Research on large flood management of tra bong river basin, quang ngai province

ABSTRACT This study aimed at proposing solutions for managing large flood in Tra Bong river basin, Quang Ngai province.. The research includes 2 main parts: create a flood inundation map

ACKNOWLEDGEMENT

Firstly, I would like to express my sincere gratitude to my advisors Assoc.Prof Hoang Thanh Tung and Dr Pham Thanh Hai for the continuous support of my M.Sc study and related research, for their patience, motivation, and immense knowledge Their guidance helped me in all the time of research and writing of this thesis I could not have imagined having a better advisor and mentor for my M.Sc study

Besides, I am especially grateful to lecturers in the Department of Hydrology and Water resources, Thuy Loi University and foreigner lectures from NICHE project who supported me for all the lectures and useful advices throughout my course

My sincere thanks also goes to my colleagues in National Meteorological Service and Hydro-Meteorological Department of Middle Centre region who supported me for data collection and analysis Without they precious support

Hydro-it would not be possible to conduct this research

Last but not the least, I would like to thank my family: my parents, my wife and children for supporting me spiritually throughout the course and my life in general

Hanoi, Nov 08th 2016

Nguyen Phu Luan

ABSTRACT

This study aimed at proposing solutions for managing large flood in Tra Bong river basin, Quang Ngai province The research includes 2 main parts: create a flood inundation map based on historical flood marks collected and flood forecasting experiment for the downstream of river by simulating flood flow with the combination

of hydrological and hydrodynamic models At the first part, 42 flood marks of the

2009 historical flood had been collected with their exactly coordinate and maximum water level The distribution and elevation of flood marks have been used to calculate with topography data (DEM) using GIS tools and the result was the 2009 flood inundation map At the second part, a rainfall – runoff hydrological model (NAM) has been firstly used to simulate flow from upstream of basin to the section at the beginning of the main river The input data were collected from 3 rain gauge stations for simulation After that, the computed water discharge got from rainfall–runoff model have been used as the upstream boundary for simulating flood flow in the main river The model has been used was 1-dimensional hydrodynamic model After simulating the flood flow in the river, the study tried flood forecasting for the downstream part Based on the flood inundation map and the result of flood forecasting experiment, the research has proposed solutions to manage large floods on Tra Bong river basin These solutions can be used to develop large flood management plans for local authorities, in order to enhance efficiency in water resources using and reduce losses caused by large flood

Abbreviation

GIS Geographical Information System

WMO World Meteorological Organization

DHI Danish Hydraulic Institute

DEM Digital Elevation Model

DMC Disaster Management Cycle

MONRE Ministry of Natural Resources and Environment

MOC Ministry of Construction

NHMS National Hydro-Meteorological Sevice

NAM Nedbør-Afstrømnings-Model (Danish, meaning rainfall-runoff

model)

1D, 2D 1-dimensional, 2-dimensional

TABLE OF CONTENTS

1 General Introduction 8

2 Description of the Study Area 8

3 Problems and Need of Study 11

4 Objectives of Study 11

5 Scope of Study 12

CHAPTER I: LITERATURE REVIEW 13 1.1 Related researches about study site 13

1.2 Flood inundation mapping 17

1.3 Geographical Information Systems in Hydrology and Water Resources 21

1.4 Flood Forecasting 23

CHAPTER II: APPROACH AND METHODOLODY: 28 2.1 Approach of study 28

2.2 Flood inundation map 28

2.3 Using GIS tools to develop flood inundation map 30

2.4 Mike 11 general description 32

2.5 Theoretical Foundation of rainfall – runoff hydrological model (NAM) 42

2.5.1 The basic parameters of NAM model 44

2.5.2 Basic modelling components 46

2.5.3 Initial conditions of the model 50

2.5.4 Model calibration 51

2.6 Flood Forecasting (Mike 11 FF): Updating procedure 51

2.6.1 Two unique features MIKE 11 FF'S updating procedure 52

2.6.2 The calibration updating parameters 55

2.7 Flood Forecasting Error 57

CHAPTER III: RESULTS AND DISCUSSIONS 59 3.1 Analysing flood features of Tra Bong river basin 59

3.1.1 Rain features 59

3.1.2 Flood features 61

3.2 Develop flood inundation mapping for Tra Bong river basin 66

3.3 Flow forming simulation using Mike NAM 74

3.3.1 Calculation layout 74

3.3.2 Data analysis 75

3.3.3 NAM model calibration and verification 77

3.4 Flood flow in the downstream using Mike 11 hydraulic model 82

3.4.1 Boundary condition 82

3.4.2 Calibration and verification of flood flow simulation model 84

3.5 Flood forecasting experiment for Tra Bong river system 90

3.6 Propose solutions of large flood managing in Tra Bong river basin 95

3.6.1 Structure methods 96

3.6.2 Non-structure methods 97

CHAPTER IV: CONCLUSION AND RECOMMENDATION 101 4.1 Conclusion 101

4.2 Recommendation 102

LIST OF FIGURES

Figure 1: Administration map of Quang Ngai Province 9

Figure 2.1: Structure of literature review 13

Figure 2.2: Conceptual framework for flood hazard and risk calculations 19

Figure 2.3: Different flood map types 20

Figure 2.4: Process for developing a flood forecasting model 25

Figure 3.1: Conceptual Framework 27

Figure 3.2: Overview of study 28

Figure 3.3: Digital Elevation Model with square grid 31

Figure 3.4: Error based on the topography data in flood inundation mapping 31

Figure 3.5: The structure of NAM model 38

Figure 3.6: Channel section with computational grid 44

Figure 3.7: The shape of the computational grid around a node which has three branches 45

Figure 3.8: The shape of the grid points and the nodes in the complete model 45

Figure 3.9: Branch matrix before reducing 47

Figure 3.10: Branch matrix after reducing 47

Figure 3.11: Three-branch node with limit for continuity equation 48

Figure 3.12: River branch with discharge boundary 49

Figure 3.13: Illustration of amplitude and phase error 53

Figure 3.14: The updating results of simulations 55

Figure 3.15: Example of measured and simulated discharge at an update location 57

Figure 3.16: Example of updating parameters 57

Figure 4.1: Chart of the possibility of flooding which reach 2nd alarm level or higher in flood season - Tra Bong river basin 64

Figure 4.2: 6 hour rainfall chart, from 19:00 27 Sep 2009 to 01:00 30 Sep 2009 – Tra Bong station 65

Figure 4.3: 6 hour rainfall chart, from 19:00 27 Sep 2009 to 01:00 30 Sep 2009 – Chau O station 66

Figure 4.4: Hourly water level process of the flood from 28 Sep to 01 Oct 2009 – Chau O station 67

Figure 4.5: 2009 Flood marks map of Tra Bong river downstream 70

Figure 4.6: 2009 Flood inundation map of Tra Bong river basin 73

Figure 4.7: Map of hydro-meteorological stations network 74

Figure 4.8: Calculation layout 75

Figure 4.9: Weight factor distribution layout of rain gauge stations in Quang Ngai province 76

Figure 4.10: Basin parameters declaring dialog 77

Figure 4.11: Parameters calibration for rainfall-runoff model (NAM) 77

Figure 4.12: Calculated flow process at Binh Minh station 78

Figure 4.13: Rainfall – runoff model (NAM) calibration, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 16 to 20 Oct, 2008 80

Figure 4.14: Rainfall – runoff model (NAM) calibration, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 28 Sep to 5 Oct, 2009 80

Figure 4.15: Rainfall – runoff model (NAM) calibration, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 13 to 19 Nov, 2010 81

Figure 4.16: Rainfall – runoff model (NAM) verification, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 14 to 20 Oct, 2011 81

Figure 4.17: Rainfall – runoff model (NAM) verification, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 5 to 9 Nov, 2011 82

Figure 4.18: Rainfall – runoff model (NAM) verification, compare observed and simulated flood discharge at Binh Minh – Tra Bong river, from 25 to 29 Nov, 2011 82

Figure 4.19: Hydraulic routing layout of Tra Bong river downstream 83

Figure 4.20: Initial conditions calibration dialog 86

Figure 4.21: Bed resistance calibration dialog 86

Figure 4.22: Bed resistance calibration for cross-sections 87

Figure 4.23: MIKE 11 (HD) calibration, compare observed and simulated water level at Chau O station – Tra Bong river, from 16 to 20 Oct, 2008 88

Figure 4.24: MIKE 11 (HD) calibration, compare observed and simulated water level at Chau O station – Tra Bong river, from 28 Sep to 5 Oct, 2009 88

Figure 4.25: MIKE 11 (HD) verification, compare observed and simulated water level

at Chau O station – Tra Bong river, from 13 to 19 Nov, 2010 90 Figure 4.26: MIKE 11 (HD) verification, compare observed and simulated water level

at Chau O station – Tra Bong river, from 14 to 20 Oct, 2011 90 Figure 4.27: Flood flow forecasting results from rainfall data at Binh Minh – The rains

from 5 to 9 Nov, 2011 92 Figure 4.28: Observed and 5 hour predicted flood water level process in Chau O

station - from 5 to 9 Nov, 2011 93 Figure 4.29: Flood flow forecasting results from rainfall data at Binh Minh – The rains

from 25 to 29 Nov, 2011 93 Figure 4.30: Observed and 5 hour predicted flood water level process in Chau O

station - from 25 to 29 Nov, 2011 94 Figure 5.1: Synthesis of large flood management solutions in Tra Bong river basin 97 Figure 5.2: Disaster Management Cycle 99

LIST OF TABLES

Table 1.1: The morphological features of Tra Bong river and major tributaries 10

Table 2.1: Overview of methods and data for high-resolution flood-risk mapping in Germany 18

Table 2.2: Predictive performance of the 3 models using independent calibration /validation data 26

Table 3.1: Technical requirements of elevation different between the contours with corresponding scales 32

Table 3.2: Quality of forecasting classification 59

Table 4.1: Rainfall causes flood rising in Tra Bong river basin 61

Table 4.2: Rainfall in the history flood on 28-30 Sep 2009 61

Table 4.3: Maximum daily rainfall in Tra Bong station, from 2006 to 2011 61

Table 4.4: Annual peak flood of Chau O station (2006-2011) 62

Table 4.5: Typical flood intensity and amplitude in Chau O station (with the peak flood higher than first alarm level) 62

Table 4.6: The typical floods in 2006-2011 period of Chau O station 63

Table 4.7: 6 hour rainfall data, from 19:00 27 Sep 2009 to 01:00 30 Sep 2009 – Tra Bong river basin 65

Table 4.8: Detail characteristics of the flood from 28 to 30 Sep 2009 66

Table 4.9: The data of 2009 flood marks in Tra Bong river basin 67

Table 4.10: The rain gauges used for hydrological calculating 75

Table 4.11: Results of parameters calibration for rainfall – runoff model 79

Table 4.12: Results of NAM model calibration and verification at Binh Minh 79

Table 4.13: Position of nodes in hydraulic calculation layout of Tra Bong river 83

Table 4.14: Analysis of model calibration efficiency and error 87

Table 4.15: Analysis of model verification efficiency and error 89

Table 4.16: Results synthesis of flood forecasting of Tra Bong river in Chau O station 95

INTRODUCTION

1 General Introduction

Large flood is a kind of disaster occur regularly and seriously annually around the world Flood has some benefits such as bring fertile soil to replace nutrient-poor soils, but it also impacts and causes enormous damage, constantly threatening people living and the economic and social development People use many methods to prevent and reduce the impact of flood including management, structures and non-structures methods Therein, large flood management is always an important objective, which requires detailed and specific research for each region

The research “Research on Large flood management of Tra Bong river basin, Quang

Ngai province” aims at approaching a new point of view in flood management, in

order to achieve high efficiency in reducing the harmful effects of flood to the economic development and environmental protection of study area

2 Description of the Study Area

Quang Ngai province is located at the latitudes 14°32'- 15°25' North, longitudes 108°06' - 109°04' East, lean on Truong Son mountain range, overlooking the Eastern sea Quang Ngai abuts Quang Nam province in the North, Binh Dinh province in the South, Kon Tum province in the West, and Eastern sea in the East Located in the middle of the country, it is 883km from Hanoi capital to Quang Ngai and 838km from

Ho Chi Minh City

Due to the steep topography and poor vegetation, so the transfer speed of flood is very high, flood’s damage is huge Every year floods have caused extensive damage and losses of life and property 600,000 people affected by flooding, especially some communes which suffered flood inundation elevation more than 3 m with the frequency of floods are 20% and 10%

Figure 1: Administration map of Quang Ngai Province

Tra Bong river is one of the 4 biggest rivers of Quang Ngai province including Tra Bong, Tra Khuc, Ve and Tra Cau Tra Bong starts from the western mountains of Tra Bong District, running through Binh Son district to the sea at the mouth of Son Tra The length is 45 km, which flows from southwest to northeast The majority of the length of the river is mountainous terrain with the elevation of 200-1300m, the rest of river flowing in alternating hills and barren plains and sandy beaches

Upstream of Tra Bong river has many tributaries including rivers and streams, as Nun stream, Ca Du stream and Tra Boi river in Tra Thuy, Tra Giang communes In downstream, the eastern of Binh Son district is relatively high, so Tra Bong river has

no longer flow rapidly as upstream Tra Bong river has four 1st level branches At the downstream, there are some small rivers and streams enter the main river before flowing into the sea

Tra Bong river basin covers most of Tra Bong district and Binh Son district The catchment area is about 697 km2 The average elevation is 196m, average slope is 10.5%, river density is 0,43km/km2

More than half of the catchment area in the upstream is high mountain jungle, forest revival, the rest in downstream is barren hills interspersed with agricultural land Sandy areas are along the estuaries and coastal account for a small part

Table 1.1: The morphological features of Tra Bong river and major tributaries

Name of rivers or

tributaries

Length of rivers (km)

Length of basins (km)

Area of basins (km2)

Average width of basins (km)

- Water level measuring station network:

Currently, there are water level stations on all main rivers of the province However, only Tra Khuc river and Ve river measure water level throughout the year, Tra Bong and Tra Cau river measure only during flood season (September to December) Specific:

The station on Tra Bong river is Chau O, in the town of Chau O, Binh Son district

- Rain gauge station network:

There are three rain gauge stations that have been collected data for use in this study are Tay Tra, Tra Bong and Chau O station

The meteorological and hydrological situations of study area including rain and flood features are described and analysed in detail in Chapter III

3 Problems and Need of Study

Every year, Quang Ngai province suffered so many disasters, especially large floods

In the year 2009, an extreme flood caused severe flooding on a large scale of downstream areas of Quang Ngai Some river and tributary basins have peak flood elevation even higher than history flood in 1999 Through the statistic data, we see the annual flood damage caused hundreds of billions VND lost, hundreds of people were killed and injured in Quang Ngai province; especially the 2009 flood caused over four thousand billions VND lost and up to 51 people dead, 506 people injured The government at central and local levels took immediate action during and after the floods to assist the local communities to cope with the crisis and to help restore food

supplies, essential services, and essential infrastructure Further more, An Integrated

Natural Disaster Mitigation Policy for Central Vietnam has been prepared which

involves both structural and non-structural measures However, the non-structural measures are more concentrated and preferable because of the current condition of fund limitation in Vietnam In these non-structure measures, improvement of flood management measures is those of highest priority Therefore, I chose this study in order not only to learn myself knowledge of flood management, but also paving the way for the further researches, which are useful for local province

4 Objectives of Study

The overall objective of this study is to research large flood management measures for Quang Ngai province of Vietnam

To obtain this objective, some sub-objectives that need to be achieved are as follows:

• Analyse flood features of Tra Bong river basin, flood season and the typical flood

• Develop flood inundation mapping for Tra Bong river basin based on flood mark’s distribution and characteristics of the major flood in 2009

• Using rainfall – runoff hydrological model and 1-dimensional hydrodynamic model to simulate flood of Tra Bong river

• Flood forecasting experiment for Tra Bong river system

• Propose solutions for large flood management

socio-• Based on data collected, analysis flood mode of Tra Bong River basin include: flood season, flood characteristics (peak flood, flood intensity, floods hydrograph and typical flood process)

• Topographic data of the basin collected as a digital elevation model (DEM) Based

on the data of flood marks coordinates and elevation of 2009 major flood, using GIS tools to take position of flood marks on the map, thereby creating flood inundation maps for the downstream area of the basin

• Using rainfall – runoff hydrological model to simulate the process of flow forming from data of rainfall gauges in the basin The result is the hydrograph at the position of last section in the upstream part of Tra Bong river

• Using hydrograph calculated from rainfall–runoff model as the upstream boundary

to simulate flood flow in the downstream The model used is 1-dimensional hydrodynamic model

• Applying the hydrological – hydrodynamic combine model to forecast flood for Tra Bong river downstream Using the measured data of several floods to verify the forecasting reliability of the model

• Propose solutions for large flood management of the basin

CHAPTER I: LITERATURE REVIEW

There are many researches on flood management using GIS tools in Vietnam and around the world This part of the study summaries the review of related literatures in structured way shown below:

Figure 2.1: Structure of literature review

1.1 Related researches about study site

The study area is the Tra Bong River Basin has some characteristics that can cause major flooding as the total rainfall in the basin is big, the basin is short and steep Characteristics of the terrain and climate are also common features of the central provinces of Vietnam in general and the river systems of Quang Ngai province in particular Thus, the study of floods in this area is relatively respected by scientists and the local government There were many studies of floods and disasters in Central provinces, Quang Ngai or even in Tra Bong river basin in recent years, which are very good references for the research objectives of the thesis

GIS in hydrology and water resources

Flood Forecasting

Integration of GIS and Hydrological/Hydraulic Models for

Large Flood Management

In 2000-2001, the Vietnam Institute of Meteorology, Hydrology and Climate Change has performed a national independent reseach: "Investigations, studies and flood warning support disaster prevention in the Central Basin" (Du, 2001) This research played an important role in hydro – meteorological research activities supporting disaster prevention for the region, as the basis for subsequent detailed studies

Within the scope of Quang Ngai province, there were some studies on flooding in recent years as follow:

Disaster Mitigation Project in Quang Ngai province was supported by the Australian Government and implemented from 2003 to 2007 One of the contents of this project was constructing hydraulic model for floodplain management Hydraulic model was a tool to verify the management plans include floodplain basins of Tra Bong, Tra Khuc,

Ve and Tra Cau rivers Constructed hydraulic model was a detail 2D model, using SOBEK software from Delft Hydraulics and some other supporting software products

In addition, the hydraulic model was also used to examine the impact of future development to flooding

In 2010-2011, Institute of Geography - Vietnam Academy of Science and Technology has implemented the research: Research on flood prevention and drainage planning of Tra Khuc river and Ve river in Quang Ngai province (Huong, 2011) This study has been taken follow the objectives:

• Setting up the database of floods and flood evolution supporting flood prevention and drainage planning of Tra Khuc river and Ve river in Quang Ngai province

• Determination of flood drainage way for estuary deltas (from the Thach Nham dam

to Dai estuary in Tra Khuc River; from Hanh Tin Dong to Lo estuary in Ve river) corresponding to a frequency of 1%, 5% , 10%

• Proposed planning options of flood prevention and drainage planning of Tra Khuc river and Ve river in Quang Ngai province

In particular, research and management agency on hydrometeorology in the region is the Hydrometeorology Department of Mid-Central region has taken some specific researches on floods related to Tra Bong river basin, as follows:

In 2000-2001, implemented the research: Develop flood zoning maps and forecasting - warning flood plans for rivers in Quang Ngai province (Lien, 2001) The study has assessed the flood regime of Tra Bong, Tra Khuc and Ve river basins, survey flood situation in 1999, thereby mapped downstream inundation on the rivers for the 1999 large flood and also flood risk maps corresponding to the design frequency This study has also built the first 32 flood marks in Quang Ngai province

In 2006-2007, implemented the research: Additional investigation and flood mark building in floodplains of main rivers downstream in Quang Ngai province (Lien, 2007) Results of study were the new flood alert levels on the three main rivers of the province included Tra Bong, Tra Khuc and Ve rivers The flood alert levels proposed

by research had been officially used due to Decision No 632/QĐ-TTg on 10/5/2010 of the Prime Minister and Decision No 1142/QĐ-UBND on 25/8/2010 of the President

of the Committee of Quang Ngai province In addition, the study continued to build additional 32 flood marks, the flood risk maps base on the flood alert levels Most of the results of study were used for flood prevention in recent years

Expanded and continued previous studies, (Chien & Thiem, 2012) had researched about flood in Tra Bong river in order to correcting flood risk map and flood warning mark following the new alarming levels In this research, the authors using GIS technology for mapping and the research aim at correcting the warning marks, which are often used in hydrological stations system, and they are very useful for farmer themself see the warning level of flood The research had a good overview about flood

in Quang Ngai province, but it did not analyze much about flood management, especially large flood However, the research helps us to have an overview on the situation of floods in the area of research and is the basis for us to compare, evaluate the accuracy of the data collected

In another research, (Hanh, 2014) has simulated flooding in downstream area of Tra Bong river The major research area is Binh Son district, where the Highway 1A running through The main content of the research is using some 1D and 2D hydraulic models to simulate flood flows and inundation calculations for the study area, which proposed solutions in terms of both construction and non-construction to prevent flood inundation for the 1A national Highway section running through Quang Ngai Research has simulated very specific and potentially apply to the local However, the content of the study did not focus on flood management issues and the main objective

is to develop solutions to prevent flooding for the highway

Research topics related to flood management in Tra Bong river basin and Quang Ngai province were mainly carried out by the local authorities as Provincial People's Committee, Regional Hydro-meteorological Department, Center for Disaster Prevention and Reduction In flood management solutions had been proposed, solutions of downstream flood forecasting from measurable rainfall data and downstream inundation mapping got high efficiency There were two main methods used in studies in the region, including:

• Traditional methods: survey, collect marks of the historic floods, based on the distribution and elevation of flood marks to build inundation maps Construct the correlation between rainfall data in the upstream and downstream flood levels, thereby develop the flood alert levels to warn local residents The researches using this method include a number of researches implemented by Hydrometeorology Department of Mid-Central region (Chien & Thiem, 2012; Lien, 2001, 2007)

• Modelling methods: Apply the combination hydrological rainfall – runoff models and hydrodynamic models to simulate the process of flow forming in the basin using rainfall data from rain gauges and flood flow progression in the river downstream Apply GIS tools to build flood map for each flood alert level to provide spatial view of the flood risk to the user Typical researches using this

method is (Disaster Mitigation Project in Quang Ngai province, 2007; Hanh, 2014;

Huong, 2011)

Each methodology had advantages and disadvantages However, despite using any method, the studies of flood in this area had great impacts on socioeconomic development in localities, especially in flood warning and forecasting on river basins The studies also provided an overview on the situation of flood and natural conditions, topography, climate and other features for the author to perform this study

1.2 Flood inundation mapping

In studies on flood management in Vietnam and around the world, the construction of flood inundation maps plays a very important role Flood inundation maps provide a spatial view of the possibility of flooding for each region, thereby plan for preventing and mitigating the damage caused by floods The construction of flood inundation maps using many methods: Investigation and survey method, statistical method, system analysis method, modelling method etc or the combination of two or more methods, depending on the requirements and objectives of the research

Studied on large floods and large flood management solutions, especially extreme floods, (Büchele et al., 2006) aim to enhance existing methods for hazard and risk assessment for extreme flood events In this study, the authors analysed the process to develop flood risk maps in three major steps:

• Regional estimation of flood discharges (basin-, site-specific hydrological loads)

• Estimation of flow characteristics in potential inundation areas (local hydraulic impacts)

• Estimation of the resulting damages (area- or object-specific risk assessment)

Furthermore, the authors also analysed two types of requirements in developing flood risk map The first are minimum requirements on data and methods for a standard quality of hazard and risk assessment on local scale The second are more sophisticated approaches which require more spatial information and more complex calculations up to fully dynamic simulations of unobserved extreme flood situations The overview of suitable approaches for these steps of hazard and vulnerability assessments is given in Table 1

Table 2.1: Overview of methods and data for high-resolution flood-risk mapping in

Germany (Büchele et al., 2006)

In the study of (Duy Kieu, 2012), the flood risk maps is developed by a combination of statistical method and system analysis method The map was developed through the following steps: 1 Zoning basin into smaller areas base on the location of monitoring stations 2 Develop criteria to assess the risk of large floods occurred in each region (the criteria are mainly based on the measured meteorological and hydrological data)

3 Scoring for each region based on the above criteria 4 Using GIS to establish flood risk maps This method has the advantages of simple, easy to calculate, easy to carry, mainly based on the systems analysis and synthesis However, the study area of the research should be relatively large and the number of monitoring stations is large enough to be able to partition flood risks in detail and with high accuracy In my research, I referred the method of statistical data analysing on floods and hydrological characteristics of (Duy Kieu, 2012) to analyse flood features of Tra Bong river basin

(Tu & Tingsanchali, 2010) had made a research to create flood risk map for Hoang Long river basin Researchers have used the method of flood flow simulation by hydraulic model combined with the use of GIS tools to develop flood risk maps The hydrologic data collected were good and the topographic data using DEM 90x90 The research used 1D Model (Mike 11) to run the simulation of creating flood flow To modeling 2D flood process, there is an addition method that called FP4 of DHI had been used

In an aggregate and systematic research, (de Moel, van Alphen, & Aerts, 2009) gave

an overview of existing flood mapping practices in 29 countries in Europe and shows what maps are already available and how such maps are used The research indicated that flood maps existed in many different forms, but in general it was possible to distinguish between flood hazard and flood risk maps Flood hazard maps contained information about the probability and/or magnitude of an event whereas flood risk maps contained additional information about the consequences (e.g economic damage, number of people affected) Within these two general types, however, there were different methods available to quantify hazards and risks, resulting in different types of flood maps (Fig 2.2)

Figure 2.2: Conceptual framework for flood hazard and risk calculations (de Moel et

al., 2009)

To represent flood situation, there were many ways to create flood maps (de Moel et al., 2009) had listed a total of six types of flood maps The maps selection depended on the requirements and objectives of the research, because each type of map represented

a different characteristic of flood and different use value The examples for all map types were shown in Figure 2.3

Figure 2.3: Different flood map types (A) historical flood map; (B) flood extent map; (C) flood depth map; (D) flood danger map; (E) qualitative risk map; (F) quantitative

risk (damage) maps (de Moel et al., 2009)

To assess flood damage to the study site takes into account the vulnerability of the protected objects, researches often build the flood inundation maps (flood hazard maps) and the flood risk maps The investigation and gathering socio-economic data to develop flood risk maps often need more time Meanwhile in order to focus on the work of flood warning, forecasting and large flood management, the flood inundation maps can relatively meet the requirements of the topic Thus in our research, the author chose to build flood elevation maps base on the distribution and elevation data

of 2009 historic flood marks This flood inundation map will help readers have spatial view on the possibility of flooding in downstream areas of Tra Bong river basin

1.3 Geographical Information Systems in Hydrology and Water Resources

In the past few decades, the development in computing capacity of the hydrology has created significant development about hydrology research and operational Geographical Information Systems (GIS) introduction has created a powerful tool to calculate and represent the result of hydrological studies, particularly flood management From the first years of application of GIS in hydrology and water resources management, (Singh & Fiorentino, 1996) in his research proposed a concept

of GIS: GIS is a general-purpose computer-based technology for handling geographical data in digital form It is designed to capture, store, manipulate, analyse and display diverse sets of spatial or geo-referenced data

In hydrology and water resources studies, concern for resource management, and environmental quality requires application of distributed models One characteristic of distributed-parameter models is that they are data intensive Multiple data types such

as hydrometeorology, topography, land use, soils, geology, streamflows, etc are commonly required Many of these data are often used to identify a hydrologic unit, because application of these models requires partitioning of the watershed into homogeneous units This often proves to be cumbersome Therefore, a spatial data analysis and manipulation tool is desirable and that tool is GIS Furthermore, the use

of computers in hydrologic modelling has become so widespread that the marriage between hydrology and GIS should be a logical step (Singh & Fiorentino, 1996)

The book “Hydrological Applications of GIS” (Gurnell & Montgomery, 2003) is based on the article was published in the journal "Hydrological Processes", focusing

on the main theme is the development and application of geographic information system (GIS) to address hydrological problems During the period before this book was published, the application of GIS in environmental modelling has flourished, utilizing the advantages of linking the data types of GIS with models

In the book, the article shows that the application of GIS in hydrology is very diverse, while scientists have simulated the hydrological processes through models, from common characteristics model to semi-distributed and distributed model The management of water resources also followed a similar path to increase the spatial resolution, especially on infrastructure With both objects, GIS is useful because it can provide data with higher resolution Furthermore, various applications of hydrology often require controlled by multiple users by access to common databases, so the database must support GIS, data quality and how management database is the "heart" for the development of GIS applications With the development of digital elevation models (DEMs), the most widely GIS application in hydrology is defining catchment flow and overflow areas based on terrain types and combinations of them with hydrological models

In recent years, the development of GIS technologies and the availability of a wide range of geodatasets provide great potential for modeling physiographic objects, in particular, river systems An adequate automatic method of river network digitizing was developed and tested using the digital elevation model based on applying the Complex Energy Index (CEI) in a research of (Gartsman & Shekman, 2016) The index form includes the independent parameterization of main geomorphologic and climatic factors of the first-order stream generation The method testing proved that it

is efficient and the results of its use can be reliably interpreted The software was developed on the basis of ArcGIS tools It implements the full algorithm of the automatic digitizing of river network using the digital elevation model It also allows advancing studies on this issue, and could be used for solving applied problems

Studies have shown that GIS is an effective tool to classify maps for each individual characteristic according to the needs of the user Thereby, the user can use the math operations to calculate between the attributes of each layer such as addition, subtraction, filtering, or set conditions to create a new map layer as required by the study These features of GIS are useful in creating flood maps of the study area, and can also display maps in accordance with the standards for presentation and publication

1.4 Flood Forecasting

The main objective of this research is to find a suitable flood forecasting method for Tra Bong river basin, in terms number of monitoring stations is limited Thus, the review part first examined flood forecasting methods that used in studies in the country and around the world to get the most appropriate selection In a report of a project implemented by Cooperative Research Centre for Catchment Hydrology – Australia (Srikanthan, Elliott, & Adams, 1994) had determined that flood forecasting methods were considered under two separate heading: rainfall – runoff methods and flood routing methods

• Real-time forecasts of discharge, obtained by rainfall – runoff modelling, are generally less accurate than those obtained by channel routing of a hydrograph observed at an upstream gauging site However, real-time forecasting methods based on rainfall-runoff modelling are necessary because in head water basins there

is no alternative since upstream stations do not exist and in some circumstances it may yield forecasts with greater lead time

• The second heading, flood routing methods offers a satisfactory means of flood forecasting for long river systems For this approach to be successful, the travel time

of flood peaks from upstream to the downstream site needs to be long enough to allow adequate period of warning Flood routing can be classified under four methods: experience methods, statistical methods, hydrologic routing methods and hydraulic routing methods

Using three of four methods mentioned above: statistical, hydrologic and hydraulic routing methods, (Duy Kieu, 2012) in his engineering doctoral thesis had researched a fully procedure of flood forecasting, in order to develop large flood management solutions for Lam river basin By statistical methods, he used the rainfall data, discharge and water level data of all monitoring stations in the basin to analyse the characteristic of climate and flood in study area Thereby, he analysed and partitioned flood risk by give marks for each sub-area of study site By the marks, using GIS tools,

he could develop the flood hazard map for the basin And applied hydrologic and

hydraulic routing methods, he simulated the flood flow in the river using some hydraulic modern as Mike 11, Mike Flood and Mike 21 In this study, the combination

of flood problems in the river system is very complex But the author stated that this is crucial to develop the scientific basis for flood management Research of the combination of large flood on Lam river system in this thesis focused on the issues: The total amount of flood water, the meeting of the flood peak flow from upstream to downstream, the origin of flood water from tributaries

Referring to the role of flood forecasting in flood management, (WMO, 2011) indicated that flood forecasting is a necessary part of flood management, given that no preventative or defence measures can be completely effective The reality of economic limits to the provision of defences, together with the possibility that the capacity of defence systems may be exceeded or they may be fail, require that other measures are

in place Provision of flood forecasting will also form part of flood management planning and development strategies, which recognize that there are occupied flood plain areas where non-structural measures can be effective

This manual book provided the basic knowledge and guidance to develop or to set up

an appropriate and tailored system for any case in which a flood forecasting and warning system is required The Manual described the various components of a flood forecasting and warning system, which were:

• Design of a flood forecasting system;

• Implementation and operation of a flood forecasting system;

• Flood warnings;

• Training

The Manual does not set out a step-by-step process for the design of a flood forecasting and warning system along the lines of a particular template or practice in any one country Rather, presented in all chapters are a number of examples of different practices and technologies, which may reflect different levels of development, ranges of needs and also capacities in a number of different situations

Figure 2.4: Process for developing a flood forecasting model (WMO, 2011)

An important part of flood forecasting is selecting a model or a combination of models

to appropriately simulate the process of flood flow forming in the river Hydrological rainfall-runoff model was selected to simulate the process of flow forming from rainfall data when the data series of flow, water levels and the number of monitoring stations is limited The structured and methodology of this model are relatively simple,

so it is not too complicated to select

To simulate the flood flow process in downstream river, a 1-dimensional (1D) or dimentional (2D) hydraulic model can be used 2D model gives more detailed results, provide a full range of information on flood flows, such as flood depth, flood velocity, flow direction, and easily combined with GIS tools to build the flood maps However,

2-2D model has high requirements on the input data, especially the topography data (usually high resolution DEM); time to run the model is long and requires a very strong computer In contrast, 1D model gives less detailed results and more difficulty combined with GIS tools for flood mapping, but it requires less data over terrain (just the data in the cross sections) and model running time is shorter In addition, to build flood maps from the results of 1D model, an additional tool (such as FP4 of DHI) can

be used to interpolate and make a map with obtained results of cross sections

To evaluate the efficiency of 1D and 2D numerical models for predicting river flood inundation, (Horritt & Bates, 2002) had made a test for 1D and 2D models of flood hydraulics (HEC-RAS, LISFLOOD-FP and TELEMAC-2D) on a 60 km reach of the river Severn, UK The three models are calibrated, using floodplain and channel friction

as free parameters, against both the observed inundated area and records of downstream discharge About the data, while HEC-RAS only needed 19 cross-sections of the channel, LISFLOOD-FP and TELEMAC-2D, the two 2D models need 50 m resolution DEM for entire channel and floodplain The results of testing are shown in Table 2

Table 2.2: Predictive performance of the 3 models using independent

calibration/validation data (Horritt & Bates, 2002)

In terms of simulations where the model is calibrated against the data it is trying to predict, the TELEMAC-2D marginally outperforms the other models apart from for the 2000 inundation data In terms of predictive performance (Table 3), the HEC-RAS model performs marginally better (on 4 out of 6 measures) than TELEMAC-2D, which shows the best performance on 2 measures and which in turn is better than LISFLOOD-FP LISFLOOD-FP does not produce the best predictive performance for any of the measures and data sets examined It is interesting to note that TELEMAC-2D is better when predicting inundation for the 1998 than HEC-RAS, the situation

being reversed for the 2000 event Finally, study had shown that HEC-RAS was the best model in flood forecasting for the study area, with the terms and requirements stated in the study

CHAPTER II: APPROACH AND METHODOLODY:

2.1 Approach of study

Figure 3.1: Approach of study

2.2 Flood inundation map

Flood maps and flood risk maps are effective tools for the preparedness of floods prevention, as well as in the emergency response phase These maps are also important materials for disaster prevention planning tasks For Central rivers in general, and

Flood forecasting experiment

Propose solutions for large flood management

rivers of Quang Ngai province in particular, the characteristics are short and steep, fast water concentration time, so these maps are fastest sources of information about the scale and level of flooding, helpful for the evacuation if flood occurs

Flood maps give the information about flooding and non-flooding areas, and also flood depth Besides information on flooding, the maps also show the other basic information, such as landmarks, transportation, resident, river Flood maps give a look from general to detail of flood level due caused by a certain rain

The use of flood maps built with some certain scenarios will help local authorities actively choose quickly responding plan when there are information of flood warnings, forecasting in downstream

Flood maps can be classified into two basic types:

• Type 1: Map of flooding caused by historic floods, such as the floods in 1999, 2009 These maps show the real information of flood so they support the most objective forecasting and warning for residents These maps are also useful in verifying the other flood forecasting maps

• Type 2: Map of flooding caused by designing floods The basis for determining the flood on this map is the designing floods with the designing frequency of rainfall or discharge in one or more gauging stations on rivers In other words, these maps are built according to some rainfall - flood scenarios

Most of the flood maps are built on the basis of the depth of flooding Flooding depth was determined from the flood water elevation and the ground level, by the following formula:

Where:

h i : Flood water depth;

H w: Flood water elevation;

Z t: Ground level

H w and Z t have to be in the same elevation system

Thus, it is necessary to determine the flood water elevation and ground level at a position to calculate the depth of flooding there

Ground level is developed from topographic maps or field measurements It is a very important factor for the flood simulating results The detail level of the topography data will affect the accuracy of flood risk maps If the point density is not enough, not fully reflect the transformation of the terrain, the floodplain simulations will not be detail Some non-flooded areas with be considered as flooded areas, and vice versa Flood water elevation: For a map of the historic flood, flood water elevation can be investigated and actual measured, combined with computational methods For flood maps corresponding to a designing flood, the flood water elevation is determined by the hydrological and hydraulic model, and calibrated, verified with observed data

2.3 Using GIS tools to develop flood inundation map

The inundation map will be created using Arc GIS software (Arc Map 10; ESRI Inc - http://www.esri.com) Arc GIS includes very powerful tools to manage, update, analyse and publish information to create a complete Geographic Information System (GIS), allows:

• Create and edit integrated data (integrating spatial data with attribute data), allows use a variety of different data formats and even the data obtained from the Internet;

• Query spatial data and attribute data from many sources and in many different ways;

• Display, query, and analyse spatial data in combination with attribute date;

• Create specific map and print with professional presentation quality

Topographic data used in Arc Map as contours or elevation grid - often presented square or triangle grid (this type of data is often called the Digital Elevation Model – DEM)

Figure 3.2: Digital Elevation Model with square grid (Source: www.p-gis.com) Geographic factor determines the classification of submerged depth For example, if the terrain with contour different is 5 m, the flooding boundary determining will have significant error because some small but high zones between 2 contours can be considered to be flooded, but in fact, it is not (Figure 3.4)

Figure 3.3: Error based on the topography data in flood inundation mapping

In general, to reduce errors due to topography data, elevation difference between the contours should be equal to or less than the elevation difference of submerged depth classification If submerged depth classification is 1m, the corresponding elevation

difference between the contours on the map is also 1m or less According to Vietnam

Construction Standard 309:2004: Surveying work in building construction - General requirements (MOC, 2005), the scale of map must be 1/2000 or more detail (Table

3.1) The average point density for creating a 1/2000 scale topography map is from 12

to 18 points per a 1 square km

Table 3.1: Technical requirements of elevation different between the contours with

2.4 Mike 11 general description

The research focuses on using rainfall – runoff hydrological model (NAM) and dimensional hydrodynamic model (Mike 11 HD) to simulate the process of flow forming in the basin, thereby experiment flood forecasting for Tra Bong river system The Theoretical Foundations of rainfall – runoff hydrological model (NAM) and 1-

1-dimensional hydrodynamic model (Mike 11 HD) are mainly based on the book Mike

11 A Modelling System for Rivers and Channels Reference Manual (DHI, 2007)

MIKE 11 is a professional engineering software package for the simulation of flows, water quality and sediment transport in estuaries, rivers, irrigation systems, channels and other water bodies MIKE 11 is a user-friendly, fully dynamic, one-dimensional modelling tool for the detailed analysis, design, management and operation of both simple and complex river and channel systems With its exceptional flexibility, speed and user friendly environment, MIKE 11 provides a complete and effective design environment for engineering, water resources, water quality management and planning applications (DHI, 2007)

The Hydrodynamic (HD) module is the nucleus of the MIKE 11 modelling system and forms the basis for most modules including Flood Forecasting, Advection-Dispersion,

Water Quality and Non-cohesive sediment transport modules The MIKE 11 HD module solves the vertically integrated equations for the conservation of continuity and momentum, i.e the Saint Venant equations

Applications related to the MIKE 11 HD module include:

• Flood forecasting and reservoir operation

• Simulation of flood control measures

• Operation of irrigation and surface drainage systems

• Design of channel systems

• Tidal and storm surge studies in rivers and estuaries

The primary feature of the MIKE 11 modelling system is the integrated modular structure with a variety of add-on modules each simulating phenomenon related to river systems

In addition to the HD module described above, MIKE 11 includes add-on modules for:

• Hydrology

• Advection-Dispersion

• Models for various aspects of Water Quality

• Cohesive sediment transport

• Non-cohesive sediment transport

MIKE 11 HD applied with the dynamic wave description solves the vertically integrated equations of conservation of continuity and momentum (the Saint Venant equations), based on the following assumptions:

• The water is incompressible and homogeneous, i.e negligible variation in density

• The bottom-slope is small, thus the cosine of the angle it makes with the horizontal may be taken as 1

• The wave lengths are large compared to the water depth This ensures that the flow everywhere can be regarded as having a direction parallel to the bottom, i.e vertical accelerations can be neglected and a hydrostatic pressure variation along the vertical can be assumed

• The flow is subcritical

Continuity equation (Conservation of mass, in differential form):

dt dx t

Q dx

dA dt

dx x

Q Q

Where:

t

A t

Momentum equation (Conservation of momentum, in difference form):

x

F x

F x

P x

U M t

Momentum: M =ρ.H.b.U

Flux: M f =ρ.H.b.U.U

Pressure:

2 2

1

H b g

P= ρ

2

C

U g b x

Figure 3.6: Channel section with computational grid.

MIKE-11 can handle multiple branches, and in the tributaries where branches meet A node will be created in which the water level is calculated The shape of the computational grid around a node which has three branches is shown in Figure 3.7

Figure 3.7: The shape of the computational grid around a node which has three

The shape of the grid points and the nodes of a complete model is shown in Figure 3.8

It should be noted that at the border point, we create a node that we can calculate the water level

In a grid point, the relationship between the variables Zi (including water level hi or flow discharge Qi) at the point and at the neighboring point is represented by using a linear equation as follows:

j n

j j n j j n j

+ +

+

1 1

1 (3.10)

The lower number in the equation (3.10) indicates the position along the branch and the upper one indicates time step The coefficient α, β, δ and γ in the equation at the point h is calculated by implicit finite difference of continuity equation and at the point

Q by implicit finite difference of momentum equation

At every grid point along the branch, the equation (3.10) is created Suppose a branch

have n grid points; n is an odd number, so the beginning and end grid points of a

branch is always water level points (h) This creates n linear equations with n + 2 unknowns Two additional unknowns are from the equations created at the beginning and the end points, at which Zj-1 and Zj+1 respectively become the water level values at the node, whereby the upstream and downstream of the river are connected The following section describes the linear equation:

(3.11)

Hus in the first equation and Hds in the final equation are respectively water levels in the upstream node and the downstream nodes In MIKE-11 compatibility water level is applied at nodes, that means the water level at the first point of a branch is the water level at the node, whereby the last section of upstream of the branch is connected In other words, h1 = Hus This means α1= -1, β1=1, γ1=0 and δ1=0 Similarly, the final grid point at which hn = Hds so αn=0, βn=1, γn= -1 and δn= 0 In Figure 3.9, this corresponds to H = HA,n = HB,n = HC,1

If contact a branch system to the water level at the end of each branch, then we will know Hus and Hds There are only n unknowns in n equation and we can solve them using standard reduction algorithm However, because MIKE-11 can handle multiple branches so we apply another approach To explain this problem, the above equation will be presented in the matrix in Figure 3.9

Figure 3.9: Branch matrix before reducing

By standard reduction method, we can change matrix 3.9 to matrix 3.10

Figure 3.10: Branch matrix after reducing

By this matrix we can see at any grid point, variable Z (water levels and flows) is represented as a function of the water level and flow upstream and downstream:

1 1

ds j n us j j n

j c a H b H

Z (3.12)

Figure 3.11: Three-branch node with limit for continuity equation

At the intersection, the continuity equation at the node of intersection is set as follows:

2 / 1 2 / 1 1

+ +

I fl

n n

Q Q

A t

H H

(3.13)

2 , 1

1 , 1

1 , 2

, 1 , 1 ,

1

5.05

+

−+

+

−+

=

∆

C n

n B n

n A n

C n

n B n

n A fl

n n

Q Q

Q Q

Q Q

A t

H H

In equation (3.14) QA,n-1, QB,n-1 and QC,2 at time level n+1 can be substituted according

to equation (3.13), and the following equation is achieved: