimum dredging depth at the offtake has taken 5m depth and then goes to downstream reach by following the slope of the river. 77

96 Figure 5.21: Schematic diagram of Option 4 RL 20 mpwd RL 14 mpwd 5m Dredging section 7 1 RL 9 mpwd 100m Figure 5.22: Design dredged section at the mouth of offtake 78

97 Base Dredging alignment for Option 1 Option 1 Guide Bund Option 2 Dredging alignment for Option 3 Option 3 Option 4 Guide Bunds Figure 5.23: Different simulated options superimposed on base bathymetry 79

98 CHAPTER 6 RESULTS AND DISCUSSIONS 6.1 General In this thesis work various models have been simulated including the base condition and followed by various options. These options are selected analyzing the offtake feature of Old Brahmaputra. To identify the benefits from these different options, various outputs have been analyzed. The bed level changes near the mouth of offtake for different options have been judged through planform basis and cross sectional analysis. The cumulative erosion and deposition at the end of monsoon in the selected reach of Old Brahmaputra is also analyzed. The discharge and percentage of flow through the Old Brahmaputra River is plotted for different options varying the flood events. Finally, a suggestion has been given for better offtake management through these comparative analyses. 6.2 Results and Discussions Different types of analyses have been done to understand the outcomes of different option simulation. The results are discussed below in terms of bed level changes, cross section analysis, volume of siltation and erosion, discharge and percentage of flow diversion in the Old Brahmaputra River Planform/Bed level changes The model simulated planform near the offtake of Old Brahmaputra for different options varying the flood events have been shown in Figure 6.1 to Figure 6.4. For the base condition that means in the existing condition, a huge siltation has occurred at the mouth of offtake for high flood event. On the other hand, less siltation occurred in the medium and low flood events. A large char is developed near the mouth of offtake for all flood events. However, it is seen that in case of high flood event, the extent of char is larger (Figure 6.1). For Option 1, siltation occurred near the mouth of offtake at the end of monsoon for high and medium flood events. The extent of char is lesser than the base condition for the high 80

99 flood event. Siltation in the offtake is relatively less for the low flood event in comparison to base condition, shown in Figure 6.2. In case of Option 2, the placing of river training works doesn t improve the situation. The offtake is getting close again at the end of monsoon for all the flood events (Figure 6.3). There is no significant change in terms of bed level for the rest of the Old Brahmaputra River. The dredging in the vicinity of offtake i.e. Option 3 shows better result than the previous one. For medium and low flood events, the bed level near the mouth of offtake is comparatively lower than the other options. It varies in the range of 14 to 15 mpwd for those two flood events (Figure 6.4). The high flood event, 1998 carries huge sediment and deposits it near the mouth of offtake. Due to dredging in the Old Brahmaputra River upto certain reach, the rest of the channel is improved. Finally, the combination of dredging and river training works i.e. Option 4 shows less siltation near the offtake for all types of flood events. The bed level varies 14.1 to 14.6 mpwd at that location for all flood events. The extent of char is also less for 1998 flood event. The channel is getting deeper from the base bathymetry at the end of monsoon in the downstream reach of the river where dredging is not executed, shown in Figure

100 Jamuna River Initial Base Bathymetry Old Brahmaputra River Developing a char High Flood Event, 1998 Huge siltation Medium Flood Event, 2005 Bed level, mpwd Mouth became closed Low Flood Event, 2001 Figure 6.1: Surveyed and simulated bathymetry for the base condition for different flood events at the end of monsoon 82

101 Initial Bathymetry Option 1 Siltation occurred High Flood Event, 1998 Dredging alignment for Option 1 Medium Flood Event, 2005 Low Flood Event, 2001 Bed level, mpwd Siltation is less Figure 6.2: Surveyed and simulated bathymetry for Option 1 for different flood events at the end of monsoon 83

102 Initial Bathymetry Option 2 High Flood Event, 1998 Guide Bund Huge siltation occurred Medium Flood Event, 2005 Low Flood Event, 2001 Bed level, mpwd Figure 6.3: Surveyed and simulated bathymetry for Option 2 for different flood events at the end of monsoon 84

103 Initial Bathymetry Option 3 High Flood Event, 1998 Dredging alignment for Option 3 Medium Flood Event, 2005 Low Flood Event, 2001 Bed level, mpwd Deeper channel Siltation is less Channel improved Figure 6.4: Surveyed and simulated bathymetry for Option 3 for different flood events at the end of monsoon 85

104 Initial Bathymetry Option 4 High Flood Event, 1998 Guide Bunds Less siltation occurred Medium Flood Event, 2005 Bed level, mpwd Deeper channel Low Flood Event, 2001 Channel improved Figure 6.5: Surveyed and simulated bathymetry for Option 4 for different flood events at the end of monsoon 86

105 Bed level, mpwd Cross section analysis Simulated cross sections for different options have been taken at different locations for different flood events. Here, the cross sections immediate downstream of offtake are shown in Figure 6.6 to Figure 6.8. The rest of the cross section at different locations are also plotted and given in Appendix B. Figure 6.6 illustrates that the river bed is shifted towards the left bank at the end of monsoon for all options during high flood event. Though the initial bed level of base condition, Option 1 and Option 2 are same; bed level is lowered down approximately 3m for Option 1 and 2 at the end of monsoon. On the other hand, 5m dredging depth in Option 3 and 4 causes the bed level much deeper than the base condition. In high flood event, the amount of siltation is also enormous. The right side of dredge section is almost silted up at the end of monsoon in Option 3 and base initial base end option 1 initial option 1 end option 2 initial option 2 end option 3 initial option3 end option 4 initial option 4 end Distance from left bank, m Figure 6.6: Simulated cross sections immediate downstream of offtake for high flood event for different options 87

106 Bed level, mpwd Similarly, simulated cross section near the offtake has been plotted for medium flood event, shown in Figure 6.7. It is seen from the figure that the cross section for base condition, Option 1 and 2 show almost similar result (slightly deeper in Option 2) at the end of monsoon. The thalweg for these options are shifted around 3m towards left bank and silted up. Again, Option 3 and 4, both of them show similar pattern of bed level at the end of monsoon. In this case, the channel is stable in the dredge section except some minor siltation occurred at the both side of dredged section Siltation along the both side of dredged section base initial base end option 1 initial option 1 end option 2 initial option 2 end option 3 initial option3 end option 4 initial option 4 end Distance from left bank, m Figure 6.7: Simulated cross sections immediate downstream of offtake for medium flood event for different options The simulated cross section for low flood event shows almost similar pattern of bed level changes near offtake (Figure 6.8). In this case, both Option 3 and Option 4 shows stable channel along the dredge section at the end of monsoon, otherwise Option 1 and Option 2 are shifting towards the left bank and silted up. 88

107 Bed level, mpwd base initial base end option 1 initial option 1 end option 2 initial option 2 end option 3 initial option3 end option 4 initial option 4 end Distance from left bank, m Figure 6.8: Simulated cross sections immediate downstream of offtake for low flood event for different options Bed Erosion/Deposition For three different flood events, it is seen that total volume of bed erosion in the 20 km reach of Old Brahmaputra is higher than the volume of deposition for all options. In case of Option 3 and Option 4, the volume of bed erosion is much higher compare to other options and base condition. The net balance is almost 5million m 3 bed erosion for these two options during high flood event. On the other hand, for average and low flood event the net balance is approximately 1.5 million m 3 bed erosion. Figure 6.9 to Figure 6.11 represent the simulated bed erosion and deposition at the end of monsoon in the 20 km reach of Old Brahmaputra for different flood event. 89

108 Volume of bed erosion or Volume of bed erosion or volume of deposition volume of erosion net balance 10 deposition, mill m Base Opt 1 Opt 2 Opt 3 Opt 4 Figure 6.9: Simulated bed erosion and deposition at the end of monsoon in the Old Brahmaputra reach for high flood event for different options deposition, mill m 3 volume of deposition volume of erosion net balance Base Opt 1 Opt 2 Opt 3 Opt 4 Figure 6.10: Simulated bed erosion and deposition at the end of monsoon in the Old Brahmaputra reach for medium flood event for different options 90

109 deposition, mill m 3 Volume of bed erosion or volume of deposition volume of erosion net balance Base Opt 1 Opt 2 Opt 3 Opt 4 Figure 6.11: Simulated bed erosion and deposition at the end of monsoon in the Old Brahmaputra reach for low flood event for different options The volume of deposition and erosion from the mouth of offtake to 20km reach of Old Brahmaputra for different flood events are also plotted and shown in Figure 6.12 to Though the total volume of erosion or deposition (Figure 6.9 to 6.11) is maximum for Option 4, but siltation/erosion is less at the mouth of offtake for this option. On the contrary, other options with base condition show huge siltation near the mouth of offtake. 91

110 Volume of deposition/ erosion (m 3 ) Volume of deposition/ erosion (m 3 ) Base - ero Opt 2 - ero Opt 4 - ero Opt 1 - dep Opt 3 - dep Opt 1 - ero Opt 3 - ero Base - dep Opt 2 - dep Opt 4 - dep Distance from the offtake (m) Figure 6.12: Volume of deposition and erosion for high flood event for different options from the mouth of offtake to 20km reach of Old Brahmaputra Base - ero Opt 2 - ero Opt 4 - ero Opt 1 - dep Opt 3 - dep Opt 1 - ero Opt 3 - ero Base - dep Opt 2 - dep Opt 4 - dep Distance from the offtake (m) Figure 6.13: Volume of deposition and erosion for medium flood event for different options from the mouth of offtake to 20km reach of Old Brahmaputra 92

111 Volume of deposition/ erosion (m 3 ) Base - ero Opt 2 - ero Opt 4 - ero Opt 1 - dep Opt 3 - dep Opt 1 - ero Opt 3 - ero Base - dep Opt 2 - dep Opt 4 - dep Distance from the offtake (m) Figure 6.14: Volume of deposition and erosion for low flood event for different options from the mouth of offtake to 20km reach of Old Brahmaputra Sediment Transport Sediment load has been extracted at few meters downstream of the offtake of Old Brahmaputra and at the downstream location of river for each option and plotted in Figure 6.15 to Figure Figure 6.15 indicates that due to high sediment flow in 1998 flood event, huge sediment has been transported through the offtake. On the contrary, for medium and low flood event, sediment transport is found lower at the offtake. Due to increase the conveyance area of Option 3 and Option 4, velocity is getting lesser in that location (Figure 6.18) and probably for that reason the rate of sediment transport is reduced at the offtake. However, at the downstream location of river, sediment transport is increasing with the flow and the velocity is also high in this area. 93

112 Sediment transport, kg/s Sediment transport, kg/s Sediment transport, kg/s Sediment transport, kg/s base option 1 option 2 option 3 option base option 1 option 2 option 3 option Jul-98 Jul-98 Aug-98 Sep-98 Sep-98 Oct-98 0 Jul-98 Jul-98 Aug-98 Sep-98 Sep-98 Oct-98 Figure 6.15: Sediment transport for different options a) at the immediate downstream of offtake and b) at downstream of Old Brahmaputra River for high flood event base option 1 option 2 option 3 option base option 1 option 2 option 3 option Jul-05 Jul-05 Aug-05 Sep-05 Sep-05 Oct-05 0 Jul-05 Jul-05 Aug-05 Sep-05 Sep-05 Oct-05 Figure 6.16: Sediment transport for different options a) at the immediate downstream of offtake and b) at downstream of Old Brahmaputra River for medium flood event 94

113 Velocity, m/s Sediment transport, kg/s Sediment transport, kg/s base option 1 option 2 option 3 option base option 1 option 2 option 3 option Jul-01 Jul-01 Aug-01 Sep-01 Sep-01 Oct-01 0 Jul-01 Jul-01 Aug-01 Sep-01 Sep-01 Oct-01 Figure 6.17: Sediment transport for different options a) at the immediate downstream of offtake and b) at downstream of Old Brahmaputra River for low flood event base at offtake opt1 at offtake base d/s of river opt1 d/s of river opt2 at offtake opt3 at offtake opt4 at offtake opt2 d/s of river opt3 d/s of river opt4 d/s of river Distance from right bank, m Figure 6.18: Velocity distribution along the cross section for medium flood event Discharge Discharge of Old Brahmaputra River has been extracted from the simulated results for all the options. It is observed that the maximum flow is taken place during the month of September for all flood events. Figure 6.19 represents the discharge for high flood event for different options during dry period. At the end of October, it is seen that the flow is 95

114 Discharge, cumec getting null for Option 1, 2 and 3 (Figure 6.19). But there is still certain amount of flow in Option 4 at that time Base Option 1 Option 2 Option 3 Option Jul-98 Sep-98 Oct-98 Dec-98 Jan-99 Mar-99 Figure 6.19: Discharge of Old Brahmaputra River for high flood event for different options In low flood event, the flow is continuously decreasing from mid October and is getting zero at the end of November for Option 1, 2 and 3. On the other hand, there is still few amount of flow (~32m 3 /s) is available in Option 4 for this flood event (Figure 6.20). The medium flood event shows good result for Option 4. It is clearly visible from Figure 6.21 that approximately 500 cumec discharge is available at the end of October for this option and during dry season it carries a continuous flow throughout the months (Table 6.1). Table 6.1: Simulated discharge (in m 3 /s) of Old Brahmaputra at the end of monsoon for different options Base Option 1 Option 2 Option 3 Option 4 High Flood Event Medium Flood Event Low Flood Event

115 Discharge, cumec Discharge, cumec Base Option 1 Option 2 Option 3 Option Jul-01 Sep-01 Oct-01 Dec-01 Feb-02 Mar-02 Figure 6.20: Discharge of Old Brahmaputra River for low flood event for different options Base Option 1 Option 2 Option 3 Option Jul-05 Sep-05 Oct-05 Dec-05 Feb-06 Mar-06 Figure 6.21: Discharge of Old Brahmaputra River for medium flood event for different options 97

116 % of flow Percentage of flow diversion Since the flow of the Old Brahmaputra is totally dependent on the flow of the Jamuna, ratio of the flow of Old Brahmaputra to the Jamuna has been extracted and presented in Figure 6.22 to Figure During monsoon, the ratio is in the range 2% to 4 % for high flood event whereas during dry period i.e. at the end of October it diminishes drastically and reaches zero percentage for all options, except for Option Base Option 1 Option 2 Option 3 Option Jul 4-Sep 14-Oct 23-Nov 2-Jan 11-Feb Figure 6.22: Percentage of flow diversion through Old Brahmaputra River for high flood event for different options For low flood event, the flow diversion ratio is in the range of 1.5% - 2.5% during monsoon (Figure 6.23). The flow ratio decreases early for Option 1 and 2 at the end of October for this flood event. On the contrary, the ratio of flow diversion sustains few days more for Option 3 and 4. Among all the options, Option 4 shows better result. It provides a good amount of flow through Old Brahmaputra River during dry period for medium flood event (Figure 6.24). Other options are showing that after the month of November, the percentage of flow is totally fall down. 98

117 % of flow % of flow Base Option 1 Option 2 Option 3 Option Jul 4-Sep 14-Oct 23-Nov 2-Jan 11-Feb Figure 6.23: Percentage of flow diversion through Old Brahmaputra River for low flood event for different options Base Option 1 Option 2 Option 3 Option Jul 15-Aug 14-Sep 14-Oct 13-Nov 13-Dec 12-Jan 11-Feb Figure 6.24: Percentage of flow diversion through Old Brahmaputra River for medium flood event for different options 99

118 % of flow diversion High Flood Event Medium Flood Event Low Flood Event Base Option 1 Option 2 Option 3 Option 4 Figure 6.25: Percentage of flow diversion at the end of monsoon through Old Brahmaputra River Figure 6.25 represents the percentage of flow diversion at the end of monsoon through Old Brahmaputra River for different flood event with different options. It is seen that Option 3 and Option 4 show the better results in terms of flow diversion. At the end of monsoon, almost 2.5% of flow is available for these two options for medium and low flood event Water depth The water depths for different options for an average flood event are represented in Figure It is seen from the figure that at the end of November, the link between the Jamuna and Old Brahmaputra River is getting disconnected for Option 1 and 2 (Figure 6.26). However, in case of Option 3, it sustains another single month. In Option 4, there is still connection between the Jamuna and Old Brahmaputra River at the end of December. 100

119 Base Option 1 No connection exist at the end of November Option 2 Starting to discontinue the connection at the beginning of December Option 3 No connection exist at the end of December Option 4 Still connected at the end of December Water depth, m Figure 6.26: Simulated water depth at offtake during dry period for medium flood event 101

120 Discharge, m 3 /s Water level, mpwd Sustainability of Option 4 To check the sustainability of Option 4, a morphological simulation has been given for the three successive monsoon periods that means from the year 2005 to The boundary for this simulation is represented in Figure u/s discharge boundary d/s WL boundary at Jamuna d/s WL boundary at OB Feb-05 Sep-05 Mar-06 Oct-06 Apr-07 Nov Figure 6.27: Boundary for three successive monsoon periods from 2005 to Figure 6.28 shows the simulated water depth for Option 4 at different periods of simulation. It is seen that at the end of 1 st year monsoon, there is a connection between two rivers. The water depth at the mouth of Old Brahmaputra is nearly 2.3 m at the end of 1 st year monsoon. In the next monsoon, the main channel of Jamuna River is dominant along the left bank which is a favorable condition for Old Brahmaputra for extracting the flow. However, due to siltation around the mouth of offtake the depth of water is decreased at the end of 2 nd year monsoon. The connection between these two rivers is totally disappeared at the end of 3 rd year monsoon. Thus, it can be concluded that if maintenance dredging is not done prior to the monsoon period, the mouth of Old Brahmaputra getting silted and no flow will be passing during the dry period. The percentage of flow is also show the declining pattern at the end of monsoon in the following years (Figure 6.29). 102

121 Initial bathymetry for Option 4 At the peak of 1 st year monsoon At the end of 1 st year monsoon At the peak of 2 nd year monsoon At the end of 2 nd year monsoon At the peak of 3 rd year monsoon At the end of 3 rd year monsoon Water depth, m Figure 6.28: Simulated water depth of Option 4 for three successive monsoon periods. 103

122 % of flow diversion through Old Brahmaputra Apr 28-May 27-Jun 27-Jul 26-Aug 25-Sep 25-Oct 24-Nov Figure 6.29: Percentage of flow diversion for the three successive monsoon periods Summary To suggest the better offtake management, a comparative analysis of different options has been done in the above section. It is understood from the analysis that only dredging along the left anabranch of Jamuna River near Old Brahmaputra offtake i.e. Option 1 is not a feasible solution. The mouth of offtake get silted up massively at the end of monsoon and the percentage of flow diversion is lower than the base condition. In addition, there are no connection exist between the two rivers at the beginning of dry period. Likewise, Option 2 is arranged by incorporating two guide bunds at the mouth of offtake to guide the flow through the Old Brahmaputra River. In this case, the siltation pattern is almost same with the previous option. The bed level near offtake is found above 16mPWD at the end of monsoon. The volume of deposition and bed erosion in the 20km reach of river is almost balanced. The percentage of flow indicates a relatively higher value from Option 1 during monsoon period, but in dry period not such improvement is observed. Option 3 is selected in such a way that dredging will be done from the vicinity of offtake to the certain downstream reach (here 5.5km reach) of river. Option 1 is also 104

123 superimposed with this option to having a continuation of flow in the vicinity of offtake and then finally a connection between them has been maintained by following a gradual slope. This option shows much better results from the last two options. The bed level near mouth of offtake is comparatively lower, and in a range of 14 to 15 mpwd for medium and low flood events. The dredging in the offtake makes the channel deeper along the whole reach of Old Brahmaputra which is sustained at the end of monsoon. The net volume of bed erosion is relatively higher than the volume of deposition. In low flood event, the flow is continuously decreasing from mid October and is getting zero at the end of November. The flow diversion is comparatively well at the beginning of dry season but at the end of December it closes to zero. The last option is Option 4, which is the combination of dredging and river training works. Dredging upto the certain reach of river make the river deeper and confined. On the other hand, placing of two guide bunds in the offtake generate less siltation near the mouth of offtake and help to convey the flow easily. The bed level varies 14.1 to 14.6 mpwd near the offtake. The cross section plot for Option 4 shows stable channel along the dredge section at the end of monsoon. The net volume of bed erosion is relatively higher than the volume of deposition like Option 3 and the rate of sediment transport is the lowest comparing to all other options. Approximately 500 cumec discharge is available at the end of October for this option and during dry season it carries a continuous flow throughout the months for average flood event. Moreover, it provides a good amount of flow through Old Brahmaputra River during the whole dry period for this flood event. On the contrary, other options are showing that after the end of November, the percentage of flow is totally fall down. So, among all options, it can be said that Option 4 shows superior result. The summary results in terms of bed level, cross section, sediment transport, discharge, percentage of flow diversion and water depth are shown in Table

124 Table 6.2: Summary results for different options at the end of monsoon Option 1 Option 2 Option 3 Option 4 Bed level (mouth of offtake) in mpwd Above 16 Above Cross section Silted up Silted up Stable Stable Sediment transport at offtake (kg/s) Discharge (m 3 /s) % of flow diversion Water Depth (m), at the end of December No connection between two rives No connection between two rives No connection between two rives Still connected 106

125 CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS 7.1 General Due to high sedimentation in the offtake, the main Old Brahmaputra River had already been lost its conveyance capacity of water originated from the Jamuna River during the lean season. The main objective of this study is to suggest a better offtake management approach by simulating various options in the model in order to augment the dry season flow through offtake. Two-dimensional morphological model has been utilized to analyze the feasibility of application of engineering options like dredging, river training works in order to keep Old Brahmaputra alive. The results obtained from the present study can be summarized as the following conclusions. 7.2 Conclusions The following conclusion can be drawn after summarizing the present studyi. The flow diversion during dry period is continuously decreasing in the last few ii. iii. decades. During , this ratio was nearly 5% (of main Jamuna flow) at the end of monsoon which was reduced to 0.5% at the end of dry period. In the recent decade this values are significantly reduced and at the end of dry period the flow diversion is almost zero. The declining pattern of Old Brahmaputra river flow indicates that sedimentation has taken place on the river bed. Historical images show that the position of offtake is continuously swinging within the region of 15 km stretches of the river ( ). Moreover, in many occasions it was found that the location of the off-take was not well defined or there might present more than one opening of a single river. Therefore, it is very difficult to identify the mouth of offtake. From the bed level and cross-sectional analysis, it is observed that only dredging along the left anabranch of Jamuna River near Old Brahmaputra (Option 1) is not a suitable solution. The mouth of offtake get silted up massively at the end of monsoon and the percentage of flow diversion is lower than the base condition. 107

126 iv. Placement of river training work does not improve the flow condition either. At the end of monsoon, the bed level near offtake is still high, about 16 mpwd. The amount of flow diversion is 2% in the monsoon for low flood event but after the end of November it detached from Jamuna. v. Dredging along the left anabranch of Jamuna as well as in the vicinity of offtake to the certain reach of Old Brahmaputra (Option 3) is showing quite better result than the previous options. The availability of water is persisted upto the mid December for average flood event and then disconnected. The flow diversion is varies from 2 to 2.4 % at the end of monsoon for average and low flood events. The cumulative bed erosion is almost 1.5 million m 3 higher than the deposition for medium and low flood events. vi. The last option (Option 4) is the combination of dredging and river training works. Combined effects of these engineering works demonstrate that the highest amount of flow diversion (2.5% during monsoon and 1.5% during dry period) among all other options for an average flood event. Moreover, it carries certain amount of flow (32 m 3 /s) throughout the months during dry season. The cross section at different reaches for Option 4 shows a deep and stable channel along the dredge section at the end of monsoon. The net volume of bed erosion is relatively higher than the volume of deposition like Option 3 and the rate of sediment transport near the offtake is the lowest comparing to all other options. vii. If ranking is done among all other options, Option 4 should be in rank 1, then comes Option 3. Option 1 and Option 2, both of them are not suitable solution to sustain the offtake. 7.3 Recommendations The recommendations for future investigation and study can be as follows: i. The mouth of offtake is not defined. As it is changes in every year, so it is quite difficult to identify the mouth of offtake for preparing the model domain. In this study, model domain has been fixed by observing the position of mouth of offtake in the bathymetry of In future study, this model domain may be extended for closer resolution. 108

127 ii. As Jamuna is a braided river, the char movement is very rapid in this river. Therefore, the dredging alignment or river training works proposed in this thesis should be modified according to the planform of the river. iii. The dredging in Jamuna should be done in an intelligent way by understanding the proper timing. During the onset of monsoon, dredging will be done to initiate the scouring and finally in monsoon nature continues and finishes these morphological processes. iv. As Jamuna is an alluvial river in a deltaic region of Bangladesh, so morphological activities are very dynamic in this river. Capital dredging may not be sustainable for a long period, though it is effective for one monsoon only. So frequent maintenance dredging is required. Sustainability of dredging works at the mouth/offtakes can be further studied in a separate scope. v. Data availability of Old Brahmaputra River and it s offtake is scarce. Extensive hydrometric and hydrologic and sediment data collection program need to be undertaken for more detailed further study related to the river offtakes of the country. 109

128 References Alam, M.K. and Hossain, M.M., (1988), Sediment Transport in the River Jamuna, Journal of Institution of Engineers, Bangladesh, Vol. 16(2), p Ali. F., (2010), A Study on siltation at the intake reach of the Old Brahmaputra River, M.Engg Project, Bangladesh University of Engineering and Technology (BUET), Dhaka, Bangladesh. Agunwamba, J.C. and Nnaji, C.C., (2009), The Effects of offtake Angle on the Velocity Distribution and Rate of Siltation of Canals, Nigerian journal of technology, Vol. 28 No.2, pp Bari, M.F. (1978), Applicability of Sediment Transport Formulas for the Ganges and Jamuna, M.Sc.Engg.Thesis, Department of Water Resources Engineering, BUET, Dhaka. Boskalis-GRC, (2000), Old Brahmaputra offtake area: Feasibility of Pilot Dredging, Prepared for BWDB, Dhaka, Bangladesh. Bray, R.N., Bates, A.D. and Land, J.M., (1997), Dredging: A Handbook for Engineers, 2nd edition, New York, John Wiley & Sons, Inc. Bristow, C.S., (1999), Avulsion, river metamorphosis and reworking by under-fit stream: a modern example from the Brahmaputra River in Bangladesh and a possible ancient example in the Spanish Pyreness, Sepc. Publis. int. Ass. Sediment, 28, BWDB (2010), Gorai River Restoration Project (Phase-II), Project executed by Bangladesh Water Development Board, Dhaka, Bangladesh. CEGIS, (2007), Long-term Bank Erosion Processes of the Jamuna River, Prepared for JMREMP, BWDB, Dhaka, Bangladesh, pp.74. Coleman, J.M. (1969), Brahmaputra River: channel process and sedimentation, Sedimentary Geology3, pp

129 Dey, K. C, Mahmood, S. and Matin, M.A. (1998), Exponent of flow velocity foe the transport of sediments in alluvial rivers, Journal of Civil Engineering,, The Institution of Engineers, Bangladesh, Vol. CE 26 No 1 (1998), p DHV Haskoning (2012), Mathematical modelling for off-take management of Gorai River, Design Report, Annex 2, Prepared for World Bank. EGIS, (2001) Planform Analysis of the Old Brahmaputra River in connection with the Sherpur Bridge Project, Prepared for SWMC. FAP 2, (1993), North West Regional Study, the Regional Plan Final Report, Prepared by Mott Macdonald Int. in association with Hydraulic Research Ltd, House of Consultants Ltd under assignment to Overseas Development Administration. FAP 3, (1993), North Central Regional Study, Supporting Report II.1: Hydrometeorology, BCEOM, Desh Upodesh Ltd., BETS Ltd. FAP24, (1996a), River Survey Project, Morphological Process of Jamuna River, Special Report 24, GoB/FPCO, Prepared for Water Resources Planning Organization, Dhaka, FAP24, (1996b), River Survey Project, Morphology of Gorai Offtake, Special Report 10, GoB/FPCO, Donor: Commision of the European Communities (CEC). Delft Hydraulics, Danish Hydraulic Institute, Osiris, Hydroland, Approtech. FAP24, (1996c) River Survey Project, Hydrology, Final Report, Annex 3, GoB/FPCO, Prepared for Water Resources Planning Organization, Dhaka. FAP24, (1996d) River Survey Project, Morphological Process of Jamuna River, Special Report 18, GoB/FPCO, Prepared for Water Resources Planning Organization, Dhaka. Fergusson, J., (1863), On Recent Changes in the Delta of the Ganges, Proc. of the Geological Society, Quarterly Journal, London, Vol. XIX, Goodbred Jr., S.L. and Kuehl, S.A., (2000a), The significance of large sediment supply, active tectonism, and eustasy on sequence development: Late Quaternary Stratigraphy and evolution of the Ganges-Brahmaputra delta, Sedimentary Geology 133,

130 Goodbred Jr. S.L., Kuehl, S.A., (2000b), Enormous Ganges-Brahmaputra sediment discharge during strengthened early Holocene monsoon, Geology, Vol. 28, Hossain, M.M. (1989), Geomorphic Characteristics of the Ganges (Padma) up to Brahmaputra Confluence Final report, R02/89, IFCDR, BUET, July 1989, pp.158. Hirst, Major F. C., (1916), Report on the Nadia Rivers, Bengal Secretariat Book Depot,Calcutta, pp. 39 Imteaz, M. A. and Hassan, K. I., (2001), Hydraulic Impacts of Jamuna Bridge; Mitigation Option Proceedings on 6 th conference on Hydraulics in Civil Engineering: The State of Hydraulics, Barton, A.C.T. Institution of Engineers, Australia, pp ISPAN, (1995), The dynamic physical environment of riverine charlands: Brahmaputra-Jamuna, Irrigation Support Project for Asia and near East (ISPAN), Prepared for Flood Plan Coordination Organisation (FPCO), Dhaka, Bangladesh. IWM, (2003), Geo-referencing of BWDB X-section pillars, Prepared for Bangladesh Water Development Board, Bangladesh, pp.233. IWM, (2004), Dewanganj- Raumari Road Bridga Project at Pollahkandi, Hydraulic Mathematical Modelling including Morphological Prediction of Old Brahmaputra River in connection with the Hydraulic Design of Proposed Bridge, Final Report. IWM, (2007), Hydrological and Hydro-morphological Study for the Balipara Bridge at 13th Km of Trishal-Nandail-Tarail Road on the Old Brahmaputra River, Final Report. IWM, (2011a), Morphological mathematical modelling for planning, design, monitoring and quality control of dredging for Gorai River Restoration Project, Phase- II, Progress Report-4, Prepared for Bangladesh water Development Board, Dhaka, Bangladesh. IWM, (2011b), Mathematical Modelling for Offtake Management of the New Dhaleswari River and Hydraulic Monitoring of New Dhaleswari-Pungli-Bangshi- Turag-Buriganga System, Prepared for Bangladesh water Development Board, Dhaka, Bangladesh. 112

131 Kleinhans, M., Jagers, B., Mosselman, E. and Sloff, K. (2006), Effect of upstream meanders on bifurcation stability and sediment division in 1D, 2D and 3D models, Published on International Conference on Fluvial Hydraulics, Lisbon, Portugal, Eds. R.M.L. Ferreira, E.C.T.L. Alves, J.G.A.B. Leal and A.H. Cardoso, Taylor and Francis/Balkema, London, UK, p Lagasse, P.F. (1986), "River Response to Dredging", Journal of Waterway, Port, Coastal and Ocean Engineering, Vol. 112, No. 1, January, 1986, Paper No Lane, E.W., (1955), Importance of Fluvial Morphology in Hydraulic Engineering, ASCE, Proceedings, Vol. 81, Paper 795, Lane, E.W., (1957), A study of the shape of channels formed by natural streams flowing in erodible material, US Army Engineer Division, Missouri River Corps of Engineers, Ohama Nebraska, M.R.D. Sediment Series 9, pp.106. La Touche (1910), Relics of the great ice age of the plains of Northern India, Calcuta, pp Lu, Q. and Nairn, R.B. (2010), Prediction on Morphological Response of Dredged Sand-Borrow Pits", Mamun, M.Y., Hossain, M.M. and Sikder, M.S. (2012), Offtake morphology of a river-a case study of Arial Khan River of Bangladesh, Proceedings in the 2nd International Conference on Water Resources (ICWR), Malaysia, 5-9 November, Makaske, B., (2001), Anastomosing rivers: a review of their classification, origin and sedimentary products, Earth Surface Reviews, Vol. 53, Mead, C. T. (1999), An investigation of the suitability of two-dimensional mathematical models for predicting sand deposition in dredged trenches across estuaries, Journal of Hydraulic Research, Vol. 37, No. 4. Morgan, J. P. and McIntire,W.G., (1959), Quaternary Geology of the Bengal Basin, East Pakistan and India, Bulletin of the Geological Society of America, Vol. 70,

132 Mosselman, E. (2004), Morphology of River Bifurcations: Theory, Field Measurements and Modelling WL,Delft Hydraulics & Delft University of Technology Mohiuddin, F. A. and Rahman, M. U., (2011), The Gorai River offtake management to ensure the lean season flow, 3 rd International Conference on Water & Flood Management, ICWFM, Dhaka. Monsur, M.H., (1995), An Introduction to the Quaternary Geology of Bangladesh, Published by Rehana Akhter, Dhaka, Bangladesh, pp. 82. Obasi, N.I., Agunwamba, J.C., Egbuniwe, N., (2008), Influence of offtake angles on flow distribution pattern at concave channel bifurcation, Nigerian Journal of technology, Vol: 27 No.2, pp Obasi, N.I., Oloke, D.A. and Agunwamba, J.C., (2012), Effect of offtake angles on spatial distribution of silt material at concave bifurcation, International Journal of Engineering Research and Applications (IJERA), Vol. 2, Issue 5, September-October 2012, pp Roosjen, R. and Zwanenburg, C. (1995), Research on bifurcations in Rivers M.Sc Thesis, Department of Hydraulic Engineering, Faculty of Civil Engineering, Delft University of Technology (DUT). Sarker, M.H., Kamal, M.M. and Hassan, K., (1999), Identifying the morphological changes of a distributary of the Ganges in response to the declining flow using remote sensing, Proc. 20 th Asian Conf. on Remote Sensing, Hong Kong, pp Schumm, S.A., (1977), The Fluvial System, John Wiley & Sons, NY. Sultana.N. (1989), Critical Review of Sediment Transport studies of Jamuna river, M.Engg.Thesis, BUET, Department of Water Resources Engineering, Dhaka. SWMC and DHI (1999), Navigabiliy of Old Brahmaputra River Survey and Modelling of River Condition, Final Report, Prepared for Mitsui Engineering and Shipbuilding Co. Ltd. SWMC, (2001), Road Bridge over Old Brahmaputra at Khurshid Mohal, Hydraulic and Geo-Morphological Feature Study by Mathematical Modelling, Final Report. 114

133 SWMC, (2001), Sherpur Bridge Modelling, Hydraulic Mathematical Modelling of the Old Brahmaputra River at the Proposed Sherpur Road Bridge Project,Final Report. Thorne, C. R., Russell, A.P.G. and Alam, M.K., (1993), Planform pattern and Channel Evolution of the Brahmaputra River, Bangladesh, Published in Geological Society, London, Special Publications 1993, v.75;pp Umitsu, M., (1993), Late Quaternary Sedimentary Environments and Land Forms in the Ganges Delta, Sedimentary Geology, 83, Van Rijn, L.C. (1986), Sedimentation of dredged channels by currents and waves, Journal of Waterway, Port and Ocean Engineering, ASCE, Vol. 112, No. 5. Winkley, B.R., (1994), Response of the Lower Mississippi River of Flood Control and Navigation Improvements, In The Variability of the Large Alluvial Rivers, Schumm and Winkley (eds.), pp , Publ. by ASCE press, NY. 115

134 APPENDIX-A Historical cross section at different location of Old Brahmaputra River 116

135 Figure A-1: BWDB transect line of cross section of Jamuna and Old Brahmaputra River. 117

136 Bed level, mpwd Bed level, mpwd Distance from right bank, m Figure A-2: Superimposition of cross section OB-1 of Old Brahmaputra River in different year Distance from right bank, m Figure A-3: Superimposition of cross section OB-2 of Old Brahmaputra River in different year 118

137 Bed level, mpwd Bed level, mpwd Distance from right bank, m Figure A-4: Superimposition of cross section OB-3 of Old Brahmaputra River in different year Distance from right bank, m Figure A-5: Superimposition of cross section OB-4 of Old Brahmaputra River in different year 119

138 Bed level, mpwd Bed level, mpwd Distance from right bank, m Figure A-6: Superimposition of cross section OB-5 of Old Brahmaputra River in different year Distance from right bank, m Figure A-7: Superimposition of cross section OB-6 of Old Brahmaputra River in different year 120

139 Bed level, mpwd Bed level, mpwd Distance from right bank, m Figure A-8: Superimposition of cross section OB-7 of Old Brahmaputra River in different year Distance from right bank, m Figure A-9: Superimposition of cross section OB-8 of Old Brahmaputra River in different year 121

140 Bed level, mpwd Bed level, mpwd Distance from right bank, m Figure A-10: Superimposition of cross section OB-9 of Old Brahmaputra River in different year Distance from right bank, m Figure A-11: Superimposition of cross section OB-10 of Old Brahmaputra River in different year 122

141 APPENDIX-B Simulated cross section at different location of Old Brahmaputra River 123

142 Bed level, mpwd Bed level, mpwd 24 X-Sec Xsec base initial base end option 1 initial end -high end -medium end -low Distance from left bank, m Figure B-1: Simulated cross section No 1 (j=165) for Option 1 24 X-Sec Xsec base initial base end option 1 initial end -high end -medium end -low Distance from left bank, m Figure B-2: Simulated cross section No 2 (j=178) for Option 1 124

143 Bed level, mpwd Bed level, mpwd X-Sec 3 base initial base end option 1 initial end -high end -medium end -low Distance from left bank, m Xsec-3 Figure B-3: Simulated cross section No 3 (j=230) for Option 1 24 X-Sec base initial base end option 1 initial end -high end -medium end -low Distance from left bank, m Xsec-4 Figure B-4: Simulated cross section No 4 (j=308) for Option 1 125