THE URBAN SPATIAL STRUCTURE OF EMPLOYMENT AND ITS IMPACTS ON TRANSPORTATION AND URBAN DEVELOPMENT: A CASE STUDY OF MEGACITY

Ahsanul Kabir

A thesis in fulfilment of the requirements for the degree of Doctor of Philosophy

Faculty of the Built Environment University of New South Wales Sydney, Australia

July 2013

i

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………......

Date ……………………………………………......

ii Abstract

In recent times, the Asian region has become a central theme in debates on economic growth and urbanization. Asian urbanization shows a distinctive character with accumulation of resources to a few selected cities and, in turn, rapid growth to megacity status. The nature and the scale of urbanization towards these megacities has posed complex and demanding challenges for planners and policymakers dealing with issues of sustainable urban development. One of the most demanding challenges is inadequate understanding of the urban spatial structure – allocation of activities in the urban landscape – of megacities. There are several schools of thoughts on the urban structure of North American, European and Australian cities and what types of urban form may generate more longer term urban and oriented sustainable solutions. Much of the western debate is based on solid empirical analyses that utilize very rich data from census sources, journey-to-work statistics and land-use distributions – data that enables modeling of future urban growth and urban structure patterns for analysis of more optimal outcomes. By contrast, similar empirical analyses of Asian megacities are few and far between; mainly due to the unavailability or absence of similar data that is relatively up-to- date for enabling explorations of urban spatial structure. This thesis attempts to make a contribution to this critical gap by utilizing various official data sets that have recently been made available by consulting teams and authorities in the megacity of Dhaka, .

The main focus of this thesis is an examination of the urban spatial structure of the megacity of Dhaka, and it argues that the fundamental underlying spatial structure of this city is likely to be very different to that of western cities due to different levels of development and socio-economic conditions. The nature and patterns of this underlying urban spatial structure are assessed through a detailed examination of employment distribution and its travel related impacts, in particular the functional relationships between the residential locations of workers, their job destinations and their actual journey-to-work trip patterns. This, in turn, allows assessment of implications for present and future sustainability of the city, and whether a different

i spatial structure – from a policy and planning perspective – may accommodate more sustainable growth into the future. An attempt is made to identify if the urban spatial structure of Dhaka is characterized as being monocentric, polycentric or falls along a continuum between the two extreme spatial structures, and to compare this with the identified structure of some other Asian megacities.

One key aim of this research is to contribute to the debates on the role of future urban spatial structure in generating a sustainable future for Dhaka. The approach taken in the thesis is that the optimum way of making this contribution is to explore and understand the land use-transport dynamics of Dhaka megacity and the potential development options for the city. Several growth scenarios are generated for the year 2024, and their travel consequences are estimated using transportation models. This research utilizes very rich and diverse datasets for Dhaka which includes the national census, data collected for the urban master plan and data collected and generated as part of the strategic transportation studies carried out in the late 2000s.

The results strongly indicate that the megacity of Dhaka is characterized by an extended monocentric urban structure as opposed to a polycentric one. There do exist very high density employment centres but these are located in relatively close proximity to the main city centre and found in high density areas that surround the main city. The three employment centres located in the outer ring dominate their respective catchment areas but have relatively low densities of employment. These three centres have the potential to develop into fully functional polycentric centres in the future.

The results also suggest that Dhaka is likely to experience higher levels of congestion and pollution unless job decentralization occurs to the 19 identified employment centres along with more sustainable transportation arrangements. Estimates are generated of significant changes in travel costs and vehicle emissions under different decentralization scenarios. The analysis suggests that spatial decentralization of employment might increase average commuting length to the employment centres in the outer ring as more people would travel to outskirts. However, it would reduce up to around 30% of vehicle emissions from commuting in the main city. In other words, job decentralization redistributes emissions and,

ii hence, would improve air quality in the main city. This study calls for proper policy and planning initiatives to expedite job re-distribution to potential job centres outside the main city and to enhance the growth of these centres to serve the megacity in a more sustainable way.

iii TABLE OF CONTENTS

Abstract ...... ii List of Figures ...... ix List of Tables ...... xiii List of Abbreviations ...... xv Acknowledgements ...... xvi

Chapter 1: Introduction

1.1 Introduction ...... 1

1.2 Urbanization and Asian Megacities ...... 5

1.3 Urbanization in Bangladesh ...... 7

1.4 Statement of the Problem ...... 12

1.5 Aims, Study Area and Scope of the Research ...... 13

1.6 Goals, Objectives and Research Questions ...... 15

1.7 Structure of Thesis ...... 18

Chapter 2: Urban Form, Travel and Employment Centres of Mega Cities

2.1 Introduction ...... 21

2.2 Urban Growth and Urban Form: Concerns of Growth Management ...... 22

2.3 Urban Spatial Structure and Transportation: Issue of Sustainability ...... 28

2.4 Modelling Urban Spatial Structure and Sub-Centre Identification ...... 39

2.5 Applications of Urban Structure Studies ...... 46

2.6 Issues In Methodology and Measurement of Polycentricity ...... 49

2.6.1 Employment Density Peaks ...... 51

2.6.2 Reference Threshold Method ...... 52

iv 2.6.3 Rank Size and Clustering Methods ...... 55

2.6.4 Parametric Method ...... 57

2.6.5 Nonparametric Method ...... 57

2.6.6 Other Methods ...... 59

2.7 Conclusion ...... 61

Chapter 3: Urban Evolution and Planning

3.1 Introduction ...... 63

3.2 Prominence of Dhaka and its Economic Activities ...... 64

3.3 History of Urban Evolution in Dhaka ...... 69

3.3.1 Settlement Before and After the (1608-1757) ...... 70

3.3.2 Dhaka under British Rule (1757 – 1947) ...... 73

3.3.3 Dhaka as the Capital (1947 – 1971) ...... 75

3.3.4 Dhaka as the Capital of Bangladesh (since 1971) ...... 76

3.4 Planning Initiatives for Dhaka Megacity...... 77

3.4.1 Dacca Town Planning Report (1917) ...... 77

3.4.2 East Pakistan Planning Sub-Committee (1948) ...... 77

3.4.3 Dacca Master Plan (1959) ...... 78

3.4.4 Dacca Integrated Urban Development Project (1981)...... 83

3.4.5 Greater Dhaka Metropolitan Area Integrated Transport Study (1994) ...... 84

3.4.6 Dhaka Metropolitan Development Plan (1995) ...... 86

3.4.7 Strategic Transport Plan (2005) ...... 89

3.5 Urban Development, Spatial Growth and Urban Form of Dhaka ...... 93

3.6 Transport in Dhaka ...... 98

3.7 Environmental Conditions in Dhaka ...... 104

3.8 Conclusions ...... 106

v Chapter 4: Research Design

4.1 Introduction ...... 108

4.2 Conceptual Framework and Research Questions ...... 108

4.3 Study Area: Dhaka Megacity ...... 112

4.4 Data and Data Sources ...... 116

4.4.1 Demography ...... 117

4.4.2 Travel Data and Transport Model ...... 118

4.4.3 Landuse and Building Footprint ...... 118

4.5 Methods and Tools ...... 119

4.5.1 Data Pre-processing ...... 122

4.5.2 Spatial Structure of Employment ...... 122

4.5.3 Employment Centre Identification ...... 124

4.5.4 Growth Scenarios ...... 129

4.5.5 Commuting Distribution ...... 132

4.5.6 Travel Impacts ...... 134

4.5.7 Estimation of Vehicle Emissions ...... 135

4.6 Conclusion ...... 138

Chapter 5: Spatial Structure of Employment in Dhaka

5.1 Introduction ...... 140

5.2 Employment Profiles in Dhaka ...... 141

5.3 Spatial Clustering of Workforce ...... 148

5.4 Job Location and Distribution ...... 156

5.5 Employment Centres in Dhaka ...... 161

5.6 Conclusion ...... 165

vi Chapter 6: Spatial Interaction of Employment Centres in Dhaka

6.1 Introduction ...... 167

6.3 Employment Centres and Commuting Behaviour ...... 167

6.4 Features and Spatial Impacts of Identified Employment Centres ...... 175

6.5 Residential Workforce Profile of the Employment Centres ...... 179

6.6 Conclusion ...... 180

Chapter 7: Exploring Growth Options for a Sustainable Dhaka

7.1 Introduction ...... 182

7.2 Modelling Future Growth Scenarios for Dhaka ...... 183

7.3 New Employment Landscape ...... 185

7.5 Commuting Patterns ...... 188

7.4 Travel Time and Cost ...... 191

7.6 Transport Vehicle Emissions ...... 192

7.7 Conclusion ...... 197

Chapter 8: Policy Implications for Sustainable Urban Structure

8.1 Introduction ...... 200

8.2 Urban Development in Dhaka Megacity and Relevant Policy Issues ...... 202

8.2.1 Demographic and Economic Plan ...... 203

8.2.2 Urban Development and Planning ...... 204

8.2.3 Transport and Infrastructure ...... 206

8.2.3 Environment ...... 207

8.3 Urban Governance for Dhaka Megacity ...... 209

8.4 Urban Centre Hierarchy and Local Capacity ...... 211

8.5 Conclusion ...... 213

vii Chapter 9: Conclusions

9.1 Summary of Research Findings ...... 214

9.3 Limitations of the Thesis ...... 220

9.3 Scope for Future Research ...... 222

References 223

Appendix A: List of Roads and Highways Under Construction and Proposed Projects ...... 250

Appendix B: Maps of Current Situation ...... 254

Appendix C: Growth Scenarios ...... 262

viii LIST OF FIGURES

Figure 1.1: Megacities in the World ...... 6

Figure 1.2: Population growth in Bangladesh ...... 8

Figure 1.3: Population Explosion of Dhaka Megacity ...... 10

Figure 1.4: Study Area ...... 14

Figure 2.1: Spatial Distribution of Population in Selected Large Cities ...... 25

Figure 2.2: Theoretical Urban Forms ...... 26

Figure 2.3: Spatial Configuration of a Hypothetical Extended Metropolitan Region ... 27

Figure 2.4: Annual CO emission estimates for 2000 (Gg year−1)...... 31

Figure 2.5: Schematic Trip Patterns for Different Urban Forms ...... 35

Figure 2.6: Monocentric Model ...... 40

Figure 2.7: Bid Rent Curve ...... 40

Figure 2.8: Rank Size Distribution of Logarithmic Employment Density ...... 56

Figure 2.9: Dynamics of Employment Density Changes ...... 56

Figure 3.1: Urban Population in Major Cities of Bangladesh ...... 65

Figure 3.2: Households by Main Sources of Income ...... 66

Figure 3.3: Employment Distribution in Dhaka...... 68

Figure 3.4: Growth of Dhaka ...... 70

Figure 3.5: Urban Growth of Dhaka ...... 71

Figure 3.6: Dhaka During Pre-Mughal Period (a) and Mughal Period (b) ...... 72

Figure 3.7: Dhaka During the British period (1757-1947) ...... 74

Figure 3.8: City Boundary Proposed by 1959 Master Plan ...... 81

Figure 3.9: Landuse Proposal in 1958 Master Plan ...... 82

Figure 3.10: Modal Split of Dhaka Commuters ...... 85

Figure 3.11: Master Plan Boundary (1959 and 1995) ...... 87

ix Figure 3.12: Dhaka DMDP Area Landuse Scenario ...... 88

Figure 3.13: Landuse Scenarios Considered in STP ...... 90

Figure 3.14: STP Recommended Mass Transit Network for Dhaka (2004-2024) ...... 92

Figure 3.15: Urban Forms in Dhaka ...... 95

Figure 3.16: Bus Routes in Dhaka ...... 100

Figure 3.17: Some Characteristics of Roads in Dhaka ...... 102

Figure 3.18: Motor Vehicle Number in Dhaka ...... 103

Figure 3.19: MPI-based Total Pollution Level in Megacities...... 105

Figure 4.1: Major Components of the Research ...... 110

Figure 4.2: Location and Administrative Boundaries of Dhaka ...... 115

Figure 4.3: Traffic Analysis Zones of the Study Area ...... 117

Figure 4.4: Analytical Framework for Scenario Building ...... 121

Figure 4.5: Illustration of Moran's I ...... 123

Figure 4.6: Illustration of Getis-Ord ...... 124

Figure 4.7: Effects of Spatial Units Size and Population on Employment Concentration ...... 126

Figure 4.8: Potential Employment Centre Identification Process ...... 129

Figure 4.9: Trip Distribution Process ...... 133

Figure 4.10: Volume Delay Function Used in Traffic Assignment ...... 134

Figure 4.11: Transport and Emission Forecasts ...... 136

Figure 4.12: Emission Estimation ...... 137

Figure 5.1: Employment Sectors of Different Regions of Dhaka Mega City ...... 143

Figure 5.2: Gross workforce density in different zones of Dhaka ...... 146

Figure 5.3: Workforce-Population Ratio in Different Regions of Dhaka...... 147

Figure 5.4: Job Clusters by Getis-Ord statistics ...... 153

Figure 5.5: Anselin Local Moran Index for Workforce of Major Employment Sectors 155

Figure 5.6: Employment-Population Ratio ...... 157

x Figure 5.7: Employment Density Map of Dhaka ...... 158

Figure 5.8: Density for Ranked Ordered Job Centres ...... 158

Figure 5.9: Job Density at Difference Distance from the CBD ...... 159

Figure 5.10: Job-Population Ratio for Rank Ordered Job Centres ...... 160

Figure 5.11: Job Clustering with Distance from CBD ...... 161

Figure 5.12: Employment Centres in the Dhaka Megacity Area ...... 163

Figure 5.13: Job Centres in Different Regions of Dhaka ...... 164

Figure 6.1: Employment Centres in the Dhaka Metropolitan Area ...... 168

Figure 6.2: Average Commuting Distance to Job Clusters ...... 170

Figure 6.3: Commuting Patterns to Manikganj Employment Centres ...... 170

Figure 6.4: Commuting Patterns to Employment Centres ...... 171

Figure 6.5: Commuting Preference Functions for Identified Centres in the Main City 172

Figure 6.6: Preference Function for Commuting to Identified Centres in the Inner Ring ...... 173

Figure 6.7: Preference Function for Commuting to Identified Centres in the Outer Ring ...... 174

Figure 6.8: Floor Area of the Identified Job Centres ...... 176

Figure 6.9: Total Floor Area in Identified Employment Clusters in Dhaka ...... 177

Figure 6.10: Floor Area Ratio of Identified Job Clusters...... 178

Figure 6.11: Work Force Population in Different Employment Centres ...... 180

Figure 7.1: Jobs in Different Regions and in Job Centres ...... 183

Figure 7.2: Total Employment under Different Scenarios ...... 185

Figure 7.3: Employment to Population Ratio ...... 186

Figure 7.4: Job Density Histogram for Scenarios of line 1 ...... 187

Figure 7.5: Job Density Histogram for Scenarios of line 2 ...... 188

Figure 7.6: Commuting Pattern in 2024 ...... 189

Figure 7.7: Variation in Average Commuting Distance under Different Scenarios ...... 190

xi Figure 7.8: Total Passenger Kilometres under Different Scenarios ...... 190

Figure 7.9: CO2 Emissions in Different Regions of the Megacity (Baseline)...... 194

Figure 7.10: CO2 Emissions in the Main City under Different Scenarios...... 195

Figure 7.11: CO2 Emission (in metric tonnes) under Different Scenarios ...... 196

Figure 8.1: Policy Domain for Megacity Development ...... 203

Figure 8.2: Non-Motorized Transport Rate and CO2 Emission ...... 208

Figure 8.3: Image of ...... 211

Figure 8.4: Landcover of Manikganj ...... 212

Figure 8.5: Landcover of Narsindi ...... 213

xii LIST OF TABLES

Table 1.1: Comparison of Urban Densities in Large Cities ...... 4

Table 1.2: Primacy of Dhaka City in the National Urban Context ...... 11

Table 2.1: Ranking of Megacities Based on Ambient Air Quality Measurements and Multi Pollutant Index ...... 30

Table 2.2: Studies on Relationships Between Urban Form and Transport ...... 37

Table 2.3: A Comparative Analysis of Governance of four Asian Megacities ...... 44

Table 2.4: Different Methods Applied By Researchers to Identify Employment Centres ...... 50

Table 3.1: Alternative Transport Strategies Being Evaluated in the Strategic Transport Plan ...... 91

Table 3.2: Rapid Growth of Population and Area of Dhaka ...... 93

Table 3.3: Modal Share and Trip Purpose ...... 99

Table 3.4: Travel Patterns in Dhaka ...... 99

Table 4.1: Urban Form in Several Asian Megacities ...... 110

Table 4.2: Dhaka as Defined by Various Administrative Authorities ...... 113

Table 4.3: Summary of Database ...... 116

Table 4.4: Employment Figures by Travel Zones ...... 125

Table 4.5: Employment Distribution Under Different Scenarios ...... 131

Table 5.1: Jobs Distribution of Jobs by Types of Industries ...... 142

Table 5.2: Distribution of Jobs by Zones in Dhaka and by Major Job Industries...... 144

Table 5.3: Global Moran's I Statistics ...... 149

Table 5.4: Getis-Ord General G Statistics...... 150

Table 5.5: Correlation Matrix ...... 150

Table 5.6: Eigenvalues for Components Scores ...... 151

Table 5.7: Rotated Component Matrix...... 151

xiii Table 5.8: Employment Centres within the Dhaka Megacity ...... 162

Table 5.9: Job Centres in Different Regions of Dhaka ...... 164

Table 6.1: Commuting Pattern to Job Centres ...... 169

Table 7.1: CO2 Emission under Different Scenarios ...... 184

Table 7.2: Distribution of Employment Growth under Different Scenarios ...... 184

Table 7.3: Travel Distance at Different Development Scenarios for 2024 (Line 1) ...... 191

Table 7.4: Travel Distance at Different Development Scenarios for 2024 (Line 2) ...... 192

Table 7.5: Changes in Vehicle Emissions under Different Scenarios For 2024 ...... 193

Table 8.1: Megacity Dhaka: Governance Principles and Present Institutional Arrangement ...... 210

xiv LIST OF ABBREVIATIONS

ADB Asian Development Bank BBS Bangladesh Bureau of Statistics BRT Bus Rapid Transit BRTA Bangladesh Road Transportation Authority BRTC Bangladesh Road Transportation Corporation CBD Central Business District CNG Compressed Natural Gas DCC DIT Dhaka Improvement Trust DITS Dhaka Integrated Transport Agency DMDP Dhaka Metropolitan Development Plan DUTP Dhaka Urban Transport Project GDP Gross Domestics Products JICA Japan International Cooperation Agency MRT Mass Rapid Transit MV/MT Motorized Vehicles / Non-Motorized Transport RAJUK Rajdhani Unnayan Kartipakkha (Capital Development Authority) STP Strategic Transport Plan UN United Nations UNDP United Nations Development Programme UTP Urban Transport Plan WB World Bank

xv ACKNOWLEDGEMENTS

There are so many wonderful persons who helped with this dissertation. I am very grateful to my supervisor Dr. Bruno Parolin for his patience and assistance throughout my dissertation, from the point of research proposal development to the writing up of my thesis. I am thankful to him for the encouragement he has always given me to seek wisdom for my research. I never forget his kind and useful meetings on the different stages of my research. Without his help, this dissertation would never have been completed.

I am especially thankful to Professor Robert Freestone, my co-supervisor, for his input to this research, and helping me out in structuring the thesis chapter.

Thanks to Donna Liu (Transportation Modeller at Parsons Brinckerhoff, Sydney) for her help in TransCAD modelling and converting EMME data to TransCAD. My gratitude is extended to Dr. Ken Doust and Dr Charles Cheung for allowing me to use their software for preference function modelling.

My sincere gratitude goes to many nice people and friends who have helped me in various ways. Bivas Das, Fazlur Rahman, Mosheur Rahman along with many others helped me in data collection. Dr Shakil Akther took the pain to send me reports and books whenever I asked for. Professor Akhter Hossain Chowdhury with his huge planning experience on Dhaka city helped me in several occasions.

I am grateful for the research opportunity in the Faculty of the Built Environment in UNSW.

I never forget the great patience and share of my wife Ripa and our two children Tawsif and Rayya, to whom this thesis is dedicated.

Thanks to all not listed in this page but deserving of credits.

xvi Chapter 1 Introduction

1.1 Introduction

More than half of the world’s total population now lives in urban settlements ( United Nations, 2009). By the year 2025 the urban population will represent more than two- thirds of the global population (United Nations Population Funds, 2011) and more than 90% of the new urbanites will locate in developing countries (Roy, 2009). One of the salient features of urbanization in developing countries is the emergence of large cities and accumulation of higher growth in larger cities (Nelson, 1976). In 1950, New York was the only megacity (population more than 10 million) in the world. By the year 2000, the total number had increased to at least 21, with 11 megacities located in Asia (United Nations, 2002). Five of these 21 megacities (, Dhaka, , and ) are in South Asia – the least urbanised region in the world (31 percent)(World Bank, 2012). During this time the urban population of Asia had increased five times to reach 1.22 billion while their proportion to the total population increased from 17.1 to 34.9 percent (United Nations, 2002). The driving factors for this rapid growth of large cities are: rural to urban migration (Yokohari, 2000), reclassification of metropolitan area (Food and Agriculture Organization, 2002), globalization, drought, systematic failure of rural agriculture (Khan, 2008), and perceived opportunity in the city to mention a few. Unmanaged urbanization has created massive urban infrastructure backlogs and haphazard , restraining cities from developing capacity to respond to population growth (GlobeScan & MRC McLean Hazel, 2004). These large urban agglomerations have become a significant issue to planners and policy makers who are concerned about the present and future sustainability of these cities (World Bank, 2009). However, we have very little understanding about the urban spatial structure and employment patterns of these cities, and how these cities interact with their

Chapter 1: Introduction 1 extended surrounding areas to affect travel patterns and greenhouse gas emissions – an important component of sustainability (Hakim & Parolin, 2009). In response to this challenge, this thesis investigates the spatial nature of the large metropolitan cities in the western world and in Asia, and examines the spatial structure of the fastest growing megacity in Asia – the city of Dhaka in Bangladesh – in order to contribute to an understanding of pathways for sustainable growth of this megacity.

Several empirical studies on urbanization and urban development patterns in the different regions of the world have focused on the identification of spatial structure and its physical features as part of efforts to promote urban and transport sustainability (Hakim, 2009). The role of the central business district (CBD) in city formation for cities is indispensable. Traditionally, the CBD holds the largest share of jobs and economic activities, and land uses around it are a reflection of its characteristics. Transport connectivity at the same time reflects the role of centres, their job numbers, nature, overall economy and also the technological availability for transportation infrastructure. However, with changing technological, economic and geographical perspectives, the influence of the CBD is changing as well. Anas et al. (1998) in their seminal work on urban structure summarised that “the monocentric paradigm is not an adequate tool for the analysis of complexities of urban spatial structure of modern cities” (Anas et al. 1998, p.1427). In urban areas throughout the world, the dominance of the central business district in employment and commerce is being challenged by auxiliary centres of activity (Giuliano & Small, 1991; Anas et al., 1998; Crane, 2000; McMillen, 2001). However, the pattern of megacity development takes different forms in different regions of the world and the urban transition and spatial imprint of the economy are not uniform for all continents of the world (Fuch, 2004; McGee, 2004).

There have been several attempts to examine the urban spatial pattern of US cities but very little effort is noticed for the cities in the Developing World. It has long been debated as to whether developing countries have their own set of dynamics, or whether

Chapter 1: Introduction 2 the experiences of poorer nations only reflects those of their more affluent counterparts in an earlier stage of development (McGee, 2001). It is unlikely that cities of developing countries will replicate exactly the same pattern and processes of older urban areas in the developed world, which had their peak under very different technological conditions, and in most cases evolved much more gradually than the exploding metropolises of the Third World (Dowall & Treffeisen, 1991; Henderson, 2002). Asia, being the centre of new urban growth, may have its own spatial pattern of urbanization that is significantly different from that experienced by western cities. However, very little evidence is available from studies that are based on empirical data analysis. McGee’s work on the extended metropolitan region is an attempt to fill that gap. McGee (1967) was the first to identify the “Southeast Asian City” as a separate category. His theoretical refinements identified a unique urban structure in Southeast Asia and his research suggested that volatile globalization and rapid urbanization transformed the Southeast Asian city’s urban core into an Extended Metropolitan Region (EMR) (McGee, 1991). Several studies conducted by the Eastern Asia Society for Transportation Studies (EASTS) also explore the spatial pattern of economic activities in large Asian cities, but have concluded with mixed results for several Asian megacities (Alpkokin, 2007; Bhandari, 2007).

The main focus of this thesis is on an examination of the urban spatial structure of the city of Dhaka. This thesis argues that there is a fundamental underlying spatial structure in this city that is likely to be very different to that of western cities. Moreover different regions within Asia have achieved different levels of economic development and that translates into their spatial imprints (Table 1.1). The nature and patterns of this underlying urban spatial structure are assessed through a detailed examination of employment distribution and its travel related impacts. The nature and patterns of the urban spatial structure of the megacity of Dhaka will also allow assessment of implications for present and future sustainability of the city, and whether a different spatial structure – from a policy and planning perspective – may be required in the future to accommodate more sustainable growth. This thesis attempts to comprehend

Chapter 1: Introduction 3 how this megacity interacts with itself and with other regions; through a focus on the spatial patterns of job locations, the spatial pattern of places of workers residences and the mode they use to get to those jobs. By exploring this job and residence linkage this research attempts to identify the key elements of spatial structure, particularly whether it is monocentric or polycentric or falls along a continuum between the two extreme spatial structures.

Table 1.1: Comparison of Urban Densities in Large Cities

Cities Urban Density (person/ha) Australia/New Sydney, Perth, Melbourne, Brisbane, Wellington 15 Zealand North America Atlanta, , Denver, Houston, , New 15 Western Europe Graz, Milan, Vienna, Bologna, Brussels, Rome, Co- 55 penhagen, Helsinki, Amsterdam, Oslo, Lyon, Nantes, Barcelona, , Madrid, Marseilles, Stockholm, Berlin, Bern, Frankfurt, Geneva, Hamburg, Zurich, Dusseldorf, London, Munich, Manchester, , Newcastle, Stuttgart, Glasgow, Athens China , , , , , 146 Ningbo High income Asian , Sapporo, Tokyo, Hong Kong, Singapore, Taipe 150 cities Low income Asian , , Mumbai, , Kuala 204 cities Lumpur, Jakarta, , Source: Kenworthy and Hu, 2002 One key aim of this research is to contribute to the debates on the role of future urban spatial structure in generating a sustainable future for the Asian megacity, and for Dhaka in particular. It is now widely accepted that a relationship exists between the shape, size, density and uses of a city and its sustainability. However, “consensus is lacking about the exact nature of this relationship” (William, Burton & Jenks 2000, p.1). Increasing attention is now being paid to the sustainability of urban areas. Sustainable urban growth aims to minimise the environmental impact of urban development. Transport being a key element of the urban system, interactions between transport and urban form are intricately woven. In a sustainable urban landscape transport is expected to increasingly provide good outcomes in both the communities’ economic productivity and social well being – the three pillars of sustainable development (Doust, 2008). The position taken in

Chapter 1: Introduction 4 the thesis is that the optimum way of making this contribution is to explore and understand the land use - transport dynamics of Dhaka megacity and the potential development options for the city. Dhaka, being the primate city of Bangladesh, is important to the economy and represents the hope and aspirations of the country. The unprecedented growth of Dhaka during the last four decades has attracted attention to plans for the city that can accommodate the future population growth while leading the economic growth of the nation.

1.2 Urbanization and Asian Megacities

By the year 2030, each of the major regions of the developing world will hold more urban than rural dwellers, and by 2050 fully two-thirds of their inhabitants are likely to live in urban areas (Gupta & Rayadurgam, 2008; Montgomery, 2008). These facts are particularly relevant for South Asia, home to over 1.6 billion people, of which a third live in urban areas (United Nations, 2011). By 2015 this region will account for 5 of the world’s 12 biggest urban agglomerations (all with more than 15 million people); namely, Mumbai, Delhi, Kolkata, Dhaka and Karachi (Figure 1.1). By the same year (2015), over 540 million South Asians will live in towns and cities and by the year 2030 this figure will exceed 813 million inhabitants (United Nation, 2005). These trends have led to an assessment of the capacity of these cities as a central theme of sustainable urban development in the region.

Chapter 1: Introduction 5

Figure 1.1: Megacities in the World Source: Mappery, 2000

An important characteristic of urbanization in Asia has been the emergence of megacities – large urban agglomerations of more than 10 million inhabitants. Several of these megacities have evolved and become primate cities over the past few decades, accumulating the lion’s share of urban growth. Gupta and Rayadurgam (2008) claim that there is ‘excessive’ primacy in low-income countries because of urban bias or other political economy reasons. In 1950, Tokyo was the only megacity in Asia. But by 1975 the biggest urban agglomeration of the world was Asian and by 2003, 11 of the 20 global megacities were Asian (Hugo, 2003; UN, 2003). About 75% of the world’s population lives in the developing countries and 58% in Asia (Heinke, 1997). By 2015, there will be 22 megacities worldwide of which 12 will be in Asia, including 7 of the 10 largest cities. These figures highlight the concern of policy makers of the world in giving emphasis to resolving problems associated with megacities in the Asian Countries.

Large cities in the Developing World face greater urbanization challenges than developed countries faced some 40 to 60 years ago. The urbanization process was much

Chapter 1: Introduction 6 slower in developed countries. “This gradual pace combined with relatively high gross domestic product (GDP) and better education rates, allowed time for the development of the political and economic institutions and market instruments essential for an efficient form of urbanization and a reasonable quality of life” (Henderson, 2002, p.90). For the developing countries, all forces of urbanization come all of a sudden and the cities are not prepared to accommodate this unprecedented growth. Due to invariable centres of economic growth, they often grow at accelerating rates influenced by global forces and become unmanageable (Cohen, 2006). A major challenge for megacities in developing countries is the management, planning, control and governance of these megacities to resolve this crisis for a better future (Meier & Quium, 1991). Most of them now suffer serious problems such as lack of sufficient housing, shortage of energy, chaotic traffic, severe shortage of water, pollution of air, water and land, unemployment, crime and corruption (Heinke, 1997; International Federation of Surveyors, 2010). Moreover, people from rural areas are still attracted to these cities in ever greater numbers because of lack of employment opportunity in rural areas and a perceived better life in the city. In these over populated megacities, central and local governments are strained beyond their financial capabilities to build the necessary facilities for its urban citizens (Hugo, 2003; Meier and Quium, 1991). Urbanization is closely linked to economic development (Moomaw & Sattar, 1996) although towns and cities are “engines of growth” for the rapidly growing economies, unplanned and unmanaged urbanization poses a serious threat to that very same growth engender social tensions and inequalities (Gupta, 2008).

1.3 Urbanization in Bangladesh Bangladesh is one of the most densely populated countries in this world with an area of 147,000 km2 (United Nations, 2003; Bangladesh Bureau of Statistics, 2001). It is predominantly an agrarian country experiencing rapid urbanization and economic transformation in its recent history (Figure 1.2). Urbanization is one of the most significant development issues in Bangladesh (Hossain, 2006). While developing nations in Asia show rapid rates of urbanization, Bangladesh still remains less urbanised

Chapter 1: Introduction 7 (27.9%) (Ahmed, 2012). Although the absolute urban population, as well as the number of cities and towns in the country, has increased rapidly over the last few decades, both the process of urbanization and the dynamics of growth of urban centres in Bangladesh are quite different from other developing countries in South and Southeast Asia (Rahman, 2004). Rapid population growth at a national scale triggers urban growth as the rural economy cannot absorb this growing pressure. The resulting urban growth which plays an increasingly significant role in economic development is challenging planning bodies in devising ways to develop and implement appropriate strategies to ensure sustainable urban development. The large population growth of Bangladesh in comparison to its resources, and frequent natural hazards, has made it more challenging to offer a minimum standard of living. Much of the country is prone to annual flooding that delimits the spatial growth for a country with limited economic strength. In 1975, only 9.9% of the country’s 75 million people lived in urban areas; this has now increased to 28% in 2010 (Ahmed, 2012). The present population density in the main Dhaka city is now believed to have reached 100 thousand people km2 (Cox, 2012). United Nations (2003) projections estimate that on current trends the urban population of Bangladesh will be around 40% by 2030.

250 Projected Urban Population 200 Rural Population

150 Total Population

100 (Populationin Millions) 50

0

Figure 1.2: Population growth in Bangladesh Source: United Nations, 2003

Chapter 1: Introduction 8

In Bangladesh, urbanization takes the form of rapid growth of urban population, largely due to natural growth, rural-urban migration, and changes in the definition of urban area (, 2006; Ahmed, 2012). Along with limited scope in the rural economy, there are pull factors like jobs, amenities, and education in the major cities to attract the growth. The hierarchy of the urban system includes both induced and spontaneous growth centres (Rouf & Jahan, 2001; Hossain, 2008). A large number of spontaneous growth centres have emerged due to rapid population growth as well as the growing importance of certain local resources that induce economic growth. Most administrative and organizational decisions, capital investments, and technological innovations generally originate in the primate city and induced growth centres such as divisional and district headquarters. However, trickle down effects are not spread in a spatially uniform way, perhaps because of spatial variations in environmental conditions and limitations of the socio-economic infrastructure of the country (Rahman, 2004). However, significant population growth is observed in other large cities. There are 64 districts in Bangladesh with varying levels of urbanization ranging from 60% to below 10%. The district of Dhaka, one of the 64 districts, is over 90% urban (Islam, 2006). Dhaka holds 38% of the national total urban population whereas and , the second and third largest cities, have only 11% and 4% of population respectively (United Nations, 2003). There also exists a wide spatial inequality in the rates of urbanization, quality and quantity of social infrastructure and economic growth and development. Islam (2006) also stated that for these inequalities development planning is always aimed at the sectoral level and the spatial form of planning is always neglected.

Chapter 1: Introduction 9

Figure 1.3: Population Explosion of Dhaka Megacity Source: United Nations (2003)

With a population of over 18 million, Dhaka is the capital and largest urban agglomeration in Bangladesh. It is also one of the largest cities in the world with a present rank of 11th largest city. Dhaka is also the fastest growing megacity in the world with a growth rate of 3.2% (Figure 1.3). Table 1.2 shows that its growth has been rapid and higher than the national urban population growth. Although income growth is higher and the incidence of poverty is lower than the rest of Bangladesh, Dhaka still is a low income city of which has a large proportion of poverty stricken inhabitants when compared with most megacities of the world. Holding the prospects for better income opportunities than most parts of Bangladesh, Dhaka is growing much faster than the rest of the Bangladesh owing to migration. This rapid urbanization is putting pressure on the city's limited land, already fragile environment and weak urban services.

Chapter 1: Introduction 10 Table 1.2: Primacy of Dhaka City in the National Urban Context

Year Population of Percentage of Area of Dhaka Percentage of Density Dhaka National (km2) National Urban (/km2) (millions) Population Population 1951 0.41 - 85.45 - 4800 1961 0.71 - 124.45 - 5700 1974 1.77 3.0 335.79 28.2 5270 1981 3.45 3.8 509.62 26.0 6770 1991 6.84 5.8 1352 30.5 5060 2001 10.71 8.0 1352 37.4 7920 2011 15.5 10 1352 54 11460 Source: Islam, 2006; BBS, 2011

Thirty percent of the urban population contributes about 60% of the GDP (Ahmed, 2012) and Dhaka city contributes about 13% of this amount (Ahmed, Ahmad & Mahmud, 2005). Income and literacy rates are higher in Dhaka than the rest of the country and the incidence of poverty is like most megacities in the developing countries (Bangladesh Bureau of Statistics, 2011). However, the poor quality and quantity of infrastructure, and constraints in spatial expansion capacity, have turned the capital into an unlivable place. Moreover, poor city management, lower efficiency in administration and management, and corruption in service provision, are exacerbating the problems. Urban traffic is reaching nightmare proportions, often causing massive delays in covering small distances with associated productivity losses. To accommodate this huge population, the city has become over-crowded with insufficient civic services, utilities, transport and environment (Sharma, 2009). Several international agencies have rated the quality of living in world cities and Dhaka scored very poorly on most of them. The United Nations’ City Development Index (CDI) ranked it 7th worst; Economist Intelligence Unit (EIU) ranked it 3rd worst tied with ; Asia week (2000) placed it 39 out of 40 cities (Ahmed, Ahmad & Mahmud, 2005).

Unequal urbanization in the country, and concentration of urban services, employment and business opportunities in Dhaka, has made it a megacity. However, it has been

Chapter 1: Introduction 11 recommended in several planning documents that there is a need to decentralise economic activities and growth from Dhaka and enhance the economic base of other urban areas (Minoprio and Spencely & P.W. Macfarlane, 1959). But no effective mechanism has been practiced nor developed.

1.4 Statement of the Problem

Urbanization in Bangladesh is increasing at higher rates than in the past and growth is highly concentrated in the major cities. Dhaka, the capital of Bangladesh, is the fastest growing megacity in the world with its nearly 20 million inhabitants. The economy of Dhaka, its infrastructure, frequent flooding and the physical geography of the city as a whole do not allow for expansion of the city with respect to its population growth. This has resulted in very high densities, severe congestion and a highly polluted city.

It has been observed that Dhaka experienced the highest growth (7%) between 1975 and 2000. Dhaka had only 4.2% of urban population in 1950 but this has reached 25% by 2000. By 2030 it is estimated to be up to 45%. But the city and its infrastructure are not prepared to accommodate this growth. Currently the city is experiencing its highest congestion level, while pollution levels are at their worst. Movement within the city has become difficult for residents and no long term planning vision has surfaced as yet. A detailed study of the city’s spatial structure is needed in order to develop scenarios and public policy plans to accommodate future growth to ensure economic progress, transportation solutions and overall sustainability for the future.

There are different models of urban development trajectory, mainly associated with Western cities. These models range from monocentric, polycentric or dispersal. There are also models of Asian cities (for example Extended Metropolitan Region by McGee). Several studies have also established polycentric or monocentric structure in Asian cities. But what remain unknown are the nature, type and characteristics of the urban spatial structure of Dhaka megacity – one of the fastest growing megacities in the world.

Chapter 1: Introduction 12 1.5 Aims, Study Area and Scope of the Research

This study aims to analyse the spatial structure that is the distribution of jobs and housing, of Dhaka megacity to provide a foundation for the understanding of sustainability of the city in the future. Travel patterns are also investigated and linked to scenarios for sustainable urban form in the future. Dhaka is challenging due to its large population and overwhelming urban problems and the spatial structure perspective of this thesis hopes to make a contribution to identifying potential growth avenues – a perspective not previously examined in the research literature on the city of Dhaka. This study applies very rich data intensive spatial modeling to identify the spatial structure of employment and to assess its impacts on commuting, environment and the over-all spatial footprint.

Dhaka has been defined many ways with different spatial extents in different studies and projects, and that makes the data comparison and longitudinal analysis difficult. There is the main Dhaka city for which Dhaka City Corporation provides services. The devastating flood of 1988 affected 65% of Dhaka’s dwellers and highlighted the need for an embankment around Dhaka City. The Flood Action Plan that was proposed led to the construction (in 1989) of an embankment along the Balu River to save the Eastern low lying area from flooding. This embankment as such extended the city boundary. The 1995 Dhaka Metropolitan Development Plan (DMDP) was the first plan that took initiatives to plan for a megacity spreading over 1528 km2. However, this was found to be insufficient and the Strategic Transport Plan (STP) (The Louis Berger Group Inc. and Bangladesh Consultants Ltd, 2005) argued that Dhaka has a much wider impact on travel and transportation in the surrounding area and they considered the outer ring as part of Dhaka. This study considers the STP defined area as the Dhaka mega region (Figure 1.4). This study divides Dhaka megacity into four regions considering population density and infrastructure provisions. The main city consists of the old part of Dhaka and the jurisdiction of Dhaka City Corporation area. The areas on the east developed spontaneously and became a part of Dhaka city. This area is named as ‘extended’ region. The master plan prepared for Dhaka in 1995 defined metropolitan

Chapter 1: Introduction 13 area with a 1528 sq. km2 area. This area is named as ‘inner ring’. The spatial coverage of the strategic transport plan (2005) recommended a bigger area (7440 sq. km2) is named here as “outer ring”.

Figure 1.4: Study Area Source: Prepared by Author

This study is limited to the identification of the spatial patterns of jobs, exploration of commuting patterns and examination of sustainability by generating scenarios using several models of urban growth and transportation analysis. Urban structure is analysed from an employment perspective – its concentration, characteristics and spatial distribution. As a result, only journey to work are investigated. For future growth it is assumed that all projects in the planning and implementation phase will be completed on

Chapter 1: Introduction 14 time and there will not be any abrupt change in policy or urban expansion, migration or population growth.

1.6 Goals, Objectives and Research Questions

The dynamic interaction between landuse and transport is intensively scrutinised to identify the spatial structure of the city and its role in transportation. Through understanding the concentration of economic activities and their associated travel behaviour, the main objective of this research is to explore sustainable growth alternatives to accommodate the future growth of the city. The key aim of this research is to understand the spatial structure of Dhaka and identify what implications this has for travel and sustainability. Once the dynamics are understood, the thesis assesses the implications of future growth and contributes as to how the planning authorities could be managing further city growth.

The first objective of this thesis is to explore the international experiences and suggestions in the literature about sustainable urban form, job centres and transportation arrangement for large cities. The most pertinent research questions are: 1.1 What are the well accepted and most recognised approaches available in literature to address urban structure for large cities? 1.2 What would be an appropriate methodology to identify job centres for Dhaka megacity?

The second objective of this research is to understand the spatial structure of Dhaka and explore the nature and characteristics of the employment clusters and corresponding travel behaviour in Dhaka city. The relevant research questions are:

2.1 What is the job spectrum of Dhaka? Is there a spatial pattern of job concentration and, if so, what is the nature of that concentration? 2.2 What general patterns of employment concentration exist for Dhaka? Do we see a monocentric pattern of job concentration or a polycentric one? Does it match

Chapter 1: Introduction 15 with large cities of the developed world or other Asian cities, or show a unique pattern?

The third objective is to identify job centres for Dhaka megacity and to explore their spatial travel pattern. The following research questions examine this objective:

3.1 What are the prominent and potential employment centres for Dhaka megacity? 3.2 What are the spatial characteristics of these centres and how do they characterise in different regions in Dhaka megacity?

The fourth objective is to investigate the employment characteristics of identified centres and analyse their travel consequences. To achieve this objective, the following questions are posed:

4.1 Given the identified urban spatial structure of Dhaka, what are the salient characteristics of commuting trips? 4.2 How can the empirical findings on the spatial structure of employment and journey to work patterns be assessed against theories and practices from the literature? 4.3 How does the identified journey to work patterns explain the spatial structure of employment? 4.4 Do the employment zones in different regions of Dhaka exhibit substantial variations in their commuting patterns? 4.5 Is there any relationship between major land use, job density and journey to work patterns?

The fifth objective is to generate growth scenarios for Dhaka megacity by distributing growth using different development strategies. The following research questions address this objective:

5.1 What is the estimated growth for Dhaka by the year 2024?

Chapter 1: Introduction 16 5.2 What growth strategies are discussed in planning reports that can be adopted to accommodate future growth? 5.3 What are the major strategies needed to achieve sustainability? 5.4 Which scenario generates better transportation outcomes for the megacity?

The sixth research objective is to compare alternative scenarios and explore the relevant policy implications. Relevant research questions are:

6.1 What would be the policy implications of implementing the employment decentralise growth as presented discussed under different scenarios for Dhaka in the future? 6.2 Are local landuse and development plans compatible with this proposed growth scenarios?

This research utilises very rich data sets for the Dhaka megacity and is one of the fortunate opportunities of having such a diverse range of relevant data that has been generated as part of longer studies conducted for the city. This research uses zonal (traffic analysis zone) level travel data at single point in time. Transport and travel data come from several surveys in 2004 conducted for the Strategic Transport Plan for Dhaka. For demographic figure, year 2001 census data were adjusted. For land and building use information this research relies on surveys for the Detailed Area Plan for Dhaka Metropolitan Development Plan. This is the first time that these data sets are accumulated into a single platform and used for research purposes. They have been used earlier by consultants working for various departments. The lack of longitudinal data at the same detailed scale constraints analysis over time, but the study overcomes this by generating 10 scenarios encapsulating two growth strategies.

Chapter 1: Introduction 17 1.7 Structure of Thesis

Chapter 1: Introduction Chapter one introduces the research by providing background on the importance of urbanization in Asian cities and in the city of Dhaka. It identifies the objectives of the research and research questions to be addressed in the thesis.

Chapter 2: Urban form, Travel and the Employment Centres of Megacities This chapter provides the context of the thesis. It presents a background literature study focusing on the importance of understanding the dynamics between urban form, land use and transportation from a city development and planning perspective. It shows that urbanization has taken different forms in different regions of the world. As a result, a uniform approach to address this issue is not sustainable. Different urban forms and their roles in transportation arrangements have been analysed. This chapter reviews different methods developed and applied in large cities to identify employment concentrations and to evaluate corresponding transport efficiency.

Chapter 3: Urban Evolution and Planning History of Dhaka Chapter 3 presents the urbanization process in the study area. This chapter provides the background to understand the current spatial structure of the megacity. It presents the urban development and planning histories for the city. It shows the development and expansion process of Dhaka’s employment landscape that was developed over the years as a result of piecemeal development plans, long term planning initiatives, political influence, geographical impacts and market forces. A historical overview of the evolution of the urban form of Dhaka is presented commencing from the Mughal era in the 17th century. The transformation of early alley and river based settlements to a highly congested road based transport system is discussed.

Chapter 1: Introduction 18

Chapter 4: Research Design This chapter explains the methodological aspects developed and adopted in this study to explore spatial structure and travel behaviour. Data sources are listed and discussed, and methods adopted to address the research questions are presented. Based on a literature review presented in the previous two chapters, a GIS based method has been developed to analyse the employment landscape and to identify job centres in Dhaka. This chapter also explains the analytical framework for future growth assumptions, methods adopted to create different urban development scenarios, and impacts on commutation and environment.

Chapter 5: Spatial structure of employment in Dhaka This chapter describes the spatial structure that is location of residences and economic activities, of Dhaka megacity and highlights identified employment centres. It studies the centres in term of their job characteristics, labour catchment and commuting behaviour. There are 19 job centres identified in the Dhaka megacity including several potential centres in the outer ring.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka The main city has the most employment. At the same time spatial specialization has been reflected as job characteristics and associated commuting pattern are found to be different in different regions of Dhaka megacity. This chapter contains a complete investigation of such job distributions and their association with surrounding land use and travel patterns. This chapter inquires about travel pattern with residential work force and job related floor area.

Chapter 7: Exploring Growth Options for a Sustainable Dhaka This chapter investigates urban growth scenarios for Dhaka. It envisions different urban expansion models and proposes an analytical framework to compare scenarios under different growth conditions. The scenarios are built on different decentralization models emphasizing future growth around employment

Chapter 1: Introduction 19 centres. The results are compared to changes in travel time, distance and emissions. Results show that growth in the outer ring would significantly reduce the commuting distance and time.

Chapter 8: Policy Implications for Sustainable Urban Structure This chapter addresses the policy area involved in sustainable urbanization plans. As a continuum to previous chapters, this chapter focuses the policy issues that need to be addressed to achieve sustainable outcome for Dhaka megacity.

Chapter 9: Conclusions Chapter 9 presents the findings of previous chapters. It emphasises the environmental impacts of the proposed growth strategy, especially the consequences of transportation emissions and the compatibility of the proposed growth with local land use and transport plans.

Appendix A: List of Under Construction Project This thesis generates scenarios with the assumption that all under-construction transport projects in Dhaka megacity will be completed in due time. This appendix lists those projects and their estimated completion time. It has been taken from the Strategic Transport Plan Project report (2005).

Appendix B: Maps of Base Year Data This appendix displays some maps with base year data. It has commuting desirelines maps for all employment centres and 3D maps showing building footprints, building height and usage.

Appendix C: Maps of growth Scenarios Appendix C displays detailed maps of the different growth scenarios. It has employment density maps and commuting desire line maps for all growth scenarios for 2024

Chapter 1: Introduction 20 Chapter 2 Urban form, Travel and the Employment Centres of Megacities

2.1 Introduction

The main purpose of this chapter is to explore the literature on urban spatial structure of large cities and to extract some key concepts and methodologies that could be applied to the city of Dhaka – the case study for this research. It investigates both western and Asian literature and emphasises research on urban structure and transport sustainability of the Asian megacities. One of the goals of this chapter is to compare and contrast this body of literature for relevant ideas and shed light on concepts that can be considered for empirical analysis to examine whether Dhaka is characterized as being monocentric, polycentric or has some other form of urban spatial structure. This review presents a critical look at the western literature and methods, and its applicability for Dhaka. At the same time, it discusses the relevance of other more recent methods that are applied to Asian large cities.

With growing population, urbanization, congestion and environmental degradation, the cities of the developing world are struggling to find a better way to make them more livable (Ichimura, 2003). Attempts are made to build an urban area having efficient use of space, be effective in energy use, have reasonably located amenities and services, and tolerable pollution and waste. But megacities in the developing world seem far away from this desired model and are often characterized by congestion, pollution, and poverty to mention a few. The research reviewed here examines the spatial pattern of development and allocation of economic activities across cities in connection with their role in transportation and urban sustainability.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 21 Since major economic centres are also centres for employment, it examines how these centres generate and draw city traffic, and its impact on the transportation system and urban structure. It is also essential to explore the reflection of existence of economic centres in transport and landuse policies. As a result, the literature covers a wide spectrum of urban development and growth management issues focusing on urban centre and sub-centre formation, their characteristics and contributions to the transportation system. To accommodate the various thoughts and debates on urban sub-centre formation in different parts of the world, the literature covers the causal pattern of urban land development and transportation both in the developing and developed world. It also discusses the methods used by researchers to model urban spatial structure and sub-centre identification methods for Dhaka.

This chapter starts by highlighting the growing importance of urban spatial structure on travel and sustainability. It focuses on the transformation or changes in urban structure and urban form to cope with thriving urban population growth. The chapter also discusses and summarizes different approaches and methods used to identify employment centres for large cities around the world.

2.2 Urban Growth and Urban Form: Concerns of Growth Management

Urbanization translates its development impact through urban form. On the flip side, urban form depends on the nature of urbanization, economy, transportation arrangement, local climate and cultural history. Urban form is a composite of characteristics related to land use patterns, transportation system, and urban design (Handy 1996). Kevin Lynch defines urban form as “the spatial pattern of the large, inert, permanent physical objects in a city” (Lynch, 1981, p.41). Urban form directly affects the environment, habitat, ecosystems, and water quality through land consumption and conversion through different development processes (Jabareen, 2006). However, there is a lack of theory that helps us to evaluate whether a given urban form contributes to sustainability or to compare different forms according to their contribution to sustainable development objectives (Jenks, et al., 2000).

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 22 A city can be of different size and type but most cities have a central area where most jobs are located - commonly known as the ‘Central Business District’ (CBD) (Murphy, 2009). City dwellers from all corners, and from its surroundings, commute to the city centre for their everyday jobs. The size of a CBD is generally proportionate to city size; it grows with the city. Urban structure of a traditional city can be presented by E. W. Burgess’s (1925) concentric models. As the CBD grows the city dwellers in the inner ring move outward because of increasing rent and availability of land. The growth of a city can take different forms guided by the nature of growth, geography, economy, spatial policy and infrastructure. The variation in urban form as presented in Figure 2.1 results from all these dynamics. In addition, urban form affects travel behaviour which in turn affects air quality, soil contamination, global climate, and noise (Roo, 2000; Oliveira & Silva 2011).

Patterns of urban development and population distribution dictate urban form. Figure 2.1 illustrates the urban form of several metropolitan cities. Population distribution in European cities shows a spreading pattern. Shanghai shows the strong influence of the main city on its population distribution in a compact nature. Jakarta demonstrates the high influence of the main city along with two peaks in the south and the east, which are the indications of sub-centre formation. The visual inspection on the form of cities assists one to comprehend the nature of the transportation system a city might require. However, factors like economy, demography and topography are important.

The dream for a sustainable urban form is as old as modern human civilization. Several models were developed over generations to find an urban form that can balance the locations of job, housing, amenities, urban services and the environment. One of the most influential town planning models of the late 19th and early 20th centuries was the ‘Garden City Movement’. Belief in a new form of community with all the advantages but none of the disadvantages of industrialising England towns in the 19th century, led Ebenezer Howard developed his Garden City ideas. He published his influential book, Tomorrow: A Peaceful Path to Reform, in 1898. It was reissued as Garden Cities of Tomorrow in 1902. Along with the decentralised, cooperative society, Howard envisaged the improvement of the urban environment

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 23 (Freestone, 1989). Le Corbusier work on Chandigarh city in India is a prominent example of the same nature. Le Corbusier conceived the master plan of Chandigarh as analogous to human body, with a clearly defined Head (the Capitol Complex), Heart (the City Centre), Lungs (the Leisure valley, innumerable open spaces and sector greens), Intellect (the cultural and educational institutions), Circulatory system (the network of roads,) and Viscera (the Industrial Area). However, it is found increasingly difficult to maintain the plan and threshold population due to enormous growth pressure from the surroundings (Malik, 2003).

Three cities in the Indian subcontinent took initiatives to adopt modern urban planning approaches to plan their capital cities in the 1950s (Maria & Imran, 2006). Chandigarh, Dhaka and Islamabad all employed western architects and planners to make their ‘ideal’ city an inspiration for future urban development. Chandigarh and Islamabad were new towns to be built on agricultural land (Kalia, 1987; Malik, 2003) and Dhaka’s growth would be controlled in a strategic manner (Minoprio and Spencely & P.W. Macfarlane, 1959). However, within twenty years all these cities experienced unprecedented growth and urbanization (Maria & Imran, 2006). The planned population thresholds were exceeded quickly. The economic growth could not follow the unprecedented population increase, and planning for infrastructure and services fall far behind the demand. The urban planning bodies could not keep pace or did not have resources to handle such growth.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 24

Figure 2.1: Spatial Distribution of Population in Selected Large Cities Source: Bertaud, 2001

The size and functionality of the ‘business districts’ of mega cities around the world has become an interesting topic amongst urban researchers and they have reported that the size of the CBD is not expandable to infinity. To accommodate the increasing jobs and economic activities several other centres have evolved as the CBD becomes larger. They can be specialized centres or a miniature of the main CBD. This particular state of having more than one prominent business district is termed polycentricity (Riguelle et al., 2007). Several urban forms have been documented as the result of the polycentric nature of large cities. The multi-nuclei city (Harris & Ullman, 1945), edge city (Garreau, 1991), ultra city, 100 miles city (Sudjic, 1995), corridor city, satellite town, and network city (Meijers, 2007) are a few examples (Figure 2.2).

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 25

Figure 2.2: Theoretical Urban Forms Source: www.reading.ac.uk/PeBBu/state_of_art/urban_approaches/high_density/high_density.htm

Harris and Ullman (1945) developed their famous model of urban growth which postulated that large cities tend to develop a number of nuclei which serve as focal points for agglomerative tendencies, some more important than others; this is their so-called multiple nuclei model, tendencies still seen in cities today. Joel Garreau (1991) took this further and observed concentrations of urban activities (job, office space, retail space) around freeway intersections or adjacent to airport or shopping. They often are not separate legal entities but are governed as part of surrounding counties in the United States. However around the world these ‘edge cities’ have become a conspicuous suburban phenomenon.

Most of these spatial models were developed to explain or to propose growth for the cities in the Western world. However, one does not see research that addresses the appropriate urban form for rapidly changing megacities in the developing countries where economy, infrastructure, jobs, and transport options are different. T. G. McGee’s work (since 1967) on the extended metropolitan region (EMR) concept is an attempt to fill that gap. McGee (1967) was the first to identify a unique urban form for the large cities in Southeast Asia in which penetration of urban activities into predominantly agricultural areas had created a mixed landuse across in a vast area (Hakim, 2009). McGee’s (1991) theoretical refinements identified a unique urban structure with five regions (Figure 2.3):

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 26 (1) Major cities: large cities or primate cities; (2) Peri-urban areas: hinterland of the major cities or labour catchment of the major cities; (3) Desakota zones: predominantly agricultural areas with increasing penetration of urban activities. These zones are commonly located along transport corridors connecting major cities and characterized by intense urban-rural mix; (4) Densely populated rural areas: where villages of wet-rice agriculture are commonly found in Asian countries; and (5) Sparsely populated frontier areas, which are remote areas that have not been well developed.

Major cities Peri-urban desakota Densely populated rural Sparsely populated frontier Smaller cities and towns Communication routes

Figure 2.3: Spatial Configuration of a Hypothetical Extended Metropolitan Region Source: McGee (1991)

The debate about whether particular urban forms, in terms of their shape, density, configuration and so on, can have an impact on the sustainability of cities has a relatively long and rich history (see for example, Breheny, 1992; Williams et al. 2000; Roo & Miller, 2000). Within this debate, researchers and planning practitioners have considered the impact of urban form on a number of elements of sustainability, such as social equity (Bramley & Power, 2009), accessibility, ecology

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 27 (Holder, 2004), economic performance (Cervero, 2001), pollution, global warming and health (Anderson et al., 1996 & Grazi et al., 2008). However, the issue which has attracted the most attention both academically and in practice is the impact of city form on travel pattern (Camagni, 2002). In particular, this field of enquiry has concentrated on the ‘best’ urban forms to facilitate sustainable transport solutions, generally seen as reducing trip lengths and times, reducing reliance on the car, enabling efficient public transport, encouraging walking and cycling and reducing transport-related emissions, pollution and accidents (see for example Rickaby, 1985; Black & Katokas, 1987; Newman & Kenworthy, 1991; Stead & Marshall, 2001; McMillan, 2003; Arbury, 2006; Munshi et al., 2008; Zhao et al., 2009). However, urban form and transportation infrastructure are not the only factors to be statistically related to metropolitan-wide travel patterns. Various socio-demographic and housing and urban history-related variables are equally associated with inter-metropolitan differences in travel patterns (Coevering & Schwanen, 2006).

2.3 Urban Spatial Structure and Transportation: Issue of Sustainability

Spatial structure of cities and their transport implications are central to an understanding of the transport and land use dynamics. High urban densities can be both positive and negative in their transport implications depending upon policy settings and choices (Dimitriou & Gakenheime, 2011). There is general agreement in the literature that current levels of car use, fuel consumption and emissions are unsustainable (Williams, 2005; Næss 2006). Transport is the second largest CO2‐ emitting sector after electricity production. The sector as a whole accounted for approximately 13% of overall green-house-gas (GHG) emissions and 24% of CO2 emissions (OECD, 2010). The annual growth rates of CO2 emissions by the transport sector in the developing world and in economies in transition are projected to be even higher, namely 3.4% and 2.2% respectively (Grazi & Bergh, 2008). Road transport emissions account for two thirds of transport emissions (Zhao, 2009). Following the higher urbanization rate in developing countries transport emissions are higher in the cities of the developing countries. The annual growth rate in CO2

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 28 emissions by the transport sector in developing countries is projected to be 3.4% (Grazi et al., 2008).

An assessment on the multi-pollutant index (MPI) that is based on emission statistics reveals higher pollution levels in cities of the developing countries (Table 2.1 and Figure 2.4). Asian megacities like Karachi, Dhaka, Delhi have more than twenty times total suspended particular (TSP) than New York and Osaka (Table 2.1). Those cities are also ranked poorly for other pollutants like SO2 and NO2. Many of those megacities in the Developing World do not have many large scale industries, and road transportation is considered as the main means of emissions. Dhaka has worst MPI of all the megacities in 2000 (Table 2.1) but a lower per capita emission rate (Figure 2.4). This is due to extremely high population density and lower rates of car ownership in Dhaka as compared to other cities of the same scale.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 29 Table 2.1: Ranking of Megacities Based on Ambient Air Quality Measurements and Multi Pollutant Index

−3 −3 −3 Megacities in 2000 TSP (µg m ) SO2 (µg m ) NO2 (µg m ) MPI [Rank] [Rank] [Rank] [Rank]

Tokyo 40 [15] 19 [9] 55 [11] −0.27 [16]

Mexico City 201 [10] 47 [4] 56 [10] 0.52 [10]

New York-Newark 27 [18] 22 [7] 63 [7] −0.23 [14]

São Paulo 53 [14] 18 [10] 47 [12] −0.29 [17]

Mumbai (Bombay) 243 [9] 19 [9] 43 [14] 0.39 [11]

Kolkata (Calcutta) 312 [6] 19 [9] 37 [15] 0.59 [9]

Shanghai 246 [8] 53 [3] 73 [5] 0.87 [8]

Buenos Aires 185 [11] 20 [8] 20 [18] −0.01 [13]

Delhi 405 [4] 18 [10] 36 [16] 0.92 [7]

Los Angeles-Long 39 [16] 9 [13] 66 [6] −0.25 Beach-Santa Ana [15]

Osaka- 34 [17] 19 [9] 45 [13] −0.37 [18]

Jakarta 271 [7] 35 [6] 120 [3] 1.24 [5]

Beijing 377 [5] 90 [2] 122 [2] 2.01 [2]

Rio de Janeiro 139 [13] 15 [11] 60 [8] 0.11 [12]

Cairo 593 [2] 37 [5] 59 [9] 1.93 [3]

Dhaka 516 [3] 120 [1] 83 [4] 2.40 [1]

Moscow 150 [12] 15 [11] 170 [1] 1.07 [6]

Karachi 668 [1] 13 [12] 30 [17] 1.81 [4]

Source: Gurjar, 2008

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 30 In the transport sector, these emissions can be reduced by three different pathways. First, by urban planning approaches (Schipper et al., 2009). Location of residence and functions or activities need to arrange in such a way that it requires less travel. Second, improvement in vehicle technology to emit less pollutants. Third, improvements in the fuel the vehicle consumes. However, updates in research do not provide evidence for any drastic improvement in vehicle and fuel for less emission in near future as a result emphasis is paid on urban structure and travel behaviour to minimize travel and maximize the use of non-motorized and public transport.

Figure 2.4: Annual CO emission estimates for 2000 (Gg year−1). Source: Gurjar et al., 2008

Asian cities are generally densely settled largely because their inhabitants mostly have low mobility in contrast to affluent western cities (Barter, 1999). Although they vary widely in terms of their motorization rate, geography, public transport policies, and transport achievements (Townsend, 2003), their compactness does maintain accessibility through non-motorised modes and low-cost public transport (Barter, 2000). Studies on travel patterns for Dhaka reveal that work-based travelling accounts a significant share of trips (32%) (The Louis Berger Group, Inc. and Bangladesh Consultants Ltd, 2005). To address this spatial phenomenon, major employment centres are found concentrated around transport nodes and vice-versa (Maat & Stead, 2005). As a result, location of employment centres stands out as an

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 31 influential element in urban structure and the design of cities. These employment centres, especially in large cities, have different characteristics and possess different levels of strength in attracting traffic, but their contribution to the transport sector is significant. Depending on job type, an employment centre has a commuting footprint for its employees. Employees may come from suburbs of similar socio-economic configuration or from different ones; depends on the type of jobs. As a result, the total commuting footprint for any employment centre depends on the type of economic activity, job numbers and the overall urban fabric – allocation of residential areas and its connectivity (transport) with jobs. As a result, the spatial planning process attempts to minimize the friction between workplace and residence. These phenomena emphasise the spatial planning, facilitating urban form to contribute to the mitigation of environmental problems caused by transport (Zhao et al., 2009). Many empirical studies have already found that as well as transport services, urban form can have a significant effect on travel patterns, even when residents’ socioeconomic factors are taken into account (see for example, Cervero & Kockelman, 1997; Crane, 2000; Boarnet & Crane, 2001; Buchanan et al., 2006; Coevering & Schwanen, 2009).

Several studies have analysed urban form to explore relationship with travel behaviour. A summary has been provided in Table 2.2. These studies classify the sub-centres (mostly neighbourhood level) based on some criteria and compare their travel behaviour. For example, Stead and Marshall (2001) use nine criteria (i. distance of residence from the urban centre; ii. settlement size; iii. mixing of land uses; iv. provision of local facilities; v. density of development; vi. proximity to transport networks; vii. availability of residential parking; viii. road network type; ix. neighbourhood type) to examine five measures of travel patterns (i. travel density; ii. journey frequency; iii. modal split; iv. travel time; v. transport energy consumption). However, most studies consider density as a major characteristic of urban form and find characterizing travel behavior for it (Table 2.2). one reason for this is because density information is readily available in most cases and residential density affects travel behavior by getting people in close proximity to destinations, consequently

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 32 increasing the number of possible destinations that can be reached within the same range of distance (Grazi & Bergh, 2008).

The majority of studies exploring the relationship between urban form and transport find a statistically significant but usually small effect of density on travel behaviour (Handy 1996; Levinson & Kumar 1997; Boarnet & Sarmiento, 1998; Ewing & Cervero, 2001). However, on the contrary, a number of theoretical studies suggest a negative relationship between commuting distance and residential density (Fujita, 1989). For example, Newton & Kenworthy (1999) used a sample of 32 cities in four continents and find a negative statistical correlation between residential density and transport demand - suggesting lower transport emissions by higher density communities. The main reason is that workers choose a residential location (conditional on workplace location) by making a trade-off between generalized commuting cost (or distance or duration) and housing price. In high density areas, workers benefit from a shorter commuting distance but face higher housing prices, whereas in low density areas they enjoy lower prices but travel longer distances. In addition changes in modal split, notably a shift from car to more energy efficient transport modes like walking biking and public transport will contribute to a reduction in transport related Green House Gas (GHG) emissions.

Several researchers have found that the features of compact land development on the urban fringe, such as high density, a high level of jobs-housing balance and compact physical pattern, would enhance accessibility to facilities and services, and thereby reduce the overall need for travel and the distance and duration of motorized journeys (for example, Cervero, 1995; Levinson & Kumar, 1997). A study conducted by Grazi et al. (2008) using housing survey data in the Netherlands showed that the density of urban form has a significant impact on mobility

(commuting distance and ) and thus on CO2 emissions. They found that workers living in the highest density locations tend to travel less than workers in the least dense location. In more densely populated areas workers tend to change from car to other travel modes, notably public transport (metro and tram). Accordingly, the predicted CO2 emissions per passenger are lower in the highest density areas. On the other hand, Gordon et al. (1989) finds a positive relation between metropolitan

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 33 residential densities and commuting times. Again studies by Cervero and Wu (1998) suggests that sub-centres do not have significant impact on commuting distance or time. The varieties in findings may cause due to data limitations, the use of different density measures, and differences in statistical methods. For example, some studies do not account for socio-demographic variables, whereas others only include attitudes and preferences of respondents.

A large amount of literature with empirical evidence is devoted to analysing the impact of city size and structure on GHG emissions via commuting (Lefevre, 2009; Brehany 1995; Muñiz & Galindo, 2005). Monocentric development invites agglomeration of economic activities, and thus reduces the level of pollution from commodity shipping between cities; in addition, it increases commuting densities and congestion within the city, hence the associated polluting emissions (Legras & Cavailhès, 2011). Multi-centric job concentration on the other hand provides local jobs and reduces commuting length. But it may incur cost from inter-city freight and passenger movement. Bertaud (2001) presented schematic trip patterns for four different urban forms (Figure 2.5`):

(1) monocentric model: the labour market remains unified and sustainable commuting can be achieved through rail or other mode of public transport. Population density tends to follow a negative slope from the CBD. (2) polycentric model (urban village): this urban agglomeration composed of several ‘self-supporting urban-village’. Ideally everyone can walk or cycle to work. According to Bertaud, this is the ideal condition that never observed in any city. (3) polycentric model (dispersed): the polycentric city with random movement appears when satellite towns are developed to accommodate increasing population and often people from these satellite town commute to main city and vice-versed. (4) Mono-polycentric model: This type of city is the result of a development in initially monocentric large cities whose structures have gradually evolved into a polycentric pattern. The CBD loses its

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 34 primacy and activity clusters generating journeys are distributed throughout the built-up urban area.

Figure 2.5: Schematic Trip Patterns for Different Urban Forms Source: Bertaud, 2001

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 35 Different debates are going to depict the nature of travel as it can help in designing neighbourhoods and allocating resources to the space. Ewing and Cervero (2001) reviews a wide range of research work and explores the relationship between travel behaviour with neighbourhood design to a certain degree. Another very popular discussion is job-housing balance. Jobs-housing balance refers to the (dis)parity between the number of jobs and housing units within a geographical area (Cervero & Wu 1998). According to Cervero and Wu (1998), the simplest measure is the jobs- housing ratio, i.e., the ratio of the number of employees to the number of resident workers. An area is considered imbalanced when the number of resident workers far exceeds the number of jobs, and therefore some of the residents have to travel long distances to their workplaces.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 36 Table 2.2: Studies on Relationships Between Urban Form and Transport

City Urban Form Transport Unit of Significant Author Variables Variables Analysis Relationship 31 large cities, USA Landuse Mix VMT, average Urban area Landuse (at the Landuse density speed, Modes, aggregate level) is Barnes, 2001 vehicles minute, not a major total travel leverage point in time/day, determining overall travel choices UK Size, Energy Urban area Urban containment or Breheny, 1995 Density consumption compact city has failed to reduce energy consumption Los Angeles, USA Distance to CBD % of Job, % of Traffic Zones Job-housing resident worker, balance has minor Giuliano & Small, 1993 excess job effect on commuting Christchurch, NZ Population density Distance travelled Traffic Zones Low density areas generates and Buchanan et al., 2006 receive higher trips Portland, USA Landuse Trip frequency Traffic Zones Landuse and travel Residential area linkage are Ewing et al., 1996 multidimensional and difficult to comprehend. USA Gross population Trip rate, vehicle Census tracts Population density density trips, walk trips, is an determinants Dunphy & Fisher, 1996 trip frequency of travel

Seattle, USA Gross population, Trip distances Census tracts Landuse mix at the Employment density, (Commuting, trip origins and job/housing shopping), travel destinations time increase in transit Frank & Pivo, 1994 and walking

San Francisco Bay Area, Accessibility, landuse VMT, car Census tracts Mix landuse USA mix, landuse entropy, ownership, trip reduce trips and population density, rate, mode choice vehicle use Cervero & Kockelman, job density 1997

Greater Oslo, Norway Job density Modal split, Location Data Central and high Energy density job district Næss & Sandberg, 1996 consumption attracts shorter travel hence consume less energy compare to peripheral and low density job district

San Diego, California, USA Household Trip frequency, Location Data Landuse explains characteristics, modal choice travel behaviour Crane & Crepeau, 1998 Neighbourhood characteristics

( continued to next page )

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 37 Sydney, Australia Distance from CBD Commuting at local and distance regional level Black, Paez & Suthanaya, different scale of 2009 urban form generates optimal travel pattern.

Source: compiled by author

These researches in many cases advocate for a ‘contained’ and compact urban area development with a mix of uses in close proximity (Williams, 2005). This is due to the assumption that such forms reduce travel demand because people can work near their homes and make use of local services and facilities. Such forms can also provide population densities high enough to support public transport services. However, the urban planning literature does not suggest or discard any particular form, but suggest that achieving sustainability should not be pinned on just the one option (Arbury, 2005). As Williams et. al. (2000) conclude, ‘The search for the ultimate sustainable urban form perhaps now needs to be reoriented to the search for a number of sustainable urban forms which respond to a variety of existing settlement patterns and contexts’ (Williams et. al. 2000, p.7).

At the same time, there is considerable uncertainty about the possible success of such ‘compact’ development to contribute to sustainable mobility at all (Modarres, 2003; Williams, 2005; Breheny, 1995; Rickaby 1987). Even if it is possible to find a consensus on which urban form is the most beneficial in terms of sustainable transport, there are still questions about our ability to implement substantial changes in the physical fabric of cities through the planning system (Williams, 2005). Existing urban form changes relatively slowly, and opportunities for newly planned towns and cities are limited in most developing countries.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 38 2.4 Modelling Urban Spatial Structure and Sub- Centre Identification

Urban spatial structure is the arrangement of land use in an urban area. The way that urban space is arranged affects many aspects of how cities function and has implications for accessibility (Guiliano & Small, 1993), environmental sustainability (Camagni et al., 2009), social equity (Bramley and Power, 2009) and economic growth (López & Muñiz, 2011). It can be defined as the degree of spatial concentration of the population and employment (Anas et al., 1998). Urban spatial structure is also concerned with the arrangement of public and private space in cities and the degree of connectivity and accessibility. Urban structure of a city is the outcome of the economic interactions between firms and households, which favour spatial concentration by reason of agglomeration economies (Fujita & Ogawa, 1982). ‘Agglomeration economies’ or ‘economies of agglomeration’ on the other hand, refers to benefits that firms obtain when locating near each other. Depending on the degree of concentration or dispersion of economic activities, these agglomeration economies vary among intra-metropolitan locations. Some of these locations might become main employment centres, whereas others remain as non-central. Sociologists, economists, and geographers have developed several models that describe and explain the distribution of economic activities and connectivity of different types of people and businesses that exist within the urban setting. Among others, the "trade-off" model of land use, first developed by Von Thünen in the early nineteenth century and modernized by urban economists such as William Alonso (1964), Mills (1967) and Muth (1969) formulates an optimal land use pattern by evaluating a trade-off between transportation costs and land rents. While originally applied to agricultural land use, the Von Thünen model is commonly used to explain urban land use patterns. This model, originally conceived for a monocentric city (Figure 2.6), explains that rent for a household (firm) is a function of its distance from the city centre. As a result, to achieve locational equilibrium, households bid up for land according to expenses saved in commuting. Thus the closer is the residing place to the CBD (where all employment is supposed to be) the higher is the rent transferred to land (which capitalizes into higher prices), emerging in this way a land rent gradient (Figure 2.7). In a monocentric model, most of the jobs are located at the

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 39 city centre, and peripheral areas concentrate on housing in low density urban schemes since the price of a unit of land decreases as the distance to the CBD increases (Figure 2.7). Although urban location models have been developed mainly to explain residential location, they can also be applied to employment location (Duarte et al., 2011). It is observed that in all these models, urban structure and transportation play a pivotal role in determining the location of an activity because an area's relative location is determined mainly by transportation systems (Shin, 2002).

Figure 2.6: Monocentric Model Figure 2.7: Bid Rent Curve

However from the viewpoint of reality a monocentric urban landuse pattern seems to be untenable (Fujita & Ogawa, 1982). The effects of city size growth especially after the World War II and diversification and globalization of economic activities, have questioned the theoretical and empirical appropriateness of the monocentric urban pattern to explain and describe the spatial organization of modern metropolitan areas (Gordon et al., 1989; Helsley and Sullivan, 1991; Anas et al., 1998; Baumont, 1999). In recent years, the monocentric model has become even less relevant because metropolitan areas are increasingly decentralized with the central business district becoming less important in terms of employment, and new subcentres are emerging at the edge of the cities (Riguelle, 2007). In other words, employment is decentralizing to the suburbs and clustering within the suburbs. In a large city, subcentres arise from the trade-off between external scale economies in production

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 40 and diseconomies of scale in transportation (Helsley & Sullivan, 1991). These changes in urban development patterns present an interesting challenge to urban researcher and planners interested in the spatial structuring of the mega cities.

‘Agglomeration of economy’ is a term which often refers to the cause of subcentre formation. Agglomeration economies are related to cost reductions and increasing outputs when firms locate close together in some areas (López & Muñiz, 2011). In simple terms, employment decentralization is explained by a trade-off between transportation and economies of agglomeration (Shukla & Waddell, 1991). Close proximity to other businesses (forward-backward or linked or similar) and services places industries and businesses in the centre of the advantage of scale. Thus, economic activities tend to concentrate in the centre. Transport costs of goods decline at the peripheral areas as firms locate closer to workers to minimize commuting costs (Gordon et al., 1989; Guiliano & Small, 1993). Decentralization would continue until firms reach the equilibrium between transportation cost savings and other advantages, such as economies of agglomeration at the centre (Shin, 2002). Moreover, some locations in the periphery are better in terms of accessibility, with the proximity of a port, a motorway or an airport. This improved accessibility for their labour force and consumers is a factor in decentralization for firms as well as for populations (Giuliano & Small, 1999). Thus, transportation investments at peripheral areas would have a decentralization effect by improving the relative locational advantage of the areas. Both the locational decentralization of economic activities and sub-centre growth are related. As a result, urban structure is found as the outcome of agglomeration economies operating in different parts of metropolitan space. It encourages the spatial concentration of economic activities and jobs and, eventually, the formation of main centres and the emergence of monocentric and polycentric spatial structure (Fujita & Ogawa, 1982; Anas et al., 1998; López & Muñiz, 2011).

In addition to agglomeration economies, several factors explain this sub-centre formation. One is increasing land scarcity in the core and the resulting high costs, which are a particular problem for land consuming activities. Traffic congestion and

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 41 pollution are usually confined to congestion in the core and provide incentives for decentralization.

Improved accessibility by public transportation acts as "pull" factors for sub-centres. There are also "push" factors in older centres. As a city centre grows and ages, various other negative externalities such as insufficient parking, air pollution, crime, and high rents--undermine economies of agglomeration (Gordon et al., 1989). In suburban areas, industries may find better places with lower land values, better transport links with highways or water ways, and closer proximity to preferred housing for the workers. In addition, telecommunication and information technology is considered a fundamental force in restructuring production and distribution processes (Ioannides, 2008). New technology tends to weaken the advantage of agglomeration at older city centres. The transition to services and high-tech industries makes firms more mobile and more dependent on quality workers and environmental amenities (Gottlieb, 1995).

Employment decentralization is also related to widespread automobile availability, transportation systems and parking provision (Rodrigue et al., 2003). A well designed public transportation system presumably gives rise to better access for suburban workers. Yet, the effect would be differential among social groups. As articulated by the "spatial mismatch" hypothesis (Kain 1992; Ihlanfeldt and Sjoquist 1998), the decentralization of jobs may negatively affect economic opportunities for low-income minorities that reside predominantly in the inner city and places opposite of growth axes. On the other hand, some negative features of urban growth like traffic congestion and environmental deterioration--may lead suburbanites to the anti-growth movement (López & Muñiz, 2010).

There are different perspectives by the studies on urban spatial structure. Most studies are designed to explore the spatial structure of a city or to know whether it is monocentric, polycentric or dispersed, or to compare the changes in spatial structure, or to use the result of urban structure as inputs to . A major point of difference between studies on polycentricism is the scale of analysis. Polycentrism can either refer to intra-urban patterns of clustering (such as in Los

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 42 Angeles, London or Paris) or to inter-urban patterns (such as the Dutch Randstad or the Flemish Diamond) (Kloosterman & Musterd, 2001). The confusion is exacerbated by the fact that employment centres within a given urban region can form an interdependent system, with a size distribution and a pattern of specialisation similar to that of cities in a larger regional or national economy (Anas et al., 1998; Riguelle, 2007). But, even if we find similar patterns at different scales, the causes and the consequences of their development may vary.

Talukder (2006) identified 10 essential criteria of urban governance for polycentric growth management and found that often megacities in developing countries do not satisfy them (Table 2.3) and Dhaka is not exception.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 43 Table 2.3: A Comparative Analysis of Governance of four Asian Megacities

Governance Principles Governance Institutions/ Settings of Asian Megacities and their Extended Metropolitan Region

Metro(EMR)-Manila Capital Tokyo Metropolitan Bangkok Metropolitan JABOTABEK

1 A Geographical Area with clearly YeRegions……M etro-Manila No…Region but Tokyo No…..BangkRegion (BMR)ok Metropolitan No…Region.JAB OTABEK is a defined responsibility of Governance Development Authority (MMDA) Metropolitan Region is being Region (BMR) is a Planning planning term and exists in

that covers the Extended is authorized for planning, governed from a large and Term but is not yet a JABOTABEK supervising, and coordinating local governance system political reality. Metropolitan 2 StrategicMetropolitan planning Region fu nction that can Yesdeve lopment and certain Yes No …though Central Yes…..BAPPENAS/ provide a vision for how the city can Metro-Manila Development Central Government and Government has a Structural Central Government

address its land use Authority (MMDA) TMG have strategic plan Plan for BMR through through JABOTABEK problems sustainably for all Tokyo Region central government agencies Plan, but implementation kept with sectoral 3 Statutory planning function that can No Yes No No…DKI Jakarta, control development to ensure Tokyo Metropolitan BMA and Provinces of Districts and

‘common good’ outcomes 7 cities and 10 Municipalities Government (TMG), and Nonthaburi, Pathumthani, Samut of Bogor, Bekashi , consistent with the strategic plan have statutory powers but not at Local Governments of Prakan, Samut Sakhon and Tangeran---some statutory MMDA level Saitama, Chiba and Nakhon Pathom have some power but none at regional 4 Development facilitation function No Partly….Yearly Summit of No Nosca le that can provide investment MMDA has some development Mayors and national More than 50 agencies share National government

coordination, partnership for facilitation powers government processes responsibility of planning, processes only infrastructure and local financing and managing validation structure development programs and operation of basic services ( continued to next page )

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 44 5 An urban redevelopment authority to NO – Urban renewal, zoning, and Partly No…. No

guide and monitor further land use planning done by MMDA TMG has responsibility for development in the core built- but can’t implement 6 Aup t rareasanspa inrent the l ocalcity process that can Yes Yesurban renewal but without No Weak

help define ‘common good’ NGOs/UPOs provide strong civil Strong local process of Weak links to civil society only Kampong governance is sustainability outcomes from society presence engaging neighbourhoods in place but is used mostly development with all in city governance 7 A coordination mechanism to ensure Partly – MMDA coordinate metro- Mayors Summit once a year Yes…..a multi-ministerial No…for top..howe downv er, DKI

planning and development wide development planning and for Regional Issues Bangkok Metropolitan Regional Jakarta has a Regional are integrated coordination but have no provides significant Development Committee Council to advise the implementation powers so coordination (BMRDC) oversees and Governor of DKI Jakarta integration occurs in theory coordinates urban development only programs in BMR. But not in 8 A way of raising the finance for the No Yes NoExt ended Bangkok No

above process including from land Some financing mechanism

9 Ade vstrongelopment link in to the national MMDA and National Economic Yesusing land No……only through the No…. just the occasional

government system to enable good Planning Department play an TMG and Local Department of Interior and other national intervention

political support important role Governments of 3 sectoral Ministries 10 New professional skills in No No No No

sustainability and local participation Source: Talukder, 2006

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 45 2.5 Applications of Urban Structure Studies

There are various debates on the structure of contemporary large cities. Duarte et al., (2011) and López & Muñiz (2011) claim that currently most large cities in the world are polycentric. Employment decentralization tends to take a polycentric form as firms cluster in a number of sub centres. Suburban employment centres are often characterized by concentrations of low to mid-rise commercial buildings with extensive parking lots centreed on a large shopping mall near highway intersections. But Gordon and Richardson (1986) in their famous studies claimed that polycentricity might be an intermediate spatial stage between monocentric and dispersed development. Bertaud (2002) on the other hand ascertain that in reality, no city is purely monocentric or polycentric but lies somewhere along the continuum. However, there is general consensus that in the last three to four decades urban areas have experienced strong but heterogeneous sprawl which has led to the structuring of polycentric structures (Anas, Arnott & Small, 1998). Polycentric urban areas combine many of the advantages of big and small urban areas. It would offer the two major economic advantages of urban systems: the presence of agglomeration economies, which results in increasing returns for companies, and a potential reduction of transport cost, which leads to a reduction in salaries and land rent (McMillen, 2003). Moreover, a polycentric urban model has social and environmental benefits arising from an improvement in transport planning and drop in commuting (Gordon et al., 1986). Whereas the traditional central business district (CBD) offers firms the advantage of significant agglomeration economies, it also requires high wages to compensate for expensive and time-consuming commutes (McMillan & Smith, 2004).

Studies on urban structure are nothing new. It has been historically practiced to comprehend the function of different centres of towns and often analyse them with their size, population, traffic etc. Central Place Theory, first proposed by Walter Christaller (1935), provides an explanation for a hierarchical organization of service and retail centres. In this model, the size, location, and function of centres are determined by market threshold and the consumer's willingness to travel. High-order goods and services, such as specialized goods and services, are purchased less often and may involve a higher cost of travel. Such services are offered in a few larger

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 46 centres. Conversely, low-order frequent goods and services, such as items offered by convenience stores, are purchased more often and regularly. Related to the notion of central place theory, Giuliano and Small (1991) found that density and frequency of centres declined with distance from the downtown area in the Los Angeles region. Many of the centres were specialized by industry. Pivo (1993) found that the proximity of office space roughly decreased as size and density increased in the Toronto region. McDonald and Prather (1994) found that in the Chicago urbanized area the density of suburban employment centres was influenced by distance to the CBD and three major employment centres (O'Hare Airport, Schaumburg, and central DuPage County). Cervero and Wu (1997) identified three tiers of hierarchical employment centres encircling downtown San Francisco in the Bay Area.

Many studies have identified various types of employment centres by using physical and economic factors to foresee implications for metropolitan development and transportation. Baerwald (1982) identifies two general physical forms of suburban "downtowns" in Minneapolis: "clusters" which usually focus on a regional shopping centre, and "corridors" which string out along axial freeways. He also identifies four general factors affecting cluster and corridor development: locational tendencies of different land uses, transportation systems in and around the concentration, historical factors and the timing of development, and other factors (e.g., social-demographic patterns, local government policies, and entrepreneurial prerogative).

Researchers have employed multivariate statistics to identify different types of employment centres and their contribution to the overall job market. Maoh and Kanaroglou (2004) using temporal data with the multivariate explanatory approach successfully identified job and population sub-centres at the outskirts of Hamilton, Ontario. Using cluster analysis, Giuliano and Small (1991) identified five types of sub centres in the Los Angeles region: specialized manufacturing, mixed industrial, mixed service, specialized entertainment, and specialized service. As indicated by the names, the classification largely relied on the type of business activities and some spatial factors (e.g., distance from the CBD and employment density).

As with the decentralization of jobs and population, traffic congestion grows around centres, researchers have examined the location and land use characteristics of

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 47 different types of sub-centre in order to relate them to transportation problems such as car dependence, traffic congestion, air pollution, and many more. Using factor analysis and cluster analysis for America's 50 largest suburban employment centres, Cervero (1989) identified six types of centres: office parks, office and concentrations centres, large-scale mixed-use developments, moderate scale mixed-use developments, sub-cities, and large-scale office growth corridors. The classification was largely based on detailed physical factors such as density, land use mix, floor area ratio, lot size, and many other factors.

Aguilera and Mighot (2001) in their study of seven very different French urban areas emphasize the existence of employment sub centres which, in addition to the city centre, are the main destination of the workers. Two kinds of sub centres are highlighted: the suburban sub centres, vast and close to the centre with which they make up a greater centre characterized by a high jobs-housing proximity, and outlying sub centres, smaller and further out but well situated along the main transport axes. The latter encourage a certain proximity to jobs for the workers living in those sub centres, but still depend strongly on the jobs located in the suburban sub centres.

McMillan and Smith (2004) reviewed and tested Fujita and Ogawa’s (1982) model of urban spatial structure. The central theoretical prediction is that the number of employment sub centres rises with population and commuting costs. Simple Poisson regressions for a sample of 62 large American urban areas provide strong support for the theory, and these two variables alone account for nearly 80% of the variation in the number of sub centres. The results imply that an urban area with low congestion develops its first subcentre when its population reaches 2.68 million and its second sub centre at a population of 6.74 million.

Alpkokin et al. (2005) in examining the rank of job clusters identified four sub- centres for and their commuting patterns which shows very distinctive features. Using a similar method Bhandari et al. (2007) identified four tiers of urban centres in Delhi, India. They are: tier 1- mature old centres; tier 2 - zones as rather developed sub-centres; tier 3 - zones as emerging as a sub-centre and tier 4 - zones as not necessarily now.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 48

There is other research work on urban structure that explores urban spatial structure to search for sub-centres and later relates the attributes of sub-centres with travel, transportation, emissions, rent and many more variables. Table 2.4 provides a short list of these, particularly those that look into urban form variables. An expanded list of studies can be found in Ewing and Cervero’s (2001) seminal work on travel behaviour and urban form.

2.6 Issues In Methodology and Measurement of Polycentricity

Polycentricism in mega cities is recognized by several researchers and its existence is well accepted for the major cities in the western world; however, empirical methods for identifying sub centres have lagged behind in term of its application for sustainability. These methods fall into five major categories: i) employment density peaks. ii) cut-off or reference threshold methods iii) clustering methods iv) parametric models and v) non-parametric models.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 49 Table 2.4: Different Methods Applied By Researchers to Identify Employment Centres

Methods Developed by Definition / concept Adopted by Employment Density peaks McDonald (1987) Employment sub-centres are area with Gordon et al. (1986); significantly higher employment McDonald & McMillen density than surrounding areas. (1990);

Reference Thresholds Giuliano and Small Employment centres should have Small & Song (1994); (1991) minimum 10 employment per acre and Cervero & Wu (1997); minimum 10,000 in total. McMillan & McDonald (1998); Giuliano & Small, (1999); Garcia-López. Muñiz (2010)

Clustering Method Alpkokin et al., (2005) They plot the log of employment Bhandari et al.( 2007); density against the rank of zone. Vichiensan (2005) Identified the cluster within given range of ED-rank curve.

Parametric Methods McDonald and Prather The sub-centres are identified through (1994) the analysis of positive residuals in a negative exponential function of employment density.

Non-parametric Methods McMillen and Based on non-parametric statistics this McMillen, 2001; Craig & McDonald (1997) model (locally or geographically Ng, 2001 weighted regression) detects “peaks” that locally adjust the density function and prioritize the effect of neighbouring sub-centres on the adjustment process .

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 50 There are many ways of identifying subcentres, but the choice of method is important given that the number of subcentres and their boundaries is quite sensitive to measurement definition (Anas et al., 1998). This choice is often guided by the availability of data. There is great heterogeneity of data in space and time. The spatial units of aggregation can be very different from one place to another and from one period to another, leading to difficulties in the comparisons. Moreover, the number of subcentres depends on the size of the unit of observation (McMillen, 2001a). There is also a lack of homogeneity in the way economic activities are classified.

2.6.1 Employment Density Peaks

McDonald (1987) first suggested a procedure for identifying employment density “peaks”. He showed a clustered pattern of employment sub-centres and used two criteria to guide sub-centre definition. First, considering the traditional definition of an urban area, a sub centre is defined as an area with significantly higher employment (not population) densities than all adjacent areas. Second, a sub centre should be large enough to have a significant effect on the overall land use structure of the urban area, by increasing population density, land prices, and even housing prices. Sub-centres produce clusters of positive residuals in the estimated function. McDonald finds that O’Hare Airport is the only dominant subcentre in the Chicago metropolitan area.

For the first criterion of identifying employment density “peaks”, there is a requirement to analyze employment density to detect local disruptions with the aid of a geographic information system (GIS). Alternatively, the employment-to-population ratio can be used to detect the areas that have higher relative concentrations of economic activity (Duarte, 2001), but this has been criticized by Giuliano and Small (1991) who argue that population should not be taken into account in defining a subcentre. The identification of employment ‘peaks’, in addition to thorough visual inspection, can be carried out by means of statistical techniques, such as those used by Gordon et al. (1986). Gordon et al. (1986) restricted the number of subcentres to

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 51 those areas with high t-values; this line of research was continued by McDonald and McMillen (1990) and Craig and Ng (2001).

McDonald’s novel approach poses several problems in practice. The procedure suffers from a statistical problem since it recognizes only the highest employment density compared to its surrounding; it may ignore centres which have adjacency to another large sub-centres. This definition does not address the size of subcentres that cover two or more zones (Giuliano & Small, 1991). The results are also sensitive to the unit of analysis (McMillen, 2001a). McDonald and Prather (1994) found additional sub-centres in Chicago using a smaller track of the same area.

2.6.2 Reference Threshold Method

Although numerous methods that have been developed for employment subcentre identification, the threshold methodology is one of the most used due to its objectivity and replicability (see Giuliano & Small, 1991; Song, 1994; Cervero & Wu, 1997; Forstall & Greene, 1997; McMillen & McDonald, 1997, 1998; Bogart & Ferry, 1999; Anderson & Bogart, 2001; Shearmur & Coffey, 2002; or Giuliano et al., 2007; Garcia-López & Muñiz, 2010 among others). The method specifies a minimum level of job density and a minimum amount of employment: an employment centre is a location with a density level greater than the density cut-off and with an amount of employment greater than the employment cut-off.

Greene (1980) defined employment sub-centres as areas with an employment density higher than twice the average of the group of zones of the metropolitan system. Another example is that of Cervero (1989) who opted for a fixed critical mass (2000 local jobs) and a stock of office floor space (1,000,000 square feet). Giuliano and Small (1991) built on McDonald (1987) by accepting employment density criteria. They define a sub-centre as a continuous set of zones, each with density above certain cutoff, that together have at least a specified total employment. They selected the sub-centres based on two criteria; first an agglomeration of contiguous tracts having a minimum employment density of 10 jobs per acre, and when combined

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 52 having total employment of at least 10,000. By applying this method, Guiliano and Small (1991) identified 32 sub-centres in the Los Angeles area.

This is considered to be the most acceptable method for sub centre identification which clusters adjacent tracts to build sub-centres based on employment density and total employment thresholds. It has been employed in subsequent work by Bogart and Ferry (1999), Giuliano and Small (1999), Cervero and Wu (1997, 1998), McMillen and McDonald (1998), Small and Song (1994) and Parolin and Kamara (2003). However, several studies used adjusted cutoff to adjust the character of the study area. Many researchers have sought variations of this method to extract urban structure information. Bogart and Ferry (1999) used the same total minimum employment of 10,000 jobs but a lower employment density of 5,000 jobs per square miles or about 8 jobs per acre to find 9 employment centres in Cleveland, Ohio. Cervero and Wu (1997) changed both the total employment and the density criteria. They used total number of jobs greater than or equal to 95,000 and the density is greater than or equal to 7 jobs per gross acre. They identified 22 employment centres in San Francisco using 1990 Census Transportation Planning Package Urban Element (CTPP-UE) data. Rather than lowering total employment and density criteria as Bogart and Ferry (1999) and Cervero and Wu (1998), McMillen and McDonald (1998) raised the minimum cutoff points to 20,000 total employees and 20 workers per acre to avoid unreasonably large sub-centres. They found about 20 employment sub-centres in Chicago in 1980 and 1990. Muñiz and García-López (2008) went a step ahead and suggested that sub-centres are zones with a density higher than the metropolitan average and having at least 1% of metropolitan employment and minimum density above metropolitan average. For Barcelona Metropolitan Region (BMR) they identified 18 sub-centres within a 12 km radius from the main city. Cladera et al. (2009) suggest to impose another criteria on the top of employment density and total employment to incorporate the interaction between different zones and termed it as the ‘functional approach’. In the functional approach, commuting flow are estimated by a gravity model (chapter 4 for more detail) and used as interaction values for the BMR.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 53 The reference threshold approach received criticism at first for the apparently arbitrary nature of the density and total employment cutoffs (McMillen, 2003; Redfearn, 2007). However, over time, their results held up well against other approaches and what first appeared arbitrary has been referred to as “local knowledge” (Redfearn, 2007). What the early criticism pointed to is the fact that the researcher can “choose” the number of centres by choosing the thresholds: the number of sub-centres found can be increased by decreasing the threshold and vice versa. However, “local knowledge” turns out to be useful, and the method has been replicated numerous times. Applying this method Giuliano and Small (1991) identify 32 sub-centres in the Los Angeles area. The same cut off points gave unreasonably large sub-centres in northern Chicago, leading McMillan and McDonald (1998) to raise their cut-offs to 20 employees per acre and 20,000 total employees. Local knowledge must guide the choice of cut off points through trial and error, limiting analysis to familiar metropolitan areas.

Another limitation of this approach is that the cutoffs do not vary over the data set (McMillen, 2001b), which means that the minimum subcentre size is the same near the CBD as in distant suburbs. This characteristic of their procedure is not desirable. If a subcentre is defined as an area with larger employment density than surrounding areas, small and relatively low density employment centres may not be identified. Since densities tend to decrease with distance to the CBD, the minimum cutoffs should tend to decrease also. Then the question becomes how to vary the cutoffs.

Giuliano and Small’s procedure is also sensitive to the unit of analysis. They had 1146 tracts covering an area of 3536 sq. mile of Los Angeles while McMillen and McDonal (1998) used 14290 tracts for 3572 sq mile of Chicago. Datasets with small tracts are more likely to have pockets with low employment density (McMillan, 2001a,) which reduces the number of sub-centres. Tresserra (2011) criticized that the larger the number of spatial units in the metro area, the highest is the difficulty to achieve the criteria of a sub-centre, and the most homogeneous is the density function across units, the higher is the probability that a large number of units are above average density (Duarte, 2011).

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 54 As McMillen and Lester (2003) noted, this method has advantages when comparing results for a single city over time, mainly the use of the same cut-offs for different years. However, if we use numerical thresholds and we study the city in a period of intense growth of employment, then there is a trend to identify an excessive number of employment centres. As Muñiz et al. (2008) noted, we can avoid this problem using statistical thresholds, which are more flexible and correct the inflationary effect on the number of centres identified.

Another common criticism to this cluster method is its strong assumptions regarding physical symmetry of the city (Redfearn, 2006). Moreover, this simplistic approach can produce results that are not statistically reliable, as the cut off determination is not statistically derived rather a rule of thumb determine by the researcher. Nevertheless, they excessively rely on local knowledge to calibrate the thresholds of critical mass and density, and this can be a problem when trying to compare different metro areas with different local experts.

However, despite all those limitations the ‘Reference threshold method’ has become most popular and has been adopted by most because of its simplicity and flexibility to adopt with employment patterns.

2.6.3 Rank Size and Clustering Methods

Research undertaken as part of the Eastern Asia Society for Transportation Studies (EAST) initiative has proposed a multi-state approach to identify and analyse employment centres. Pelin Alpkokin (2005) and others in the EAST group devised an alternative approach for large cities in developing countries to match with their data inadequacy characteristics. In this method, the logarithm of employment density has been plotted against the rank size of cities. “Cluster analysis” is then used to identify employment clusters (Figure 2.8). In this method, after plotting the rank size distributions as a two-dimensional graph where gross employment density is on the y-axis and the ranks on the x-axis, the change in job location patterns and the emergence of some new sub-centres may be examined by plotting the same data for different time points. This is then used to explain the pattern of growth.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 55

C l u st e r 1 C lu Ln (Emp. Density) s t e r 2 C l u s t e r 3

Rank of zones

Figure 2.8: Rank Size Distribution of Logarithmic Employment Density Source: Alpkokin et al., 2005

This method is useful when comparing job distribution over two time periods. It can analyse the dynamics of change (Figure 2.8). If the increment of employment growth is exactly the same in every zone then the two distributions are parallel (Figure 2.9a). Other theoretical patterns are possible: smaller increments in the big centres and larger increments in the smaller zones – decentralization (Figure 2.9b); larger increments in the big centres and smaller increments in the smaller zones – centralization (Figure 2.9c); and the possibilities of absolute declines in employment in the larger zones (or in the smaller zones).

Figure 2.9: Dynamics of Employment Density Changes Source: Alpkokin et al. (2005)

Alpkokin et al. (2005) used gross employment density instead of net density as an indicator with the assumption that different land-use functions other than employment may also have a potential capacity for accommodating economic activity in the future. Secondly, instead of a cut-off value, clusters of zones along the rank size distribution are used to identify employment zones. The findings of the EAST researchers highlight the presence of polycentricity in Istanbul (Alpkokin et al., 2005) and New Delhi (Bhandari et al., 2007). However, Vichiendsen (2005)

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 56 adopted the same approach for the mega city of Bangkok and found it is best characterised as an expanded monocentric city with strong economic influences exerted by the main centre.

2.6.4 Parametric Method

This method identifies sub-centres through the analysis of residuals of a regression in a negative exponential function of employment density. McDonald and Prather -bx+k (1994) tested a regression model based upon the model Di = D0e , where D0 is the density in the CBD and x the distance to the CBD. The sub-centres were delimited through the analysis of positive residuals, exceeding 95 per cent confidence of the regression equation (Cladera et al., 2009). This method makes strong assumptions regarding physical symmetry of the study area (Redfearn, 2007).

The negative exponential function of the parametric methods identifies the “local peaks” of job density, without the need to establish fixed reference thresholds. Nevertheless, even the most rigorous parametric methods based on regression models have indicated a basic limitation of the one-dimensional consideration (employment density gradient) of the metropolitan area differentiated only by distance to the CBD. The analysis of parametric methods is incapable of fully comprehending the wealth, variety and complexity of urban systems, especially of those generated over long historical periods (Cladera et al., 2009).

2.6.5 Nonparametric Method

The misspecification of monocentric models and the lack of statistical tests for the clustering methods have motivated the development of several different types of flexible, regression-based models—with McMillen (alone and with others) authoring the major share (McMillen & McDonald 1997; McMillen 2001a; McMillen 2001b; McMillen 2003; McMillen & Lester 2003; McMillen & Smith 2003). The best example of this is McMillen’s (2001a) model.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 57 The most commonly used non-parametric methodology consists of adjusting locally weighted regression (LWR) models (Cladera et al., 2009). McMillen (2001) proposed a two-stage methodology for the sub-centre detection. The first stage identifies candidate subcentres as significant positive residuals in a smoothed LWR employment density function. Sub-centres are those sites that provide significant explanatory power in the second-stage, semi-parametric employment density function estimation (McMillen, 2001a). Other non-parametric applications include those of Craig and Ng, (2001) who have used ‘employment density quantile functions’ through smoothed splines applied to 5 percent of the denser census tracts for the delimitation of sub-centres.

Nonparametric approaches offer significant advantages over simple linear regression procedures. Nonparametric estimators are flexible, allowing the slope of density functions to vary across the metropolitan area. As an example, suppose that employment density declines more rapidly on the north side of the city than on the south. The standard linear regression estimator used by McDonald (1987) imposes the same gradient on both sides of the city, which tends to produce positive residuals on the north side and negative residuals to the south (McMillen, 2003). This functional form misspecification increases the probability of finding a sub-centre on the north side of the city even if none exists.

Craig and Ng’s procedure is not as sensitive to the unit of analysis as the McDonald and Giuliano and Small procedures. Though larger tracts lead to smoother employment density functions, a large sub-centre will produce a rise in the function whether the data set includes acres, quarter sections, or square miles. The procedure is readily reproducible by other researchers and requires scant knowledge of the local area. Much of the arbitrariness of the Giuliano and Small procedure is eliminated because the local rise that defines a sub centre is subject to tests of statistical significance. However, the Craig and Ng procedure requires some local knowledge to choose which sites are sub centres within rings around the CBD, and the imposition of symmetry around the CBD is unsuited to cities that are distinctly asymmetric due to varied terrain or multiple sub-centres.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 58 Redfearn (2007) developed a nonparametric method but with difference from what McMillen did. McMillen’s approach highlighted outliners obscuring the local job landscape and that becomes potential centre. But Redfearn estimated locally weighted employment density surface with a subsample that is kept sufficiently small to keep intact the structure of local employment. Local maxima on the density surface become candidate centres. Using this locally weighted regression based model Redfearn (2007) identified 41 job centres in Los Angeles.

2.6.6 Other Methods

Exploratory Spatial Data Analysis (ESDA) ESDA is a set of techniques that explores the pattern of spatial relationships among zones in order to identify employment clusters. It defines a potential employment centre as an area having significantly higher employment and employment density than neighbouring zones. It applies global spatial autocorrelation, local spatial autocorrelation and spatial heterogeneity. Baumont et al. (2004) applied this technique in the city of Dijon, France to identify employment sub-centres which are statistically significant. Hakim (2010) relied on ESDA and factor analysis to identify major components of the spatial structure of employment in the Jakarta Metro Area (JMA). He revealed the hierarchy of the employment clusters in JMA, in which the Jakarta urban core and the regional CBD are at the top of the hierarchy (in terms of the spatial extent and the number of jobs), followed by manufacturing corridors, local government regions, desakota, and the agricultural areas.

GIS and Spatial modelling

Most approaches need to analyze the spatial relationship between objects, such as the adjacency of census tracts. However, the spatial relationship has been identified manually and inefficiently because both models do not have appropriate functions to handle it. Similarly, the empirical studies on the effects of subcentres on surrounding density and housing price also have difficulties in conducting spatial analysis (Pan & Ma, 2004).

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 59 Sophisticated mathematical modeling and statistical techniques have been used for sub centre identification but GIS based spatial modeling and cartographic techniques were rarely applied, though sub-centreing process has a very potential spatial dimension to explore. Few attempts were made but most of them are translation of empirical model to GIS.

Pan and Ma (2004) proposed a GIS to enable the Guiliano and Small (1991) method. They argued that the minimum cut-off method by Giuliano and Small (1991) has solid statistical meanings under the assumptions that employment density by tract in a region is distributed normally. They used the standard score or Z-score value for the employment density. Using the same 10:10 cutoffs, they identified 11 employment centres in 1990 and 12 in 2000 for Houston metropolitan area. Parolin and Kamara adopted a similar method for Sydney and identified 28 and 41 employment clusters in 1981 and 1996 respectively.

Modarres (2003) applied raster interpolation technique Krigging using the census track centroids, weighted by the total number of employments for Los Angeles County. He attempted to identify sub-centres as an aid to evaluate the public transport connectivity among the sub-centres.

Wang (2000) adopted the raster surface modeling technique in ArcInfo GIS software to create an employment surface from (log of) employment density data. He followed the modified Giuliano and Small like cut offs to identify employment centres in Chicago.

Number of resulted sub-centre of a metropolitan may largely depend on the method adopted. Tressera (2011) found 23 potential sub-centres in Barcelona Metropolitan Region (BMR) while applying parametric method. Threshold approach (based on Muñiz & García-López (2008)) resulted 11 and functional approach 13. In addition to method selection, unit of analysis is another factor. Moreover, all dataset do not necessarily support all methods. As a result, method selection should be carefully performed while looking into data characteristics and pattern in urban structure. It is noticeable that most case studies are from the developed world and many from the

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 60 US cities. This is because recognition of the polycentric nature of urban form is evident in many north American cities and its various data availability. The availability of city specific data for many developing world nations also hinders research in this area (Kasarda & Crenshaw, 1991).

2.7 Conclusion

Urbanization is increasing with an accelerating rate in developing countries in Asia. Developments of agricultural technology, population and urban pull factors have caused extensive migrations to the cities. Insufficient rural-urban linkage has transformed the fringe rural areas as captive workforce for the megacities. Availability of cheap land and labour, forces the industries to move towards the urban fringe and creates clusters of activities. Some of these sub-centres have been created by the government as an initiative to decongest the city but several have evolved organically and forced by agglomeration economies. Since planning control is generally limited in many cities in developing countries, these centres are developing via market forces and are heavily influenced by local socio-economic circumstances. Study initiatives require one to understand the pattern and character of those sub-centres and modelling their role in the space.

Metropolitan areas in the western countries have been characterized by polycentric structure in last decades (Tresserra, 2012) and an extensive body of literature has highlighted polycentric urban structure as the central theme for sustainable transportation (Aguilera, 2011). However, there has been a lag in developing a methodology that can be replicated for booming megacities in Asia in particular. This chapter reviews the prominent methods developed, modified and applied for large cities in North America, Europe, Australia and Asia and seek wisdom to develop a method that can be applied for Dhaka megacity with available data. It is revealed that availability of data in appropriate format is a key in method selection for sub-centre identification. Commonly used factors are population density, job ratio, nature of jobs etc. The methods and criteria deemed successful for North American metropolitan cities are ruled out in Asia by huge urbanization and

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 61 population growth and found not applicable. Chapter 4 of this thesis steer the discussion of method development for Dhaka megacity. This chapter and the next chapter on the urban evolution and planning context of the Dhaka Mega city provide a basis for the research approach to address the research questions on spatial structure and travel impacts of Dhaka, as formulated specifically in previous chapter.

Chapter 2: Urban Form, Travel and the Employment Centres of Megacities 62 Chapter 3 Urban Evolution and Planning History of Dhaka

3.1 Introduction

The earlier chapters present an overview of the research, set out the research aims and discuss the potential avenues for approaching the research problem. The previous chapter discusses the concepts and lessons learnt from empirical studies on the urban form and transportation relationship, and presented an analytical discussion on sub-centre identification methods. This chapter focuses on the megacity of Dhaka, which is the study area for the empirical analysis of this thesis. This chapter provides a basis to understand the study area and it aims to discuss some critical background on the urban growth and development of Dhaka – the demographics, landuse and transport dynamics that give an understanding of the area. It investigates and traces the development of settlement, major infrastructure and economic activities, particularly employment, in this rapidly growing megacity and examines its distribution and concentration patterns and its role in the growing urban landscape.

This chapter begins with a brief introduction on the role of Dhaka in the national economy. It describes the geographical and physical setting of Dhaka, the urban growth and development trends of the capital city and depicts the urban development, transportation and travel patterns that have emerged with the urbanization process. A historical overview on urban structure evolution and planning is discussed covering a wider time frame that includes establishment of the town by Mughals in early seventeenth century, its role and development as a colonial trading post until the mid-nineteenth century, early independence from the British,

Chapter 3: Urban Evolution and Planning History of Dhaka 63 its function as a provincial capital of Pakistan, and rapid growth following independence in 1971 that made it the national capital of a sovereign country.

Human settlement in Dhaka can be traced back as early as the 12th century (Ahmed, 1986). It is likely that the fertile land of the delta brought human settlement on this land. The Mughals recognized the importance of the land and established a city in the early seventeenth century (Karim, 1989). The Mughals established the city as their capital and strategic centre to rule the region, and as a business centre for handicrafts (Mohsin, 1989). Dhaka remained as British colony for two centuries and its growth went into demise to some extent during this period. The independence of the sub-continent after World War II in 1947 gave it a fresh impetus for development as it became the capital of East Pakistan – a province of Pakistan. The creation of Bangladesh as a sovereign country in 1971 further boosted development and it quickly grew to primate city status. The city in its present form translates its experiences through crowding, congestion, environmental degradation, land grabbing, mismanagement, corruption and political rivalry (Mahmud, 2007). However, as a home for millions and hope for more, it holds the opportunities for a better living place – a sustainable megacity that could lead this poverty stricken country towards a brighter future (Ashraf, 2010).

3.2 Prominence of Dhaka and its Economic Activities

Bangladesh is predominantly an agrarian country experiencing rapid urbanization and economic transformation. In 1974, only 8.8% of the country's 76 million people lived in urban areas; agriculture provided over half the national gross domestic product (GDP) and accounted for about three quarters of the labour force. By 2003 the level of urbanization had increased to nearly 25% and the contribution of the agricultural sector to GDP had decreased to less than 20% (United Nations, 2003). The share of agricultural workers as a proportion of the total labor force, however, had not fallen significantly, remaining at over 60%. The contribution of the urban sector to national GDP grew to nearly 42% in 1998–1999 from only 26% in 1972–

Chapter 3: Urban Evolution and Planning History of Dhaka 64 1973 (Bhattacharya, & Deb, 2006). This share is probably close to 50% at present (Islam 2006).

Dhaka is centrally located, and the most accessible city, from different parts of the country. Dhaka's status as a primary city is historical and is likely to continue into the future (Islam 2006). The total urban area of Dhaka spans about 7440 square kilometers (The Louis Berger Group, I. and B. C. Ltd, 2005), The concentration is also uneven in terms of economic activities. According to 1991 statistics Dhaka contained 36.5% of the total urban population. The other two major cities Chittagong and Khulna, have 12% and 4% of national urban population respectively. Following independence, Dhaka gained extra momentum in urbanization (Figure 3.1). The primacy of Dhaka is stronger functionally than in terms of population. Administrative headquarters and civil employment, financial and banking services, international commerce and business (except port functions) are all largely concentrated in Dhaka. Education, cultural, research activities are also concentrated in the capital area (Islam, 2006).

20 18 Chittagong 16 Dhaka 14 12 Khulna 10 8 6 4

Urban Population in Millionnds 2 0

Figure 3.1: Urban Population in Major Cities of Bangladesh Source: BBS, 2011

A disproportionately large concentration of industrial and various public sector investments have been made in the area. In spite of the declared government policy of decentralized administrative and economic development introduced in the early

Chapter 3: Urban Evolution and Planning History of Dhaka 65 ‘80s the actual development in the huge export oriented ready-made garments industries sector during the eighties shows an overwhelming concentration in Dhaka city. More than 80% of the 4000 garments industries of Bangladesh are located here (Mohsin, 1989; Islam, 2006). Concentration is also obvious in the social service, trade, commerce, and finance sectors. For example, 51 of the country’s 59 private universities established in the last decade are located in Dhaka City. The situation is similar with respect to medical facilities. Nonetheless, some degree of industrial concentration is also taking place in the second largest city, Chittagong, because of its port (Islam 2006).

Because of this concentration of income opportunities, per capita income in the city is much higher compared to the national level or with other cities in Bangladesh. In 1999-2000, average per capita income in the city was about US$872, with a national average of US$363 (Ahmed, Ahmad & Mahmud, 2005). Primarily as a result of the large job market in Dhaka, the city received large numbers of migrants every year. The manufacturing and garments industries located around the city has influenced the job market significantly by employing the female workforce in large numbers (Karim, 2012).

450

400 Rural 350 300 City and 250 200 Other Urban

Household number (in Thousands)(in number Household 150 100 50 0

Figure 3.2: Households by Main Sources of Income Source: BBS, 2001

Chapter 3: Urban Evolution and Planning History of Dhaka 66 Dhaka’s labour force has expanded much faster than the national average. There were an estimated 3.5 million active participants in the Dhaka labour force in 2000, representing 59% of the working age population in the city (aged 10 years and over). The service sector accounts for two thirds of total employment as compared with only 25% for the country as a whole (BBS, 2001). However, the majority of the labour force is engaged in informal economic activities which do not have any precise statistics, but is generally thought to be around half of the total (BBS, 2010). People engaged in transport, business and others categories are generally considered to be in the informal sector (Figure 3.2). The central business district (Motijheel and Gulistan) is characterised by concentrated employment opportunities (Figure 3.3). However, a few urban centres, located some distance from the central business district (Savar, Narayanganj), also have large numbers of jobs.

Chapter 3: Urban Evolution and Planning History of Dhaka 67

Figure 3.3: Employment Distribution in Dhaka. Source: PPK Consultants, 1994 Chapter 3: Urban Evolution and Planning History of Dhaka 68 3.3 History of Urban Evolution in Dhaka

Dhaka has come to its present state through experiencing different rulers of different cultures and through a wide range of governance structures, administration, natural calamities, poverty and an enormous number of other problems (Ahmed, 1986). In order to realize the form of the city and its spatial consequences, it is essential to understand the historical evolution of the city through knowing its major expansion, growth strategies, planning decisions and reasons behind it. Dhaka was a small rural settlement on relatively high and flat land surrounded by flood affected swampy land until the end of the 16th century. In the last 400 years the city experienced several rulers having different perspectives on city development and expansion. In the last 60 years in particular Dhaka experienced phenomenal spatial and population growth (Figure 3.4). The Mughals established their capital in Dhaka in 1610 and developed the city as a business hub along with their defence headquarters. The British took over control of Dhaka in 1757 and considered Dhaka as a regional trading centre and market. Dhaka became the provincial capital of East Pakistan in 1947. During this time Dhaka received large numbers of Muslim migrants from India. To accommodate this sudden growth several area development projects were undertaken between 1950 and 1960. As the capital of Bangladesh since 1971, Dhaka leads the nation in urbanization and city primacy (Figure 3.4).

Chapter 3: Urban Evolution and Planning History of Dhaka 69 8000 350

7000 300

6000 250

5000 Population (000) 200 4000 Area(Km2) 150

3000 Km inSq. Area

Poipulation (in Thousand) (in Poipulation 100 2000

1000 50

0 0 1600 1700 1800 1814 1824 1838 1867 1872 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001

Figure 3.4: Growth of Dhaka Source: Adapted from Islam, 1996b ; BBS, 2001

3.3.1 Settlement Before and After the Mughal Empire (1608-1757)

Dhaka grows since last four centuries (Figure 3.5). Dhaka came into prominence after the Mughals appointed Islam Khan as the first Mughal viceroy of Bengal in 1608. Islam Khan established his capital in Dhaka in 1610 with a view to subjugate the landlords of Bengal (Ahmed, 1986; Karim, 1989). It was the geographical location of Dhaka, the topographic advantages of being situated on higher ground in a low lying region, and above all its strategic position on the water-routes of the country which convinced the Mughals to establish their capital (Chowdhury & Faruqui, 1989). During the Mughal era, Dhaka attained great commercial importance and became a trading centre for the whole of South East Asia.

Chapter 3: Urban Evolution and Planning History of Dhaka 70

Figure 3.5: Urban Growth of Dhaka Source: Shankland Cox Partners and Others, 1981

Mughal Dhaka was developed as a manufacturing centre for a wide variety of cottage products –some of these items included shell works, bamboo mats, fine , refined butter and cheese along with a number of agricultural products that were exported to other parts of the country. The easily washable fabrics of the Dhaka region attracted the European trading companies to settle and establish their factories in the city (Mohsin, 1989). Inland and external trade expanded and handicraft production, especially the production of sharply increased during this time (Islam, 1989a). Due to its commercial importance Dhaka attracted the European traders - the Portuguese, the Dutch, the English and the French, and also the – they all came and established their trading posts in Dhaka in the 17th century. They established their factories in the Tejgaon area (Figure 3.6b), which continued to enjoy commercial importance during the next century (Chowdhury & Faruqui, 1989). However, the greatest development of the city took place under (1662-1679). The city then stretched for 20 km in length and 12 km in breadth and is said to have contained nearly a million people (Ahmed, 1986).

Chapter 3: Urban Evolution and Planning History of Dhaka 71

(a) (b) Figure 3.6: Dhaka During Pre-Mughal Period (a) and Mughal Period (b) Source: Islam, 1996b

Mughal Dhaka was divided into several functional areas. The ‘bazzar’ or market place was the focus of development. The market was well located to serve both the upper class and the lower class residential areas. It was also close to the which served as the principal means of communication (Ahsan, 1989). An important aspect of the city’s life were the cottage industries. They were located close to the market area and the artisans also lived there. In most cases, the same house was used as a factory and a residence (Islam, 1996a). Within the industrial area there were different zones for different crafts. All these specialized industrial and trading areas, and some surrounding other localities, housed a major part of the city’s working class population consisting of artisans, labourers and petty traders.

Rivers and canals were used as a means to transport goods and people from surrounding villages. Within the city, development was compact and walking was the main mode of travel. There was very little vehicular traffic and this accounts for the absence of any well-developed road system in Mughal Dhaka. There were horses, but these are mainly owned by the Mughal army and very rich merchants. The city was divided into a number of neighbourhoods which was a cluster of houses webbed with intricate narrow lanes (Islam, 1996a). The commercial importance in the late 17th century helped to allocate some land in Tejgaon to establish factories

Chapter 3: Urban Evolution and Planning History of Dhaka 72 for the Europeans and hence settlements. Bungalows were sparsely located around the Tejgaon area where European businessmen mainly lived. This area later became connected by a train line in the late 19th century (Ahsan, 1989).

Dhaka lost its status as a capital town in 1706 when the Mughal capital was shifted from Dhaka to , Calcutta. At this time it went into decline and experienced a long slumber for more than a century (Ahmed, 1986). However, Karim (1989) claims that the export of the European companies from Dhaka increased enormously during this time and the factories of the companies attracted many people connected with manufacturing, trade, commerce and money lending. So, in spite of shifting the capital to Murshidabad, Dhaka as a city and commercial centre did not decline significantly, though city expansion and development was definitely contained.

Dhaka lacked any kind of corporate or municipal institutions during the Mughal period (Gupta, 1989). The Mughals constructed forts and other administrative establishments on newly cleared land at strategic locations, and canals were excavated for quick deployment of Mughal forces. They did not have any specific plans for city expansion but the city naturally expanded towards the north because, being on low land other directions were not suitable for any development (Karim, 1989).

3.3.2 Dhaka under British Rule (1757 – 1947)

With the end of Mughal rule and the inception of British power around 1760, Dhaka began to decline in importance and contract in size. The city experienced disastrous famines, floods and fires. Calcutta was growing in importance and it was difficult for Dhaka to compete with Calcutta which was the capital of British India. During this period Dhaka suffered physical shrinkage as well (Chowdhury & Faruqui, 1989). From 1800 to 1867 the population dropped from 200,000 to less than 60,000 (Islam, 1996).

Chapter 3: Urban Evolution and Planning History of Dhaka 73 The second half of the 19th century marked the beginning of the physical renewal of Dhaka. The city did not expand but the old Mughal town underwent changes. As a result, medieval Dhaka was finally transformed into a modern city with metalled roads, open spaces, street lights and piped water supply’ (Ahmed 1986, p. 129). In the process of consolidation of Dhaka’s commercial dominance in Eastern Bengal, the Narayanganj-Dhaka-Mymenshing railway was opened in 1886. The rail was laid almost parallel to the Mugla road from through Tejgaon to the Phulbari area and encircled the Ramna area to save the greenery (Figure 3.7). The inner part of the city also witnessed widespread rebuilding activities of roads and houses, but not following any definite plan.

Figure 3.7: Dhaka During the British period (1757-1947) Source: Islam, 1996b

The Appointment of Sir Patrick Geddes in 1906 to make a plan for the city highlights its recognition as an important city for the British in India. Sir Patrick Geddes, following his week-long visit in the locality, proposed an outline for the development of the city. He emphasized the protection of natural greenery in the Ramna area and the widening of roads for increased traffic (Geddes, 1906). Geddes emphasis was to conserve the character of any area while making plans to

Chapter 3: Urban Evolution and Planning History of Dhaka 74 accommodate growth. All that Geddes appreciated about Dhaka and his dream about the city development did not translate into reality due to administrative and bureaucratic barriers. Some of his ideas tickled down to future generations but without any recognition (Hyder, 1994).

3.3.3 Dhaka as the East Pakistan Capital (1947 – 1971)

Since becoming the capital of East Pakistan in 1947, the landscape of Dhaka City has been undergoing rapid change; the city expanded northward and the high class residential areas were constantly endeavouring to keep themselves at the northern periphery of the city by creating ‘new Dhaka’ (mainly, because, higher lands were available in the North and low lands in East and West were vulnerable to annual floods). The Motijheel area, once desolate and lying on the fringe of marshes and swamps, was earmarked as a commercial area in 1954 (Ahsan, 1989). To cater to the ever increasing residential needs of the new capital, the Dhanmondi area, which was adorned with paddy fields, came to be developed as a residential area after 1955. The Mirpur road formed an axis, and the highland on either side of the road came to be occupied right up to Mohammadpur and Mirpur. In the 1960s these two areas came to be developed by the government mainly to accommodate the migrant Muslim population (Figure 3.8).

The need for a planning body was felt at this time and the Dhaka Improvement Trust (DIT) was created in 1956 for the planning and development control of the city. DIT developed the Ghulsan Model Town in 1961, Banani 1964, 1965 and Baridhara model town in 1972. Although these satellite towns were developed to accommodate middle income groups, the inappropriate allocation procedure and rapid increase of land price has spin-off the lower and middle income group and transformed into high class residential areas. These satellite towns have now matured and cater for a large number of jobs – they are important centres for the city (Kabir & Parolin, 2010).

First Master Plan (Dacca Master Plan 1959) was initiated during this era. A brief description of that plan has been discussed in section 3.4.4 of this thesis.

Chapter 3: Urban Evolution and Planning History of Dhaka 75

3.3.4 Dhaka as the Capital of Bangladesh (since 1971)

Independence in 1971 enhanced urbanization in Bangladesh. Dhaka, being the largest city and having the most amenities, received the lions share of this growth. The driving factors for this rapid rural urban migration are surplus labour in agriculture, globalization, drought, systematic failure of rural agriculture and perceived opportunity in the city (Khan, 2008). With the increase in population, the highlands spreading towards the north came to be occupied and built up. The intervening ditches, swamps and marshes were filled up, not in any planned manner but as the exigency arose, and private initiative played a dominant role in the process (Huq and Alam, 2003). The Rajdhani Unnayan Kartripakkha (RAJUK) had been emerged through the ongoing crisis of planned and controlled development of Dhaka City. RAJUK established in April 30, 1987 by replacing Dhaka Improvement Trust (DIT). The prime intension of the organization was to develop, improve, extend and manage the city and the peripheral areas through a process of proper development planning and development control (RAJUK, 2013). The planning and development authority DIT/RAJUK chose the highlands on the Dhaka-Tongi axes as sites for its residential projects. The city did not experience any serious effort to reclaim land under a well-planned scheme to give the city a homogenous and cohesive growth (Chowdhury & Faruqui, 1989; Talukder, 2006).

Up until 1990 the main city was limited between the Balu and Turag rivers in the east and west due to the area beyond this limit being low, swampy and flood affected. But with rapid population growth and the advent of private developers, the low lying areas were filled in and elevated and different housing schemes offered. As the land become scarcer, the vertical development for both commercial and residential sectors becomes a common pattern.

Chapter 3: Urban Evolution and Planning History of Dhaka 76 3.4 Planning Initiatives for Dhaka Megacity

Dhaka is the prominent city in the region. This flood prone fertile delta on which the city is located has historically been a high density area due to agricultural production and business potential. City expansion is highly contained by its geographical limitations. Though the presence of physical planning is historically not strong, it has had significant impact in shaping the city and to bringing it to its current situation.

3.4.1 Dacca Town Planning Report (1917)

During British colonial times, Sir Patrick Geddes was commissioned to make a development plan for Dhaka city. Geddes produced a proposal after a week-long visit to the city. He mentioned the inadequacy of his own ‘diagnostic survey’ in Dhaka, because of the time constraint. But he showed the way to deal with particular aspects of the city, and proposed that further surveys should be conducted in the same line to complete the process of diagnostic survey and subsequently, to produce a more authentic master plan for the city. Geddes emphasis was to conserve the character of any area while making plans to accommodate growth (Geddes, 1917). Dhaka city was divided into zones in this plan, which offered an outline for development of the old town area with colonial offices and residential buildings around Ramna Green (Safi, 2010). But that plan was never adopted formally or no efforts were made for its implementation (Ameen, 1998). However, the influence of the guidelines that it chalked out is very prominently evident in the Dhaka University area (Hyder, 1994).

3.4.2 East Pakistan Planning Sub-Committee (1948)

There appears to have been a lack of comprehensive planning to accommodate the sudden growth that occurred resulted after independence from the British in 1947. Due to the absence of any planning body, the (or East Pakistan) government created a planning division in 1948 under which a sub-committee

Chapter 3: Urban Evolution and Planning History of Dhaka 77 prepared a physical plan for the city’s expansion (Islam, 1996b). However, this plan lacked any formal background studies. The plan was a sort of development scheme concentrated on about 50 km2 area around Dhaka in 1948. It suggested improvements of roads, new road construction, new residential areas (Dhanmondi for high class residential accommodation, Azimpur and Motijheel for government employees), an industrial area (Tejgaon), a ‘New Market’ shopping centre and hotel development (Shahbagh). Mohammadpur and Mirpur were designated to rehabilitate non-Bengali refugees arriving from India. These represented the first introduction of the Western concept of ‘sites and services’ and the ‘core house’ for the general masses (Islam, 1996b). Tejgaon was developed to encourage factory development and to provide employment to basically non-Bengali immigrants. These schemes were later adopted in the 1959 master plan. These new physical developments, though piecemeal in their efforts, reflect European concepts of urban planning and design with functional zoning of land use (Mowla, 2007).

All these areas with their heterogeneous landuse zones, have developed centres of activities. Residential plan was partially implemented but the impacts of these development projects have been quite significant on the spatial structure of the city (Ameen 1998; Mowla, 2007). Residential area like Dhanmondi and Mohammadpur become desired destinations for the middle and high income groups. However these areas quickly transformed into mixed use areas. Activities such as private schools, universities, private hospitals and offices moved to these areas. The development projects initiated by the planning sub-committee in 1948 later transformed into the nuclei of urban development, and areas around these projects have been developed as commercial or industrial areas.

3.4.3 Dacca Master Plan (1959)

The first comprehensive master plan was made by an expatriate firm, Minupria and Macfarlane, in 1958. The objective of this master plan was to establish planning principles rather than to lay down a detailed and inflexible scheme. It identified two

Chapter 3: Urban Evolution and Planning History of Dhaka 78 main problems for the development of the city: (i) shortage of land above flood level on which to build, and (ii) the congestion in the old central area of Dhaka (Minupria and Macfarlane, 1959). The plan was prepared over a 20 year planning horizon (1958 to 1978) with an estimated population increase of 40% (1.75% per annum). The plan defined the landuse pattern, zoning, water bodies, flood prone areas and buildable zones. They assumed that population will increase from 575,000 to 816,000 in the main city and from 1,035,000 to 1,466,000 in the . However, the population growth in reality far exceeded the number assumed in the planning process.

The plan recommended two major changes to the boundary of the area under the jurisdiction of the ‘Dacca Improvement Trust’ (Now known as RAJUK, Capital Development Authority)(Figure, 3.8). Dhaka did not have much scope to expand to the south, east or west; so they therefore proposed that the northern boundary should be extended to include all the land in Tongi suitable for building and concluded that this should be sufficient to accommodate the growth of Dhaka until 1978. The plan also recommended taking control of the land on the south bank of the Buriganga River: “any unsuitable development there could, on account of its proximity to Dhaka, be most detrimental to the town’s amenities” (p.5 Minupria and Macfarlane, 1959).

Other prominent suggestions were:

• In order to discourage the excessive growth of Dhaka, it was highly recommended to formulate a national planning policy for Bangladesh aimed at encouraging the expansion of industry and commerce in other towns. • Dhaka’s growth as a capital city, administrative and university centre is stimulated by setting up substantial new industries; serious considerations should be given to steering them elsewhere. • They estimated that 70,000 people will require factory employment and 1400 acres will be required for this factory land.

Chapter 3: Urban Evolution and Planning History of Dhaka 79 • This plan suggested an additional 4481 acres of land zoned for industry. At 50 persons/ acre this would provide employment for 224,050 persons. • They estimated that housing will be required for 402,700 persons and suggested several new housing schemes for them (Figure 3.9). • They detailed out the strategic location of these sites. For example, since Tongi is located far to the north and did not have good transport links with Dhaka, they suggested it be developed as a self-contained new town having a balanced community of houses and a nucleus of industry. Mirpur, on the other hand, was suggested as a residential satellite having good commuting options to the main city.

Chapter 3: Urban Evolution and Planning History of Dhaka 80

Figure 3.8: City Boundary Proposed by 1959 Master Plan Source: Based on Shankland Cox Parters and Others, 1959

• They identified the unique living standard for the old city dwellers and suggested extended residential areas on reclaimed land in the

Chapter 3: Urban Evolution and Planning History of Dhaka 81 south (Keraniganj and Postogola) to accommodate the growth of the old part of the city.

Figure 3.9: Landuse Proposal in 1958 Master Plan Source: Based on Shankland Cox Parters and Others, 1959

Chapter 3: Urban Evolution and Planning History of Dhaka 82 However due to unusual spatial growth, increases in population and changes in socio-political conditions RAJUK failed to fulfil their tasks (Ameen, 1998). The implementation process is marked more by breach and deviation than adherence to the plan (Nahrin, 2008). RAJUK had made efforts to develop new areas for residential, administrative and commercial purposes. It has made changes in landuse pattern and chalked out detailed structure plans which would never be implemented. Again adhoc decisions from bureaucratic quarters and natural forces influenced the growth of the city (Ameen, 1998; Mowla, 2007).

3.4.4 Dacca Metropolitan Area Integrated Urban Development Project (1981)

In 1981, with Asian Development Bank assistance, an expatriate firm Shankland Cox Partnership, and others, undertook the project named ‘Dhaka Metropolitan Area Integrated Urban Development Project’ (DMAIUDP). The impetus for this plan was the storm water drainage and flood problems of the Dhaka metropolitan area, and its purpose was to provide a long term growth strategy for urban expansion (Shankland Cox Partners and Others, 1981). The ‘master plan’ concept had by this time gone out of fashion and it was almost impossible to bring vast unplanned developments under any sort of rigid control. It is noticeable that this study referred to projects rather than plans. It was an excellent study and considered nine options and several recommendations. They realized that flooding and drainage are major problems of the city (Ameen, 1998). With scanty resources it is not possible to make low areas flood free, instead the focus was on acquiring developed land and growth would take place in that direction in a more linear pattern. The recommendation was for a north- south corridor (mass-transit) with branches towards Savar and, ultimately, to Aricha. Institutional re-arrangements were also suggested. The strategy plan took a long- term view of the city up to 2001. The recommendations of this strategy plan were never taken seriously (Zaman & Lau, 2000) and the reports only served to become a reference work. Nevertheless, many of the assumptions of the plan proved to be accurate, and these later provided a comprehensive basis for the future urban growth of Dhaka.

Chapter 3: Urban Evolution and Planning History of Dhaka 83

3.4.5 Greater Dhaka Metropolitan Area Integrated Transport Study (1994)

The Greater Dhaka Metropolitan Area Integrated Transport Study (DITS) was an initiative of the Government of Bangladesh with assistance from UNDP. The main aims of the study were to collect information about the demand for transport services and the infrastructure to deliver these services in Greater Dhaka, and to prepare an immediate action plan for the effective management of the existing traffic and transport system (PPK Consultants and Others, 1994).

The study was concurrent with the Dhaka Metropolitan Development Plan (DMDP) project and considered the DMDP area (Figure 3.13) as the DITS study area which was divided into 151 traffic zones. Information was collected on travel behaviour, travel demand, and public transportation usage. This has revealed that about 60% of trips are on foot (Figure 3.10) and almost half of the remaining trips were performed by non-motorized transport (NMT). However, this trip pattern shows the compact nature of urban form and higher level of mixed landuse. Residential and employment landuse mixture creates shorter trips for work that seem to be satisfied by NMT modes (walk, bicycle and tri-cycle).

Chapter 3: Urban Evolution and Planning History of Dhaka 84 3 Walk Rickshaw Autorickshaw 17.3 Autotempo Bus Car

60 motor cyle bi-cycle water transport

Figure 3.10: Modal Split of Dhaka Commuters Source: PPK Consultants and Others, 1994

This is the first transport study of this scale undertaken for Dhaka, and its survey based methods generated data to estimate travel demand and propose possible solutions for its improvement. The study identified insufficient road space, mixed traffic on the same lanes, and the absence of required public transport as the main reasons for the transport situation in Dhaka. After two years of rigorous efforts on data collection and analysis the team recommended short and long term solutions. To ease the traffic congestion and improve the transport services they suggested public transport (bus) routes, improvement of selected intersections (with geometric design), level separation and making footpaths for pedestrians. For the long run improvement, the DITS suggested some institutional reform as well. This study proposed to establish a separate organization (transport authority) that would coordinate and operate the transportation system.

This study had a significant impact on contemporary urban transport decision making in Dhaka as the transport problems were prominent at that time. The Mohakhali and Khilgaon flyover were constructed, a few intersections were improved, more bus routes were introduced and several attempts were made to free the footpaths for pedestrians. Those well appreciated steps were found to be insufficient as no effective actions were taken to reduce or alter demand, and no large scale public transport projects were constructed in light of rapidly increasing

Chapter 3: Urban Evolution and Planning History of Dhaka 85 travel demand (Ameen, 1998). The Dhaka Transport Coordination Board (DTCB) was formed in 1998 based on this proposal. However, its operational jurisdiction remains overlapped with Dhaka City corporation (DCC), RAJUK, Bangladesh Road Transport Authority (BRTA) and the Roads and Highways (R&H) department.

3.4.6 Dhaka Metropolitan Development Plan (1995)

In 1995, another expatriate group, Mott Macdonald, in association with Culpin Planning Ltd. and others undertook the project: Dhaka Metropolitan Development Plan (DMDP). The plan was presented as a structure plan and has four major components: (i) Strategic growth options (ii) Structure plan (iii) Urban area plan, and (iv) Detailed area plan. The DMDP Structure Plan (SP): The DMDP Structure Plan provides a long-term strategy for the 20 years to 2015 for the development of the greater Dhaka. This plan refers to the 590 square miles (1528 km2) area administered by RAJUK. Urban Area Plan (UAP): The DMDP Urban Area Plan (UAP) provides an interim mid-term strategy for the 10 year to 2005 and covers for the development of urban areas within metro Dhaka management area. Detailed Area Plan (DAP):The DMDP Detailed Area Plans provide more detailed planning proposals for specific sub-areas of Dhaka. However they do not initially cover the entire Dhaka structure plan area (590 square mile). While all sub-areas will eventually require a DAP, it has been formulated only for only priority areas and still awaiting for approval to enforce.

The plan assumed that with the implementation of the Flood Action Plan, a vast area, including the present low areas, will be flood free and new peripheral expansion will take place on the east and west by encroachment on suburban land and with the agricultural land, with the northern area being mainly for wealthier groups of the population following existing trends of growth (Figure 3.11). More roads and highways are recommended to link with the sprawling new developments, leading to a vision of low rise, low density city form, with long journeys to work (Zaman & Lau, 2000). These proposals were meant to reduce traffic congestion in the older parts of the city.

Chapter 3: Urban Evolution and Planning History of Dhaka 86

Figure 3.11: Master Plan Boundary (1959 and 1995)

It was earlier disclosed that the Bangladesh Water Development Board (BWDB) would include all of FAP 8A priority components (as given in JICA 8A study) in its short-term capital programme and JICA was expected to determine the available funding sources, but several discrepancies became apparent regarding the procurement of its funds and its implementation potentials (Ameen, 1998). As mentioned earlier, the population of the city increases by approximately 250,000 persons each year, thus any delays in the execution of any project by two to four years, for example, would result in a further influx of about a half or a million more people. If this (FAP) was executed as proposed, a substantial amount of flood protected land of approximately 7000 ha would become available, adjacent to the existing urban core.

Chapter 3: Urban Evolution and Planning History of Dhaka 87

Figure 3.12: Dhaka DMDP Area Landuse Scenario Source: DMDP, 2001

Whatever the quality of the project documents (DMDP), there is no body to take responsibility of follow its outlines. The project was sponsored by the Ministry of Public Works and Housing in collaboration with RAJUK. But RAJUK does not

Chapter 3: Urban Evolution and Planning History of Dhaka 88 seem not to be seriously involved in the plan; their efforts appears to be with development of Yusufganj (Purbachal) high land close to Dhaka, which unfortunately has been ignored by the strategy plan.

3.4.7 Strategic Transport Plan (2005)

The Strategic Transport Planning (STP) project was initiated in 2004 to make a long term strategic plan for greater Dhaka to cater for the transport demands of the future and to assist with the encouragement of coordinated transportation planning with sustainable urban development (The Louis Berger Group, I. and B. C. Ltd, 2005). The STP study investigated a wide ranging series of issues, starting from a consideration of population and economic growth patterns. The study proposed a series of possible visions for the likely growth and settlement patterns for the 20 year planning horizon up to 2024. These land use scenarios were basically different ways of accommodating the huge population increase predicted from an existing population of 17 million to a future population of 36 million (in 2024). The consultant team considered three land use scenarios (1: urban corridor – strong central spine scenario; 2: growth pole-satellite community scenario; 3: dispersal settlement development scenario) and resulted in a selection of just one land use option; namely, a Growth Pole Scenario incorporating satellite communities (Figure 3.13). In this scenario, the vision sees the creation of a series of satellite communities on the fringe of the main city. The satellite communities would be developed with their own employment base so that they would not be “commuter cities” but rather would be self-contained communities with their local transportation systems and other facilities (The Louis Berger Group, I. and B. C. Ltd 2005). STP also attempted to identify properly designed connections between these cities and Dhaka.

Chapter 3: Urban Evolution and Planning History of Dhaka 89

a) Scenario 1: urban corridor – strong central spine b) Scenario 2: growth pole-satellite community c) Scenario 3: dispersal settlement development scenario scenario; scenario; Figure 3.13: Landuse Scenarios Considered in STP Source: The Louis Berger Group, I. and B. C. Ltd (2005),

Chapter 3: Urban Evolution and Planning History of Dhaka 90 Following the acceptance of the Growth Pole Scenario (Figure 3.13b), the study looked into the levels of future transport activity which are likely to occur in the area. The investigation showed that the level of transportation activity by 2024 would be about 2½ times the activity in 2004 in both the case of vehicle trips and transit trips. In short, it is estimated that the future transportation system would need to be designed to accommodate a daily peak hour demand of about 1.3 million public transport trips and almost 300,000 vehicle trips (The Louis Berger Group, I. and B. C. Ltd, 2005).

A number of transportation strategies were considered (Table 3.1) to satisfy the estimated demands using different emphases for public and private transport links. It was clear that road-based systems would not be sufficient to satisfy demand for 36 million people by 2024 and the team arrived at the final recommendations for the integrated system. Strategy 2b is recommended as the preferred strategy because, in the Consultants’ views, “it represents the best balance between public transport provision and individual transport to serve the future needs of Dhaka and also offers the optimum flexibility in mass rapid transit bearing in mind full knowledge of the context of reasonably expected financial resources” (The Louis Berger Group, I. and B. C. Ltd (2005), page 9-4).

Table 3.1: Alternative Transport Strategies Being Evaluated in the Strategic Transport Plan STRATEGY LEVEL OF BRT METRO

NAME ROADS COMPONENT COMPONENT

INVESTMENT 1a Roads+ All BRT No METRO 1b Roads+ Part BRT Part METRO 1c Roads+ No BRT All METRO 2a Roads++ All BRT No METRO 2b Roads++ Part BRT Part METRO 2c Roads++ No BRT All METRO 3a Roads+++ All BRT No METRO 3b Roads+++ Part BRT Part METRO 3c Roads+++ No BRT All METRO 3d Roads+++ No BRT No METRO Source: The Louis Berger Group, I. and B. C. Ltd (2005)

Chapter 3: Urban Evolution and Planning History of Dhaka 91 In the Preferred Strategy (2b Modified), there are 3 Bus Rapid Transit (BRT) lines and 3 Metro lines as proposed by the STP team (Figure 3.14) connecting major generators and those provide good coverage of the travel demand. To implement such an extensive citywide metro takes long time and involves huge investment. Therefore, STP recommended several BRT lines that should start to be implemented as early as possible whilst the Metro system is being developed.

Figure 3.14: STP Recommended Mass Transit Network for Dhaka (2004-2024) Source: The Louis Berger Group, I. and B. C. Ltd (2005)

Chapter 3: Urban Evolution and Planning History of Dhaka 92 However, the STP study and its proposed urban growth model have several criticisms. Bari and Efroymson (2007) claimed that STP adopted a top-down approach in order to develop alternative transport strategies which are pro-poor and biased towards expensive infrastructure investment. They claimed that scenario 1a is technically and economically more feasible than adopted modified strategy 2b.

3.5 Urban Development, Spatial Growth and Urban Form of Dhaka

In the last 60 years, the population of Dhaka city has increased more than 30 fold, yet the area of the city has only expanded 18 times (Table 3.2). Due to unprecedented urbanization, Dhaka has grown on its own in a haphazard manner and the topography (especially flood proneness) of the area has dictated the direction of growth. The development of Dhaka has been influenced by the rivers and floods which have been accommodated in its spatial plans over the past 50 years. The dynamic nature of river morphology and the lower load bearing capacity of the soil made the construction expensive and uncertain. This is the main reason that Dhaka did not expand` toward the south beyond the Buriganga River until the construction of a bridge in the 1980s and 1990s.

Table 3.2: Rapid Growth of Population and Area of Dhaka

City Megacity

Year Area Population Area Population 1951 31.1 276,033 85.45 411,279 1961 72.5 368,575 124.45 718,766 1974 103 1,403,259 335.79 2,068,353 1981 103 2,475,710 509.62 3,440,147 1991 256 3,612,850 1,352.87 6,487,459 2001 300 5,378,023 1,528 9,912,908 2011 1,528 12,800,000 7,440 17,300,000 Source: Compiled from Islam, 1996b; Talukder 2006; BBS, 2011; The Louis Berger Group, I. and B. C. Ltd (2005)

Chapter 3: Urban Evolution and Planning History of Dhaka 93 The urban expansion around the city has been restricted by the rivers at its south and by low-lying lands adjacent to the rivers at its west and east, resulting in the city’s growth along the northern routes in an irregularly elongated manner. Development intensity is higher in the old and main city reaching its density up to 140,000 persons per km2 (The Louis Berger Group, and B. C. Ltd, 2005). For a city with very few high rise buildings, this density creates very compact development. There is also an acute shortage of available land for development in or around Dhaka. In addition to land scarcity, other reasons for this situation are malfunction in the operation of the land market and the performance of the development control body that is strategically not able to restrict development.

In Bangladesh, development and growth of cities depends on the actions of numerous individuals without any reference to plans or planning. There are many different local traditions and decisions (political and community) to influence buildings and settlement morphology along with natural phenomenon (Islam, 1996b). Dhaka city grew from a rural settlement to become a megacity without much planning effort (Islam 1996b; Huq & Alam, 2003; Nilufar, 2010). Nilufar (2010) identified four distinct urban patterns in Dhaka (Figure 3.15) which reflects the age of the city, its rulers, planning philosophy and, above all, planning control. First is the historical core or ‘Old Dhaka’ which still contained the Mughal layout of the human scale city. Narrow lanes, very compact buildings, mixed use – all remain the same and remind one of the glorious past. The second category of urban area comprises the formal planned area or satellite towns conceived and planned since 1950. Dhanmodi, Gulshan, Mirpur, Baridhara are examples. Until 1980 only Capital Development Authority (Rajdhani Unnayan Kortipokkha, RAJUK) was involved in planning and designing for this type of development but now private developers are also providing the same. There is a third category combining the two – a fusion of old and formal patterns – which occupies the most part of development in Dhaka (Mowla, 2007). Shanty towns and small temporary houses falls in to this fourth category where about 30% of Dhaka dwellers live.

Chapter 3: Urban Evolution and Planning History of Dhaka 94

Old Dhaka Gulshan Residential (planned)

Organic Slum

Figure 3.15: Urban Forms in Dhaka Source: Google Earth (from Altitude 600 meter)

In the organic or informal patterns, the intensity of development is high. There is a preponderance of mixed use: not intrinsically harmful, but leading to conflict and congestion. There are also low proportions of non-residential uses- parks, roads, commercial and industrial areas – compared to predominantly residential use. The roads network in these informal areas is irregular and insufficient in many cases and there is the absence of any room for footpaths. These areas grow organically and are often characterized by congestion in peak hours.

As a part of the planning effort to accommodate growth several satellite towns were developed. The newly developed areas with institutional support (residential and industrial estates) when compared to spontaneous infill, show a striking difference in general texture and morphology (Mowla, 2007). Although present urbanization follows the Mughal tracts, growing northwards, the pattern of most of the existing development appears to follow colonial and quasi-colonial footsteps (Zaman & Lau, 2000). Historically, the main communications routes are in high demand for development, therefore unplanned growth and development occurs naturally along communication channels. In most cases these developments occur along the major

Chapter 3: Urban Evolution and Planning History of Dhaka 95 corridor which trapped village enclaves or undeveloped land behind it. Development stripped of agricultural land and forced villagers to move elsewhere or to change profession. Some of the villagers have moved to villages further away to find agricultural land, some have come to the city and become involved in the informal sector that does not require institutional support, much investment or skills. A similar situation is observed in other South Asian megacities (Mowla, 2007). These enclaves, on the other hand, characteristically lack urban amenities but provide cheap land and rent. Their irregular roads developed along earlier winding village paths (hallot), have become congested with traffic and seldom connect well with the super-imposed street system of surrounding urban extensions. When densified, the settlements resemble the mahallas (locality, smaller than suburb) of old towns indicating that the basic rural morphology is at the root of indigenous urban settlement at Dhaka (Geddes 1917). But the morphology in the planned part shows a regular grid pattern in most cases. Very often due to development in the surrounding area, these planned neighbourhoods become a transit point to connect several privately or organically developed residential areas.

Residential neighbourhood development is one of the popular practices of the planned development, where a low land (marshy or agricultural) is chosen and converted to residential plots with minimum provision for urban services. Though most of them are for low to middle income groups, in reality they accommodate the rich (bureaucrats, govt official and successful business personnel). Government housing (other than housing for high level bureaucrats) built after 1947 was traditionally located at some distance from the place of work; often on geographically isolated sites for which there was no competition. There was no attempt to integrate them with nearby areas and such housing was seldom provided with adequate and well-designed community or social facilities. However, these developments soon formed the nucleus for further organic developments in respective areas.

An important part of Dhaka has been occupied by the defence organizations. Dhaka Cantonment and the old airport (now used by the Air Force) have occupied 530 acres of prime land in the centre of the city. The headquarters of Border Security has taken

Chapter 3: Urban Evolution and Planning History of Dhaka 96 another 250 acres of flood free high land, contributing to the scarcity of land and traffic congestion. The cantonments, as a legacy of European rule, continues to be the single largest land area in Dhaka city. The location and shape of the cantonment has had profound effects on urban growth patterns in the past as well as in present times. The building and infrastructure inside those closed areas is not necessarily comparable with the surrounding areas.

In central city locations, the land value increased at a rate well above that of other financial indices. For example, the land value in Dhanmondi residential area is as high as USD3500 per square meter which can only be compared with very exclusive residential neighbourhoods in rich countries. In the expanding periphery, agricultural lands are sub-divided in preparation for urban uses and prices there, too, have spiraled upward. The speed and extent of the impact on the land market has adverse effects on the expansion of the city. This overwhelming increase in land price has led to a flourishing land developments and land speculation. Moreover, traditional land ownership had been a symbol of the social elite as the purchase of urban land conferred a social status (Islam, 1996a). There are also economic reasons to purchase land. For those with money to invest, urban land becomes the preferable commodity (Willcox, 1996). In addition, much of the money repatriated by Bangladeshi working overseas ends up in the urban land market. So does a substantial portion of the wages and profits paid to employees and business-men inside the country. It has been noticed, that any planned development (residential or commercial) lead by government planning and urban development organizations attracts land developers to acquire cheap land and convert it to a residential neighbourhood. Once it comes to market, it sells out quickly and with time they acquire the development approval. In many cases, zoning has been altered to accommodate these so called ‘housing schemes’. The ineffective laws and regulations followed by a corrupt administrative mechanism have allowed this encroachment (Mahmud, 2007).

Dhaka megacity is growing following its historical development footprint. At times there were plans, but these were not implemented. Rather, several discrete development projects were implemented without much control on the surrounding

Chapter 3: Urban Evolution and Planning History of Dhaka 97 landuse. As the land become scarce, people began to be interested in the areas surrounding the planned development an private developers are under taking projects to develop the surrounding areas. Some of these developments were in violation of planning documents (Nahrin, 2008). Development planning was inactive until 90s to control these privately initiated development projects. Even after all those, the urban strategy does not encompass a much larger; it only practices planning control within the inner zone. As a result, development in the periphery is moving rapidly. Low lying agricultural areas are being converted to residential plots without much access to major roads, without any urban amenities and without any future to consider them as part of Dhaka megacity.

3.6 Transport in Dhaka

Dhaka is one of the least motorized megacities in the world (less than 2%) (World Bank, 2012). The rapid rise in population along with mushrooming urban expansion have generated considerable travel demand. Since the supply of infrastructure and services are less elastic, it has caused numerous transport problems in Dhaka. It has resulted in deterioration in accessibility, service levels, safety, comfort, operational efficiency and urban environment (Hoque & Hossain, 2004). The congestion and pollution problems are rapidly growing due to the combined effects of rapid motorization and urbanization.

Dhaka megacity, as a case study, with a population of 17.3 million generates 23 million trips on a daily basis (The Louis Berger Group, I. and B. C. Ltd., 2005) without any metro or effective mass transit, and 8.4 million of these trips (36%) are by rickshaw (non-motorized tri-cycle) (Table 3.3). Higher rates of school trips (25%) also reflect a higher proportion of young persons who, on average travel 3.5 km (Table 3.4). It is notable that 48% of total trips are performed by walking and by rickshaw and their average travel distance is 1.45km and 2.34km respectively. The short length of these trips is explained by higher urban density and could be the result of a compact urban form and predominance of mixed land use (Acharya & Morichi, 2007).

Chapter 3: Urban Evolution and Planning History of Dhaka 98 Table 3.3: Modal Share and Trip Purpose

Travel Mode Home Based Home Based Home Based Non-Home All Trip Work Education Other Based Purposes Walk 14% 19% 6% 6% 12% Rickshaw 29% 43% 10% 24% 36% Transit 41% 29% 31% 41% 34% Motorised 16% 9% 23% 29% 18% (non-transit) All Modes 31% 25% 36% 8% 100% Source: The Louis Berger Group, I. & B. C. Ltd (2005)

Fifty two percentages of trips depend on different types of motorized vehicles (Table 3.3). Without any effective public transport system, 1500 buses are running on 150 routes serving 7.5 million trips (The Louis Berger Group, I. & B. C. Ltd., 2005). However, most bus routes are limited within the main city and its extended areas (Figure 3.16). Passengers from the outer ring who want to come to the main city for everyday jobs or medical services or for shopping are required to make several transfers.

Table 3.4: Travel Patterns in Dhaka

Trip rate by Purpose (per HH per Trip rate by Primary Travel Mode (per day) HH per day) Home to Work 2.7 Walk 1.2 Home to Education 1.1 Rickshaw 2.9 Home to Other 3.9 Transit 3.7 Non Home Based 0.75 Motorized (non-transit) 0.7 All Purpose Trip 8.5 All Travel Modes 8.5

Trip length (Km) by Purpose Trip Length (km) by primary travel mode Home to Work 5.81 Walk 1.45 Home to Education 3.5 Rickshaw 2.34 Home to Other 6.03 Transit 8.39 Non Home Based 6.44 Motorized (non-transit) 7.76 All Purpose Trip 5.37 All Travel Modes 5.37

Source: The Louis Berger Group, I. and B. C. Ltd (2005)

Chapter 3: Urban Evolution and Planning History of Dhaka 99 *width of the line represent number of buses on that road segment. Numbers in red font in

Figure 3.16: Bus Routes in Dhaka Source: The Louis Berger Group, I. & B. C. Ltd (2005)

Chapter 3: Urban Evolution and Planning History of Dhaka 100 Given the absence of any urban rail system, the transportation system is completely road based. Existing road space is often insufficient to meet demand (around 7% of total land area) (Hossain, 2004). Though walking is a major mode, footpaths are often missing in the road geometry or occupied by construction materials or by street vendors. Another issue with roads are with utility services. Water, sewerage and telecommunication cables are installed beneath roads and there is no coordination body for their installation and repair works (Photo A in Figure 3.17). Water logging is another issue in the city. Due to high sedimentation in the surrounding rivers, the water accumulates during the monsoons and heavy rainfall causes local floods and inconvenience to passengers and damage to infrastructure (Photo B in Figure 3.17). Moreover, intersection management, mixed traffic and the encroachment of roads are important issues to resolve to implement a fair traffic movement.

Chapter 3: Urban Evolution and Planning History of Dhaka 101 A. Road Works B. Water logging

C. Trapped in Road due to a Political Rally D. Phulbari Bus Stand – a common scenario

Figure 3.17: Some Characteristics of Roads in Dhaka

Chapter 3: Urban Evolution and Planning History of Dhaka 102

Insufficient road space, mixed traffic of varying operating speeds, lack of traffic discipline and, above all increasing travel demand has brought about a worsening situation on the roads of Dhaka (Hossain, 2004). Over the past few decades, the motor vehicle population has increased many-fold in Dhaka but the number of buses or bus routes has not improved accordingly (Figure 3.18). This leads to an unsustainable transport situation as more cars are increasingly contributing to road congestion and air pollution (Akther, 2009).

200 20 Motor Car 180 18 Jeep/St. W 160 16

140 14 Taxi

120 12 Bus

100 10 Minibus 80 8 Truck 60 6 Vehicle Number (in Thousands) (in Number Vehicle Population (in Million) (in Population 40 4 Others

20 2 Urban Population 0 0 UP TO 2004 2005 2006 2007 2008 2009 2010 2011 2012 2003 Figure 3.18: Motor Vehicle Number in Dhaka Source: BRTA, 2012

It is observed from Table 3.3 that the share of mass transit is very low (34%); and walking still remains the dominant travel mode. It has become increasingly clear that adequate and efficient mass transit services would play an important role in combating the ever-worsening problems of traffic congestion. One cannot ignore the fact that transport is a derived demand. So, the problem lies with the distribution and characterization of landuse that generates traffic. Significant attention is required to pay to address the landuse scenarios that would likely generate less traffic, shorter trips.

Chapter 3: Urban Evolution and Planning History of Dhaka 103 STP (2005) suggested a combined public transport-system with Mass Rapid Transit (MRT) and Bus Rapid Transit (BRT). As the majority of the population lives below the poverty line, a metro may not be readily economically feasible. Buses with well designed routes could serve the demand for the time being. A major problem with bus routing is that, due to the large volume of traffic, it cannot move at its desirable speed and many of the roads are not suitable for bus operation (only 250 km out of a total of 2500 km of roads are suitable for city buses) (The Louis Berger Group, and B. C. Ltd, 2005,). The geometry of the roads, scale of landuse and flatness of the land has encouraged dependence on non-motorized transport (NMT). Again, sharing the same lane, NMT is found to reduce the operational speed of roads. But the affordability of people and the dependency of more than a million people and their families on rickshaws has made it a political issue (rather than a planning issue) to allow it on busy roads (Saito, 1993). However, a project has been initiated to implement (modified) line-6 of the MRT proposal by the year 2017. For achieving sustainable urban transport in Dhaka, the Strategic Transport Plan (STP) also suggested the development of six major corridors as mass transit routes including three Bus Rapid Transit BRT corridors. Also there are project to improve water transport especially for freight transport around Dhaka. Small scale initiative is taken to start MRT connecting high demand areas.

3.7 Environmental Conditions in Dhaka

The low trends of industrialization in the country and relatively lower trends of motorization does not keep the air pollution at lower level in the megacity of Dhaka. A study on megacities ranked Dhaka as having the worst level of air pollution amongst the world’s cities, followed by Beijing and Cairo (Figure 3.19). Vehicle emissions from road transport is seen as the major cause of air pollution. Air Quality Action Plan (Khan, 2005) for Dhaka summarised the attributes of air pollution in Dhaka: - Over population and inadequate infrastructure - Large fleet of old vehicles - Quality of fuels

Chapter 3: Urban Evolution and Planning History of Dhaka 104 - Unregulated burning of biomass, and - Road congestion.

Figure 3.19: MPI-based Total Pollution Level in Megacities. Source: Gurjar, 2008

Air quality of Dhaka has significant impact from its waste management practices, transportation system and building construction works. Until recently, air pollution from transportation systems occurred because of the use of low lead gasoline without proper catalytic converters (Karim, 2001). In recent times, there have been several policy measures to improve emission levels. Since 1999, Bangladesh has introduced unleaded petroleum. From 2005, EURO-I and EURO-II standard are set for diesel and petrol for transportation. From the year 2003, all two-stroke three-wheelers are banned in the Dhaka mega city (Ministry of Environment and Forests & Ministry of Communication, 2010). Also, importing reconditioned cars now comes under severe scrutiny. Nearly half of the passenger cars in Dhaka have been converted to compressed natural gas (CNG) from petrol (Wadud & Khan, 2011). This has reduced the SO2 level significantly. At the same time, all buses are now running on CNG. The CNG conversion has improved air quality but there remains emissions of methane (CH4) which is 10 times more detrimental than carbon-monoxide for global warming. These increasing emission are concentrated in the main city. So policy

Chapter 3: Urban Evolution and Planning History of Dhaka 105 measures need to be taken to reduce vehicle emissions and to make sure that emission concentration does not reach to a harmful level.

3.8 Conclusions

Dhaka is the major urban centre and primate city of Bangladesh. Incredibly, it grows on average by more than 2000 persons every day. High population density, poor infrastructure, and low employment rates have been enduring concerns. National population growth has been slowed down to 1.34% recently (BBS, 2011), but the growth rate for Dhaka still remains around 6%, mainly due to voluminous migration from rural areas and other towns. Around half a million persons are added to its demographic profile every year that, in term, requires new land for housing, infrastructure and all other urban amenities. To meet the increased demand every year, around 150 thousands motorised vehicles are adding to already congested city streets (BRTA, 2012). City governance cannot cope with these increasing demands at the same rate due to financial constraints, corruption, and shortage of skilled man power and influence of political stakeholders (Mahmud, 2007). This is accentuated with weak institutions, inconsistent policy, delays in decision making and unawareness. Though the first Master Plan in 1959 advocated decentralization of urban activities from Dhaka to other towns by creating industries and manufacturing, this was never seriously implemented. Dhaka grows organically with new ‘attractions’ for jobs, education and health care. It is the highest attraction place having all the headquarters of government and non-government departments, universities (60 out of 82), medical colleges (32 out of 38), good schools, and medical facilities (UGC, 2012).

Dhaka has exhausted all avenues of growth. It has become a ‘city of inconveniences’ (Ameen, 1998, p. 170). It has floods, water logging, pollution, power failures, traffic congestion, political disturbances, strikes, vandalism and many other problems. Despite this, Dhaka has a rich and vibrant cultural life. It has a prosperous economy and an enormous workforce. For the capital to lead the nation in economic growth and living conditions, it is imperative that all planning stakeholders agreed on its

Chapter 3: Urban Evolution and Planning History of Dhaka 106 development and guide it in such a way that growth potential and the quality of life is addressed properly in policy formulations and, hence, in development initiatives. Dhaka needs to plan to accommodate its expansion; not only the spatial aspects but economic activities as well without sacrificing the quality of life. Physical, social and economic plans should be integrated to accommodate the nation’s goal towards development. The urban structure of the megacity, the allocation of land-use and activity and its intensity reflects the historical past of the city. Comprehensive action is urgently required that recognizes the character of an area and allocation of controlled activity in a sustainable manner.

Chapter 3: Urban Evolution and Planning History of Dhaka 107 Chapter 4

Research Design

4.1 Introduction

This chapter presents the research design for the thesis. It includes formulation of research questions based on the literature review conducted in the previous two chapters, and methods employed to achieve the goals of the research. Data and methods used for the empirical analysis are discussed. The overall flow of the empirical analysis is presented as a framework showing the relationships amongst the research questions, data, methods and results. The research design provides a detailed and rational guide for the empirical analysis conducted in the next three chapters.

This chapter emphasises the methodology developed for empirical analysis to address the research questions. It describes the methods and tools, and focuses the lessons from different literature discussing the transport pattern and its relevant urban spatial pattern. First, the conceptual framework is set. Second, the study area and data sources for the case study are introduced. Third, analytical methods are described and the operationalization processes for a number of tools and measurements are explained to justify its appropriateness for the applications to represent each concept identified in the research objectives.

4.2 Conceptual Framework and Research Questions

A range of literature in Chapter 2 has shown that urbanization has taken different forms in different societies. Asian megacities which might have expected to follow the path of the developed world have shown different characteristics under different levels of technological development and socio-economic settings. Large cities in the

Chapter 4: Research Design 108 western world are often categorised as monocentric, polycentric or somewhere in between. But there is a lack of research on the same categorisation of Asian cities. The limited amount of research on the urban spatial structure of large Asian cities displays mixed results (Table 4.1). T.G McGee identifies East Asian large cities of having a different and distinctive spatial structure that can be summarized as expanded monocentric region (for details please see Chapter 2, Section 2.2). Research from the Eastern Asia Society for Transportation Studies (EASTS) uses longitudinal employment and travel data to explore the spatial variation of temporal change in jobs and reveals the development of clusters of jobs outside the CBD, and claims to have evidence of polycentricity in several cases. Alpkokin et al. (2005) finds growing sub-centres outside the central business district of Istanbul, Turkey. Bhandari et al. (2007) also concludes the same for Delhi, India. However, Vichiensan (2007) finds that Bangkok is still monocentric in nature. Hakim and Parolin (2009) reveals a strong role of CBD in job distribution of Jakarta Metropolitan Area (JMA) and manufacturing along the transport links radiating out the centres. All these studies convey a common message of the fact that job concentration has a reciprocal relation with commuting distance and this relationship influences travel behaviour, and hence impacts emissions and sustainability (Figure 4.1).

Chapter 4: Research Design 109 Table 4.1: Urban Form in Several Asian Megacities City Population Urban Structure Author (in million)*

Delhi, 23.5 Polycentric employment distribution. Bhandari et al. India Development of satellite town along the highway (2007) and railway Karachi, 21.9 Monocentric Qureshi and Lu, Pakistan 2007 Bombay, 21.0 In transition from mono to polycentric Shirgaokar, 2012 India Bangkok, 14.1 Extended monocentric or early stage of Vichiensan, 2007 Thailand polycentric development

Jakarta, 25.6 Extended Metropolitan Region Hakim and Indonesia Parolin, 2009; Hakim, 2010 Tokyo, 33.6 Monocentric with multiple sub-centres Sorensen, 2001 ; Japan Kawabata, 2003 Istanbul, 13.6 Polycentric Alpkokin et al., Turkey 2005 Shanghai, 15.8 Dense monocentric to organic Polycentric Haixiao, 2000 China Manila, 21.7 Monocentric with multiple sub-centres Morichi and Philippines Acharya, 2012 *http://www.citypopulation.de/world/Agglomerations.html

Figure 4.1: Major Components of the Research Source: Author

Chapter 4: Research Design 110 Urban spatial structure refers to the arrangement of urban land uses – i.e. allocation of land for housing, industries, manufacturing and other employment and amenities (Næss, 2006). Urban spatial structure is generally defined as the spatial distribution pattern of urban economic activities and residences which influences spatial distribution of commuting (Sohn, 2005). The way land uses are arranged in a city affects many aspects of how cities function and has implications for accessibility, environmental sustainability and social equity (Bentley, 1995). It is an imperative of this research to explore the spatial structure of Dhaka and to identify the relationships between employment locations and trip patterns. The first degree of investigation identifies its job nature in term of concentration patterns of population and employment. Broadly speaking, it ascertains whether the city is monocentric or polycentric, or shows any other pattern in term of the spatial concentration of jobs and worker residents locations.

Journey from residence to workplace is a major concern for any large city. In Dhaka, home to work trip accounts for 32% of total trips (The Louis Berger Group, I. and B. C. Ltd, 2005). Commuting occurs in specific time bands, and being a major share of trip making and due to its impact on the performance of the transport system, commuting influences residential and workplace locations and emissions (Grazi, 2008; Prashker et al., 2008). Commuting behaviour deals with travel distance, mode choice, and trip chaining, fare and travel time. Travel behaviour shows the spatial imprint of residential employment and other major landuse distributions on the urban landscape. It carries the impact of transit availability, affordability, distance travelled, time required, comfort and even safety. Travel behaviour links the landuse interaction between residential area and employment, or work destination reflects the influences of spatial distribution of employment, type and the socio-economic conditions of the employees. Insufficient and unaffordable (time, cost, or comfort) transport may fail to provide sustainable linkage between workplaces and residences. It may also create congestion, noise and emissions and these are a threat to sustainable urban habitat.

The dynamic interaction between landuse and transport is intensively scrutinized to identify the spatial structure of the city and its role in transportation. Through

Chapter 4: Research Design 111 understanding the concentration of economic activities and their associated travel behavior, the main objective of this research is to explore sustainable growth alternatives to accommodate the future growth of the city. The key aim of this research is to understand the spatial structure of Dhaka and identify what implications this has for travel and sustainability. Once the dynamics are understood, the thesis assesses the implications of future growth and further understanding of how planning authorities ought to be managing further growth of the city.

The first objective of this thesis is to explore the international experiences and suggestions in the literature about sustainable urban form, job centres and transportation arrangement for large cities. The research explores appropriate methodology to identify job centre for Dhaka megacity. The second objective of this research is to understand the spatial structure of Dhaka and explore the nature and characteristics of the employment centres in Dhaka city. One of the important objectives is to investigate the employment characteristics of identified centres and analyses their travel consequences. The research then generates growth scenarios for Dhaka megacity by distributing growth using different development strategies and compare the commuting and vehicle emission.

4.3 Study Area: Dhaka Megacity

Defining the extent of the megacity of Dhaka is a challenging task especially when there is a need to aggregate data from several sources that possess different spatial and operational definitions of the city. As urbanization has progressed rapidly in recent times, its spatial definition has also grown. Moreover, different organizations working on Dhaka define the boundary with their own definition and utilize different city extents; therefore, in turn, data accumulation becomes more difficult. The situation adds additional complexity as the boundary changes with the organizations. For example, the extent of Dhaka varies from 360 to 1530 km2 according to different organizations operating in Dhaka. Very often the terms ‘city’, ‘megacity’ and ‘urban agglomeration’ are interchangeably used in reports and literature and their statistics are compared and aggregated inappropriately (Table 4.2). For example, Ahmed and

Chapter 4: Research Design 112 Ahmad et al. (2005) mentioned that the size of Dhaka city is more than 2000 km2with 34000 people per km2 (page 2), but a national daily (Prothom Alo, 2009) with a reference to a national population research centre claims to have 27,700 inhabitants per km2. That makes the data comparison and combination very difficult and confusing. The spatial dimension is the main focus of this research so, for this study, it is of utmost importance to define Dhaka consistently in terms of its spatial extent.

Table 4.2: Dhaka as Defined by Various Administrative Authorities

Administrative Body Name Area Population Density

(in Km2) (per Km2)

Dhaka City Corporation Dhaka City 3601 6,000,000 16,666

Bangladesh Bureau of Statistics (BBS) Statistical 3,9662 9672763 2,438 Metropolitan Area (SMA)

RAJUK (Capital Development Dhaka Metropolitan 1,5283 9,300,000 6,086 Authority) Development and Planning (DMDP) Area

Dhaka Transport Coordination Board Strategic Transport 7,4404 17,090,417 2,297 (DTCB) Plan Area (17,300,000

Source: 1Dhaka City Corporation 2Bangladesh Bureau of Statistics, 2001 3DMDP, 1997 4The Louis Berger Group, I. and B. C. Ltd (2005),

Dhaka city can be considered as the main city, which contains the old part of the city and a comparatively high density area enclosed by the Buriganga, Turag and Balu Rivers (Figure 4.2). The planning and development authority (RAJUK) in Dhaka has defined their jurisdiction by a plan in 1995 which comprised Dhaka city, several other municipalities within the and its surroundings in all directions with a higher level of urbanization. It is commonly known as the Dhaka Metropolitan Development Plan (DMDP) area. Further, the Strategic Transport Plan (STP) for Dhaka completed in 2005 defined an even larger area for Dhaka. The STP area is an

Chapter 4: Research Design 113 extension of the DMDP area covering the whole Dhaka division with an assumption that areas outside the city and DMDP area have significant contributions and connections with traffic in Dhaka (Figure 4.2). Bangladesh Bureau of Statistics (BBS), the government statistics department, has a different spatial coverage for Dhaka (Table 4.3).

Chapter 4: Research Design 114

Figure 4.2: Location and Administrative Boundaries of Dhaka Source: The Louis Berger Group, I. and B. C. Ltd (2005), DMDP (2005),

Chapter 4: Research Design 115 The mega city is comprised of different levels of urbanized area. It has been divided into four regions according to their urbanization pattern (Figure 4.2). The Main city has the highest population density, economic opportunities and infrastructure. The east side of the city was developed organically. Low lying agricultural land was divided into small residential plots and quickly became a part of the main city. The DMDP plan defined Dhaka as the ‘inner-ring’ region. Inner ring is the focus of current urban development and infrastructure is growing in that region. The outer ring, on the other hand, is the vast area, mainly agriculture based on the rural community. It geographical distance is not far from the main city but, due to poor transport links, it is not developing according to its potential.

4.4 Data and Data Sources

Data required for this study have been collected from the national census and the different planning projects of the study area (Table 4.3). Transport and travel behaviour data are collected from the Strategic Transport Plan Project (STPP) for Dhaka City. This database includes travel survey data for 2004 and various GIS maps for travel analysis. Landuse information has been collected from the Dhaka Metropolitan Development Plan (DMDP) project. Demographic and employment data are collected from publications from the Bangladesh Bureau of Statistics (BBS).

Table 4.3: Summary of Database Data Source Description

Strategic Transport Plan Project, • GIS format: Traffic Analysis Zone map, road network 2004 • Tabular Format: Household survey data covering 45000 persons for travel data • Matrix: travel distance, cost, time between all zones (217) • Household / travellers interview Detailed Area Plan, Dhaka • GIS format: landuse map, building footprint, building use, road network, number of floors, and Metropolitan Development Plan, other geographic features. 2005.

Bangladesh Bureau of Statistics, • Occupation at origin

2001 • Demographic data

Source: Prepared by author

Chapter 4: Research Design 116 The spatial unit of this research is the traffic analysis zone (TAZ) as used in the STPP. The scope of the study area covers Dhaka City and its surrounding areas including Nayanganj, Tongi, Savar, Munshiganj, Narshindi, and Manikganj. The study area appears large enough for the analysis since it covers most locations of employment and residences for Dhaka megacity. STPP divided the area into 217 TAZs in Dhaka megacity: 154 within the Dhaka Main City, 19 covering the remainder of Dhaka district and 44 covering the surrounding districts of Gazipur, Manikganj, Munhsiganj, Narayanganj and Narshindi (Figure 4.3).

Figure 4.3: Traffic Analysis Zones of the Study Area Source: The Louis Berger Group, I. and B. C. Ltd (2005)

4.4.1 Demography

Demographic information was obtained from the 2001 national census and some 2004 adjusted data from STPP. Basic demographic and workers data were used from the census. These came with different area coverage and, in most cases, larger spatial units. These were then disaggregated and adjusted to fit with TAZs of the STPP data. Job numbers in an area was not readily available from the census. The number of persons by occupation category was available and used in this study.

Chapter 4: Research Design 117 4.4.2 Travel Data and Transport Model

Transport and travel data were obtained from the STPP. This data includes inter- zonal trip numbers between all pairs of TAZs by income group by purpose. Other information in matrix format are the inter-zonal travel times and distances in km. This information is stored in 217x217 matrices and aggregated according to the purpose of analysis. Given that most commuting occurs between home and job locations, the coverage of GIS data encompasses not only the employment but also the home of each employee. If a job location is outside of this GIS coverage, the employment location is excluded from the analysis because it is beyond the scope of this research. Along with this travel data, numerous transportation related information is obtained from the STPP main report and working papers. Transport and travel data were mainly collected from the Strategic Transport Plan Project (STPP) and considered to be the most detailed data for this purpose. Information from other sources were aggregated and disaggregated to fit with the STPP data. STTP data includes 2004 trip matrices by income group by trip purpose, transport networks and survey results (household interview, road side interview and screen line survey).

4.4.3 Landuse and Building Footprint

A strategic plan for the future 20 year development and planning control for Dhaka megacity was completed in 1995 for next 20 years. As a detailed phase of the plan, a planning scheme was developed in 2005 to prepare for the detailed area planning of the whole megacity. DMDP area is divided into major six blocks and development plans are in progress. Detailed land use maps have been collected for all blocks and a combined map has been prepared for this study.

A GIS database with building footprints is also available that describes the number of floors, use of the building and year of construction. However, building use information may not provide enough detail, and floor information is not available for all structures.

Chapter 4: Research Design 118 This study uses data from three different sources which has temporal, spatial and contextual variation which may impact the results to a minimal scale. BBS data was collected from the 2001 census, STP dataset was collected from a series of surveys in 2004 and data from DMDP was collected between 2006 and 2008. Census data was based on household interviews and there may be inaccuracies in data collection level. STP transportation model, which has been utilized in this research, is under criticism. The sample size is not sufficient and several assumptions are not realistic.

4.5 Methods and Tools

To achieve the main objectives of exploring sustainable urban structure, this study first undertakes a data mining exercise with statistics, GIS and transportation tools using MS Excel, SPSS, ArcGIS and TransCAD software to explore the spatial patterns in job distribution, population concentration. At first, tables of the number of persons by occupation (from census dataset) is merged with GIS layers to visualize and explore their spatial patterns. Job concentration is summarized at a regional level to compare and contrast different occupational concentrations for different regions (Chapter 5, section 5.2). The spatial dimension of employment clustering is then explored using different spatial statistics tools described in section 4.5.2. Employment clusters are identified and centres are selected using statistical methods described in 4.5.3. Based on identified job centres, future scenarios are generated and evaluated using simulated traffic for future years. This research also develops an analytical framework (Figure 4.4) that combines statistical, spatial and travel analysis to aid the evaluation and decision of spatial planning options.

In this study, a GIS based analytical framework has been developed that can generate alternative scenarios based on given criteria of urban development and job allocation. The Dataset includes data from the Strategic Transport Plan Project (STPP) for Dhaka (The Louis Berger Group Inc and Bangladesh Consultants Ltd, 2005), demographic data from the Bureau of Statistics (Bangladesh Bureau of Statistics, 2001), map layers from the planning and development authorities and relevant consultants. The proposed analytical framework (Figure 4.4) allows generating

Chapter 4: Research Design 119 results based on given urban growth scenario to examine whether certain growth scenarios can be achieved and under what circumstances. It would be able to visualize different scenarios based on different conditions and assumptions.

After much investigation on the spatial patterns of employment concentration (section 4.5.2) an appropriate methodology is developed to identify employment centres in Dhaka megacity (section 4.5.3). The resulted centres are used as a growth ‘seed’ for scenario building and assigned employment growth to these centres. This study assumes 15% employment growth between 2004 and 2024, as estimated in the STPP, and assigned this growth to the selected areas according to their contribution to the job market. It is assumed that any centres contributing higher will have higher growth. It can be argued as an obstacle for polycentric growth since large centres are getting more growth. But there are allocation limits set in the model for regional distribution that will ensure that large centres will not get more growth in all cases and smaller centres will not be always ignored. These scenarios start by assigning higher growth (50% of new jobs) in the main city and lower growth in the outer ring. Later, the growth is decreased in the city and increased it gradually in the outer ring (Table 4.4). The SDSS generates two lines of scenario with the same development ratios but different spatial pattern of development. Line one assumes that employment growth will be contained within the selected 19 employment centres, which requires much stronger growth development control. Line two of the scenario assumes that after a certain stage of saturation in selected zones, trickle-down effects will be triggered and employment growth will speeded in the adjacent zones. This model is capable of utilizing the maximum limit of growth in an area that may come from development control policy. Since no such policy is practiced for Dhaka it is kept open.

Chapter 4: Research Design 120

Figure 4.4: Analytical Framework for Scenario Building

Chapter 4: Research Design 121 4.5.1 Data Pre-processing

Information required for this thesis is collected from several sources at different scales. As a result, pre-processing is required to bring them to a common scale and format in order to make it suitable for desired analysis. This pre-possessing includes data entry, digitization, format conversion and, most importantly, preparing data at TAZ level. The TAZ boundary map is digitised based on the Dhaka maps from other transport planning reports. Census data is also dis/aggregated to match with the TAZ boundaries. Analysis of the data commences with data mining, in order to have an understanding of the spatial pattern of employment.

4.5.2 Spatial Structure of Employment

Exploring the spatial structure of jobs is the first step in having an understanding of the distribution of employment in Dhaka. Two aspects of employment are examined. First, location of workers residences and second, the locations of jobs. Data from the national census is used and connected to GIS maps for spatial analysis. ArcGIS has been used to map the workforce-population ratios in different zones of Dhaka (section 5.2). Moran’s I and Getis-Ord statistics were computed to measure the clustering pattern and to examine the co-location of the workforce. Hakim (2010) identified regional job clusters outside the Jakarta urban core using Getis-Ord statistics. Three dimensional maps were generated in ArcGIS-ArcScene to illustrate the differences in job figures across Dhaka. The ranksize distribution of employment zones (Section 5.4) helps to identify employment peaks. Several studies by the EASTS group use this technique to identify job peaks (Alpkokin, 2005 and Vinchesen, 2005). However, they compared employment data from two points of time to identify sub-centres based on differences.

Global Moran's I

Moran’s I is a classical spatial autocorrelation tool to measure and analyze the degree of dependency among observations in a geographic space (Getis and Ord, 1992). It evaluates whether the pattern expressed is clustered, dispersed, or random (Figure

Chapter 4: Research Design 122 4.5). The tool calculates the Moran's I Index value to evaluate the significance of the clustering of workers residences.

Figure 4.5: Illustration of Moran's I Source: Environmental System Research Institute, 2012

The Getis-Ord general statistic is calculated based on equation 4.1 as suggested by Getis and Ord (1992).

n n

∑∑ zzw jiij n i=1j = 1 I = n … … … … (Equation 4.1) So 2 ∑ zi i=1

Where, zi is the deviation of an attribute from feature i from its mean (xi- X ) and xj are attribute values for features I and j, and wij is the spatial weight between feature i and j.

Getis-Ord Statistics

Moran’s I examines whether clustering exists or not, where as the Getis-Ord statistic in particular measures the concentration of high or low values for a given study area. This statistic illustrates the degree of spatial concentration of high or lower values (Figure 4.6).

Chapter 4: Research Design 123

Figure 4.6: Illustration of Getis-Ord Source: Environmental System Research Institute, 2012

The Getis-Ord general statistics is calculated based on equation 4.2 as suggested by Getis and Ord (1992).

n n

∑∑ )( xxdw jiij i=1j = 1 G = n n … … … … (Equation 4.2)

∑∑ xx ji i=1j = 1

Where, xi and xj are attribute values for features I and j, and wij is the spatial weight between feature i and j.

These two spatial statistics tools are employed to understand any clustering pattern in the various job categories. In addition, correlation analysis, and principal component analysis are performed on resident worker numbers (Table 5.6 and Table 5.7). The purpose of this quantitative analysis in to find any co-location of resident workers in the study area.

4.5.3 Employment Centre Identification

Existing employment centre identification is the central theme of this research. The recent tendency in the literature is to define sub-centres as a contiguous set of zones, each with a minimum density where, all together, a minimum total employment cut off is reached. The methodological problem is that this cut-off has varied for different cities (McMillen, 2001). A further classification is then achieved according to some locational and employment specific characteristics, such as edge city, or post-World War II suburban development, that are not necessarily ranked (McMillen

Chapter 4: Research Design 124 and McDonald, 1998). In terms of job centre identification methodology, a well- accepted clustering method with some minor modifications to adjust for the dataset and characteristics of the study area is adopted for this research. However, unlike Giuliano and Small (1991 and 1993), no minimum cut-off value of jobs is set for centre identification. A statistical approach was adopted instead (Figure 4.7). Several researchers who use clustering methods rank the candidate employment centres based on their total employment and later refine their centres by using gross employment density (Giuliano and Small, 1991 and 1999; Alpkokin, 2005; Bhandari et al., 2008; Duarte et al., 2011), the employment-population ratio (Forestall and Green, 1997), z-score of total employment (Parolin and Kamara, 2003), or z-score of gross employment density (Pan and Ma, 2004). The total number of jobs is not an important criterion in this case as the traffic zones are of varying size (0.14 to 300 sq. km.) and, as a result, a bigger zone may have a large number of jobs while having a low low density. The number of jobs in a zone does not by itself portray its prominence as a job centre but its spatial pattern (size, distribution and location) in respect to its neighbouring zones are crucial. However, gross job density and job- population ratio have unique characteristics which are contrasted in some cases in identifying job centres. This study adopts an approach that combines both criteria to select job centres.

Variations in job distribution among zones are readily noticeable in the Dhaka region (Table 4.4). Zones vary from 104 jobs per km2. to 95 thousand per km2, and the same applies with the job-population ratio values. Rather than using the raw job number, the corresponding z-score is calculated for centre identification.

Table 4.4: Employment Figures by Travel Zones

Minimum Average Max Number of Job 3,338 44,015 383,586 Gross Job Density 104 25,247 95,214 Job Population 0.126 0.625 1.490 Ratio Source: Estimated Based on STPP data

Chapter 4: Research Design 125 Z-scores are calculated for both the gross employment density (ZGJD) and job- population ratio (ZJPR). Z-score values determine the variation of values from their arithmetic mean. A positive z-score refers to a higher than average value. Figure 4.7 illustrates the process of Z-score calculation and its possible impacts on the outcome.

Higher ZGJD values emphasise those zones where more jobs are located in a smaller area. This criterion identifies smaller job centres. Business, shopping, corporate office, markets etc can be extracted using this criteria. On the contrary, the job- population ratio focuses on those zones with more jobs and less population. Dedicated business parks, industrial estates or manufacturing may be identified using this criteria.

Assume A and B are two different zones with a given job, population and other attributes. Zone A is a mixed landuse, high density, zone and Zone B is an industrial estate. They have an equal number of jobs but different area and population. The z- score of total employment will be equal as it is not sensitive to size of the area (unit of analysis) but the z-score of gross employment density and employment- population ratio varies by area and population respectively. So, it may have a different Z-score for job density and job-population ratio.

− XX A Z = i B i σ

Population 200 50 Z-Score

Jobs 100 100 Total Jobs A=B

Area 10 20 Gross Employment A>B Density

Gross 10 5 Jobs/Population A

Job Ratio 0.5 2

Figure 4.7: Effects of Spatial Units Size and Population on Employment Concentration Source: Author, 2012

Chapter 4: Research Design 126 In considering gross job density, an industrial estate may not be recognized as a job centre because an industrial estate occupies a larger space compared to its number of employees. But since it has less or no residential population, its job-population ratio would be higher. On the other hand, zones with taller buildings or a higher floor area ratio (FAR) may have a higher gross job density.

The job centres were identified based on the following approach. First, “potential” employment concentrations were identified for use in further analysis. A “potential” concentration was defined as a zone with a job density and / or a job-population ratio significantly higher than the regional mean. Employment data were standardized by calculating Z-scores for gross job density and job-population ratio. Positive z-score values show zones above the regional mean. All zones with standardized employment values (Z-scores) above 0, in either case, were considered as “potential” concentrations. This approach identified 143 zones.

Z-scores of the associated job numbers are used as it has a solid statistical meaning. Under the assumption that zonal employment density in a region is normally distributed, the Z-score values can be used to describe the difference between the density of a certain zone and the average density of all the zones.

i ZGJD , the Z-score associated with the employment density X in zone i, is given by:

i i − XX ZGJD = … … … (Equation 4.3) σ Where, X is the mean and σ is the standard deviation of the employment density variables. Given a threshold T on employment density, the above formula is transformed as follows:

− XX Z = i >= T … … … (Equation 4.4) i σ or

i >= +σ * TXX ...... (Equation 4.5)

Chapter 4: Research Design 127 A zone is selected when its density is significantly greater than or equal to the mean density. Under the assumption that employment density follows a normal distribution, selection of those zones that have density higher than 90 percent of zones, the threshold of the Z-score value T is set as 1.6. Therefore, the minimum cut- off density ( X +1.6 *σ ) was used to identify those zones in the 90th percentile. 45 zones were selected using this criterion. Similarly, the selection of zones that are in the top 10% using the job-population ratio criterion requires a minimum cut-off ratio set at ( X +1.0 *σ ); as a result 23 zones are selected. At the same time, zones that show positive z-scores from both have been selected; 55 zones fall into this category. After candidate zones are highlighted as clusters of contiguous zones with employment density over the minimum density threshold on the map, a GIS-based procedure is used to identify employment subcentres among the candidate sites.

In summary, potential zones are defined mathematically by the following two criteria:

i) Z-Score is higher than 1σ in either case (ZGJD > 1.6 or ZJPR > 1.0)

ii) Z-score is higher than 0 in both cases (ZGJD > 0 and ZJPR > 0)

This process identified 95 zones, all within the Main city and inner ring of the Dhaka metro area. The employment profile in the outer ring is so low that it cannot compete with employment numbers in the Main city and surrounding zones. So, any zone with above average job numbers is considered a potential employment centre in the outer ring; this process identified 3 additional zones. There are 98 zones in total that comprise the employment centres of the Dhaka metro area.

Chapter 4: Research Design 128 217 TAZ TAZ

Z > 0 & Z > 0 GJD JPR Job Z > 1.6 or Z > 1.0 Centre GJD JPR Selection Job > average (outer ring)

Potential 98 TAZ Zones

Dissolve (Adjacent TAZs)

12 Centres in Main City Potential 19 Job Centres 4 Centres in Inner Ring Zones 3 Centres in Outer Ring

Figure 4.8: Potential Employment Centre Identification Process

Zones with higher z-scores were considered and adjoining zones with similar z- scores are merged to finally determine a total of 19 employment centres (Table 5.9 and Figure 5.14).

4.5.4 Growth Scenarios

GIS is used to model this employment landscape and run the transport module to estimate the changing travel pattern for different scenarios. For sustainable

Chapter 4: Research Design 129 polycentric growth it examined the performance of the need to develop employment zones in the outer ring as people can find their job locally. It is assumed 15% employment growth will occur from 2004 to 2024, as estimated in STPP. Scenarios A to E are created considering 50% to 0% growth in the city employment figure and 25% to 75% growth in the outer ring (Table 6.1).

One of the objectives of this research is to explore different growth scenarios in order to compare their corresponding travel consequences. It is accomplished by assigning new jobs to the selected centres at different rates. It is estimated that there will be more than a million additional jobs by 2024 (The Louis Berger Group, I. and B. C. Ltd, 2005); STP working paper 8: Model Development and Calibration). These new jobs are distributed to different regions of Dhaka (Figure 4.2) according to the ratio for any specific scenario (Table 4.4). The regional allocation of jobs is then again re- distributed to the identified centres within that zone based on their current ratios in the job market. This implies that employment centres within a region will maintain their current ratio of job distribution within the zones in future. It is based on the assumptions that big job centres attract more jobs than smaller centres and at a proportionate rate of its current jobs. Ten scenarios are created based on two urban development strategies. Strategy-1 assumes that job growth can be contained within the identified centres. This can be achieved by strategic economic planning with strong development control. Strategy-2, on the other hand, assumes that it would be difficult to limit new jobs within specific job centres. It suggests a spill-over effect and assumes that after a certain level of job saturation all zones adjacent to job centres will be affected by new jobs. Scenarios are named from A to E and 2A to 2E for strategy-1 and strategy-2 respectively.

The extended part of Dhaka city developed organically on the east of the main city due to its close proximity to the city. The urban form, road and other infrastructure patterns resembles much with the other organic and unplanned part of Dhaka. The extended part is relatively a small area. Having similar characteristics of the main city and due to its adjacency, extended part of Dhaka and main city are grouped together for scenario analysis.

Chapter 4: Research Design 130 Table 4.5: Employment Distribution Under Different Scenarios

Distribution of Employment Growth in Different Scenarios

A B C D E

Main City and Extended City 50% 40% 30% 25% 0%

Inner Ring 25% 35% 35% 25% 25%

Outer Ring 25% 25% 35% 50% 75%

The new employment number in a zone can be calculated by:

Ji ic NJRJJ ∗+= n … … … … (Equation 4.6)

∑ Ji i=1

Where J c is total number of jobs in a centre by 2024. Ji is current job number in the corresponding zone. NJR represents the total number of new jobs in the region estimated by a scenario.

This technical aspect of this model logically connects ArcGIS and TransCAD with the database. Regional employment growth is estimated based on scenarios and then distributed to the employment centres of the region. The new employment landscape is generated in GIS to assist with visualization of the distribution (Appendix B). This new employment is then considered as an ‘attraction’ for calculating trip attraction. Production (number of job) in each zone is calculated based on the predicted demography and economic growth adopted in recent plans. Transport network has been built with the assumption that all ongoing, committed and proposed transport infrastructure projects will be completed on time. Gravity model is then applied in TransCAD environment to estimate interzonal commuting movement (Appendix B). The model generates the number of trips, travel time and travel distance between all pairs of zones for all scenarios.

However, one of the limitations of this approach is that achieving controlled job growth, especially outside the Main city would be difficult. This is because under

Chapter 4: Research Design 131 current trends of infrastructure development, the outer ring may not have sufficient infrastructure to accommodate this large growth (especially scenario D and E). Only with strong policy implementation and large investment for outer ring may able to make it. Secondly this top down approach lacks consideration of local development plans. Appropriate local empirical data and local plans are not available for this study and beyond the scope of this research. The whole perspective of this scenario based approach is to explore the decentralization of growth. But assumptions to distribute growth within the region contradict the decentralization to some extent by assigning more jobs to the large centre. However, on the other hand, it supports the approach to develop regional centres and a polycentric model of growth.

4.5.5 Commuting Distribution

Growth scenarios distribute new jobs according to strategies and attractions of the job centres. The new employment landscape is then entered into a TransCAD based gravity model to estimate inter-zonal trips and commuting parameters (Figure 4.9). This trip estimation, commonly known as “Trip Attraction”, is performed in TransCAD while taking into consideration that all under-construction, committed and planned transport projects will be completed by the year 2024 (APPENDIX A). Appropriate calibration values were taken for home to work type trips. Since trip attraction is used to generate values for all ten scenarios, GIS-DK (developers Kit) was used for the repetitive functions.

Chapter 4: Research Design 132

Figure 4.9: Trip Distribution Process Sourec: Author, 2012

In a gravity model, the flow between zones is calculated from the following equation:

i ⋅ dfP ij )( AT iij ∗= n . . . (Equation 4.7)

∑ z ⋅ dfP zj )( z=1

where: Tij = the forecast flow produced by zone i and attracted to zone j Pi = the forecast number of trips produced by zone i Ai = the forecast number of trips attracted to zone j dij = the impedance between zone i and zone j f(dij) = the friction factor between zone i and zone j

Gravity model estimates of interzonal trips are stored in matrices. Desire lines are created to represent resulting interzonal trip patterns for each of those gravity based matrices for further analysis of travel pattern (section 6.5). Each line of a desire line represents the number of trips between each pair of zones.

The transport network from the STP project is used to create all scenarios. It would be more appropriate to utilize a different network adjusted for corresponding scenarios. Again data limitations remain as the bottleneck. Several scenarios (especially D and E) assume much emphasis on the outer ring, but the network does not comply with that. Having more links and improving connectivity would generate different commuting and pollution scenarios.

Chapter 4: Research Design 133

4.5.6 Travel Impacts

To assess the performance of transport links and impacts under different scenarios, trip assignment is applied in TransCAD. Trip assignment estimates the number of trips on each link or road segment and calculates other values (vehicle km., total flow etc). Trip assignment for a link is estimated by:

  v β  tt f 1+= α   … … … (Equation 4.8)   c  

Where: t = congested link travel time, tf = link free flow travel time, v= link volume and c = link capacity and α, β = calibration parameters.

Dhaka megacity has several road categories with different volume, speed, lane, capacity and width attributes. The STP modelling team developed volume delay functions (VDF) to depict the link performances of different categories of roads under different traffic situations (Figure 4.10).

Volume-delay functions

90 1 - not defined

21 - Multi-lane divided, wide median, with NMV lanes 80 22 - Multi-lane divided, wide median, without NMV lanes 23 - Multi-lane divided, narrow median, with 70 NMV lanes 24 - Multi-lane divided, narrow median, without NMV lanes 25 - Multi-lane undivided, with NMV lanes 60 26 - Multi-lane undivided, without NMV lanes

27 - Multi-lane undivided, without NMV lanes 50 28 - Multi-lane undivided, without NMV lanes

40 31 - Two-lane divided, with NMV lanes 32 - Two-lane divided, without NMV lanes

Speed (km/hr) Speed 30 33 - Two-lane undivided, with NMV lanes 34 - Two-lane undivided, without NMV lanes 20 40 - One-lane, 3.5 to 4.5 meter width 41 - One-lane, 3.5 to 4.5 meter width

10 42 - One-lane, 3.25 meter width

43 - One-lane, 3.0 meter width

0 44 - One-lane, 3.5 to 4.5 meter width 0 500 1000 1500 2000 2500 Volume (PCUs / hour)

Figure 4.10: Volume Delay Function Used in Traffic Assignment Source: The Louis Berger Group, I. and B. C. Ltd (2005)

Chapter 4: Research Design 134

It is noticeable that road geometry is so insufficient that even multi-lane roads turn into very slow moving roads as the volume crosses 500 passenger car unit (PCU). The situation is severe for several road categories (road types 25 and 31 for example) which have both motorized and non-motorised traffic on the same lane.

Trip assignment estimates several key attributes on links. It includes traffic volume, volume to capacity ratio (VOC), vehicle kilometres travelled (VKT) and speed, etc. These are compared between scenarios to explore the impacts of job distribution at different rates in the different regions. Trip assignment is also made for the motorized modes to estimate their vehicle emissions.

4.5.7 Estimation of Vehicle Emissions

Climate change triggered by air pollution is a major concern now a day. Transportation is the fastest growing major contributor to global climate change, accounting for 23% of energy-related carbon dioxide (CO2) emissions (Asian Development Bank, 2010). Many experts foresee a three- to five-fold increase in

CO2 emissions from transportation in Asian countries by 2030 compared with emissions in 2000, if no changes are made to investment strategies and policies (Asian Development Bank, 2010). Sustainability of urbanism is sought through urban planning aids and by reducing movement to attend activities (other approaches are cleaner fuel invention and modernization of vehicle for fewer emissions).

This study estimates transport emissions from different urban development scenarios. It explores and compares their transportation emissions. In most Asian countries and cities, the majority of information necessary to assess air pollutant emissions cannot be found. These are collected only in individual projects. Road transport is associated with emissions of carbon dioxide (CO2), carbon monoxide (CO) and oxides of nitrogen (NOx), as well as particulate matter (PM). These emissions have a high negative impact on human health, particularly in densely populated urban areas. Often emissions are estimated by top down approach, mainly by looking into vehicle numbers and type and oil consumption. But this often leads to a gross estimation as

Chapter 4: Research Design 135 fuel consumption by vehicles and their vehicle kilometre is unknown. Authorities in developing Asian countries can not adequately “measure carbon”. Existing aggregate data tell us only approximately how many vehicles of each kind have been at one time registered nationally or by state. In almost no Asian developing country are data collected or official estimates made of how far vehicles by type and fuel are driven in a year or how much fuel each vehicle-fuel combination consumes (Schipper et al., 2009). There are also no regular national travel or commodity flow surveys. At the state or metropolitan level are occasional travel surveys and traffic counts but little reliable data on fuel consumption and almost no data on vehicle use.

One of the most acceptable approach to estimate emission is named ASIF, which assumes emissions (G) in the transport sector are a product of the level of travel activity (A) in passenger kilometers (or ton-km for freight), across all modes; the mode structure or percentages by mode (S); the fuel intensity of each mode (I), in liters per passenger-km; and the carbon content of the fuel or emission factor (F), in grams of carbon or pollutant per litre of fuel consumed (Schipper et al., 2009).

Figure 4.11: Transport and Emission Forecasts (Schipper et al., 2009).

Chapter 4: Research Design 136 A simplified version of this approach was adopted to estimate future emission for Dhaka (Figure 4.12). Motorized total vehicle kilometer (VKT) for each of the roads (links) are estimated from the trip assignment result. It is further disaggregated to vehicle types and emissions are stimated.

Figure 4.12: Emission Estimation Source: Author

Chapter 4: Research Design 137 Gravity model estimated commuting number between all pair of zones for all scenarios. The trip matrices from the gravity model do not have estimation on the usage of different modes of transportation. Recent, studies suggested that about 43% of commuting performed by non-motorized transportation (walking 14% and rickshaw 29%) (STP, 2005). It is assumed that shorter trips will be by walk and rickshaw. So, longer trips (above 57% percentile) are taken into consideration for emission estimation. These motorized trips are further dis-aggregated by modes. Modal share from STP has been used here. It is assumed that, in recent future, there will not be much change in the modal share. The car ownership trend shows an increasing pattern in car numbers but it is unlikely to impact the modal share as urbanization rate is higher (Bose, 2007). Emission factor from each of these modes are taken from (assets.dft.gov.uk/publications/road-vehicle-emission-factors-2009/ emissionfactors.xls) for fuel type used in Bangladesh.

Estimating total emission for Dhaka is beyond the scope of this research. This study attempts to estimate emissions from commuting movement and compare it among scenarios. Several thoughtful assumptions were made where empirical data was not available in appropriate format. Generalization is made on vehicle engine size and fuel types categories. Moreover, whole emission is estimated for road transport. No estimation is made for road construction, rail movement or others for other sources. Since there is no train metro system in iteration and no plan in near future, all estimation is limited to road based emissions.

4.6 Conclusion

Research objectives and questions are addressed within the framework of empirical analysis. Methods of empirical analysis are borrowed from different studies on similar field or with similar objectives. However, Modification has been proposed in cases, to adjust with the case study, data available and for better accuracy. GIS analysis, transport modelling and statistical analysis are exercised in an integrated way as output from one software can be utilized in another environment. To examine impacts of scenarios, GIS and transport models are run using iterative and conditional process which involved model builders (in ArcGIS) and application of

Chapter 4: Research Design 138 GIS-DK (in TransCAD). Transport emissions are estimated and surface of emissions are generated. However, due to the absence of air flow and direction modelling, the air pollution movement won’t be much accurate. The emission results can not thus be used as the ‘total’ air pollution rather emissions from road transport.

Chapter 4: Research Design 139 Chapter 5

Urban Spatial Structure of Dhaka Mega city

5.1 Introduction

This chapter address the first set of research questions regarding the spatial structure of Dhaka mega city. Empirical analysis begins by presenting the workforce profile of the mega city and its spatial allocation. Next, the identification of employment clusters is initiated through exploration of job distribution patterns and then by applying appropriate methods, employment clusters for Dhaka are finally identified.

This chapter measures the concentration of jobs within the Dhaka mega city and evaluates whether it is more characteristic of other Asian mega cities in terms of a polycentric structure or is more of a monocentric type city, or characterizing something different. This is achieved through employment and travel data analysis using GIS and transportation modeling software. Spatial distribution and concentrations of jobs are analyzed based on employment size, type, density and its contiguity in the metropolitan area. Standard score of gross employment density and employment to population ratio are estimated. The clusters of zones with higher standard score are then compared to evaluate their role and significance. Spatial characteristics of the resulting employment clusters are investigated and contrasted between themselves.

While the focus of the thesis is on employment spatial structure, spatial associations of the employment structure on several dimensions are probed to shed more light on the spatial impacts of employment distribution in Dhaka. The results are evaluated

Chapter 5: Spatial Structure of Employment in Dhaka 140 and discussed with regard to how they have addressed the research questions set out in Chapter 1. Possible policy implications are also investigated along with the discussion of empirical results. The resulting employment spatial structure and important spatial characteristics that are, in turn, identified are used in the next chapter to investigate the commuting trips attached to these centres.

This chapter presents the research finding on job cluster identification for Dhaka. First, it uses census information, to present the analysis of spatial patterns of workforce location for different categories of employment – their contribution in overall job market, allocation in different regions, and specific clustering patterns. Secondly, this chapter portrays the job distribution pattern in the study area. The main focus is on gross job density and the job-population ratio – as these two job characteristics have been widely used for cluster identification and match well with the dataset available for this research. The resulting job-clusters are then further analysed in term of their travel impacts, building footprint and usage in order to explore the extent of spatial interaction with the area.

5.2 Employment Profiles in Dhaka

The household survey from the 2001 national census indicates that the workforce in the greater Dhaka region is approximately 5.7 million in number (BBS, 2001). The distribution of jobs by major nine industries is presented in Table 5.1. Agriculture (25.6%) and business (17.85%) are major job industries. The service sector (2.82%) is represented by formal jobs in Government, semi-government and private institutes. The other category (32.52) includes employees in private business, self employees, and informal sector jobs.

The tertiary sector of the economy (also known as the service sector or the service industry) is one of the three economic sectors, the others being the secondary sector (approximately the same as manufacturing) and the primary sector (agriculture, fishing, and extraction of natural resources).

Chapter 5: Spatial Structure of Employment in Dhaka 141

Table 5.1: Jobs Distribution of Jobs by Types of Industries Job Number of Category Jobs Percentage Primary

Agriculture 1,457,093 25.60%

Secondary

Industry 599,640 10.54%

Tertiary

Utility 20,576 0.36%

Construction 220,374 3.87%

Transport 331,079 5.82%

Hotel and Restaurant 35,039 0.62%

Business 1,015,771 17.85%

Service 160,415 2.82%

Other 1,850,714 32.52%

Total Job 5,690,701 100.00%

Source: Bangladesh Bureau of Statistics, 2001

The high proportion (25.6%) employed in Agriculture is not an anomaly, as the spatial distribution of these jobs (Figure 5.1) reveals that a major share of population of the outer region of Dhaka rely on ‘Agriculture’ as their means of living. At the other extreme, the main city largely depends on tertiary sector jobs (up to 78%). The declining national achievement in industrial development is truly reflected by its workforce population (10.54%). Businesses of different types and sizes account for a higher percentage of jobs in Dhaka (17.85%).

Chapter 5: Spatial Structure of Employment in Dhaka 142

Figure 5.1: Employment Sectors of Different Regions of Dhaka Mega City Source: Processed by author

Figure 5.1 and Table 5.2 show that while tertiary jobs dominated across all the rings of the city, a higher agriculture dependent population at farther distance from the main city. It is noticeable that a small share (11.57%) of agricultural workers live in the main city. People living in the main city are not necessarily engaged in direct agricultural practices but they are more likely to be engaged in agricultural related business. Several of them are working in government departments and institutes involved with agricultural development, administration and marketing.

Chapter 5: Spatial Structure of Employment in Dhaka 143

Table 5.2: Distribution of Jobs by Zones in Dhaka and by Major Job Industries Job Percentage Zone Area (Km2) Population Primary Secondary Tertiary Total

Dhaka City Dhaka Main City 147 5,604,755 11.57 10.34 78.09 100

Extended City 155 991,382 14.63 8.29 77.08 100

Inner Ring

West 284 630,788 24.03 11.43 64.54 100

North 369 470,122 18.91 26.24 54.85 100

East 495 748,061 31.84 15.04 53.13 100

South-East 335 1,517,537 14.20 14.36 71.44 100

South-West 177 633,215 13.43 7.93 78.64 100

Outer Ring

Western 1,673 1,694,553 57.86 2.90 39.23 100

Northern 1,129 2,268,134 55.16 4.43 40.41 100

Eastern 1,353 2,313,227 44.90 10.44 44.66 100

Southern 1,327 1,836,366 42.12 3.20 54.68 100

Source: Processed by author based on Bangladesh Bureau of Statistics, 2001

Therefore there is a tendency of an increasing primary sector workforce from the centres to the outskirts of the metro area. However there are discrepancies and these can be explained by looking at the landuse and transport connectivity in the area. For example the southern zone in the inner ring has relatively lower primary jobs and higher tertiary jobs compared to its western and eastern counter parts. Old Dhaka is very close to the south zone, which has its impact on its workforce. Keraniganj (south of Dhaka) has been developed as a low and middle income residential area. The Southern zone is the closest place from Old Dhaka offers cheap land and rents. Construction of the bridge on the Buriganga river in 1990 that transformed the landuse changes and workforce transformation in the South. Many people commute from Keraniganj to Old Dhaka and the main city for small business and the informal

Chapter 5: Spatial Structure of Employment in Dhaka 144 sector. The secondary sector workforce is highly concentrated in the inner ring, especially towards the North (26.24%). Several urban planning studies have documented the north area of Dhaka as a potential growth area (See Chapter 3) and growth in ready-made garments, light manufacturing and pharmaceuticals have now occupied this area.

The spatial distribution of the workforce shows a clear pattern of main-centre dominance (Figure 5.2). The higher workforce density in the main city reflects the extreme population density of the area. This demographic characteristic can be explained by the fact that the main city has a higher number of jobs and urban amenities that attracts families of all demographic ranges. The workforce-population ratio map (Figure 5.3) shows a very clear pattern of lower job-population ratios in the outer zones. Interestingly, the main city and the extended city have lower workforce ratios compared to the area north of inner-ring. This is because this area (North of the inner-ring) is characterized by secondary sector jobs but it does not have adequate urban amenities to attract families. As a result, families with a higher workforce ratio live there or prefer to live close to their workplace sacrificing other urban amenities. The main city, on the other hand, has a very high workforce and extremely high population, but the job-population ratio is low. The differences in job-population ratio distribution also reflect differences in different social aspects and infrastructure. This not only shows the growing economic importance of the inner-ring region but also shows insufficiency in urban infrastructure in the inner and outer ring that might attract people to live close to their workplace. This issue needs to be addressed in urban spatial policy, and appropriate measures need to be taken for development priorities and resource distribution.

The amenities in the main city attracts population from everywhere in the country. Whoever can afford it want to stay in the main city. The increasing availability of land made available by the flood protection embankments has in turn made available plenty of low developable land and has attracted industries and residential investment. However, due to lack of growth in urban amenities it has not attracted people to move from the main city. It also shows that a large number of people

Chapter 5: Spatial Structure of Employment in Dhaka 145 involved in business, industries, service, construction live in the outer ring, but eventually this causes commuting to inner ring or to main city for everyday jobs.

Figure 5.2: Gross workforce density in different zones of Dhaka Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 146

--

Figure 5.3: Workforce-Population Ratio in Different Regions of Dhaka. Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 147

The dominance of agriculture-based jobs in the outskirts indicates that urbanization did not decentralize homogenously in the megacity, not at least from an employment perspective. The historical urban growth trend as illustrated in chapter 3 shows a northward growth. The same trend is reflected in urban jobs namely business, industries, service, construction and others. As seen in figure 5.2, a larger population share live in the northern area, in Gazipur for example. The main city does not show any prevalence of any workforce concentration- this reflects mixed population as in other metropolitan cities where people of different occupational background mingled together spatially.

5.3 Spatial Clustering of Workforce

The Location of job centres has significant importance for its role in overall spatial structure, travel pattern and sustainability of the city (Vega and Reynolds-Feighan, 2008). Spatial distribution of residence location of the workforce is equally important. Firstly, because residential locations are decided by trade-offs among rent, distance from work, commuting cost and proximity to urban services. People try to minimize the cost while maximising income. Secondly, location and growth of residential areas have impact on surrounding land uses and urban form. Another important aspect of workforce location is its impact on transportation arrangement. Based on nine major categories of workers in Dhaka this research examines the spatial clustering of the workforce to see whether it has any particular pattern and how does it impact the travel, landuse and over all urban settings. The Getis-Ord, Moran’s I along with Anselin Local Moran’s Index are used to explore and explain the types and patterns of workplace.

The Dhaka mega region shows a general pattern of tertiary workforce concentration in the central part and an agriculture dependent population in the outer zone (Figure 5.2 and Figure 5.3). However, this clustering pattern assists in understanding the spatial distribution of the workforce more explicitly, which certainly has connections with the locations of economic activities. The Spatial Autocorrelation tool evaluates whether the pattern expressed is clustered, dispersed, or random. Global statistics like the spatial autocorrelation (Global Moran’s I) tool assess the overall pattern and

Chapter 5: Spatial Structure of Employment in Dhaka 148

trend of the data. Having a higher positive z-score, Moran’s index depicts spatially clustered values for all categories of workforce (Table 5.3). When the z-score or p- value indicates statistical significance, a positive Moran's I index value indicates tendency toward clustering while a negative Moran's I index value directs tendency toward dispersion (Getis and Ord, 1992).

Table 5.3: Global Moran's I Statistics Moran's Expected Variance z-score p-value Index Index Agriculture 0.190056 -0.004630 0.000070 23.191619 0.000000 Industry 0.072572 -0.00463 0.000068 9.375959 0.000000 Utilities 0.057282 -0.00463 0.00007 7.421724 0.000000 construction 0.105549 -0.00463 0.00007 13.18385 0.000000 transport 0.095602 -0.00463 0.00007 11.98765 0.000000 Hotel and Restaurant 0.05673 -0.00463 0.00007 7.339298 0.000000 Business 0.134165 -0.00463 0.00007 16.5708 0.000000 Service 0.046897 -0.00463 0.00007 6.149135 0.000000 Other 0.073366 -0.00463 0.000068 9.442515 0.000000 Total Workforce 0.180616 -0.00463 0.000071 21.99497 0.000000 Source: Processed by author

Getis-Ord Statistics measure the degree of clustering for either high values or low values using the Getis-Ord General G statistic (Getis and Ord, 1992). Having a lower p-value, (which significantly rejects the hypothesis of randomness), negative z-score and smaller observed General G index than expected index (Table 5.4), indicates that (a) the pattern is not random (b) there is a clustering of similar workforce types, and (c) low values are clustered in the study area. It shows a clustering pattern for employees of most job sectors. The higher negative z-score and smaller observed General G index than expected index, indicate that low values are clustered spatially in the study area. The lower observed value for ‘Hotel and Restaurant’ employment shows a dispersal pattern for its workforce; this accounts for the locational scatter pattern for hotel and restaurant employment types (Table 5.4). This spatial statistical test finds a general clustering pattern of workforce locations. However these statistical results does not specify the area of clustering rather a global picture is presented.

Chapter 5: Spatial Structure of Employment in Dhaka 149

Table 5.4: Getis-Ord General G Statistics

Observed Expected Variance: z-score: p-value: General G: General G:

Agriculture 0.000013 0.000130 0.000000 -6.478342 0.000000 Industry 0.000048 0.000130 0.000000 -4.179709 0.000029 Utilities 0.000074 0.000130 0.000000 -4.239074 0.000022 Construction 0.000058 0.000130 0.000000 -5.819670 0.000000 transport 0.000065 0.000130 0.000000 -5.067507 0.000000 Hotel and 0.000125 0.000130 0.000000 -0.485748 0.627146 Restaurant Business 0.000054 0.000130 0.000000 -6.353548 0.000000 Service 0.000085 0.000130 0.000000 -4.254531 0.000021 Other 0.000075 0.000130 0.000000 -5.079570 0.000000 Total 0.000037 0.000130 0.000000 -7.420692 0.000000 Workforce Source: Processed by author

The correlation matrix (Table 5.5) shows higher order correlation among tertiary jobs while agriculture has lower correlation with other sectors. It depicts a scenario of having several job categories in the secondary and tertiary sectors located together. The rotated component score (Table 5.6) and component matrix (Table 5.7) supports a similar pattern. Factor 1 is positively associated with the workforce in all tertiary sectors. In fact, factor 1 depicts the core region because it loads high on tertiary sector jobs that are commonly concentrated in the main city. At a confidence interval of 0.05, the Gi* score identifies several workforce clusters in the inner ring region.

Table 5.5: Correlation Matrix Hotel and

Agriculture Industry Utility Construction Transport Restaurant Business Service Industry .353 Utility .355 .791 Construction .483 .804 .884 Transport .439 .793 .833 .890 Hotel and Restaurant .202 .675 .755 .727 .680 Business .642 .755 .826 .871 .841 .719 Service .502 .581 .634 .672 .613 .586 .649 Other .455 .777 .828 .868 .864 .737 .914 .590 Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 150

Table 5.6: Eigenvalues for Components Scores Initial Eigenvalues Extraction Sums of Squared Loadings Component Total % of Variance Cumulative % Total % of Variance Cumulative % 1 6.589 73.209 73.209 6.589 73.209 73.209 2 .920 10.219 83.428 .920 10.219 83.428 3 .496 5.516 88.944 4 .321 3.571 92.514 5 .232 2.579 95.093 6 .166 1.850 96.943 7 .136 1.506 98.449 8 .089 .992 99.441 9 .050 .559 100.000 Extraction Method: Principal Component Analysis. Source: Processed by author

The distinctly higher eigenvalue for component-1 (Table 5.6) rejects any chance of having specialized workforce clustering. The higher component value for the secondary and all tertiary sectors jobs (Table 5.7) reveals a mixed workforce all over the Dhaka region. The higher correlation of secondary sector (industry) with tertiary and higher component values in the rotated component matrix indicates those industrial and tertiary workforces are co-located. Perhaps it can be inferred that their demand for urban services, and travel for secondary and tertiary employees are the same, as their choice of residential location.

Table 5.7: Rotated Component Matrix

Employment Component Category 1 2

Agriculture .547 .810 Industry .863 -.167 Utility .915 -.185 Construction .949 -.036 Transport .919 -.067 Hotel and Restaurant .805 -.347 Business .943 .148 Service .750 .220 Other .929 -.063 Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 151

The Hot Spot analysis for factor 1 reveals secondary and tertiary workforce clustering in different regions of the inner-ring (Figure 5.4). At a 95% significance level, a Gi*Score larger than 2.08 represents clustering. This result is surprising as the Main city and the extended part in the East do not have any clusters. However, these can be explained with the properties of Hot Spot on varying size spatial data and demographic pattern in Dhaka. First, in the main city, the workforce ratio (to population) is smaller than that in the inner-ring. As in the main city, urban facilities like schools, hospitals and restaurants are more prevalent; it has more of a non- working population. Secondly, for Hot Spot analysis (Gi*Score calculation), the number of person in the workforce has been used. The zones in the main city are small compared to zones in the inner and outer ring. This variation in size is not addressed in hotspot analysis. As a result the Gi*Score on the workforce number is inclined to find clusters among smaller zones.

Chapter 5: Spatial Structure of Employment in Dhaka 152

Gazipur

Narshindi Tongi

Manikganj Savar

Narayanganj

Munshiganj

Gi* Score > 2.08

Figure 5.4: Job Clusters by Getis-Ord statistics Source: Processed by author

To visualize the spatial clustering pattern and to explore it at the local level, Anselin’s Local Moran Index was computed and mapped (Figure 5.5). It shows the clustering of high and low values (Anselin, 1995). Figure 5.5 shows that a higher number of the agricultural workforce are located to the north-west and north-eastern side, with very low values for agricultural employment in the main city. Interestingly, the municipalities have a different employment character as they are surrounded by higher rates of agricultural employment. Main city seems to have proximity for all employment categories. The inner-ring highlights very distinctive characteristics. A High-Low (HL) category on the Anselin Index shows a higher frequency surrounded by low values. The HL value for the inner ring, for most of the categories, illustrates that the inner-ring has a high workforce for several categories but the adjacent outer ring, or areas in the main city have lower values. This clearly shows that the inner-ring is the location for non-specialist employment. However, north of the inner-ring is a higher cluster of the secondary sector workforce. The inner-ring is characterized by the same level of infrastructure development as in the main city, both has lower rents. From a bid-rent perspective, it can be concluded that

Chapter 5: Spatial Structure of Employment in Dhaka 153 people living in the inner-ring serve as a labor catchment for the main city. The Anselin Index also indicates strong concentrations of service, industry, transport and construction sector workforces to the north of the city. This probably, supports the northward growth trend of the city.

Chapter 5: Spatial Structure of Employment in Dhaka 154

Figure 5.5: Anselin Local Moran Index for Workforce of Major Employment Sectors Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 155

5.4 Job Location and Distribution

The gross employment density (Figure 5.6) shows a traditional monocentric pattern of employment concentration, thereby highlighting the well-known employment hubs such as Old Dhaka, Motijheel, Gulshan, Dhanmoni, Mirpur etc. The Job- population ratio on the other hand demonstrates extended job concentrations towards the north of the city (Figure 5.7). This is because, firstly, in the main city, more jobs are located within small land parcels compared to low density jobs in the outer areas. Old Dhaka still remains a prominent economic centre and has significant contributions in the local and national economy. Several older parts of Dhaka, namely Sadar ghat, Keraniganj, Lalbagh and Sutrapur have a significant role in the economic landscape of Dhaka. Several areas of Dhaka were developed as a result of conscious planning initiatives to accommodate future growth and facilitated by proper planning for roads and infrastructure. However, there are wide inequalities across the metro area in basic infrastructure development. As a result, most of these planned residential, commercial and industrial areas have attracted activities and have turned into economic hubs in the city. These have made the inner city very competitive for land markets. To accommodate the growth pressure, the land locked nature of Dhaka has allowed growth of the city mainly towards the North, which has been documented in several planning reports (see Chapter 3). However, the urban services (school, hospital, markets etc.) and infrastructure are not abundant to allow flourishing of residential development in these areas. But industrial and manufacturing growth has instead taken place at a greater pace. The growth of Tongi, Gazipur and Savar constituted the lion’s share of the secondary sector in the region. So, this greater northern region of the main city has evolved as a light manufacturing zone with low employment density and even lower residential population density.

Chapter 5: Spatial Structure of Employment in Dhaka 156

Mirpur Savar Gulshan

Khilgaon

Motijheel

Narshindi

Dhanmondi Narayanganj Gulshan

Figure 5.6: Employment-Population Ratio Source: Processed by author

The highest gross job density areas are Khilgaon, Old Dhaka, Motijheel, Dhanmondi, Mirpur and Gulshan. Old Dhaka still remains as the main centre for retail business. Presence of several readymade garments industry in Kilgaon made it the highest density area. On the contrary, the outer ring does not show any employment peaks. These two figures (Figure 5.6 and Figure 5.7) are evidence of having two types of job clusters in Dhaka: one with high density- too many jobs in a small area and the other with a high job-population ratio characteristics of a huge number of jobs dispersed over a wider area with residential population. The main city, which has been developed over the past centuries is characterised by high density development, while the inner ring, on the contrary, has a higher job- population ratio.

Chapter 5: Spatial Structure of Employment in Dhaka 157

Savar Gazipur

Gulshan Narsindi

OldDhaka

Munshiganj

Figure 5.7: Employment Density Map of Dhaka Source: Processed by author

Old Dhaka

Figure 5.8: Density for Ranked Ordered Job Centres Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 158

The rank-size distribution of the job centres and their gross density (Figure 5.8) shows a density gradient characteristics of a monocentric city. Some parts of Old Dhaka have the highest gross job density (95,000 jobs/Km2) area as banks, insurance and financial institutions are located in this high rise and high density area. The job density in Mohammadpur (104,000 jobs/Km2), Mirpur (90,000 jobs / Km2), Gulshan (68,000 jobs/Km2) and Dhanmondi (62,000 jobs/ Km2) are higher as well. Gross job density when compared to the distances of zones from the CBD, creates a mono- centric type gradient (Figure 5.9), which means that density is reduced with increasing distance from the CBD. This figure also exposes the compact nature of the main city and itseconomic activities as most centres are located within 10-15 km from the CBD. Beyond 20 km no centres showing distinctive job density.

100000 Old Dhaka 90000 80000 70000 Gulshan ) 2 60000 50000

(perKm 40000 Job Density Dhanmondi Narshindi, 30000 Savar Munshiganj 20000 10000 0 0 10 20 30 40 50 60 70 80 Distance From CBD (Km)

Figure 5.9: Job Density at Difference Distance from the CBD

Source: Processed by author

The job to population ratio identifies those areas which have more jobs compared to their residential population (Figure 5.10). The ratio identify Savar as the most prominent job cluster with 1500 jobs per 1000 residential population (ratio 1.49), followed by Gulshan (1.22) and Gazipur (1.18). Savar region has turned into a light industrial and manufacturing zone and does not have much residential population.

Chapter 5: Spatial Structure of Employment in Dhaka 159

The inclusion of an export processing zone (EPZ) in Savar has seeded the employment growth witnessed in the region. Gulshan on the other hand, though planned for residential purposes, become a new hub for corporate headquarters and a financial and economic centre. The close proximity to foreign missions, the higher status nature of the suburb and its accessibility to important locations have made it a lucrative place for activities which have extensive foreign involvement. These corporate and business activities have pushed out the residential population and has become a major centre.

1600 Savar 1400 Gulshan 1200 1000 800 Motijheel Dhanmondi 600 (000 (000 population)

Job Population Ratio 400 Narshindi 200 0 0 50 100 150 200 250 Rank of Employment Zones

Figure 5.10: Job-Population Ratio for Rank Ordered Job Centres Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 160

Savar Gulshan Gazipur

Narayanganj

Figure 5.11: Job Clustering with Distance from CBD Source: Processed by author

The higher job-population ratio at distances shows a promise and potential to have centres at distances. Figure 5.11 shows that job centres in the main city and those close to the CBD have 500 to 800 jobs per thousand people, which indicates a higher ratio of mixed landuse. In other words, the CBD has twice the population as the number of jobs.

The employment density and employment-population ratio highlights two characteristics of employment centres. Any area that has higher number of jobs in a unit of land will have higher gross job density, without considering their population profile. Job-population ratio on the other hand highlights the area having more jobs compare to its population. In both categories one can observe the proximity to many employment centres at close distance to CBD.

5.5 Employment Centres in Dhaka

In terms of job centre identification methodology, a well-accepted clustering method, with some minor modifications to adjust with the dataset and characteristics of the study area is adopted for this research (for detailed, chapter 4, section 4.5.3). However, unlike Giuliano and Small (1991 and 1993), no minimum cut-off value of jobs is set for centre identification. A statistical approach was adopted instead

Chapter 5: Spatial Structure of Employment in Dhaka 161

By using the z-score method on gross employment density, and the employment- population ratio, a total of 98 zones with employment above the average fof the regional mean were identified as employment centres. The contiguous zones are merged, and this identifies that there are 12 centres inside the main city and 4 inside the inner ring (Table 5.9 and Figure 5.14). However, the analysis has identified 3 other potential clusters in the outer ring which contain a significant number of jobs around the areas they represent. They are not significant as centres in comparison to the large employment centres within the main city, but they appear to be influential at a regional level and have potential for decentralized development in the future.

Table 5.8: Employment Centres within the Dhaka Megacity Job- Gross Job Population No. Job Centre Job Area Population Density* Ratio Main City 1 Old Dhaka 959694 19.51 1573292 49,190 0.610 2 Kafrul and Pallabi 560055 13.55 719962 41,332 0.778 3 Badda and Khilgaon 399062 10.01 481911 39,866 0.828 4 Mohammadpur 276593 5.88 402825 47,040 0.687 5 Gulshan 261335 8.50 226000 30,745 1.156 6 Mirpur 191091 4.65 284659 41,095 0.671 7 Tejgaon 170401 4.63 245997 36,804 0.693 8 Cantonment 124123 10.21 134765 12,157 0.921 9 Ramna 115239 2.06 167827 55,941 0.687 10 Motijheel 84216 1.48 134706 56,903 0.625 11 Dhamondi 79993 1.72 134272 46,508 0.596 12 Hazaribag 61035 1.59 89389 38,387 0.683 Inner Ring 13 Savar 854286 257.17 675737 3,322 1.264 14 Gazipur 784171 304.81 691476 2,573 1.134 15 Narayanganj 185229 20.29 364449 9,129 0.508 16 Tongi 57984 16.35 65968 3,546 0.879 Outer ring 17 Munshiganj 46756 20.45 88836 2,286 0.526 18 Narshindi 45067 13.50 96593 3,338 0.467 19 Manikganj 26153 27.03 67377 968 0.388 * Number of jobs per square km. Source: Processed by author

Job density in the employment centres of the main city is significantly higher than any part of the metropolitan region (Table 5.9), with job density declining rapidly as

Chapter 5: Spatial Structure of Employment in Dhaka 162 one moves towards the outskirts. This is characteristics of a high density monocentric region which has a steep development gradient from the main city. It is noticeable that all employment clusters in the city have higher job density (ranging from 38 to 56 thousands/km2). It also reveals that tertiary jobs are located within the city and the space occupying secondary sector has been pushed to the inner ring which is not too far from the city but still at a distance. The inner ring is characterised by secondary sector jobs with a density of 3 to 9 thousands jobs km2. However having a higher job-population ratio, these clusters have been selected as job centre. In the outer ring, municipalities were for having relatively higher job concentrations compared to their agriculture surroundings. Gross job density ranges from a thousand to three thousand jobs per km2.

Figure 5.12: Employment Centres in the Dhaka Megacity Area Source: Processed by author

Chapter 5: Spatial Structure of Employment in Dhaka 163

The employment landscape reveals job spreading all over the mega city; some area may have less and more in some area. However, the identified centres contain 5.2 million jobs which is nearly 54% of total job. The main city has about 3.28 million jobs in its 12 centres located within a 154 km2 area and houses 4.6 million residents. Most of these jobs are on mixed landuse zones and having different types of jobs mingled together. The inner ring on the other hand has 1.9 million jobs within a 1400 km2 area.

5,000 Job 4,000 Population 3,000 2,000 (in (in Thousands)

Joband Population 1,000 Population 0 Job Main City Inner Ring Outer Ring

Figure 5.13: Job Centres in Different Regions of Dhaka Source: Processed by author

Table 5.9: Job Centres in Different Regions of Dhaka Gross Job Job-Population 2 Region Job Number Area (km ) Population Density Ratio Main City 3282837 84 4595604 39179 0.71434286 Inner Ring 1881670 599 1797629 3143 1.046750838 Outer Ring 117976 61 252805 1935 0.466667779 Total 5282483 743 6646039 7106 0.79483183

Old Dhaka has a history of being a market place for the city. Over the past centuries its importance has not lessened but has been transformed into the retail hub of the country. The area is characterized by high-density old buildings with narrow roads winding through the area. Most building are of mixed use. Generally, shops or godowns or businesses are on the ground floor and residential uses on the top floors.

Chapter 5: Spatial Structure of Employment in Dhaka 164

Most buildings are 2 to 4 stories in height. Old Dhaka does not have a higher gross employment density nor job-population ratio. There are several reasons for this. First, because of the mixed nature of the buildings job-population ratio is low (0.52) for any employment sub-centres. Secondly, the presence of retail businesses in this area requires very large warehouse which does not create many jobs at all. Most importantly, its urban form and fabric carries the history and heritage of past centuries and characterized by narrow winding streets and low rise buildings. However, its attractiveness in not completely captured by these numbers. Because of the hub of retail business, it attracts a large population from all over the Dhaka region and other parts of the country; this causes enormous congestion problems in the narrow road system.

Motijheel is another important job cluster with very high job density in the region (56 thousand/km2) as the area is characterized by high rise financial institutions. Commercial buildings are located along the main road but a significant part of the zone has residential buildings (figure B.6, appendix-B). These reduced the job ratio to 0.625. Another, prominent job cluster is Dhanmondi (46,500 jobs/km2), which was originally planned and developed for residential purposes in the 1950s. It has several shopping centres, universities, clinics, many schools and small and medium sized offices. Mohammadpur, Mirpur and Gulshan are also planned residential areas with equally important job clusters. Dhaka cantonment is also identified as a job cluster with a high job-population ratio. Military cantonment itself has a large number of officers and staff and many people from different parts of the city commute to the cantonment for their work.

5.6 Conclusion

The focus of this chapter has been on the spatial structure of economic activities within the Dhaka Mega city and its impact on the travel behaviour of resident workers. Specifically, the chapter has examined questions of how the location and dispersal of employment, and growth of business centres, affects the characteristics of travel and transport sustainability and identifies the key issues for policy makers

Chapter 5: Spatial Structure of Employment in Dhaka 165 to address in attempting to move to a more sustainable environment. An overview of the dynamics of employment concentration in the Dhaka metropolitan area is presented here by identifying major employment centres. A modified clustering method has been used on data for 2005 to identify 19 employment centres that play a significant role in accommodating the majority (54%) of employment of the metropolitan area. Gross job density and the job-population ratio measures and their spatial characteristics were used as the basis to identify significant employment centres. It can be concluded that the Dhaka mega city has a spatial form that is not polycentric but, rather, one that can be called an ‘extended monocentric’ structure. The role of the main city and some parts of the inner areas are paramount with the highest levels of employment density concentration. Some minor centres in the outskirts of the metropolitan area do not play a regional role in terms of employment concentration but have local significance and the potential to be major centres in the future with the right policy signals.

The methodology adopted here was carefully designed to ensure that employment cluster of any significance should be left out. It was found that Dhaka is characterised by an extended monocentric pattern of spatial structure, unlike many other mega cities in Asia that appear to be developing polycentric characteristics. However, some locations, especially those towards the north and north-west parts of the metropolitan area, have a high potential to become major employment centres in the future. There is no doubt that from the first attempt at assessing the travel data that past ad-hoc metropolitan policies of decentralisation and centre formation have not proved successful (see chapter 4). The lack of adequate road and public transport based infrastructure in these inner and outer areas has undoubtedly contributed to persistence of the extended monocentric pattern and, as a consequence, to severe sustainability problems across the entire metropolitan area. One clear implication of these findings for metropolitan planning strategy in Dhaka is that the major job clusters can act as potential locations for a well-planned polycentric pattern of metropolitan growth - where jobs will be closer to where workers live and closer to villages and settlements on the outskirts of the metropolitan area. If these centres could be supported by adequate infrastructure, especially public transport, it is likely that urban and transport sustainability may increase over time.

Chapter 5: Spatial Structure of Employment in Dhaka 166

Chapter 6 Urban Spatial Structure of Dhaka Mega city

6.1 Introduction The previous chapter explores the spatial pattern of employment distribution in Dhaka megacity and identified 19 employment centres (Figure 6.1). This chapter investigates the spatial interaction of those employment centres. The resulting job- clusters are analysed in term of their travel impacts, building footprints and usage in order to explore the extent of spatial interaction with the area.

6.3 Employment Centres and Commuting Behaviour The employment landscape reveals job spreading all over the mega city. However, the identified centres contain 5.2 million jobs which are nearly 54% of total jobs. The main city has about 3.28 million jobs in its 12 centres located within a 154 sq. km2 area. Most of these jobs are on mixed landuse zones with different types of economic activities. The inner ring, on the other hand, has 1.9 million jobs within a 1400 sq. Km area. Jobs in the inner ring are relatively low in density and dispersed. Location of several garments, pharmaceutical and manufacturing industries in Savar and Gazipur could be the reason of such pattern of employment distribution in these areas.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 167

Figure 6.1: Employment Centres in the Dhaka Metropolitan Area

The workforce concentration has been shown to be characterised by differences in their spatial locations (Chapter 5, Section 5.3). People employed in tertiary sector are mostly located in the main city without much occupational or specialization clustering. An analysis on their commuting distance also reveals the same. There are as many as 12 job clusters in the Main city with 3.28 million jobs for which people commutes on average, less than 5 km (Table 6.1). The commuting distances are higher for the inner ring and outer ring workers. The desire lines of commuting patterns to these identified centres present respective labour catchment areas and their influence on spatial structure (Appendix B). Figure 6.2 shows average commuting distance and maps in Appendix B (Figure B1, B2 and B3) show the travel patterns of six representative zones from the selected employment centres. Motijheel and Gulshan attracts employees from all parts of the city but the major catchment concentration is within the Main city, with an average commuting

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 168

distance of less than 8 km. Savar, while located outside the Main city and within the inner ring, attracts a large number of trips from its surrounding areas. This is due to the presence of large numbers of garment industries in Savar and people from neighbouring small settlements/villages commute to Savar for their daily job. Narshindi shows that only relatively few people outside the municipality commute here, and most are from neighbouring villages.

Table 6.1: Commuting Pattern to Job Centres Vehicle Kilometre Average Trip Region No of EC Job Travelled Length (Km.) Main City 12 3,282,837 16,119,749 4.91 Inner Ring 4 1,881,670 19,625,495 10.43 Outer Ring 3 117,976 1,214,481 10.29 Total 19 5,282,483 36,959,725 7.00 Source: Processed by author

Three different commuting patterns are recognized in three regions: main city, inner ring and outer ring. These differences are related to their job types, urban form and land uses. Old Dhaka attracts a higher number of shorter trips. Perhaps mixed landuse around the old part has a relation with it. At the same time, longer trips are experienced from all over the Dhaka to Gulshan. This is because Gulshan has become a preferable place for corporate offices, multi-national companies and foreign companies. The suburb itself is very expensive and as a result, many people cannot afford to live at a closer distance. Job clusters in the inner ring such as Savar, Gazipur and Tongi highlight different patterns. Being industrial in their job profile means they have extended labour catchment areas. It is evident from their desire line diagram (Appendix B) that people from the rural settlements in the outer ring are commuting to these job clusters in the inner ring.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 169

Average Commuting Distance

45 40 35 30 25 Inner Ring Outer Ring 20 15 MainCity Distance (Km.) Distance 10 5 0 Savar Tongi Badda… Kafrul… Mirpur Ramna Gazipur Tejgaon Gulshan Mohamm… Motijheel Narshindi Hazaribag Dhamondi Old Dhaka Manikganj Munshiganj Cantonment Narayanganj Figure 6.2: Average Commuting Distance to Job Clusters Source: Processed by author

Most of the employment centres attract employee within 10 km distance (Figure 6.2). However, centres outside the main city require longer trips. Manikganj has 29 thousand jobs and most of the employees are coming from surrounding smaller urban centres (Thana) or rural areas. Though very few in number, there are trips from inner ring centres like Narayanganj, Joydebpur (Figure 6.3)

Figure 6.3: Commuting Patterns to Manikganj Employment Centres Source: Processed by author

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 170

Narayanganj shows a distinctively higher commuting distance (Figure 6.3 and Figure 6.4). This area is characterized by industries where workers are coming from distant rural areas. Narayanganj is located on the south-east corner of the Dhaka region. Its labour catchment may be extended to the Chittagong region. Since this study considers only commuters from the Dhaka region, this may have led to an increase in the distance figure.

Figure 6.4: Commuting Patterns to Narayanganj Employment Centres Source: Processed by author

The plotting of preference functions provides further opportunity to analyse the commuting pattern of identified employment centres. The preference function of the commutes to the job clusters in the Main city shows a general trend of having 5% to 50% commuting from a closer distance and the rest from different distances (Figure 6.5). It identifies a few clusters for which employees come from adjacent zones. Hazaribagh, Motijheel and Cantonment are examples. Dhanmondi, on the other hands shows 30% local employees with the rest commuting a relatively long distance. Mohammadpur has a similar pattern.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 171

Plot of Employment Location Specific Preference Function for commuting to the Main City 1.0 OldZone Dhaka 34 0.9 KafrulZone and 43 Pallabi 0.8 BaddaZone &Khilgaon 9 0.7 MohammadpurZone 19 Gulshan 0.6 Zone 8 MirpurZone 49 0.5 TejgaonZone 58 0.4 CantonmentZone 50 0.3 RamnaZone 27 0.2 MotijheelZone 15 Dhamondi Y[Cum. Pro. ofWorkforce Captured] 0.1 Zone 24 HazaribagZone 30 0.0 0 0.2 0.4 0.6 0.8 1

X [Cumulative Proportion of Workforce Reached]

Figure 6.5: Commuting Preference Functions for Identified Centres in the Main City Source: Processed by author

The preference function plot shows a different pattern for the commute to inner ring clusters (Figure 6.6). Gazipur and Savar have up to 80% of their employees from surrounding localities. But Narayanganj has a different pattern with very low local commutes. The reason for this would be good accessibility from distance areas, shortage in local workforce and location of backward and forward industries. These characteristics reveal the fact that only zone location (whether it is inner or outer zone) of a centre does not necessarily define its commuting characteristics. Rather these are largely dependent on the nature of the job, skill level of the local labour force and transport connectivity.

The preference function for the employment centres in the outer ring shows very distinctive characteristics (Figure 6.7). On an average 80% of the employees are local or commuting from short distances. There are long trips associated with each of the outer ring employment centres. But they are very little in number. All three job centres in the outer ring are district headquarter. They have some administrative job,

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 172

local business and may have some light manufacturing. Employees of these jobs are local residents.

Plot of Employment Location Specific Preference Function for Zones in the Inner Ring

1.0 0.9 SavarZone 94 0.8 0.7 0.6 GazipurZone 99 0.5

0.4 NarayanganjZone 125 0.3 0.2 TongiZone 101

Y [Cum. Y [Cum. Pro. ofWorkforce Captured] 0.1 0.0 0 0.2 0.4 0.6 0.8 1 X [Cumulative Proportion of Workforce Reached] Figure 6.6: Preference Function for Commuting to Identified Centres in the Inner Ring Source: Processed by author

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 173

Plot of Employment Location Specific Preference Function Employment Centres in the Outer Ring 1.0

0.9

0.8

0.7 ZoneManikganj 113 0.6 ZoneMunshiganj 117 0.5 ZoneNarshindi 137 0.4

0.3

0.2 Y [Cum. Y [Cum. Pro. ofWorkforce Captured] 0.1

0.0 0 0.2 0.4 0.6 0.8 1 X [Cumulative Proportion of Workforce Reached]

Figure 6.7: Preference Function for Commuting to Identified Centres in the Outer Ring Source: Processed by author

The commuting pattern (Appendix B) in combination with preference functions (Figure 6.5, 6.6 and 6.7) provide an opportunity to understand the movement of people for their everyday work. It creates the commuting profile of an area which, in combination with landuse and type of job, can be an important input in structuring transport infrastructure and policy formulation for a sustainable city. These, in turn identify the demand for likely transport solutions, its required coverage and optimum routing. At the same time, these may help in formulating policy, strategy and planning for accommodating urban growth.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 174

6.4 Features and Spatial Impacts of Identified Employment Centres

Job clusters have a higher concentration of jobs and as a result, it attracts more traffic than other areas. Accessibility and connectivity of employment zones are important issues and have enormous impact in planning. For Dhaka, the job clusters of different regions shows different travel characteristics, which is associated with the nature of employment and the types of business located there. An investigation on the land and buildings on which jobs are located can help in explaining further the likely impacts on planning.

The employment zones used in this research are of varying size depending on their job types and location. Job clusters in the Main city are small in area but have a higher density of floor space. On the other side, job clusters outside the Main city occupy more land in comparison to its floor space. Savar has 885 hectares of floor space on 25717 hectares of land but Old Dhaka which is comprised of several small zones has a total area of nearly 1400 hectares of floor space on 1950 hectares of land. The competition for land in the Main city pushed the space consuming primary sector to the outskirts of the city which is limited by the availability of flood free land and accessibility to the city.

The building footprints (Appendix B) give a sense about the density of buildings in Dhaka. It is noticeable that open space is minimal in most of these employment centres in Dhaka and mixed use buildings are located along the main roads. Building clusters of Dhanmondi, for example, shows that there is a co-location tendency among non-residential buildings (Figure B.7, Appendix B). Buildings for private universities, clinics, shopping centre tend to cluster around the intersection of main roads or along the major roads. The statistical association between jobs and work related floor space is very high (R2=0.92) when fitted with a 4th order polynomial curve, indicating the existence of several clusters in term of the nature of floor space consumption by the job clusters (Figure 6.8). From Figure 6.8 it can be observed that old Dhaka remains an outlier, the opposite of other job centres in the Main city, though their job-floor space ratio would be linear. Savar and Gazipur again fall in to

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 175

the same cluster. Gulshan and other planned areas show another type in their floor space consumption per job. A detailed look at their floor space occupancy against jobs, residential use and floor area ratio (FAR) would provide a give better understanding about the nature of jobs in different clusters.

600

Old Dhaka 500

400 R² = 0.9216 300

Savar 200 Narayanganj Gulshan Work AreaFloor Related (Ha) Gazipur 100

Cantonment 0 0 200 400 600 800 1000 1200 Jobs ('000)

Figure 6.8: Floor Area of the Identified Job Centres Source: Processed by author

The Main city has many tertiary sector jobs and is characterised by mixed landuse, compact development and short commute trips. The analysis of building footprint and usage supports these findings. In general, in the Main city, job-clusters represent a smaller geographic area but the floor area ratios (FAR) are higher in comparison to job clusters in the outskirts (Tongi, Narayangang, Gazipur, Savar) (Figure 6.9). Dhanmondi (1.23), Motijheel (0.996) and Ramna (0.903) have very high FAR as these areas are characterised by dense development and the presence of many tall commercial buildings. The total floor area is very high in Old Dhaka but its FAR is low as the older part of Dhaka carries the traditional old low rise buildings that form very narrow alleys. Hazaribagh has a low floor area ratio as the area has been developed around the vast Martyr's Column - remembering the Martyrdom of the

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 176

independence in 1971. Interestingly the Cantonment has plenty of open space and medium size buildings (up to 4 stories) resulting a lower FAR value.

Floor Area Ratio (FAR) 0 0.2 0.4 0.6 0.8 1 1.2 1.4

Tongi Narayanganj Total_FA(Ha) Gazipur FAR Savar Hazaribag Dhamondi Motijheel Ramna Cantonment Tejgaon Mirpur Gulshan Mohammadpur Badda and Khilgaon Kafrul and Pallabi Old Dhaka

0 500 1000 1500 Total Floor Area (FA) in Ha.

Figure 6.9: Total Floor Area in Identified Employment Clusters in Dhaka Source: Processed by author

75% of the total floor area is used for residential purposes. However, an area with lower residential floor areas indicates higher rates of job related floor space. Narayanganj and Tejgaon have lower residential space (about 50%). Job-clusters in the Main city maintain both higher residential floor space and higher floor area, indicating high density mixed use urban form (Figure 6.10).

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 177

Residential Floor Area 100% 80% 60% 40% 20% 0%

Tongi Narayanganj Gazipur Savar Hazaribag Dhamondi Motijheel Ramna Cantonment Tejgaon Mirpur Gulshan Mohammadpur Badda and Khilgaon Kafrul and Pallabi Old Dhaka GrossJobDensity 0 10000 20000 30000 40000 50000 60000 Gross Job Density Residential Figure 6.10: Floor Area Ratio of Identified Job Clusters Source: Processed by author The investigation on travel pattern, floor area and landuse of job clusters suggest that there are at least three types of job clusters in Dhaka. First job clusters in the outer ring are small job centres with smaller catchments. Most of the workers are local employees or from villages in very close proximity to these clusters. The job category includes the informal sector, small business and very few formal jobs. Second, job clusters in the inner ring are characterized by primary sector jobs. The growth of ready-made garments exported in the last twenty years has established many linked industries in the inner ring. They occupy large land amounts of but do not require higher FAR. The main labour catchments of these industries are rural areas. However, for management and business purposes many people travel from the city. Third, the inner ring has higher FAR values and higher job density. Several job clusters in the Main city have higher residential floor space (Dhanmondi- 84%; Mirpur 88%).

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 178

6.5 Residential Workforce Profile of the Employment Centres Residential workforce residing in different employment centres may help to understand their local employment opportunities and commuting pattern. The residential workforce distribution in Figure 6.11 shows that inhabitants occupied in the household and other activities are prominent in all areas. Opportunities for jobs are limited and a large portion of female workforce is only engaged in household works. The category ‘others’ include informal sector workers who work in the small business, shopkeepers, street vendors and other jobs that does not require any formal training. For the industrial worker one can notice a few clustering. Followed by old Dhaka, Kafrul and Badda have industrial workers. Old Dhaka has light manufacturing, printing and metal fabrication jobs. These employees are local in many cases considering the mixed landuse pattern in the Old Dhaka. However, the nature and skills of business in Old Dhaka is unique in draws traffic from all corners of the megacity (Figure B.1, Appendix B). Kafrul, Pallabi, Kilgaon and Badda on the hand are the key locations for early day’s garments industry and their linkage industries. Most of these industries are relatively old and grew before the expansion of Dhaka megacity. It is noticeable that these industries attract workers from distant rural hinterlands (Figure B.2, Appendix B). Other clusters of industrial workers are found in Savar, Gazipur and Tongi, all of which are recognized locations for manufacturing and industries. The manufacturing and industrial growths in these areas have encouraged the in-fill development by the workforce of the relevant industries and perform shorter trips for commuting (Figure B.1, Appendix B). The Outer ring centres show higher unemployment rate and higher population in household and agriculture sector. As results, these centres attract local trips (Figure 6.3 and Figure B.2 in Appendix B).

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 179

Manikganj

Narshindi

Munshiganj

Tongi

Narayanganj

Gazipur

Savar Other Hazaribag Service Dhamondi Business Motijheel Hotel&Restaurant

Ramna Transportation

Cantonment Construction Utilities Tejgaon Industry Mirpur Agriculture Gulshan HouseHold Mohammadpur

Badda and Khilgaon

Kafrul and Pallabi

Old Dhaka

0 50 100 150 200 250 Workforce Population (in thousands)

Figure 6.11: Work Force Population in Different Employment Centres Source: Processed by author

6.6 Conclusion

The focus of this chapter has been on the spatial structure of economic activities within the Dhaka Mega city and its impact on the travel behaviour of resident workers. Specifically, the chapter has examined questions of how the location and dispersal of employment and growth of business centres, affects the characteristics of travel and transport sustainability and identifies the key issues for policy makers to address in attempting to move to a more sustainable environment. An overview of

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 180

the dynamics of employment concentration in the Dhaka metropolitan area is presented here by identifying major employment centres. A modified clustering method has been used on data for 2005 to identify 19 employment centres that play a significant role in accommodating the majority (54%) of employment of the metropolitan area. Gross job density and the job-population ratio measures and their spatial characteristics were used as the basis to identify significant employment centres. It can be concluded that the Dhaka mega city has a spatial form that is not polycentric but, rather, one that can be called an ‘extended monocentric’ structure. The role of the main city and some parts of the inner areas are paramount with the highest levels of employment density concentration. Some minor centres in the outskirts of the metropolitan area do not play a regional role in terms of employment concentration but have local significance and the potential to be major centres in the future with the right policy signals.

The methodology adopted here was carefully designed to ensure that employment cluster of any significance should be left out. It was found that Dhaka is characterised by an extended monocentric pattern of spatial structure, unlike many other mega cities in Asia that appear to be developing polycentric characteristics. However, some locations, especially those towards the north and north-west parts of the metropolitan area, have a high potential to become major employment centres in the future. There is no doubt that from the first attempt at assessing the travel data that past ad-hoc metropolitan policies of decentralisation and centre formation have not proved successful (see chapter 4). The lack of adequate road and public transport based infrastructure in these inner and outer areas has undoubtedly contributed to persistence of the extended monocentric pattern and, as a consequence, to severe sustainability problems across the entire metropolitan area. One clear implication of these findings for metropolitan planning strategy in Dhaka is that the major job clusters can act as potential locations for a well-planned polycentric pattern of metropolitan growth - where jobs will be closer to where workers live and closer to villages and settlements on the outskirts of the metropolitan area. If these centres could be supported by adequate infrastructure, especially public transport, it is likely that urban and transport sustainability may increase over time.

Chapter 6: Spatial Interaction of Employment Centres in Dhaka 181 Chapter 7 Exploring Growth Options for a Sustainable Dhaka

7.1 Introduction

This chapter extends enquiries on the sustainable spatial structure of Dhaka mega city by investigating different growth options in terms of their job distribution, journey to work patterns and transportation emissions. The focus of this chapter is on addressing the third research question in regard to investigating the impact of the spatial structure of employment in Dhaka and addresses the impacts of different growth options for urban form and transportation. The chapter begins by investigating the degree of spatial interaction in terms of journey to work trips across different regions of Dhaka. Then, the journey to work dimensions of travel distance and travel time are compared across the employment clusters identified in Chapter 6. Investigations of the impacts of the spatial structure of employment on travel are continued with the application of travel models that aim to explore how the job distribution affects travel decisions. The chapter also discusses the vehicle emissions in different growth scenarios. The results are discussed in the context of transport sustainability and aim to put forward possible policies that can promote sustainable urban habitat aided by better transport conditions for the city.

Chapter 7: Future growth Options for Sustainable Dhaka 182 Chapter 5 presents the spatial distribution of employment in Dhaka megacity and identifies 19 employment centres including 16 established employment hubs and 3 potential centres in the outer ring of Dhaka. Chapter 6 provides a detailed look of these centres and explores the commuting travel patterns of these centres and provides an assessment of their travel behaviour and interaction with other areas of the megacity. This chapter presents 10 growth scenarios for Dhaka for the year 2024 by assigning different numbers of jobs to these 19 identified centres. In addition to different number of jobs, two different urban growth assumptions are made which have been discussed in Chapter 4. This chapter also presents a spatial analysis of these growth scenarios which include the employment landscape, travel behaviour and commuting vehicle emissions under different scenarios. This analysis finds that decentralized employment growth would minimize overall commuting and enhance a more sustainable growth outcome for the megacity.

7.2 Modelling Future Growth Scenarios for Dhaka

This study indicates that more than half of the total jobs are located within the main city (Figure 7.1). The selected 19 centres have more than 60% of jobs and 37% of total jobs are located within the 12 centres in the main city. These figures illustrate the influence of the main city in the economy of Dhaka.

11.76% 1.40%

Main City 25.78% 51.33% 18.19%

Extended City 37.37% Inner Ring 1.15% 11.13% Outer Ring Jobs in Different Zones of Dhaka Jobs in Employment Centres

Figure 7.1: Jobs in Different Regions and in Job Centres Source: processed by author (based on, The Louis Berger Group and Bangladesh Consultants Ltd, 2005)

Chapter 7: Future growth Options for Sustainable Dhaka 183 It is suggested that rather than dispersed growth, the spatial distribution of population and employment needs to be concentrated in current and potential centres (Grazi et al., 2008). Different scenarios are built on this assumption. For sustainable urban structure, the scenario based analysis examines their performances (travel cost, vehicle emission etc.). It is assumed that nearly 15% of employment growth would occur from 2004 to 2024, as estimated in STPP. Scenarios A to E are created considering 50% to 0% growth in the city’s employment figures and 25% to 75% growth in the outer ring (Table 7.1).

Table 7.1: CO2 Emission under Different Scenarios

Distribution of Employment Growth Under Different Scenarios (%)

Regions A B C D E Main and Extended 50% 40% 30% 25% 0% City Inner Ring 25% 35% 35% 25% 25% Outer Ring 25% 25% 35% 50% 75% Source: Author

Table 7.2: Distribution of Employment Growth under Different Scenarios

Distribution of Employment Growth Under Different Scenarios (%) Regions A B C D E

Main and Extended 701,857 561,486 421,114 350,929 0 City

Inner Ring 350,929 491,300 491,300 350,929 350,929

Outer Ring 350,929 350,929 491,300 701,857 1,052,786 Source: Author

According to estimates by STPP, another 1.4 million jobs would be added to Dhaka’s economy by 2024. The model used in the present study distributes these new jobs to the 19 centres. Under different scenarios, the main city and extended regions are expected to have between 0 and 0.7 million new jobs up to 2024 (Table 7.2). On the other hand, the inner ring and outer ring are expected to accommodate

Chapter 7: Future growth Options for Sustainable Dhaka 184 between 0.35 and 0.49 million more jobs and more than 1 million jobs respectively (Table 7.2). By emphasizing growth in the inner and outer ring, this study investigates the potential impacts of job decentralization.

However, total job numbers in different regions of the city under different growth scenarios suggests that even if no growth occurs in the main city, distribution of more than a million jobs (75%) in the outer ring would not reduce the prominence of employment centres in the main city over the next 20 years. But it would help to crystalize employment centres at distances from the main city and finally decentralization would occur (Figure 7.2). It is noticeable that in scenario-E, which assumes 0% growth in the main city, the total number of jobs would be nearly 5 million in the main city by 2024.

7 Main City

6 Extended City 5 Inner Ring 4 Outer Ring 3 (Millions)

Numberof Jobs 2

1

0 Baseline Scenario A Scenario B Scenario C Scenario D Scenario E

Figure 7.2: Total Employment under Different Scenarios Source: Author

7.3 New Employment Landscape

The baseline employment scenario reveals a continuing employment growth in the main city which leads to an increase in congestion and hence pollution from transport. The employment allocation under different scenarios reveals that as little as 25% growths in the outer ring may create high employment centres in those areas (Figure B.1, Appendix C). However, in reality, growth cannot be contained within a

Chapter 7: Future growth Options for Sustainable Dhaka 185 certain zone. There would be a spill-over of jobs and surrounding zones would be affected. Line-2 scenario is based on the assumption that for each employment centre, adjacent zones would be affected by the centres when they reach a certain stage. Employment landscape maps are generated from the results (Figures in Appendix C), which indicate that the polycentric nature of growth can be attained for Dhaka megacity by allocating a higher proportion of employment to the outer ring. The employment density maps at the same time convey the spatial pattern of job concentrations (Figure 7.4). These show that jobs in the main city are less space occupying (higher density). On the other hand, employment centres in the inner ring and outer ring offer more space per job (low density). Perhaps some space occupying industry categories can be allocated to these centres. However, appropriate infrastructure development and relevant policy initiatives must be met.

2024 Baseline 2024 Scenario E

Figure 7.3: Employment to Population Ratio Source: Author

Figure 7.4 and Figure 7.5 indicate how decentralization of jobs from the main city to the inner ring and outer ring could alter the job landscape of the megacity. In the ‘Do

Chapter 7: Future growth Options for Sustainable Dhaka 186 Nothing’ baseline approach there would be more area with very high job concentration. Decentralization reduces these pressures from the main city and distributes jobs to other zones which have better proximity to employees. Higher order decentralization (scenarios E and 2E) creates mid-range job density (50 to 60 thousand jobs per km2) centres (Figure 7.4 and Figure 7.5) and would assist to crystalize a polycentric growth pattern.

60

Baseline 50 Scenario-A

40 Scenario-B Scenario-C 30 Scenario-D

Percentage of Employment of Percentage 20 Scenario-E

10

0

Job Density (per km2)

Figure 7.4: Job Density Histogram for Scenarios of line 1

Chapter 7: Future growth Options for Sustainable Dhaka 187 60

Baseline 50 Scenario-2A

40 Scenario-2B Scenario-2C 30 Scenario-2D Percentage of Employment of Percentage 20 Scenario-2E

10

0

Job Density (per km2)

Figure 7.5: Job Density Histogram for Scenarios of line 2

7.5 Commuting Patterns

Analysis of commuting desire lines indicates that there is a significant impact on the spatial pattern of travel under different scenarios (Figure B.3, Appendix C). In the baseline scenario, numerous trips are coming to the main city from all jurisdictions of the megacity (Figure 7.6). Important centres like Savar, Tongi and Gazipur appear to have specific labour catchment areas but centres in the main city attract employees from all over the megacity. These patterns also differentiate the nature of jobs in industry based centres located in Savar, Tongi, Narayanganj and Gazipur from tertiary sector employment in the main city.

The allocation of employment to the centres enables them to act as hubs of commuting destinations. Commuting maps of several scenarios depict the formation of centres. For example, in scenario-E, Manikganj, Munshiganj and Narshindi attract large numbers of commuters from their adjacent zones respectively (Figure 7.6).

Chapter 7: Future growth Options for Sustainable Dhaka 188

Figure 7.6: Commuting Pattern in 2024

Chapter 7: Future growth Options for Sustainable Dhaka 189 The distances that commuters travel do not indicate much sensitivity to job allocation in several of the scenarios. For all scenarios, the analysis indicates that more than 40% of workers commute a maximum of 10 km. But at the same time, more than 10% of workers would need to travel 30km or more to go to their jobs (Figure 7.8). However, total passenger kilometres travelled (PKT) shows variations under different scenarios. It shows that under different scenarios there would be a decrease in PKT for the main city and an increase for jobs in the inner ring and outer ring (Figure 7.7).

30% Baseline A 25% B C 20% D 15% E

10%

5%

0% 0 - 5 5 - 10 10 - 15 15 - 20 20 -25 25 - 30 30+ Commuting Distance (in km.)

Figure 7.7: Variation in Average Commuting Distance under Different Scenarios Source: Author

180 160 140 120 100 Outer Ring 80 Inner Ring 60 40 Extended City

Passenger Kilometer (in (in millions) Kilometer Passenger 20 Main City 0

Scenarios

Figure 7.8: Total Passenger Kilometres under Different Scenarios Source: Author

Chapter 7: Future growth Options for Sustainable Dhaka 190 It is interesting to observe that total PKT is nearly indifferent across the different scenarios. The main reason for this due to job decentralization in the distant centres in the outer ring, more long distance trips have been generated replacing the shorter distance trips in the main city. The travel patterns also supports this findings as it highlights that more people commute to distant centres from the main city as are increases the number of jobs in the outer ring.

7.4 Travel Time and Cost

Results from the different scenario analysis (discussed in chapter 4) indicate that there would be large increases in travel distance unless major job re-allocation is performed (Table 7.3 and Table 7.4). Under the current trend, the main city and its surroundings will keep on attracting more jobs and agglomeration would continue. The city is likely to increase in all possible directions and commuting distance and time would increase. As a result, the congestion effects would increase and travel times would increase significantly as well. Under baseline scenario, people living in the outer zone would depend on the main city at an increasing rate. However, with the increasing allocation of jobs in the outer ring, the travel distance comes down significantly. With the proper allocation of jobs, more people would find jobs in relatively close proximity of their homes and may not need to travel long distances.

Table 7.3: Travel Distance at Different Development Scenarios for 2024 (Line 1) Average Commuting Distance in Km ( Scenario - line 1) Regions Baseline A B C D E

Main City 18.21 12.30 12.24 12.10 11.95 11.76 Extended City 20.61 15.27 15.21 15.06 14.90 14.70

Inner Ring 22.21 16.76 16.96 16.87 16.54 15.80 Outer Ring 45.93 29.50 29.56 28.79 27.88 26.85

Tables 7.3 and Table 7.4 indicate that a decrease in travel times there would occur as the outward growth of employment increases. The distribution of jobs would have significant impacts, especially for workers commuting to the outer region. Since the decentralization of jobs would likely to bring more workers to the outer region

Chapter 7: Future growth Options for Sustainable Dhaka 191 (region 4) for their job, the reduced travel distance for employees in region 4 would most likely enhance its attractiveness to employees.

Table 7.4: Travel Distance at Different Development Scenarios for 2024 (Line 2) Average Commuting Travel Distance in Km ( Scenario - line 2)

Region 2A 2B 2C 2D 2E Main City 12.30 12.24 12.10 11.95 11.66 Extended City 15.26 15.21 15.06 14.90 14.58

Inner Ring 16.92 17.17 17.07 16.70 16.54 Outer Ring 29.34 29.15 28.63 28.21 27.69

Commuting patterns to Munshiganj, under different scenarios, have been shown as an example (Appendix C). The commuting diagram (Appendix C) shows that with decentralization of employment, Munshiganj creates more local commuting, mainly from surrounding areas. So the labour catchment area is localized. This provides a potential in term of sustainability, as more people are likely to find a job in their locality. The reduced pressure benefits the transport network towards the main city.

7.6 Transport Vehicle Emissions

Road transport is widely recognized as a significant and increasing source of air pollution world wide (Colvile, et. al, 2001). Motorization in the developing countries are growing fast but still it is relatively low (ADB, 2003). However, several factors make vehicle emissions more harmful than in the developed world. These include fuel quality, age and condifition of vehicles, fuel inefficient vehicles (Bose, 2007) and insufficient roads and road geometry that creates excessive congestion (Bose, 2007). The most significant transport emissions to the atmosphere by mass are carbon dioxide (CO2) and water vapour (H2O) from the complete combustion of fuel.

Other pollutants are: nitrogen oxides (NOx), Sulphus dioxide (SO2), carbon monoxide (CO), Methane (CH4) and particulate emissions.

Vechicle emissions are estimated in the different scenarios modelled in this study. As the scenarios do not assume any changes in travel behaviour or modal choice, or

Chapter 7: Future growth Options for Sustainable Dhaka 192 any other policy implication, there are no significant differences in PKT under the different scenarios. The same is reflected in total vehicle emissions. However, the results (Tale 7.5) highlight that scenario-2C and scenario-2E would produce less emissions and that scenario-D and scenario-E produce emissions at a higher level. This is likely to happen due to the slightly low or high PKT levels in the differerent scenarios.

Table 7.5: Changes in Vehicle Emissions under Different Scenarios For 2024 Total Emissions (% of Scenarios change) Baseline 0.00 A 0.01 B 0.01 C 0.01 D 0.03 E 0.03 2A 0.00 2B 0.00 2C -0.01 2D 0.00 2E -0.01

Chapter 7: Future growth Options for Sustainable Dhaka 193 The ‘Do-nothing’ scenario indicates that CO2 vehcile emissions in the main city of Dhaka megacity would be more than 2000 tonnes per day, followed by 1800 tonnes per day in the inner ring (Figure 7.9). However, there is an impact on the spatial distribution of emissions. As different scenarios distribute jobs and, hence, traffic in different proportion, the result would be a significant decrease in pollution in the main city (Figure 7.10). However, the inner ring and outer ring would have higher rates of reduced pollution.

Main City 2,500 2,000 1,500 1,000 500 Outer Ring 0 Extended City

Inner Ring Unit : tonnes/day

Figure 7.9: CO2 Emissions in Different Regions of the Megacity (Baseline) Source: Author

Chapter 7: Future growth Options for Sustainable Dhaka 194

2,500

2,000

1,500

1,000 Emission ( Tonnes/day) ( Emission 2 500 CO

0

Scenarios

Figure 7.10: CO2 Emissions in the Main City under Different Scenarios Source: Author

Vehicle emissions across different regions highlight variations under different scenarios. Figure 7.10 indicates that job decentralization at a different rate would redistribute vehicle emissions. The gradual decrease in vehicle emission in the main city with the increasing employment decentralization to the outer ring indicates a direct relationship. Figure 7.11 illustrates the CO2 vehicle emissions under different scenarios across all regions of the city. It highlights clearly that CO2 emissions in the main city would be reduced by 30% simply by redistributing employment growth especially in scenario-2E. It also confirms the general findings in the literature that emissions are a function of activities (Schipper et al., 2009). If economic activities are spatially distributed, then vehicle emissions are likely to be distributed and, hence, their concentration can be minimized.

Chapter 7: Future growth Options for Sustainable Dhaka 195 Main City Main City 3,000 3,000 2,000 2,000 1,000 1,000 Outer Ring 0 Extended City Outer Ring 0 Extended City

Inner Ring Inner Ring A 2A

Main City Main City 3,000 3,000 2,000 2,000 1,000 1,000 Outer Ring 0 Extended City Outer Ring 0 Extended City

Inner Ring Inner Ring B 2B

Main City Main City 3,000 3,000 2,000 2,000 1,000 1,000 Outer Ring 0 Extended City Outer Ring 0 Extended City

Inner Ring Inner Ring C 2C

Main City Main City 3,000 3,000 2,000 2,000 1,000 1,000 Outer Ring 0 Extended City Outer Ring 0 Extended City

Inner Ring Inner Ring 2D D

Main City Main City 3,000 3,000 2,000 2,000 1,000 1,000 Outer Ring 0 Extended City Outer Ring 0 Extended City

Inner Ring Inner Ring E 2E

Figure 7.11: CO2 Emission (in metric tonnes) under Different Scenarios

Chapter 7: Future growth Options for Sustainable Dhaka 196 7.7 Conclusion

The results of the analysis presented in chapter 7 highlights that concentrating jobs growth solely on the main city would aggravate the transportation situation and the environment by increasing commuting distance, time and congestion. This research suggests an analytical framework that could generate different scenarios of transportation based on given criteria for employment growth. As a case study, 10 growth scenarios have been generated for 2024, based on different scales of employment decentralization and commuting. Key findings of the analysis are:

- The Dhaka main city has a lion share of jobs concentrated in its employment centres namely Old Dhaka, Gulshan, Dhanmondi, Motijheel etc. Currently the main city has 51% total megacity employments and its 12 employment centres have 37%. The influence of the main city is so strong that even without any additional jobs in the main city by 2024; the city would still keep its primacy.

- Employment decentralization would eventually reduce the average commuting distance. As the jobs decentralize, employees from distant locations would prefer jobs within shorter distance. However, in higher decentralized scenarios (for example scenario E) there would be reverse- commuting (people from the main city would travel to employment centres in the outer ring). This is because, employment growth in the main city is restricted in those scenarios and growing city dwellers need to travel to the outer ring centres for their job. Aguilera Mignot (2004) witnessed the similar longer commuting distance for the outlying sub centres in French cities.

- With the increasing employment decentralization, potential employment centres outside the main city would find a central position for commuters (Figure 7.6) and would act as regional transport hub.

- Travel analysis suggested no significant difference in total vehicle kilometre under different scenarios but their spatial distribution would vary under

Chapter 7: Future growth Options for Sustainable Dhaka 197 different scenarios. Decentralized employment would induce travelling towards potential employment centres in the outer ring.

- The study provide evidence that employment decentralization reduces both the shorter (up to 5km) and longer trips (30+ km.); and increases medium size trips.

- Vehicle emission is proportional to vehicle kilometre travelled. As decentralization in employment growth increases longer trips, number of motorized commuting would increase with employment decentralization. However, though total commuting vehicle emission would increase but it would be spatially segregated. As a result, their impact on air quality would improve. In other words, polycentric employment clusters could work well as an urban development strategy to decentralize jobs, minimize transportation costs (time, distance), pollution and strengthen sustainability.

Several studies have suggested that a highly concentrated metropolitan employment is not as desirable a spatial structure as one based on strong suburban or sub-regional centres. Polycentric urbanization is widely regarded as successfully reducing traffic congestion (Lin et al, 2012). Polycentric urban growth can redirect traffic away from central areas to outlying employment centres. As employment opportunities move to suburbs or outer areas, the labour force tends to follow, which allows many employees to enjoy reduced commuting time and less traffic congestion in the main city (Gordon, Kumar & Richardson, 1989; Giuliano & Small, 1993; Gordon & Richardson, 1996).

The modelling undertaken in this study altered the distribution of employment in different employment centres without changing the transport network and population distribution. Investment in infrastructure planning and development to connect workers and employment in the inner ring and outer ring would reduce commuting length significantly. At the same time, employment allocation to the currently unattractive centres would make them attractive place to live for many. Eventually

Chapter 7: Future growth Options for Sustainable Dhaka 198 there would be demographic movement in the whole megacity which could enhance centres in the outskirts and could reduce longer commuting. These “what if ..” scenarios need to be discussed in light of policy debates and explored for the potential of sustainable urban development in Dhaka megacity.

Chapter 7: Future growth Options for Sustainable Dhaka 199

Chapter 8 Policy Implications for Sustainable Urban Structure

8.1 Introduction This thesis investigates the urban spatial structure of Dhaka megacity and explores several scenarios of urban development for the future of the city. Chapter 5 has examined the spatial distribution of employment and chapter 6 investigated the clustering pattern of employment and identified the employment centres playing a significant role in the job landscape of the megacity. Chapter 7 proposed several employment growth scenarios for the year 2024 and examined their spatial interactions, travel behaviour and vehicle emissions from commuting. This chapter addresses the relevant policy involvement in the sustainable urbanization of Dhaka. As a continuum to previous chapters, this chapter focusses on the policy issues that various stakeholders need to address in order to achieve a more sustainable outcome for Dhaka megacity.

The empirical evidence presented in this thesis suggests that Dhaka is generally characterised as having a monocentric structure in term of its spatial pattern of employment. There are several new employment centres that have developed as a result of planning initiative but these are too close to the main city and have been engulfed by urban development from the main city to be considered as a polycentric structure. More specifically, the combination of these newer employment centres and

Chapter 8: Policy Implications for Sustainable Urban Structure 200 the main city has generated what can be defined as an extended monocentric spatial structure for the megacity. This type of urban structure in significantly different from that found in north American metropolitan cities and to some extent from the other Asian megacities. Brennan and Richarson (1987) claim that polycentricity is the pathway to sustainability for the Asian megacity. However, there is not much empirical evidence for this claim in the present urban structure of Dhaka and, unfortunately, Brennan and Richardson (1987) provided no guideline on how to move the Asian megacity towards the path of polycentricity. The question of what policies attentions are required to move Dhaka in the direction of polycentricity is the basis of this chapter.

In 1959 the Dhaka master plan first recognized the need for decentralized growth in the city. However, the plan did not spell out about the details of the growth and its spatial allocation but only provided an outline about the future growth of the city. The master plan also recommended strong decentralization at a national scale. Divisional headquarters and other large cities were recommended for job decentralization. Along the same lines, Dhaka Strategic Transport Plan (STP)(2005) also advocated for polycentric growth under the banner of “regional development scenario” (Chapter 3, Figure 3.13). But, their suggested centres are closely located (within 3 to 5 kilometer distance) from each other and close to the main city. As a result, these newer centres do not contribute to polycentric development; rather they were engulfed by a rapidly expanding main city and expands it.

For decentralization to occur and for polycentricity to develop centres should ideally be developed at some distance from the main city whereby they could function either as smaller versions of the CBD to serve the locality or be developed as specialized centres which would later attract more activities to make them more self-contained – a characteristic of polycentric centres in some metropolitan cities in the developed world (Parolin and Kamara, 2003). This type of spatial outcome would require coordinated and strategic planning and policy implementation.

Chapter 8: Policy Implications for Sustainable Urban Structure 201 8.2 Urban Development in Dhaka Megacity and Relevant Policy Issues Megacity dynamics are complex and to improve and maintain the standard of living and quality of life of the inhabitants requires sound policy measures encompassing all relevant sectors (Figure 8.1). This includes urban development, transportation, housing, economic plans and strategies and many more (Figure 8.1). It is now widely accepted that a relationship exists between the shape, size, density and uses of a city and its sustainability. However, there is no overall consensus about “the exact nature of this relationship” (William, Burton & Jenks 2000, p.1). Increasing attention is now being paid to the sustainability of all urban areas. Sustainable urban growth aims to minimize the environmental impact of urban development by reducing emission, congestion and other costs. Transport being a key element of urban systems, interactions between transport and urban form are intricately woven. In a sustainable urban landscape transport is expected to increasingly provide good outcomes for environmental stewardship, economic productivity and social well being – the three pillars of sustainable development (Doust, 2008). Sustainability cannot be achieved while sacrificing any one of them. It requires development along all sectors as improvements in one can be supported by the others. Tackling urbanization is the most challenging task for a megacity. Rapid urbanization affects urban planning, transportation, environmental management, demographic and economic plans. Failure to accommodate the growth in one sector may have impacts on the success of others. For example, failure in addressing an urban planning issue may impact negatively on transportation which may significantly reduce economic productivity, increase travel costs and vehicle emissions. This thesis argues that urban development for a megacity is a complex process and involves many factors including the dynamics of demography, environment, transportation and urban development. However each of these has an internal dynamism in themselves being regulated by contingent factors, action and policies. Figure 8.1 is presented to illustrate the connectivity.”

Chapter 8: Policy Implications for Sustainable Urban Structure 202

Figure 8.1: Policy Domain for Megacity Development Source: Author

For Bangladesh urban development policies and initiatives are often disjointed and non-compatible. There is no umbrella planning body governing the development or growth management of the megacity of Dhaka. Institutes taking care of the urban development, transportation, utilities have their own plans and often act on an adhoc project basis. These organizations seldom collect and share information, which is a recent development, and do not make any integrated plans. So it is an imperative to develop an urban governance structure that can coordinate a plan for the megacity of Dhaka.

8.2.1 Demographic and Economic Plan

There is a strong relationship between economic growth and the level of urbanization (Kawsar, 2012). Evidence is found for a bell shaped relationship between primacy and economic development (Junious, 1999). However, the developing countries in Asia are not achieving the optimum level of agglomeration benefits from urbanization due to the absence of proper public policy; conversely they are facing various challenges like traffic congestion, environmental degradation, growth of

Chapter 8: Policy Implications for Sustainable Urban Structure 203 urban slums and regional inequality (Junious, 1999). Although urbanization has been occurring very rapidly for the last three decades in Bangladesh, the country is not getting the optimum agglomeration and spill over benefits from this growth (Kawsar, 2012). In contrast, its large cities are facing various negative externalities like urban congestion, environmental pollution, unplanned urban growth and widening of income inequality. One of the underlying reasons for these negative trends is the lack of comprehensive urban development policies that could have incorporated sustainable strategies both in urban and regional development (Kawsar, 2012).

The National Population Policy of Bangladesh (Government of Bangladesh, 2008a) focusses on population growth and addresses the issues to curb its rapid rise. However, there is no indication about its desirable spatial distribution and no nation- wide economic plan exists for the population in coming years. STP (2005) suggested that there would be around 1.5 million additional workers in Dhaka by 2024. This is based on recent trends. As indicated in ‘Dacca master plan in 1959’ and in subsequent discussions, Bangladesh requires a policy guideline that would spell out the spatial distribution of jobs outside of Dhaka megacity and strengthens the economic base of other areas of Bangladesh (Minoprio and Spencely and P.W. Macfarlane, 1959).

The agricultural sector in Bangladesh is hrinking and surplus labourers are migrating to the large cities. Limited opportunities in the regional cities, in turn, have forced this huge labor force to migrate to the megacity capital. A nation-wide population distribution strategy coupled with economic plan would need to ensure sure that jobs are created at the local level and dependence on the capital city is reduced. This national scale strategy would trigger local urban development at regional levels.

8.2.2 Urban Development and Planning

Urban development and planning is a vital task to make a megacity functioning properly. Most megacities in developing countries have master plans that include some recommendations about the future directions of megacity growth. These master plans are often found to be unsuccessful in serving their purposes (Brennan and

Chapter 8: Policy Implications for Sustainable Urban Structure 204 Richardson, 1989). The reasons given are: erroneous population projection, inability of the planning body to control development, and often they are too rigid and inflexible and do not allow readjustment in the light of changing conditions (Brennan and Richardson, 1989)

The national urban strategy estimates the growth and develops strategies for its optimal spatial distribution. In Bangladesh, an urban strategy at national scale has been absent. Though there are estimates of urbanization based on natural growth, there is no spatial plan to accommodate this urbanization. Only recently, the National Urban Sector Policy for Bangladesh has been drafted with the aim of ensuring regionally balanced urbanization (Government of Bangladesh, 2011). The policy reveals and recognises the primacy of Dhaka, and other cities and towns are taken into focus for their infrastructure, housing and transportation development. However, the master plan for Dhaka prepared in 1959 also recommended decentralization of employment to other large cities (Minoprio, Spencely, Macfarlane, 1959). This research explores several employment growth scenarios and concluded that decentralized growth would give better urban development outcomes by improving travel time (cost), distance and emissions. This decentralization needs to be addressed both at a national scale and for the megacity of Dhaka.

- The Dhaka urban development master plan currently considers decentralization within the inner ring region. But this study provides evidence that for better sustainable outcomes, the outer ring should be brought under the Dhaka master plan area in order to facilitate important centres for further development.

- At the same time densification within main city should be strictly maintained. The population and job density have already reached very high levels, and vulnerability in many cases. it is of the utmost importance that the development control organizations be strengthened to stop any development that may cause more congestion and delay in the main city.

Chapter 8: Policy Implications for Sustainable Urban Structure 205 - The spatial analysis of the various scenarios undertaken in this thesis shows that decentralization of future growth could sustain Dhaka. However, to make the employment centres sustainable, feasible and independent, requires comprehensive area development plans in order to attract other activities such as residential development and urban amenities.

8.2.3 Transport and Infrastructure Success in urban economic development depends on sufficient good-quality infrastructure. Developing infrastructure is critical for generating growth, supporting jobs, strengthening connectivity and linkages, and increasing a city’s competitiveness (Asian Development Bank, 2008). The investigations undertaken in this thesis show that potential employment centres located in the outer region are poorly connected with the main city and with the surrounding catchments. Providing transport links and transportation services to connect these centres would benefit their future growth in many ways. This research reveals that in most of the decentralized scenarios examined (scenarios D and E), the outer ring attracts longer trips from the inner ring region and the main city. The travel pattern diagram shows that these centres do not attract many commuters from their surrounding rural hinterlands. Accessibility is an issue in that case. If surrounding hinterlands can be connected effectively with these potential employment centres, they could serve as local hubs. So, in term of transportation related policy issues there is a need to address for the following for a sustainable Dhaka:

- Encourage non-motorized transportation. Data from the STP study finds that the average commuting distance is 5km and a large share of people travel less than a kilometre for their everyday job, school and other work. If provisions are made for pedestrian movement over shorter distances and for the rickshaw and bicycle for medium distance, this could significantly reduce pressure from the motorized modes.

- Provide link for the potential centres in the outer ring region. The outer ring lacks connectivity with its surroundings and with other centres. This

Chapter 8: Policy Implications for Sustainable Urban Structure 206 connectivity could reduce the commuting distance and at the same time could reduce dependency on the main city.

- Efficient public transport is fundamental to efficient urban growth (Asian Development Bank, 2008). The current public transport network shows that mass transport service is limited within the main city. Encouraging growth in the outer ring would require good public transportation arrangements between the outskirt centres and their surrounding rural hinterlands and also between the main city and the outer region employment centres.

- A mass rapid transit (MRT) system has been on the discussion table for a long period of time; however, progress has recently been made. The proposed first phase is planning to serve only the high demand area (Figure 3.14). This MRT needs to expand to bring the potential centres under its coverage. This way, the distance barrier of the outer region centres can be minimized.

8.2.3 Environment In contrast to pollution levels in Dhaka, its car ownership is very low (Gurjar, 2008 ; World Bank, 2012). This result is due to congestion in roads, vehicle conditions and impurities in fuels. The banning of two stoke motors in Dhaka in 2003 improved air quality by reducing particulate matter. However, the increasing CNG (compressed natural gas) based vehicles are injecting NO2 and CO2 in the air.

Transportation studies on Dhaka megacity show a significant shift from non- motorized to motorized modes of travel. A 29% shift to motorized modes is recorded between 1994 and 2004 (PPK Consultants Pty Ltd et al., 1994 ; The Louis Berger Group, I. and B. C. Ltd, 2005). The increasing motorized rate increases vehicle emissions significantly (Figure 8.2). Measures must be taken in both ways to reduce motorized trips and to increase non-motorized trips. Walk ways need to be essentially free for pedestrian traffic and there should be non-motorized transport arrangements for shorter trips.

Chapter 8: Policy Implications for Sustainable Urban Structure 207

CO Emission at Different NMT Rates 2 7000

6000

5000

4000

3000

2000 Current NMT Rate

1000 Emission Emission from commuting Vehicle (in tonnes)

2

CO 0 40 45 50 55 60 65 70 75 80 Non-motorized Travel Rates (%)

Figure 8.2: Non-Motorized Transport Rate and CO2 Emission The travel analysis of the scenarios has provided evidence that decentralization could reduce ‘vehicle kilometre travelled (VKT). Moreover to reduce the traffic emissions from the same VKT several policy measures could be taken which include emission based policy, fiscal policy and urban sector policy. Fuel and vehicle standards could be strengthened. Higher taxes on private cars may reduce motorized trips, and incentives in public transport may also have positive impacts. The treatment of road geometry could reduce congestion and waiting time and thus emissions even further.

Bangladesh’s urban particulate matter concentration is twice the world average and a major cause for concern. Road transport contributes a lot towards this problem, which hastens the death of thousands each year. (Government of Bangladesh, 2008b).

Chapter 8: Policy Implications for Sustainable Urban Structure 208 8.3 Urban Governance for Dhaka Megacity

In Dhaka governance has developed in a centralized and sectoral way rather than emphasising on the place (Islam, 2012). This has fractured and become totally inadequate as a means of solving the deep environmental, social, and economic problems of the megacity (Talukder, 2006). Talukder (2006) identified 10 essential criteria of urban governance for polycentric growth management and found that often megacities in developing Asia do not satisfy them (Table 2.3 in Chapter 2) and Dhaka is not exception (Table 8.1). The second key issue is the appropriate institutional framework in managing the growth of Megacities. Municipal administrations cannot perform their task effectively because their jurisdictions are confusing and often smaller than effective urban area. Another concern about these institutions are paucity of funding and skilled personnel and thus they are usually fully occupied with the day-to-day problems of routine administration and service provision.

Chapter 8: Policy Implications for Sustainable Urban Structure 209 Table 8.1: Megacity Dhaka: Governance Principles and Present Institutional Arrangement Governance Principles Evident in Present Dhaka Institutional Arrangement?

1 A geographical area with clearly defined responsibility of No governance that covers the Extended Metropolitan Region 2 Strategic planning function that can provide a vision for how the Partly….RAJUK has a structural plan but without any comprehensive vision city can address its land use problems sustainably. to effectively address land use problems 3 Statutory planning function that can control development to No….RAJUK has been engaged to supplement its structural plan with a ensure ‘common good’ outcomes consistent with the strategic statutory plan/detailed area plan for its zones and projects. Local plan governments are not resourced properly for development control. 4 Development facilitation function that can provide investment No….Development facilitation is only being done by central government coordination, partnership for infrastructure and local validation spending, partnership is not on the political agenda yet. structure 5 An urban redevelopment authority to guide and monitor further No ….urban redevelopment and/or development come under the sole development in the core built-up areas in the city authority of RAJUK. 6 A transparent local process that can help define ‘common good’ No sustainability outcomes from development with all stakeholders 7 A coordination mechanism to ensure planning and development No…though a coordination committee (CC) is constituted, however it is with- are integrated out any statutory power. DTCB is another organization to coordinate some transport projects. 8 A way of raising the finance for the above process including No from land development 9 A strong link in to the national government system to enable Yes, but to the extent that the national government runs much of Dhaka. No, good political support in the sense that there is no single organization that has strong vertical linkage for EMR of Megacity Dhaka to the national government. 10 New professional skills in sustainability and local participation No

Source: Talukder, 2006

Chapter 8: Policy Implications for Sustainable Urban Structure 210 8.4 Urban Centre Hierarchy and Local Capacity Commuting patterns in different growth scenarios illustrates that polycentric growth is more sustainable. However, it is an essential to assess the capacity and impact on any location before any employment assignment. Deliberate, policy driven and away from central business district (CBD) job centres have been attempted in several metropolitan cities (Sydney, Mumbai, Shanghai etc). It is experienced that hierarchy of urban centres need to be establish at first. A centre can be developed as specialised one with specific economic activities. With time, these centres would grow and attract other activities and would reduce dependency on main city. Alternatively, a centre can be developed as a miniature of CBD, having all function in a small scale. All these centres can be put into tiers of primary, secondary and tertiary. However, the local economy, natural and human resources, geography and geology are critical characteristics to examine the locality.

Figure 8.3: Image of Old Dhaka Source: Google Earth

In Dhaka megacity, the main city centres are full of buildings (Figure 8.3), hardly any space left to accommodate new growth unless major revitalization operations are taken into action. On the contrary, inner ring and outer ring urban centres are of very low density. For example, Manikganj district headquarter is relatively small urban

Chapter 8: Policy Implications for Sustainable Urban Structure 211 settlement with built-up area along the main highway (Figure 8.4). The road pattern shows concentrated around the centre. It is clear from the satellite image that more growth can be accommodated by expanding the roads and infrastructure.

Figure 8.4: Landcover of Manikganj Source: Google Earth

Another, outer ring centre Narsingdi looks to have dispersed settlement (Figure 8.5). There are well distributed road network except in the southern boundary. This centre would be suitable for activities that do not require larger area. On the other hand, Manikganj may accommodate space consuming light industries or manufacturing.

On the top of all site specific suitability analysis, there need to be a strategic urban growth policy that would guide the growth to allocate to centres. The strategic policy need to provide enough insights to enable to distribute activities according to local situation. As the outer ring centres are populated with mainly unskilled man power compared to main city, it requires to locate employment that suits the local people. Initially, it may not function properly if skilled tertiary sector jobs (like banking, insurance, IT) are developed in those location.

Chapter 8: Policy Implications for Sustainable Urban Structure 212

Figure 8.5: Landcover of Narsindi Source: Google Earth

8.5 Conclusion

Success of a decentralized polycentric growth depends on the urban strategic policy, local plan implementation, investment in infrastructure and strengthening the urban and transport planning and management institutes. Emphasizing the local economy and labour force is paramount. Another important issue of megacity development is integration among plans. There seems to have a lack in integrating urban growth, transportation, utilities and other plans. Suitability can be achieved and impacts can be minimized only through proper policy, integrated implementation and optimial use of resources.

Chapter 8: Policy Implications for Sustainable Urban Structure 213 Chapter 9 Conclusion

9.1 Summary of Research Findings

Urbanization shows distinctive characteristics in cities in the developing world. While urbanization has reached a mature stage in many developed countries, it has become rapid in many developing countries across all continents. One of the striking features of urbanization in developing countries is agglomerations around the large cities. Larger cities are attracting more growth than smaller ones. This growth trend has turned several large cities into megacities in recent years in the developing countries and most of them are in Asia. Urban population growth in the Asian megacity has become phenomenal in recent times and brought the Asian megacity to the centre of concern for its sustainable future. There have been several studies with varying approaches that have searched for a solution that could enable these megacities to function properly and lead the society for economic, social and environmental fulfillment. In this endeavor, one of the prominent campaigns is emphasizing the urban spatial structure and transportation arrangements of these cities. Asian megacities with distinct features from Western large cities show different characteristics (Table 9.1).

Chapter 9: Conclusions 214

Table 9.1: Urban Structure in Several Asian Megacities City Population Urban Structure Author (in million)* Delhi, 23.5 Polycentric employment distribution. Bhandari et al. India Development of satellite town along the highway (2007) and railway Karachi, 21.9 Monocentric Qureshi and Lu, Pakistan 2007 Bombay, 21.0 In transition from mono to polycentric Shirgaokar, 2012 India Bangkok, 14.1 Extended monocentric or early stage of polycentric Vichiensan, 2007 Thailand development Jakarta, 25.6 Extended Metropolitan Region Hakim and Indonesia Parolin, 2009; Hakim, 2010 Tokyo, 33.6 Monocentric with multiple sub-centres Sorensen, 2001 ; Japan Kawabata, 2003 Istanbul, 13.6 Polycentric Alpkokin et al., Turkey 2005 Shanghai, 15.8 Dense monocentric to organic Polycentric Haixiao, 2000 China Manila, 21.7 Monocentric with multiple sub-centres Morichi and Philippines Acharya, 2012 *http://www.citypopulation.de/world/Agglomerations.html

However, there is not enough empirical evidence examining the spatial structure of these cities in term of their long time sustainability. In response to these problems and gaps in the literature as stated above, this research contributes here by exploring the spatial distribution of employment for Dhaka – one of the most thriving megacities in the world. In this research, the nature and patterns of this underlying urban spatial structure are assessed through a detailed examination of employment distribution and its travel related impacts, in particular the functional relationships between the residential locations of workers, their job destinations and their actual journey-to-work trip patterns. This, in turn, allows assessment of implications for present and future sustainability of the city, and whether a different spatial structure – from a policy and planning perspective – may accommodate more sustainable growth into the future. To elaborate the research objective, several research questions are formulated and discussed in Chapter 1, from which the following conclusions can be drawn:

- Methodological issues to identify urban spatial structure. This thesis has classified the methods available in the literature into several groups and presented with their data requirements, advantages and benefits (Table 2.4). However, from critical analysis, it reveals that most of these methods are developed

Chapter 9: Conclusions 215 and tested for large cities in developed countries. Since Dhaka has much higher population and employment density, perhaps it would require different criteria, threshold or a new approach to be able to adapt with its employment distribution patterns. Chapter 4 of this thesis contains the methodology for employment centre identification. The selected method is based on Giuliano and Small’s (1991) clustering method but with added statistical robustness to overcome the criticisms.

- Spatial structure of employment in Dhaka. The spatial analysis on the census data and data from transport project demonstrates a centralized employment concentration in Dhaka showing a monocentric pattern. However, there are manufacturing and light industrial jobs in the northern region of the city which gives it an expanded monocentric trend. There are at least 16 employment centres in Dhaka megacity and 3 potential employment centres in the outer region (Chapter 5). But the numbers of jobs in those outer ring centres are too low to have any impact on the overall spatial structure.

- Commuting patterns. Commuting to job shows distinct travel pattern for different regions. Commuting trips to the main city are relatively short and outer ring captures both short and very longer trips (Chapter 6). Unavailability of mode specific travel data restraints to perform modal choice analysis. However, trip length variation indicates that walking and non-motorized (Bi-cycle and Rickshaw) are two major modes in the main city. On the contrary, outer ring employment centres have around 80% local employee who walk or rely on non-motorized modes.

- Growth scenario for future development. Dhaka is likely to experience more threats for its sustainability in term of population, congestion, emission. Several researches recommend decentralize polycentric growth as the sustainable solution for megacities (for example, Brennan and Richardson, 1989). Several documents on planning for Dhaka have also mentioned the need for decentralization (Dacca Master Plan, 1959; Dhaka Metropolitan Development Strategy, 1995; Strategic Transport Plan, 2005). This research proposed an analytical framework (Figure 9.1) that generates 10 different growth scenarios (Chapter 4, Section 4.5.4) based on different urban strategies and estimates their travel impacts and vehicle emission (Chapter 7).

Chapter 9: Conclusions 216

Figure 9.1: Analytical Framework for Scenario Building

Chapter 9: Conclusions 217

- Policy implications for sustainable growth for Dhaka. This research also looks into the required attention from policy perspective to ensure the sustainable urban growth as discussed in the thesis. It argues that to guide the urban growth and management, and to ensure polycentric growth, appropriate policies need to be strengthening and formulated. It also emphasises on the institutional issues expected to play a role in implementing the polycentric growth successfully.

The research reported in this thesis has utilised expansive spatial and attribute data on population, jobs, travel and transportation for Dhaka megacity and has empirically determined that the city maintains a monocentric pattern in its employment distribution. However, as a response to planning intervention and to accommodate growth, several centres developed but they were quickly engulfed by the rapid growth and densification of the main city, primarily due to their close proximity to the city, scant in number and scale and failure to control growth in an effective manner.

Using a unique method and rich empirical data sources, this research has identified 12 employment centres in the main city of Dhaka and 4 in the inner ring region. In addition, 3 other potential cenrtres were identified in the outer ring. These centres are seen as the centres of employment and future employment landscape as part of the scenario analysis that was generated using different urban development assumptions. These scenarios were further tested for sustainability in terms of their spatial interactions, travel responses and associated vehicle emissions footprints. Furthermore, this thesis has argued that the results could be used as evidence to promote more sustainable transport in Dhaka. The importants findings of the research are:

- The type of employment centre identification method used is a crucial and important factor in determining the number of centres in an urban area. From the literature on employment centre identification it becomes evident that the same data may generate different results, once different methods are adopted. This study adopted an approach that could recognize different types of employment clusters within Dhaka.

Chapter 9: Conclusions 218

- The results strongly indicate that the megacity of Dhaka is characterized by an extended monocentric urban structure as opposed to a polycentric one. There do exist very high density employment centres but these are located in relatively close proximity, are close to the even higher density main city and, in fact, are an integral part of the main city. The three employment centres located in the outer ring dominate their respective catchment areas but have relatively low densities of employment. These three centres have the potential to develop into fully functional polycentic centres in the future. To redirect growth from the main city to these outskirt centres, these centres need to be attractive in term of their amenities and living environment.These centres require additional attention in their detailed design and development to ensure the appropriate supply and quality of urban amenities and services to support and encourage economic activities.

- The main city of Dhaka has by far the majority of jobs concentrated in its employment centres; namely Old Dhaka, Gulshan, Dhanmondi, Motijheel etc. Currently the main city has 51% of total megacity employment and its 12 employment centres have 37% of total employment. The influence of the main city is so strong that even without any additional jobs in the main city by 2024; the city would still keep its primacy.

- This research exposes the commuting patterns of these employment centres and notes that people living in the main city travel shorter distances compared to people living in the inner or outer ring. The analysis also suggests that Dhaka is likely to experience higher levels of congestion and pollution unless significant job decentralization occurs to the employment centres in the outskirts.

- Employment decentralization increases longer commuting trips but would eventually reduce the average commuting distance. As the jobs decentralize, employees from distant locations would prefer jobs within shorter distance. However, in higher decentralized scenarios (for example scenario E) there

Chapter 9: Conclusions 219

would be reverse-commuting (people from the main city would travel to employment centres in the outer ring). This is because employment growth in the main city is restricted in those scenarios and the growing number of city dwellers need to travel to the outer ring centres for their jobs. Mignot (2004) witnessed similar longer distance commuting for the outlying sub-centres in French cities.

- The increasing prominence of Dhaka is likely to create more traffic and worsen road congestion in the future. Moreover, transport studies on Dhaka show an increasing rate of motorized trips, which would make it severe. Between 1994 and 2004 there is a 29% increase in motorized commuting.

This generates an additional 83 tonnes of CO2 emissions from vehicle commuting for every 1% increase in motorized travel.

9.3 Limitations of the Thesis There are certain factors that have set the limits of the scope and achievement of this thesis. Availability of data is one of them which is a critical factor for any study in developing countries. However, the assumptions of the research and the time limitations of the research set constraints on the amount of work undertaken. Some major limitations are:

Data Availability Updated data of appropriate format has been a major issue. Studies of this nature require detailed employment data from the census which is not available for Dhaka. Employment data are sourced from the Strategic Transport Plan (STP) study in which employment numbers are estimated based on population and growth factors. Availability of temporal data is an issue as well. Several studies on urban spatial structure of employment compared the number or density of employment to identify the growth location. This was not possible in this study. STP was the first to estimate the employment number in the year 2005 for Dhaka. Dhaka Integrated Transport Study (DITS) did an employment estimate but with a smaller area coverage and for selected commercial regions. This study also lacks the analysis of the spatial

Chapter 9: Conclusions 220 interaction between employment centres and its rural hinterlands, which is very important for the potential employment centres in the outer ring. Several studies on North American cities find relationships between employment types and commuting behavior (travel time, length, mode choice, etc.) which is helpful if planning an effective transportation system. However, in the absence of such detailed employment data, this study could not perform analysis that might contribute the detailed transport solutions for the important employment centres.

This research combines data from different sources and certain generalizations were made to put them together in one spatial unit, as most analysis is spatial or linked with the spatial units (discussed in chapter 4). For example, census and STP data have two different spatial units and they are often combined for analysis purposes. The generalization and assumption made during their aggregation was to sacrifice some data quality and this may, in turn, have an impact on the analysis results.

Population Movement

The scenario analysis distributes employment based on different assumed urban development strategies. However, these scenarios do not consider movement in population, though residential population around the potential employment centres would significantly increase as the employment numbers increase. There is no policy or strategy that spells out the spatial distribution of population and, as a result, no assumptions were made. Different assumptions on the distribution of population could be made and more scenarios could be generated. This is left for a future study.

Transport Network This study assumes that all on-going and conceived transport projects will be completed by 2024 However, in current planning practice the megacity of Dhaka is considered to be the Dhaka Metropolitan Development Plan (DMDP) area which excludes the outer ring of this study. As a consequence, one does not observe many plans for infrastructure development in the outer ring. In turn, though the scenarios of this study consider decentralization of employment in the potential centres in the outer ring, infrastructure connecting those centres is still not adequately specified.

Chapter 9: Conclusions 221

These have impacts on the results as well. Travel pattern analysis (Chapter 7) shows that, with the decentralization of jobs in the outer ring, there would be an increase in the number of longer trips. This is because the outer ring potential centres are not well connected with their surrounding areas. As a result, this generates longer trips and has an impact on vehicle emissions.

Urban sector strategy and policy The urban sector strategy lacks the vision of the spatial distribution of growth for the megacity of Dhaka. As a result, there is no policy guideline to assign growth to the centres. This study looked into demographic, employment and travel data at one point in time and distributed growth to its centres. Possible travel and commuting vehicle emissions were estimated but it remains untested to examine the validity of these growth assignments and whether it could comply with the local plans and does not exceed the capacity of the area.

9.3 Scope for Future Research This study has the potential for opening several avenues of research in the field of urban development and planning and transportation for Dhaka. Availability of the above mentioned data and waiver of the limitations would open new opportunities of research. The analytical framework and research structure developed and tested in this study could be modified to accommodate other variables to test several other growth scenarios. At the same time, the impacts of any transport projects on travel behaviour and vehicle emissions could be assessed. For example, the elasticity of any relevant policy improvement on transport mode choice could be measured.

Chapter 9: Conclusions 222

References

Acharya, SR & Morichi, S 2007, 'Motorization and Role of Mass Rapid Transit in East Asian megacities', IATSS Research, vol. 13, no. 2, pp. 6-16.

Aguilera, A & Mignot, D 2004, 'Urban Sprawl, Polycentrism and Commuting - A Comparison of Seven French Urban Areas', Urban Public Economics Review, vol. 1, pp. 93-113.

Aguilera, A 2010, 'Urban Form And Commuting: A Critical Review of Literature', World Conference on Transportation Research

Ahmad, S 2012, 'Demand for Owner, Renter and Squatter Housing in Urban Bangladesh', 48th ISOCARP Congress 2012. .

Ahmed, S, Ahmad, JK & Mahmud, A 2005, Making Dhaka Livable, World Bank.

Ahmed, SJ, Bramley, G & Dewan, AM 2012, 'Exploratory Growth Analysis of a Megacity through Different Spatial Metrics: A Case Study on Dhaka, Bangladesh (1960–2005)', Journal of the Urban and Regional Information Systems Association (URISA), vol. 24, no. 1, pp. 9-24.

Ahmed, SU (ed.) 1989, Dhaka Past Present Future, The Asiatic Society of Bangladesh, Dhaka.

Ahmed, SU 1986, Dacca: a study in urban history and development, Curzon Press, Dhaka.

Ahsan, RM 1989, 'Changing Pattern of the Commercial Area of Dhaka city', in Ahmed, SU (ed), Dhaka Past Present Future, The Asiatic Society of Bangladesh, Dhaka, pp. 396-414.

Akther, MS 2009, 'A big no to flyover and subway in Dhaka', The Daily Star, 14/11/2009. viewed 14/11/2009,

Alam, MJB & Habib, KMN 2003, 'Effects of Alternative Transportation Options on Congestion and Air Pollution in Dhaka City', Journal of Civil Engineering, vol. CE 21, no. 2, pp. 165-174.

Alonso, W 1964, Location Theory, MIT Press, Cambridge MA.

Alperovich, G & Deutch, J 1996, 'Urban structure with two coexisting and almost

References 223 completely segregated populations: The case of East and West Jerusalem', Regional Science and Urban Economics 26 (1996) 171-187, vol. 26, pp. 171-187.

Alpkokin, P, Cheung, C, Black, J & Hayashi, Y 2008, 'Dynamics of clustered employment growth and its impacts on commuting patterns in rapidly developing cities', Transport Research Part A, vol. 42, pp. 427-444.

Alpkokin, P, Hayashi, Y, Black, J & Gercek, H 2005, 'Polycentric Employment growth and Impacts on Urban Commuting: Case Study of Istanbul', Journal of the Eastern Asia Society for Transportation Studies, vol. 6, pp. 3835-3850.

Ameen, MS 1998, 'Growth Potentials of a Pre-Mature Mega City: an event of grim portent', Journal of Civil Engineering, vol. CE 6, no. 2, pp. 151-172.

Amin, ATMN 1987, 'The Role of the Informal Sector in Economic Development', International Labour review, vol. 126, no. 5, pp. 611-623.

Anas, A 1998, 'Transport and Land Use: Editor’ s Introduction', Urban Studies, vol. 35, no. 7, pp. 1015-1018.

Anas, A, Richard Arnott & Small, KA 1998, 'Urban Spatial Structure', Journal of Economic Literature vol. 37, pp. 1426 - 1464.

Anderson, WP, Kanaroglou, PS & Miller, EJ 1996, 'Urban Form, Energy and the Environment: A Review of Issues, Evidence and Policy', Urban Studies, vol. 33, no. 1, pp. 7-35.

Arampatzis, G, Kiranoudis, CT, scaloubacas, P & Assimacopoulos, D 2004, 'A GIS-based decision support system for planning urban transportation policies', European Journal of Operational Research, vol. 152, pp. 465-475.

Arbury, J 2009, 'From Urban Sprawl to Compact City – An analysis of urban growth management in Auckland', Auckland University, Auckland.

Ashraf, KK 2010, 'A New Dhaka is Possible', Forum.

Asian Development Bank (ADB) 2003, Reducing Vehicle Emissions in Asia, Policy Guidelines for Reducing Vehicle Emissions in Asia, Asian Development Bank.

Asian Development Bank (ADB) 2010, Reducing Carbon Emissions from Transport Projects, Asian Development Bank.

References 224 Asian Development Bank (ADB) 2007, Sustainable Urban Transport - Dhaka Case Study, ADB, Dhaka.

Asian Development Bank (ADB) 2008, Managing Asian Cities, Philippines.

Bangladesh Bureau of Statistics (BBS) 1991, Dhaka Statistical Metropolitan Area, Bangladesh Bureau of Statistics, Dhaka.

Bangladesh Bureu of Statistics (BBS) 2011, Population and Housing Census 2011 - Preliminary Results, Bangladesh Bureau of Statistics, Dhaka. viewed July 2011,

Bangladesh Road Transport Authority (BRTA) 2012, updated 04/12/2012, .

Banister, D 1992, 'Energy Use, Transport and Settlement Patterns', in Breheny, MJ (ed), Sustainable Development and Urban Form, vol. 2, Pion Limited, London, pp. 160-181.

Banister, D 2008, 'The sustainability mobility paradogm', Transport Policy, vol. 15, pp. 73-80.

Banister, D 2009, 'The Dilemmas of Sustainable TransportT', Projections - MIT Journal of Planning, vol. 9, no. Sustainable Transportation - An International Perspective, pp. 122-124.

Bari, M & Efroymson, B 2007, 'Dhaka Strategic Transport Plan (STP) - A Critical Review', Unpublished Discussion Paper.

Barnes, G 2001, Population and Employment Density and Travel Behavior in Large U.S. Cities Minnesota Department of Transport,

Barter, PA 1999, 'An International Comparative Perspective on Urban Transport And Urban Form In Pacific Asia: The Challenge Of Rapid Motorisation in Dense Cities', PhD Thesis, Murdoch University, Perth.

Barter, PA 2000, 'Urban Transport in Asia: Problems and prospects for high-density cities', Asia-Pacific Development Monitor, vol. 2, no. 1, pp. 33-66.

Batty, M 2001, 'Polynucleated Urban Landscapes', Urban Studies, vol. 38, no. 4, pp. 635-655.

Baumont, C & Gallo, JL 1999, 'Empirical Foundations of Multicentric Urban Models', 46th North American Congress of the Regional Science Association International.

Baumont, C, Ertur, C & Gallo, JL 2004, 'Spatial Analysis of Employment and

References 225 Population Density: The Case of the Agglomeration of Dijon 1999', Geographical Analysis, vol. 36, no. 2, pp. 146-176.

Bentley, I 1995, Responsive Environments: a manual for designers, Routledge

Bertaud, A 2001, 'Metropolis: A measure of the Spatial Organization of 7 large cities',

Bertaud, A 2002, 'Note on Transportation and Urban Spatial Structure', The World Bank's 14th Annual Bank Conference on Development Economics (ABCDE), Washington.

Bertaud, A 2003, '“Metropolitan Structures Around the World” What is common? What is different? What relevance to Marikina in the context of ?', Prasentation, in Marikina, May 2003.

Bertaud, A 2004, The Spatial Organization of Cities: Deliberate Outcome or Unforeseen Consequence?, Institute of Urban and Regional Development, University of California at Berkeley.

Bhandari, K, Peng, J, Alpkokin, P, Mukhopadhyaya, D & Gangopadhyay, S 2007, 'Policies, Commuting pattern and Accessibility in a Non-monocentric City: Case Study of Delhi', Proceedings of the Eastern Asia Society for Transportation Studies, vol. 6.

Bhattacharya, D & Deb, UK 2006, Bangladesh 2020: An Analysis of Growth Prospect and External Sector Behaviour, Centre for Policy Dialogue (CPD), Dhaka.

Bird, J 1977, Centrality and Cities, Routledge & Kegan Paul, London.

Black, D & Henderson, V 1999, 'A Theory of Urbn Growth', Journal of Political Economy, vol. 107, no. 2, pp. 252-284.

Black, J, Cheung, C, Doust, K & Shabty, O 2007, 'Metrics of Changes To Major Employment Centres: Analyses of Spatial Plans For Sydney 1948 - 2031', Journal of the Eastern Asia Society for Transportation Studies, vol. 7, pp. 31-45.

Black, JA & Katakos, A 1987, 'Optimisation methods and the classification of city structure: tehory and emperical testing', Environment and Planning B, vol. 14, pp. 93-107.

Boarnet, M & Crane, R 2001, 'The influence of land use on travel behavior: specification and estimation strategies', Transportation Research Part A, vol. 35, pp. 823-845.

Bogart, WT & Ferry, WC 1999, 'Employment Centers in Greater Cleveland: Evidence

References 226 of Evolution in a Formerly Monocentric City', Urban Studies, vol. 36, no. 12, pp. 2099-2110.

Boiteux-Orain, C & Guillain, R 2004, 'Changes In The Intrametropolitan Location Of Producer Services In Île-De-France (1978–1997): Do Information Technologies Promote A More Dispersed Spatial Pattern?', Urban Geography, vol. 25, no. 6, pp. 550-578.

Bontje, M & Burdack, J 2005, 'Edge Cities, European-style: Examples from Paris and the Randstad', Cities, vol. 22, no. 4, pp. 317-330.

Bose, RK 2007, 'Urban Transport Scenarios in South Asia : Energy and Environmental Impact of Enhanced Public Transport Systems', Transportation Research Record: Journal of the Transportation Research Board, vol. 2011, pp. 116-126.

Bramley, G & Power, S 2009, 'Urban form and social sustainability: the role of density and housing type', Environment and Planning B: Planning and Design, vol. 36, no. 1, pp. 30-48.

Breheny, M 1995, 'The Compact City and Transport Energy Consumption', Transactions of the Institute of British Geographers, vol. 20, no. 1, pp. 81-101.

Breheny, MJ 1992, 'Sustainable Development and Urban Form: An Introduction', in Breheny, MJ (ed), Sustainable Development and Urban Form, vol. 2, Pion Limited, London, pp. 1-23.

Brennan, EM & Richardson, HW 1989, 'Asian Megacity Characteristics, Problems, and Policies', International Regional Science Review, vol. 12, no. 2, pp. 117-129.

Brunsdon, C, Fotheringham, AS & Charlton, M 2002, 'Geographically weighted summary statistics — a framework for localised exploratory data analysis', Computers, Environment and Urban Systems, vol. 26, pp. 501–524.

Buchanan, N, Barnett, R, Kingham, S & Johnston, D 2006, 'The effect of urban growth on commuting patterns in Christchurch, New Zealand', Journal of Transport Geography, vol. 14, no. 2006, pp. 342–354.

Buliung, RN & Kanaroglou, PS 2006, 'Urban Form and Household Activity-Travel Behavior', Growth and Change, vol. 37, no. 2, pp. 171-199.

Burgess, EW 1925, 'The growth of the city: an introduction to a research project', in Park, RE, Burgess, EW & McKenzie, RD (eds), The City, pp 47-62, University of Chicago Press, Chicago.

Callaway, K & Espindola, C 2005, Working Paper No: 8 Dhaka UTP Model Development and Calibration, Strategic Transport Plan for Dhaka, The Louis

References 227 Berger Group, Inc, Bangladesh Consultants Ltd, Dhaka Transport Coordination Board, Dhaka.

Callaway, K 2005, Working Paper No: 9; Assumptions Used in STP Model Development, The Louis Berger Group, Inc Bangladesh Consultants Ltd, Dhaka Transport Coordination Board, Dhaka.

Camagni, R, Gibelli, MC & Rigamonti, P 2002, 'Urban Mobility and Urban Form: The Social And Environmental Costs of Different Patterns of Urban Expansion', Ecological Economics, vol. 40, pp. 199-216.

Casello, JM & Smith, TE 2006, 'Transportation Activity Centers for urban Transportation Analysis', Journal of Urban Planning and Development vol. 132, no. 4, pp. 247-257.

Céline, R & Patrice, T 2008, Commuters Graphs and Cities ’ Polycentric Cohesion, < http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.86.982>

Cervero, R & Duncan, M 2006, 'Which Reduces Vehicle Travel More: Jobs-Housing Balauce or Retail-Housing Mixing?', Journal of the American Planning Association, vol. 72, no. 4, pp. 475-491.

Cervero, R & Kockelman, K 1997, 'Travel Demand and the 3Ds: Density, Diversity, and Design', Transport research Part D, vol. 2, no. 3, pp. 199-219.

Cervero, R & Lands, J 1995, 'The Transportation-landuse Connection Still Matters ', Access, vol. 7, pp. 2-11.

Cervero, R & Wu, K-L 1998, 'Sub-centering and Commuting: Evidences from the San Francisco Bay Area, 1980-90', Urban Studies, vol. 35, no. 7, pp. 1059-1076.

Cervero, R 1989, 'Suburban Employment Centers: Probing the Influence of Site Features on the Journey-to-Work', Journal of Planning Education and Research, vol. 8, no. 2, pp. 75-85.

Cervero, R 2001, 'Efficient Urbanisation: Economic Performance and the Shape of the Metropolis', Urban Studies, vol. 38, no. 10, pp. 1651–1671.

Cervero, R 2002, 'Built environments and mode choice: toward a normative framework', Transportation Research Part D, vol. 7, pp. 265-284.

Charlton, M, Fotheringham, S & Brunsdon, C 2006, Geographically Weighted Regression, Economic and Social Research Council.

Charton, M & Fotheringham, AS, Geographically Weighted Regression, National Centre for Geocomputation, Ireland.

References 228 .

Chen, NY-p & Heligman, L 1994, 'Growth of the World's ', in Fuchs, RJet al (eds), Mega-City Growth and the Future, United Nations University Press, Tokyo.

Cheung, C & Black, J 2005, 'Residential Location-Specific Travel Preference in an Intervenning Opportunity Model: Transport Assessment for Urban Release Areas', Journal of the Eastern Asia Society for Transportation studies, vol. 6, pp. 3773 - 3788.

Choguill, CL 1982, Spatial Planning for Urban Development in Bangladesh, United Nations Centre for regional Development .

Choguill, CL 1987, 'The Accelerating Urbanization of Bangladesh', in Choguill, CL (ed), New Communities for Urban Squatters - lesson from the plan that failed in Dhaka, Bangladesh, Plenum Press.

Choguill, CL 1994, 'Crisis, Chaos, crunch? Planning for Urban Growth in the Developing World', Urban Studies, vol. 31, no. 6, pp. 935-945.

Chow, L-F, Zhao, F, Liu, X, Li, M-T & Ubaka, aI 2001, 'Transit Ridership Model Based on Geographically Weighted Regression', Transportation Research Record: Journal of the Transportation Research Board,, vol. 1972, pp. 105-114.

Chowdhury, AM & Ahmed, SU (eds), 2010, 'Transport Infrastructure of Dhaka City: Status and Challenges', 400 Years of Capital Dhaka and Beyond, International Seminar on the History, Heritage and Urban Issues of Capital Dhaka,

Chowdhury, AM & Faruqui, S 1989, 'Physical Growth of Dhaka City', in Ahmed, SU (ed), Dhaka Past Present Future, The Asiatic Society of Bangladesh, Dhaka, pp. 43-63.

Cladera, JR, Duarte, CRM & Moix, M 2009, 'Urban Structure and Polycentrism: Towards a Redefinition of the Sub-centre Concept', Urban Studies, vol. 46, no. 13, pp. 2841-2868.

Clark, WAV & Kuijpers-Linde, M 1994, 'Commuting in restructuring urban regions', Urban Studies, vol. 31, no. 3, pp. 465-483.

Coevering, Pvd & Schwanen, T 2006, 'Re-evaluating the impact of urban form on travel patterns in Europe and North-America', Transport Policy, vol. 13, pp. 229-239.

Cohen, B 2004, 'Urban Growth in Developing Countries: A Review of Current Trends and a Caution Regarding Existing Forecasts', World Development, vol. 32, no.

References 229 1, pp. 23-51.

Cohen, B 2006, 'Urbanization in developing countries: Current trends, future projections, and key challenges for sustainability', Technology in Society, vol. 28, pp. 63-80.

Cox, W 2012, 'Evolving Urban Form: Dhaka ', New Geography, viewed 03/06/2013, .

Craig, SG & Ng, PT 2001, 'Using Quantile Smoothing Splines to Identify Employment Subcenters in a Multicentric Urban Area', Journal of Urban Economics, vol. 49.

Crane, R & Crepeau, R 1998, 'Does Neighborhood Design Influence Travel? A Behavioral Analysis of Travel Diary and GIS Data', Transportation Research Part D, vol. 3, no. 4, pp. 225-238.

Crane, R 2000, 'The Influence of Urban Form on Travel: An Interpretive Review', Journal of Planning Literature, vol. 15, no. 1, pp. 3-23.

Dahiya, B 2012, 'Cities in Asia, 2012: Demographics, economics, poverty, environment and governance', Cities, vol. 29.

Daniels, R 2008, 'Jobs Closer to Home: What does it mean and how to achieve it', 30th Australian Transport Research Forum.

Das, RJ & Dutt, AK 1993, 'Rank-Size Distribution and Primate City Characteristics in India - A Temporal Analysis', GeoJournal, vol. 29, no. 2, pp. 125-137.

Davoudi, S 2003, 'Polycentricity in European Spatial Planning: from an Analytical Tool to a normative Agenda', European Planning Studies, vol. 11, no. 8, pp. 989-999.

Department of Environment (DoE) 2005, Dhaka City - State of Environment 2005, Dhaka.

Dhaka City Corporation (DCC), 2012, Dhaka City Corporation,

Dimitriou, HT & Gakenheimer, R (eds) 2011, Urban Transport in the Developing World - A Handbook of Policy and Practice, Edward Elgar, Cheltenham.

Douglass, M 1995, 'Global Interdependence and Urbanization: planning for the Bangkok Mega-urban region', in McGee, TG & Robinson, IM (eds), The Mega-Urban regions of Southeast Asia, UBC Press, Vancouver.

Doust, K & Black, J 2009, 'A Holistic Assessment Framework For Urban Development

References 230 and Transportation With Innovative Triple Bottom Line Sustainability Metrics', Projections - MIT Journal of Planning, vol. 9, no. Sustainable Transportation - An International Perspective, pp. 10-27.

Doust, K & Parolin, B 2008, 'Performance metrics for sustainable urban transport - accessibility and greenhouse gas emissions', 31st Australian Transport Research Forum.

Doust, KH 2008, 'Metrics of Environmental Sustainability, Social Equity and Economic Efficiency in Cities', PhD Thesis, School of Civil And Environmental Engineering, University of New South Wales, Sydney.

Dowall, DE & Treffeisen, PA 1991, 'Spatial Transformation in Cities of the developing world - multinucleation and labd-capital substitution in Bogota, Colombia', Regional Science and Urban Economics, vol. 21, pp. 201-224.

Dreborg, K 1996, 'Essence of Backcasting', Futures, vol. 28, no. 9, pp. 813-828.

Duarte, CM, Núñez, CA, Tresserra, JM, Mejia, EC & Prieto, CP 2011, 'Polycentrism in the Spanish metropolitan system: an analysis for 7 metro areas', European Regional Science Association Conference, European Regional Science Association. viewed 09/02/2012, .

Erickson, RA 1986, 'Multinucleation in Metropolitan Economics', Annals of the American geographer, vol. 76, no. 3, pp. 331-346.

Ewing, R & Cervero, R 2001, 'Travel and the Built Environment', Transportation Research Record vol. 1780, pp. 87–114.

Fan, Y & Khattak, AJ 2008, 'Urban Form Individual Spatial Footprints, and Travel: An examination of Space use Behavior', Transportation Research Record, vol. 2138, pp. 85-93.

Farrington, JH 2007, 'The new narrative of accessibility: its potential contribution to discourses in (transport) geography', Journal of Economic Geography, vol. 15, pp. 319-330.

Food and Agriculture Organization (FAO) 2002, Urbanization - Linking Development Across the Changing Landscape, Rome.

Fotheringham, AS, Brunsdon, C & Charlton, M 2002, Geographically Weighted Regression - Analysis of Spatially Varying Relationship, John Wiley & Sons, Ltd, West Sussex.

Freestone, R 1989, Model Communities : the garden city movement in Australia,

References 231 Nelson, Melbourne.

Fuchs, RJ (ed.) 2004, Mega-city growth and the future United Nations University,

Fujita, M & Ogawa, H 1982, 'Multiple equilibria and structural transition of non-monocentric urban configurations', Regional Science and Urban Economics, vol. 12, no. 2, pp. 161-196.

Gaigné, C, Riou, S & Thisse, J-F 2010, Are compact cities environmentally friendly? .

Garcia-López, M-À & Muñiz, I 2010, 'Employment Decentralisation: Polycentricity or Scatteration? The Case of Barcelona', Urban Studies, vol. 47, no. 14, pp. 3035-3056.

Garreau, J 1991, Edge City: Life on the New Frontier, Anchor Books, New York : Doubleday.

Geddes, P 1917, Report on Dacca City Planning 1917. Bengal Secretariat Book Depot, Calcutta,

Getis, A & Ord, JK 1992, 'The Analysis of Spatial Association by Use of Distance Statistics', Geographical Analysis, vol. 24, no. 3, pp. 189-206.

Giuliano, G & Small, K 1991, 'Subcenters in the Los Angeles Region', Regional Science and Urban Economics, vol. 21, pp. 163-182.

Giuliano, G & Small, KA 1993, 'Is the Journey to Work Explained by Urban Structure?', Urban Studies, vol. 30, no. 9, pp. 1485-1500.

Giuliano, G & Small, KA 1999, 'The determinants of growth of Employment Subcenters', Journal of Transport Geography, vol. 7, pp. 189-201.

Giuliano, G 1995, 'The Weakening Transportation land Use Connection', Access, vol. 6, pp. 3 - 15.

Giuliano, G, Redfearn, C, Agarwal, A, Li, C & Zhuang, D 2007, 'Employment Concentrations in Los Angeles, 1980-2000', Environment and Planning A, vol. 39, p. 23.

GlobeScan & MRC McLean Hazel 2004, Megacity Challenges, Munich.

Gordon, C 2012, 'The challenges of transport PPP's in low-income developing countries: A case study of Bangladesh', Transport Policy, vol. 24.

Gordon, P & Richardson, HW 1996, 'Beyond Polycentricity - The Dispersed Metropolis,

References 232 Los Angeles, 1970-1990', Journal of the American Planning Association vol. 62, no. 3, pp. 289-295.

Gordon, P, Kumar, A & Richardson, HW 1989, 'The Influence of Metropolitan Spatial Structure on Commuting Time', Journal of Urban Economics, vol. 29, no. 138-151.

Gordon, P, Richardson, HW & Jun, M-J 1991, 'The Commuting Paradox - Evidence from Top Twenty', Journal of the American Planning Association vol. 57, no. 4, pp. 416-420.

Gordon, P, Richardson, HW & Wong, HL 1986, 'The distribution of population and employment in a polycentric city: the case of Lon Angeles', Environment and Planning A vol. 18, no. 1, pp. 161-173.

Gordon, P, Richardson, HW & Yu, G 1998, 'Metropolitan and Non-metropolitan Employment Trends in the US: Recent Evidence and Implications', Urban Studies, vol. 35, no. 7, pp. 1037-1057.

Gottlieb, PD 1995, 'Residential Amenities, Firm Location and Economic Development', Urban Studies, vol. 32, no. 9, pp. 1413-1436.

Grazi, F, Bergh, JCJMvd & Ommeren, NNv 2008, 'An Emperical Analysis of Urban Form, Transport, and Global Warming', The Energy Journal, vol. 29, no. 4, pp. 97-121.

Greene, D 1980, 'Urban Subcenters: recent trends in urban spatial structure ', Growth and Change, pp. 29-40.

Greene, DL 1980, 'Recent Trends in Urban Spatial Structure', Growth and Change, vol. 11, no. 1, pp. 29-40.

Gupta, N (ed.) 1989, Urbanization in South Asia in the Colonial Centuries, The Asiatic Society of Bangladesh, Dhaka.

Gupta, S & Rayadurgam, I 2008, 'Urban growth and Governance in South Asia', South Asian Perspective, vol. (forthcoming).

Gurjar, BR, Butler, TM, Lawrence, MG & Lelieveld, J 2008, 'Evaluation of emissions and air quality in megacities ', Atmospheric Environment, vol. 42, pp. 1593-1606.

Haixiao, P 2000, 'Shanghai from Dense Mono-center to Organic Poly-Center Urban Expansion', Unpublished Paper,

References 233 Hakim, I & Parolin, B 2009, 'Spatial Structure and Spatial Impacts of the Jakarta Metropolitan Area: a Southeast Asian EMR Perspective', International Journal of Human and Social Sciences, vol. 4, no. 6, pp. 397-405.

Hakim, I 2009, 'The spatial structure of employment and its impacts on the journey to work in the Jakarta Metropolitan Area: a Southeast Asian Extended Metropolitan Region (EMR) perspective', Faculty of Built Environment, University of New South Wales, Sydney.

Hakim, I 2010, 'The spatial structure of employment and its impacts on the journey to work in the Jakarta Metropolitan Area: a Southeast Asian Extended Metropolitan Region (EMR) perspective', Faculty of the Built Environment, University of New South Wales, Sydney.

Handy, S 1996, 'Methodologies for exploring the link between urban form and travel behavior', Transportation Research, vol. 1, no. 2, pp. 151-165.

Harris, CD & Ullman, EL 1945, 'The Nature of Cities', Annals of the American Academy of Political and Social Science, vol. 242, no. Nov.,1945, pp. 7-17.

Hasan, S & Mulamoottil, G 1994, 'Environmental Problems of Dhaka City - a study of mismanagement', Cities, vol. 11, no. 3, pp. 195-200.

Hasnath, SA 1982, 'Sites and Services Schemes in Dacca: a Critique', Public Administration and Development, vol. 2, pp. 15-30.

Heinke, GW 1997, 'The Challenge of Urban growth and Sustainable Development for Asian Cities in the 21t Century', Environmental Monitoring and Assessment, vol. 44, no. 1-3, pp. 155-171.

Helsley, RW & Sullivan, AM 1991, 'Urban Structure Formation', Regional Science and Urban Economics, vol. 21, pp. 255-275.

Henderson, V 2002, 'Urbanization in Developing Countries', The World Bank Research Observer, vol. 17, no. 1, pp. 89-112.

Hill, M 2005, Urban Settlement and Land Use, Hodder Murray, London.

Holden, E 2004, 'Ecological footprints and sustainable urban form', Journal of Housing and the Built Environment, vol. 19, no. 1, pp. 91-109.

Hoque, MM & Hossain, T 2004, 'Augmentation of Mass Transit Mode in Dhaka, Bangladesh', Co-operation for Urban Mobility in the Developing World. Codatu XI.

Hossain, M & McDonald, M 1998, 'Modelling Of Traffic Operations In Urban

References 234 Networks Of Developing Countries: A Computer Aided Simulation Approach', Computer Environment and Urban Systems, vol. 22, no. 5, pp. 465-483.

Hossain, M 2004, 'Shaping up of urban transport system of a developing metropolis in absence of proper management setup: the case of Dhaka', Journal of Civil Engineering, vol. 32, no. 1, pp. 47-58.

Hossain, S 2006, 'Rapid Mass Urbanisation and Its Social Consequences in Bangladesh: The Case of the Megacity of Dhaka', Biennial Conferences of the Asian Studies Association of Australia, University of Wollongong, Australia, 26th-29th June 2006.

Hossain, S 2008, 'Rapid Urban Growth and Poverty in Dhaka City', Bangladesh e-Journal of Sociology, vol. 5, no. 1.

Hughes, MA 1991, 'Employment decentralization and accessibility', Journal of the American Planning Association, vol. 57, no. 3, pp. 288-299.

Hugo, G 2003, 'Urbanization in Asia: An Overview', African Migration in Comparative Perspective.

Huq, S & Alam, M 2003, 'Flood Management and Vulnerability of Dhaka', in Kreimer, A, Arnold, M & Carlin, A (eds), Building Safer Cities: the future of Disaster Riask, vol. Disaster Raik Management Series, The World Bank, Washington.

Hyder, Z 1994, 'Organic Cities and the Case of Patrick Geddes in Dhaka', Department of Architecture, Massachusetts Institute of Technology, Cambridge. viewed May 06, 1994,

Ichimura, M 2003, 'Urbanization, Urban Environment and Land Use: Challenges and Opportunities (An Issue Paper)', viewed 01/02/2012,.

Ihlanfeldt, KR & Sjoquist, DL 1998, 'The Spatial Mismatch Hypothesis: A Review of Recent Studies and Their Implications for Welfare Reform', Housing Policy Debate, vol. 9, no. 4, pp. 849-892.

Imon, SS 2001, 'Evaluating the options of Compact Development as a Sustainable Urban Form for the Growth of Dhaka', The Centre of Urban Planning & Environmental Management, University of Hongkong, Hong Kong.

Ingram, GK 1998, 'Patterns of Metropolitan Development: What Have We Learned?', Urban Studies, Vol. 35, No. 7, 1019± 1035, 1998, vol. 35, no. 7, pp. 1019-1035.

International Federation of Surveyors 2010, Rapid Urbanization and Mega Cities: The Need for Spatial Information Management, International Federation of

References 235 Surveyors, Copenhagen, Denmark.

Ioannides, YM, Overman, HG, Rossi-Hansberg, E & Schmidheiny, K 2008, 'The effect of information and communication technologies on urban structure', Economic Policy, vol. 23, no. 54, pp. 201-242.

Islam, I, Sharmin, M, Masud, F & Moniruzzaman, M 2009, 'Urban Consolidation Approach for Dhaka City: Prospects and Constraints', Global Built Environment review (GBER), vol. 7 no. 1, pp. 50 - 68.

Islam, MS 2007, 'Physical Density and Urban Sprawl: A Case of Dhaka City', Department of Urban Planning and Environment, Kungliga Tekniska Högskolan (KTH), Stockholm.

Islam, MS, Rahman, MR, Shahabuddin, AKM & Ahmed, R 2010, 'Changes in Wetlands in Dhaka City: Trends And Physico-Environmental Consequences', Journal of Life and Earth Science, vol. 5, pp. 37-42.

Islam, N (ed.) 1996a, Megacity Management in the Asian and Pacific Region, ADB,

Islam, N 1996b, Dhaka from city to mega city: perspective on people, places, planning and development issues, Urban Studies Programme, Department of Geography, University of Dhaka,

Islam, N 2006, 'Bangladesh', in Roberts, B & Kanaley, T (eds), Urbanization and Sustainability: Case Studies of Good Practice, ADB.

Islam, N 2012, ' Urbanization and Urban Governance in Bangladesh', 13th Annual Global Development Conference on Urbanization and Development: Delving Deeper into the Nexus. .

Jabareen, YR 2006, 'Sustainable Urban Forms : Their Typologies, Models, and Concepts', Journal of Planning Education and Research, vol. 26, pp. 38-52.

Jenks, M, Burton, E & Williams, K 2000, Compact Cities and Sustainability: An Introduction, Spon Press, London.

Jones, GW 2001, 'Studying Extended Metropolitan Regions in South-East Asia', XX1V General Conference of the IUSSP, 1.

Jones, GW 2002, 'Southeast Asian urbanization and the growth of mega regions', Journal of Population Research, vol. 19, no. 2, pp. 119-136.

Junius, K 1999, 'Primacy and Economic Development: Bell Shaped or Parallel Growth

References 236 of Cities?', Journal Of Economic Development, vol. 24, no. 1, pp. 1-22.

Kabir, A & Parolin, B 2010, 'Urban Spatial Structure and Transport Opportunity in Dhaka – A Question of Sustainability', 1st International Conference On Sustainable Urbanization, 15-17 December 2010, Hong Kong

Kabir, A & Parolin, B 2011, 'Dhaka Mega City: Growth, Congestion, Environment and Planning for a Sustainable Future', Asian Planning School Association. 19-21 September 2011, Tokyo University, Tokyo.

Kabir, A & Parolin, B 2011’Planning and Development of Dhaka – a story of 400 years’, The15th International Planning History Society Conference, 15-18 July, 2012, Sao Paulo, Brazil.

Kabir, A. Parolin, B & Freestone, R 2011, ‘The Impacts of Employment Centres on Travel Patterns in the Mega City of Dhaka’, World Planning School Congress 2011, 4-7 July Perth.

Kain, JF 1992, 'The Spatial Mismatch Hypothesis: Three Decades Later', Housing Policy Debate, vol. 3, no. 2, pp. 371-392.

Kalia, R 1987, Chandigarh, Oxford University Press, Bombay.

Karim, A (ed.) 1989, Origin and Development of Mughal Dhaka, The Asiatic Society of Bangladesh, Dhaka.

Karim, MM 1999, 'Traffic Pollution Inventories and Modeling in Metropolitan Dhaka, Bangladesh', Transportation Research Part D vol. 4, pp. 291 - 312.

Karim, MM 2001, 'Status of Air Quality and State of art Control Measures in Dhaka, Bangladesh', 94th Annual Conference and Exhibition of Air and Waste Management Association, June 24 - 28, 2001, Orlando, Florida, USA.

Karim, MM 2012, 'Agglomeration in Dhaka: How Does Ready Made Garment (RMG) Industry Make Dhaka the Fastest Growing City of the World?', .

Kasarda, JD & Crenshaw, EM 1991, 'Third World Urbanization: dimensions, theories, and determinants', Annual Reviews of Sociology, vol. 17, pp. 467-501.

Kassens, E 2009, 'Planning For Sustainable Transportation : an International Perspective', Projections - MIT Journal of Planning, vol. 9, pp. 6-9.

Kawabata, M 2003, 'A GIS-Based Analysis Of Jobs, Workers, and Job Access in Tokyo', Center for Spatial Information Science, University of Tokyo,

References 237 Bunkyo-ku, Tokyo, Japan

Kawsar, MA 2012, 'Urbanization, Economic Development and Inequality', Bangladesh Research Publications Journal, vol. 6, no. 4, pp. 440-448.

Kenworthy, J & Hu, G 2002, 'Transport and Urban Form in Chinese Cities - An International Comparative and Policy Perspective with Implications for Sustainable Urban Transport in China', DISP, pp. 4-14.

Khan, H 2008, 'Challenges for Sustainable Development: Rapid Urbanization, Poverty and Capabilities in Bangladesh', Munich Personal RePEc Archive,

Khan, MS & Barua, S 2009, 'The Dynamics of Residential Real Estate Sector in Bangladesh: Challenges Faced and Policies Sought', Journal of Management Studies, no. December.

Kidokoro, T & Hanh, LD 1993, 'Urban Explosion And Transport Crisis in Asian Mega-Cities – Overview and UNCRD Approach ', IATSS Research, vol. 17, no. 1, pp. 6-13.

Kloosterman, RC & Lambregts, B 2001, 'Clustering of Economic Activities in Polycentric Urban Regions: the case of Randstad', Urban Studies, vol. 38, no. 4, pp. 717-732.

Kloosterman, RC & Musterd, S 2001, 'The Polycentric Urban Region: Towards a research agenda', Urban Studies, vol. 38, no. 4, pp. 623-633.

Knox, PL & McCarthy, L 2005, Urbanization, Second edn, Pearson Prentice Hall,

Lefèvre, B 2009, 'Urban Transport Energy Consumption: Determinants and Strategies for its Reduction - An analysis of the literature.', SAPIENS, vol. 2, no. 3 (Cities and Climate Change),.

Legras, S & Cavailhès, J 2011, 'Urban form and sustainable development', 18th Annual Conference of the EAERE.

Lin, D, Allan, A, Cui, J & McLaughlin, R 2012, 'The Effects of Polycentric Development on Commuting Patterns in Metropolitan Areas, paper presented in the Regional Studies Association Global Conference 2012, 24-27th June, 2012, Beijing, China.

Locantore, JK, Montagu, AS, Rudy, SD & Sabina, EE 2009, 'Scenario Analysis Helps Identify Sustainable Land Use and Transportation Policies', Projections - MIT Journal of Planning, vol. 9, no. Sustainable Transportation - an International Perspective

References 238 Lynch, K 1981, A Theory of Good City Form, MIT Press, Massachusetts.

Maat, K & Stead, D 2005, 'The Effects of Proximity to Infrastructure on Employment Development: Preliminary Evidence from the Netherlands', in Williams, K (ed), Spatial Planning, Urban Form and Sustainable Transport, Ashgate, Hampshire.

Mahmud, MA (ed.) 2008, Insufficient Enforcement of Planning Standards by RAJUK Hindered the Growth Management and Planned Development of Dhaka City, Bangladesh Institute of Planners, Dhaka.

Mahmud, MA 2007, Corruption in Plan Permission Process in RAJUK: A study of Violations and Proposals, Dhaka.

Mahmud, MA 2007, Corruption in Plan Permission Process in RAJUK: A Study of Violations and Proposals, Transparency International Bangladesh, Dhaka.

Mahtab-uz-zaman, QM 1993, 'Consolidation as a Response to Urban growth - a case in Dhaka', Unpublished Masters Thesis thesis, University of HongKong, Hongkong.

Mahtab-uz-Zaman, QM, Mallick, FH, Abdullah, AQM & Ahmad, J 2006, 'In search of a Habitable Urnam Space-Built Ratio: a case staudy of building and planning regulation in Dhaka City', in Bay, J-H & Ong, B-L (eds), Tropical Sustainable Architecture - social and environmental dimensions, Elsevier.

Malik, A 2003, 'Post-Colonial Capitals of South Asia: A Critical Analysis of Chandigarh, Dhaka and Islamabad', Global Built Environment Review (GBER), vol. 3, no. 1, pp. 68-80.

Maoh, HF & Kanaroglou, PS 2004, 'Economic Clustering and urban Form: the Case of Hamilton, Ontario',

Mappery 2000, Growth of Megacities Map

Maria, SI & Imran, M 2006, 'Planning of Islamabad and Rawalpindi: What Went Wrong', 42nd ISoCaRP Congress, Istanbul, Turkey

Masuya, Y, Shitamura, M, Tamura, T, Saito, K & Black, J 2010, 'Optimal Commuting Assignment Problem with Travel Preference Functions: A study of the location of residences and employment and trip lengths in cities of Hokkaido, Japan', Journal of the Eastern Asia Society for Transportation studies, vol. 8.

McDonald, JF & Prather, PJ 1994, 'Suburban Employment Centres: The case of Chicago', Urban Studies, vol. 31, no. 2, p. 18.

References 239 McDonald, JF 1987, 'The Identification of Urban Employment Subcenters', Journal of Urban Economics, vol. 21, pp. 242-258.

McGee, T 2001, 'Urbanization Takes on New Dimensions in Asia’s Population Giants', Population Today, vol. 29, no. 7.

McGee, TG & Greenberg, C 1992, 'The Emergence of Extended Metropolitan Regions in ASEAN', ASEAN Economic Bulletin, vol. 9, no. 1, pp. 22-44.

McGee, TG 1991, 'The emergence of desakota regions in Asia: expanding a hypothesis', in Ginsburg, NJ & Koppel, B (eds), The extended metropolis: settlement transition in Asia, University of Hawaii Press, Honolulu, pp. 3-26.

McGee, TG 2004, 'Labour force change and mobility in the extended metropolitan regions of Asia', in Fuchs, RJ (ed), Mega-city growth and the future United Nations University, pp. 62-102.

McGee, TG 2008, 'Managing the rural–urban transformation in East Asia in the 21st century', Sustainability Science, vol. 3, pp. 155-167.

McGill, R 1998, 'Viewpoint: urban management in developing countries', Cities, vol. 15, no. 6, pp. 463-471.

McMillen, DP & Lester, TW 2003, 'Evolving subcenters: employment and population densities in Chicago, 1970-2020', Journal of Housing and Economics, vol. 12, pp. 60-81.

McMillen, DP & McDonald, JF 1997, 'A nonparametric analysis of employment density in a polycentric city', Journal of Regional Science, vol. 37, no. 4, pp. 591-612.

McMillen, DP & McDonald, JF 1998, 'Suburban Subcenters and Employment Density in Metropolitan Chicago', Journal of Urbnan Economics, vol. 43, no. 2, pp. 157-180.

McMillen, DP & Smith, SC 2003, 'The Number of Subcenters in Large urban Area', Journal of Urban Economics, vol. 53, no. 3, pp. 321-338.

McMillen, DP 2001a, 'Nonparametric Employment Subcenter Identification', Journal of Urban Economics, vol. 50, no. 3, pp. 448-473.

McMillen, DP 2001b, 'Polycentric Urban Structure: the case of Milwaukee', Economic Perspective, vol. 25, no. 2, pp. 15-27.

McMillen, DP 2003, 'Identifying Sub-centers Using Cotiguity Matrices', Urban Studies, vol. 40, no. 1, pp. 57-69.

Meier, RL & Quium, ASMA 1991, 'A Sustainable State for Urban Life in Poor

References 240 Societies:Bangladesh', Futures, vol. March.

Mills, ES 1969, 'Transportation And Patterns Of Urban Development - An Aggregative Model of Resource Allocation in A Metropolitan Area', American Economic Association, vol. 57, no. 2.

Ministry of Environment and Forests & Ministry of Communication 2010, ‘Developments in EST in Bangladesh’ presented in the Fifth Regional Environmentally Sustainable Transport (EST) Forum in Asia, 23-25 August 2010, United Nations Centre for Regional Development (UNCRD), Bangkok, Thailand

Minoprio and Spencely & P.W. Macfarlane 1959, Master Plan for Dacca, Minoprio and Spencely and P.W. Macfarlane, London.

Modarres, A 2003, 'Polycentricity and Transit Services', Transport Research Part A, vol. 37, pp. 841-864.

Mohsin, KM 1989, 'Commercial and Industrial Aspects of Dhaka in the Eighteen Century', in Ahmed, SU (ed), Dhaka Past Present Future, The Asiatic Society of Bangladesh, Dhaka, pp. 24-42.

Montgomery, MR 2008, 'The Urban Transformation of the Developing World', Science, vol. 319, p. 761.

Morichi, S & Acharya, SR (eds) 2012, Transport Development in Asian Megacities - a new perspective, Springer,

Mowla, QA 2003, 'Urbanization and the Morphology of Dhaka - a historical perspective', Journal of Asiatic Society, Bangladesh, vol. 48, no. 1, pp. 145-170.

Mowla, QA 2007, 'A Review of Dhaka's Urban Morphology', The Jahangirnagar Review, Part II: Social Sciences, vol. XXXI, pp. 15-29.

Mun˜iz, I & Galindo, A 2005, 'Urban form and the ecological footprint of commuting. The case of Barcelona', Ecological Economics, vol. 55, pp. 499– 514.

Muñiz, I, Garcia-López, MÀ & Galindo, A 2008, 'The Effect of Employment Sub-centres on Population Density in Barcelona', Urban Studies, vol. 45, no. 3, pp. 627-649.

Munshi, T, Brussel, M & Zuidgeest, M 2008, 'Developing a Geo-Spatial Urban Form - Travel Behaviour Model for the City of , India', CODATU XIII, Sustainable Development Challenges of Transport in Cities of the Developing World: Doing what works.

References 241 .

Murphy, RE 2009, The Central Business District - a study in urban geograpghy, Aldine-Atherton, Chicago.

Muth, RF 1969, Cities and housing, University of Chicago Press, Chicago.

Næss, P & Sandberg, SL 1996, 'Workplace Location, Modal Split and Energy Use for Commuting Trips', Urban Studies, vol. 33, no. 3, pp. 557-580.

Naess, P 2006, Urban Structure Matters, London.

Nahrin, K 2008, 'Violation of Land Use Plan and Its Impact on Community Life in Dhaka City', Jahangirnagar Planning Review, vol. 6, pp. 39-47.

Nelson, JM 1976, 'Sojourners versus New Urbanities: Causes and Consequences of Temporary versus Permanent Cityward Migration in Developing Countries', Economic Development and Cultural Change, vol. 24, no. 4, pp. 721-757.

Newman, PWG & Kenworthy, JR 1991, 'Transpport and urban form in thirty-two world's principal cities', Transport Reviews, vol. 11, no. 3, pp. 249-272.

Newton, P 2001, 'Urban Form and Environmental Performance', in Williams, K, Burton, E & Jenks, M (eds), Achieving Sustainable Urban Form, Spon Press, London.

Nilufar, F 2010, 'Urban Morphology of Dhaka city: Spatial Dynamics of Growing City and the Urban Core', International Seminar on the Celebration of 400 Years of the Capital Dhaka.

Nowlan, DM & Stewart, G 1991, 'Downtown population growth and commuting trips', Journal of the American Planning Association, vol. 57, no. 2, pp. 165-182.

Odland, J 1978, 'The Conditions For Multi-Center Cities', Economic Geography, vol. 54, no. 3, pp. 234-244.

OECD-ITF 2010, Reducing Transport Greenhouse Gas Emissions - Opportunities and Costs (priliminary findings), OECD-International Transport Forum.

Oliveira, MF & Silva, LT 2011, 'The influence of urban form in urban noise propagation', International Journal of Energy and Environment, vol. 5, no. 3, pp. 356-363.

Pacione, M 2005, Urban Geography: A Global Perspective, Routledge, London.

References 242 Pan, Q & Ma, L 2004, 'Employment Subcenter Identification: a GIS-based method ', ESRI User Conference 2004, ESRI.

Parolin, BP & Kamara, S 2003, 'Spatial patterns and functions of employment centres in Metropolitan Sydney, 1981 - 1996', State of Australian Cities.

Patterns: an international review and evaluation', European Journal of Transport and Infrastructure Research, vol. 1, no. 2, pp. 113-141.

PPK Consultants Pty Ltd, A, Declan International Corporation, C & Development Design Consultants, B 1994, Greater Dhaka Metropolitan Area Integrated Transport Study.

Prashker, J, Shiftan, Y & Hershkovitch-Sarusi, P 2008, 'Residential choice location, gender and the commute trip to work in Tel Aviv', Journal of Transport Geography, vol. 16, pp. 332–341.

Qureshi, IA & Lu, H 2007, 'Urban Transport and Sustainable Transport Strategies: A Case Study of Karachi, Pakistan', Tsinghua Science and Technology, vol. 12, no. 3, pp. 309-317.

Rahman, MM 2004, 'Regionalization of Urbanization and Spatial Development: Planning Regions in Bangladesh', The Journal of Geo--Environment, vol. 4, pp. 31-46.

Rahman, MS 2008, 'Future Mass Rapid Transit in Dhaka City: Options, Issues and Realities', Jahangirnagar Planning Review, vol. 6, pp. 69-81.

Rahman, MS, Fujiwara, A & Zhang, J 2008, 'Development of Travel Mode Choice for Commuting trips in Dhaka City', Proceedings of the International Conference on Sustainable Transport for Developing Countries (STDC) 2008, Department of Civil Engineering, BUET, Dhaka, Bangladesh . Available at SSRN: http://ssrn.com/abstract=1969672.

Rajdhani Unnayan Kartripakkha (RAJUK), 1995, Dhaka Metropolitan Development Plan (1995-2015): Structure Plan, Master Plan and Detailed Area Plan for Dhaka City, 1995, Dhaka

Rajdhani Unnayan Kartripakkha (RAJUK), 2013, 'History of RAJUK', viewed 01/06/2013,.

Rana, MMP 2011, 'Urbanization and sustainability: challenges and strategies for sustainable urban development in Bangladesh', Environment Development Sustainability, vol. 13, pp. 237–256.

Redfearn, CL 2007, 'The Topography of Metropolitan Employment: Identifying centers

References 243 of Employment in a Polycentric Urban Area', Journal of urban Economics, vol. 61, pp. 519-541.

Richardson, HW, 'Monocentric vs. Policentric Models: the future of urban economics in regional science', Western Regional Science Association, 27th Annual Meeting, California.

Richarson, HW 1988, 'Monocentric vs. policentric models: The future of urban economics in regional science', Presential Address to the Western Regional Science Association twenty-Seventh Annual Meeting, Napa Valley, California; The Annals of Regional Scince, vol. 22, no. 2, pp. 1-12.

Riguelle, F, Thomas, I & Verhetsel, A 2007, 'Measuring urban polycentrism: a European case study and its implications', Journal of Economic Geography, vol. 7, pp. 193-215.

Roberts, B & Kanaley, T 2006, 'Urbanization and Sustainability: Overview', in Roberts, B & Kanaley, T (eds), Urbanization and Sustainability: Case Studies of Good Practice, Asian Development Bank.

Robinson, IM (ed.) 1995, Emerging Spatial Patterns in ASEAN-Urban Regions: Alternative Strategies, UBC Press, Vancouver.

Rodrigue, J-P, Comtois, C & Slac, B 2003, The geography of Transport System, Routledge, New York

Roo, Gd & Miller, D 2000, Compact Cities and Sustainable Urban Development: A critical Assessment of policies and Plans from an International Perspective, Ashgate Publishing, Aldershot.

Roo, Gd (ed.) 2000, Compact Cities, Environmental Conflict and Policy Strategies: Complexity as a Criteria for Decision Making, Spon Press, London.

Rosenshein, L, Scott, L & Pratt, M 2011, 'Finding a Meaningful Model', ArcUser Online, vol. Winter,

Rostami, S 2005, 'Application of the transport needs concept to rural New South Wales: a GIS-based analysis', The University of New South Wales, Sydney.

Rouf, MA & Jahan, S 2001, 'Urban Centres in Bangladesh: Trends, Patterns and Characteristics', 37th International ISoCaRP Congress .

Roy, M 2009, 'Planning for Sustainable Urbanization in Fast growing Cities: Mitigation and Adaptation Isuues Addresses in Dhaka, bangladesh', Habitat International,

References 244 vol. 33, pp. 276-286.

Rozenblat, C & Tissandier, P 2008, 'Commuters Graphs and Cities' Polycentric Cohesion'.

Safi, S 2010, 'Implementation of Dhaka City Master Plans', The Daily Star, 23/02/2010. viewed 23/02/2010, .

Saito, F 1993, 'A Continuing Role for Rickshaws in Dhaka, Bangladesh', Canadian Journal of Development Studies, vol. 14, no. 2, pp. 281-293.

Sammons, R & Hall, P 1977, 'Urban Structure and Modal Split in the Journey to Work', Urban Studies, vol. 14, pp. 1-9.

Sanchez, TW 1999, 'The Connection Between Public Transit and Employment - the cases of portland and Atlanta', Journal of the American Planning Association vol. 65, no. 3, pp. 284-296.

Schipper, L, Fabian, H & Leather, J 2009, Transport and Carbon Dioxide Emissions: Forecasts, Options Analysis, and Evaluation, Asian Development Bank. .

Schwanen, T, Dijst, M & Dieleman, FM 2004, 'Policies for Urban Form and their Impact on Travel: the Netherlands Experience', Urban Studies, vol. 41, no. 3, pp. 579-603.

SENES Consultants Limited 2007, Dhaka Metropolitan Development Plan Strategic Environmental Assessment, Prepared for World Bank, Dhaka

Shankland Cox Partnership and Others 1981, Dacca Metropolitan Area Integrated Urban Development Project, Planning Commission, Dhaka.

Shearmur, RG & Coffey, WJ 2002, 'Urban Employment Subcenters and Sectoral Clustering in Montreal: complementary approaches to the study of urban form', Urban Geography, vol. 23, no. 2, pp. 103-130.

Shen, Q 2000, 'Spatial and Social Dimensions of Commuting', Journal of the American Planning Association vol. 66, no. 1, pp. 68-82.

Shen, Q 2001, 'A Spatial Analysis of Job Openings and Access in a U.S. Metropolitant Area', American Journal of the Planning Association, vol. 67, no. 1, pp. 53-67.

Shin, S 2002, 'Spatial Dimension of Workplace and Effects on Commuting: The Case of Metropolitan Dallas-Fort Worth', Urban and Regional Science, Texas A&M

References 245 University, Texas.

Shirgaokar, M 2012, 'Employment Centers and Middle-Class Travel Behavior: The Case of Mumbai', Transportation Research Board 91st Annual Meeting.

Shukla, V & Waddell, P 1991, 'Firm location and land use in discrete urban space : A study of the spatial structure of Dallas-Fort worth', Regional Science and Urban Economics, vol. 21, no. 2, pp. 225–253.

Small, KA & Song, S 1994, 'Population and Employment Densities: Structure and Change', Journal of Urban Economics, vol. 36, pp. 292-313.

Smith, DA 2009, Polycentricity and Sustainable Development: a Real Estate Approach to Analysing Urban Form and Function in Greater London, http://blog.casa.ucl.ac.uk/wp-content/uploads/2009/07/Smith.pdf, viewed 27/04/2010,

Sohn, J 2005, 'Are Commuting Patterns a Good Indicator of Urban Spatial Structure?', Journal of Transport Geography, vol. 13, pp. 306-317.

Sorensen, A 2001, 'Subcentres and Satellite Cities: Tokyo’s 20th Century Experience of Planned Polycentrism', International Planning Studies, vol. 6, no. 1, pp. 9 - 32.

Srinivasan, S 2010, 'Linking Travel Behavior and Location in , China - Geographically Weighted Approach', Transportation Research Record: Journal of the Transportation Research Board, vol. 2193, pp. 85-95.

Stead, D & Marshall, S 2001, 'The Relationships between Urban Form and Travel

Stubbs, JR & Clarke, G (eds), 1996, 'Megacity Management in the Asian and Pacific Region', Megacity Management in the Asian and the Pacific, Istanbul.

Susilo, YO & Kitamura, R 2008, 'Structural Changes in Commuters' daily Travel: the case of auto and transit commuters in the Osaka metropolitan area Japan, 1980-2000', Transport Research Part A, vol. 42, pp. 95-115.

Talukder, SH 2006, 'Managing Mega Cities: A case Study of Metropolitan Regional Goverenace for Dhaka', PhD Thesis, Murdoch University, Perth.

The Louis Berger Group, I & Ltd, BC 2005, Strategic Transport Plan for Dhaka, Dhaka Transport Coordination Board, Dhaka.

Tole, L 2008, 'Changes in the built vs. non-built environment in a rapidly urbanizing region: A case study of the Greater Toronto Area', Computers, Environment and Urban Systems, vol. 32, pp. 355-364.

Townsend, C 2003, 'In Whose Interest? A Critical Approach To Southeast Asia’s Urban

References 246 Transport Dynamics', PhD Thesis, Murdoch University, Perth.

Tresserra, JM 2011, 'Polycentrism and emerging sub-centres in the restructuring of metropolitan systems. The case of the Barcelona Metropolitan Region (RMB). ', 58th Annual North American Meetings of the Regional Science Association International (NARSC-RSAI), Miami- Florida.

Tresserra, JM 2012, 'Towards a methodology to identify and characterize urban sub-centres: Employment Entrophy Information versus Employment Density', Regional Studies Association European Conference 2012, May 2012, Delft University of Technology, NL, Miami- Florida.

United Nations (UN) 2003, World Urbanization Prospect: the 2003 revision data tables and highlights, Department of Economic and Social Affrairs, Population Division, United Nations, , United Nations, New York.

United Nations (UN) 2003, World Urbanization Prospects - The 2003 Revision Data Tables and Highlights, United Nations, New York.

United Nations Population Funds (UNFPA) 2011, The State of World Population 2011, New York.

University Grants Commission (UGC) 2012, University Grants Commission of Bangladesh.

Vichiensan, V 2005, 'Dynamic of urban Structure in Bangkok - based on employment cluster and commuting pattern ', Journal of the Eastern Asia Society for Transportation Studies, vol. 7, pp. 1559-1574.

Vichiensan, V, Páez, A, Kawai, K & Miyamoto, K 2006, 'Nonstationary Spatial Interpolation Method for Urban Model Development', Transportation Research Record: Journal of the Transportation Research Board, vol. 1977, pp. 103-111.

Wang, F 2000, 'Modeling Commuting Patterns in Chicago in a GIS Environment: a job accessibility perspective', Professional Geographer, vol. 52, no. 1, p. 14.

Wheaton, WC 1974, 'A Comparative Static Analysis of urban Spatial Structure', Journal of Economic Theory, vol. 9, pp. 223-237.

Willcox, DL 1996, 'Master Planning: Dhaka, Bangladesh - before 1985', Tales from the Experience of Urban Development in Asia - a chronic of learning by doing, Development Planning Unit, University College London, London.

Williams, K 2005, 'Spatial Planning, Urban Form and Sustainable Transport: An Introduction', in Williams, K (ed), Spatial Planning, Urban Form and Sustainable Transport, Ashgate, Hampshire.

References 247 Williams, K, Burton, E & Jenks, M 2000, 'Achieving Sustainable Urban Form: an Introduction', in Williams, K, Burton, E & Jenks, M (eds), Achieving Sustainable Urban Form, Taylor & Francies, London.

World Bank (WB) 2009, System of Cities - Harnessing Urbanization for Growth and Poverty Alleviation, The World Bank Urban and Local Government Strategy, Finance, Economics & Urban Department, Sustainable Development Network, World_Bank

World Bank (WB) 2007, Dhaka: Improving Living Conditions for the Urban Poor, Bangladesh Development Series, Paper No. 17, The World Bank Office Dhaka.

World Bank (WB) 2012, Urban Development Data, World Bank

World Bank 2007, Bangladesh: Strategy for Sustained Growth, Bangladesh Development Series, Paper No. 18, WB, Dhaka.

World Health Organization (WHO) 2012, OAP Database,

Yokoharia, M, Takeuchib, K, Watanabec, T & Yokota, S 2000, 'Beyond greenbelts and zoning: A new planning concept for the environment of Asian mega-cities', Landscape and Urban Planning, vol. 47, pp. 159-171.

Zaman, QMMU & Lau, SSY 2000, 'City Expansion Policy versus Compact City Demand - the case of Dhaka', in Jenks, M & Burgess, R (eds), Compact Cities: Sustainable urban forms for Developing Countries, Spon Press.

Zhao, F & Park, N 2004, 'Using Geographically Weighted Regression Models to Estimate Annual Average Daily Traffic', Transportation Research Record: Journal of the Transportation Research Board, , vol. 1879, pp. 99–107.

Zhao, P, Roo, Gd & Lu, B 2009, 'Planning for low-carbon urban growth: the impact of urban form on transport and the environment on the urban fringe of Beijing', Low Carbon Cities; 45th ISOCARP International Congress, Porto - Portugal, . .

References 248

Appendices

Appendix A: List of Roads and Highways Under Construction and Proposed Projects 250

Appendix B: Maps2004 254

Appendix C: Maps 2024 262

Appendix A

List of Roads and Highways Under Construction and Proposed Projects

2024 Traffic Sl. Type & Current Description Flow (pcus/hr No. Specification Status per direction)

1 Gulshan-1 to Badda (Pragoti Saroni) 2-lane dual 3,000 Ongoing

Argagaon Road () to Mirpur 2 2-lane dual 500 to 2,500 Under Process Section 2 through Senpara Parbata

Bangla College to Kafrul intersecting Rokeya 3 2-lane dual 1,000 to 3,000 Under Process Sharoni

Upgrading 4 Ashulia to Aricha Road (C & B More) 2-lane 500 needed

From Airport Road near Khilkhet to First Balu 5 Bridge (Isapura) via Baruna (Nikunja - 2-lane dual 2,000 Upgrading Yousufganj)

From first Balu Bridge near Tek Noadda to 6 2-lane dual 500 Upgrading Sitalakhya River near Kanchan

Under 7 Zia Colony to Mirpur Cantonment - Pallabi 2-lane dual 4,500 process

Bijoy Saroni to Tongi Diversion Road (from Proposed 8 2-lane dual 2,500 Rangs Y Jn.) alignment

9 Konakhola to Hazratpur 2-lane 1,000 Upgrading

10 Hazratpur to Hemayetpur 2-lane 1,000 Upgrading

Sonargaon (Panthapath) to Rampura Bridge Proposed 11 2-lane dual 3,000 () via duct

Tunnel connecting Shahid Gate 12 3-lane dual 4,000 Tunnel (Cantonment) & Rokeya Saroni

Merul Badda - Babur Jaiga - Balirpar - Parain - 13 2-lane dual 2,500 Upgrading Golakandial

Appendix A: List of Roads and Highways Under Construction and Proposed Projects 250 Bashaboo Jame Mosque to Trimohini 14 2-lane dual 2,000 Proposed Ghdaraghat via Shekker Jaiga Bridge

15 Tongi Ghorashal Road (via Kaliganj) 2-lane 500 Upgrading

16 Fatulla - Muktarpur - Munshiganj road 2-lane 1,000 Upgrading

Upgrading 17 Hemayetpur - Singair - Manikganj road 2-lane 1,000 proposed

Circular Road (ring road) over embankment 18 2-lane dual 1,000 to 2,000 Upgrading (Ashulia to Buriganga 2nd bridge (Babu bazar))

Connecting roads to Keraniganj, Nawabgonj & Proposed 19 Dohar from Buriganga 3rd Bridge (both side 2-lane dual 1,000 alignment approaches)

20 Dhaka Bypass 2-lane 500 Ongoing

NMT road from BDR Rifles gate (Jigatala) to New 21 New Market (problem with Bangladesh Rifles 1-lane Not in Network connection Project)

Progati Sarani (Baridhara - Beraid - Balu River 22 2-lane 500 Proposed - Murapara) to Bhulta (Nawabganj)

Eastern Bypass [Joydebpur (Board bazar) - 23 Mausaid - Barun - Baburjaiga - New 3-lane dual 1,500 to 3,000 Proposed Narayanganj] road (Chittagong Highway)

Malibagh Chowrasta Gulbagh - Janapath 24 2-lane dual 2,000 Proposed (Bishaw road)

Zikatala - Hazaribagh (Sikder Medical College) 25 2-lane 1,000 Upgrading road

Link road along rail track between Moghbazar 26 2-lane dual 2,500 Proposed and Malibagh rail crossing

Mohammadpur Bus Stand Embankment 27 2-lane dual 1,000 Proposed Berry Bandh) - upgradation

Mohammadpur Shia Mosque (near Japan 28 Garden City) - Mohammadpur Bus Stand 2-lane dual 1,000 Proposed (Widening)

Rasulpur Bridge (Embankment) - Peelkhana 29 road - Azimpur Old Grave yard Eden Girls 2-lane 1,000 Upgrading Cillege

Western By-pass from Dhaka EPZ (Savar) to Abdullapur (Dhaka - Mawa Highway) via 30 2-lane dual 500 to 1,000 Proposed Hemayetpur Bhakurta Taranagar - kalindi

Motijheel Shapla Chattar to Kamlapur Railway 31 2-lane 1,500 Proposed Station (Widening)

32 Pallabi (Mirpur) to Uttara 3rd Phase up to Tongi 2-lane dual 1,500 Proposed

Appendix A: List of Roads and Highways Under Construction and Proposed Projects 251 - Ashulia road (Dhour) alignment

Uttara Sector-10 to Embankment road to the Proposed 33 2-lane dual 2,000 West alignment

Construction of Muktarpur bridge over river Under 34 2-lane dual 1,000 Dhaleswari process

Construction of Buriganga 3rd bridge near Under 35 3-lane dual 1,000 Basila process

Construction of 3rd bridge over Sitalakkhya at Under 36 3-lane dual 2,500 Demra process

Construction of bridge over Balu river at New 37 2-lane dual 2,000 Keodata proposal

Mohammadpur Krishi Market to Mirpur Road 39 2-lane 1,000 Proposed (Sohrawardy Hospital)

Krishi Market & Baitul Aman (Y Junction) to 40 2-lane 1,500 Proposed Embankment to the west

Mirpur-14 (Sagorika) to Airport Road (Banani 42 Railway Station) along the fringe of Kurmitola 2-lane dual 2,500 to 3,500 Proposed Golf Course

Mirpur Zoo to Embankment (Berry Bund) to the 43 2-lane 1,000 Proposed west

Further 44 Elevated Express way 2-lane dual 3,000 to 5,000 Study

45 Bashabo Kadamtola Road up to Manikdi 2-lane dual 2,000 Upgrading

46 Dayaganj - Postogola (Jurain) 2-lane dual 1,500 Ongoing

Road over Dholai Khal (Box-culvert) from 47 2-lane 1,500 Ongoing Loharpul (Gandaria) to Dayaganj

Road over Debdulai Khal (Box-culvert) from 48 2-lane 1,000 Ongoing Dayaganj to Dholaipar

Road over Gerani Khal (Box-culvert) from 49 2-lane 1,000 Completed Dayaganj to Dholpur

50 Construction of Mohakhali Flyover 2-lane dual 3,000 Completed

51 Construction of Khilgaon Flyover 2-lane single 1,500 Completed

52 Jatrabari - Gulistan Flyover 2-lane dual 2,500 Ongoing

Dhalka (Postogola) - Fatullah - Narayanganj 53 2-lane dual 500 (Chashara) road improvement

Eastern Embankment along Balu river, Tongi, 54 2-lane dual 500 to 1,000 Proposed Demra (Chittagong Highway)

Uttara Sector-8 to Uttar khan Trimukh (Balu 55 2-lane 1,500 Proposed river embankment)

Appendix A: List of Roads and Highways Under Construction and Proposed Projects 252 56 Uttara Sector-4 - Dakhinkhan - Khordi 2-lane 500 Proposed

57 Nabinagar - EPZ - Chandra road improvement 2-lane dual 500 Proposed

Jatrabari - Khanchpur bridge (widening of 58 4-lane dual 1,500 Proposed polder road to 8 lane)

59 Berulia (Dhour) - Ashulia - EPZ road 2-lane dual 1,000 Proposed

Dhaka (link road) - Narayanganj road up to 62 2-lane dual 1,500 Proposed IWT Terminal with an over-pass

Pallabi to west Embankment via North 63 2-lane 1,000 Proposed Rupnagar

Further 64 Flyover at Malibagh-Moghbazar (Combined) 2-lane dual 2,500 Study

Jatrabari crossing to Demra Ghat (Sitalakhya 69 3-lane dual 2,500 Proposed 3rd bridge) road

Road connecting Buriganga 1st and 2nd bridges 70 via Subhadia & Zinjira (South of Buriganga 2-lane dual 1,000 Proposed river)

Construction of the Link Ramna Star gate to 73 2-lane dual 2,000 Completed Notre Dam College Road

BIWTA Pangaon Inland Container Terminal to 74 2-lane dual Not in Network Proposed First Buriganga Bridge Road (Toll Road)

Traffic Manageemnt works and small scale Proposed improvements

Regional Highway Development including Proposed Inventory, Safety Audit and works

Preliminary Engineering as preparatory to all

phases

Source: The Louis Berger Group, I. and B. C. Ltd (2005)

Appendix A: List of Roads and Highways Under Construction and Proposed Projects 253

Appendix B Maps

This appendix contains maps generated from spatial analysis of travel behaviour of the base year (2004).

Appendix B: Maps 254

Figure B.1: Commuting Desireline

Appendix B: Maps 255

Figure B.2: Commuting Desireline

Appendix B: Maps 256

Figure B.3: Commuting Desireline

Appendix B: Maps 257

Residential

Mixed

Commercial

Educational

Figure B.4: Building Footprint in Old Dhaka

Appendix B: Maps 258

Residential

Mixed

Commercial

Educational

Figure B.5: Building Footprint in Gulshan

Appendix B: Maps 259

Residential

Mixed

Commercial

Educational

Figure B.6: Building Footprint in Motijheel

Appendix B: Maps 260

Residential

Mixed

Commercial

Educational

Figure B.7: Building Footprint in Dhanmondi

Appendix B: Maps 261 Appendix C Maps for Growth Scenarios

This appendix contains maps generated from spatial analysis of job distribution and travel behaviour of the year 2024.

Appendix C: Maps for Different growth Scenarios 262

Employment

Baseline

Population

A Ratio Employment Density Employment

B

Figure C.1: Employment Density and Employment Population in Different Scenarios (line 1)

Appendix C: Maps for Different growth Scenarios 263

Employment

C

Population Ratio

D

Employment Density Employment

E

Appendix C: Maps for Different growth Scenarios 264

Figure C.2: Job density in different scenarios (line 1)

Baseline

2A

Appendix C: Maps for Different growth Scenarios 265

Employment

2B

Population Ratio

2C Employment Density Employment

2D

Appendix C: Maps for Different growth Scenarios 266

2E Figure C.3: Job density in different scenarios (line 2)

Appendix C: Maps for Different growth Scenarios 267

Figure C.4: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 268

Figure C.5: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 269

Figure C.6: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 270

Figure C.7: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 271

Figure C.8: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 272

Figure C.9: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 273

Figure C.10: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 274

Figure C.11: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 275

Figure C.12: Commuting Pattern in Different Scenarios

Appendix C: Maps for Different growth Scenarios 276

Table C. 1: Commuting Distance in Different Scenarios Commuting Distance Baseline A B C D E 2A 2B 2C 2D 2E 0 - 5 15.85% 15.16% 14.93% 14.77% 14.75% 14.27% 14.94% 14.75% 14.58% 14.52% 14.05% 5 - 10 26.50% 25.48% 25.21% 25.02% 24.96% 24.24% 25.39% 25.17% 25.06% 25.07% 24.50% 10 - 15 16.32% 17.39% 17.41% 17.72% 18.13% 18.66% 17.04% 17.09% 17.21% 17.34% 17.50% 15 - 20 13.58% 14.65% 14.78% 14.94% 15.02% 15.27% 14.87% 15.00% 15.15% 15.21% 15.46% 20 -25 7.60% 8.35% 8.41% 8.62% 8.82% 9.19% 8.04% 8.12% 8.20% 8.24% 8.47% 25 - 30 7.01% 7.12% 7.22% 7.26% 7.21% 7.40% 7.16% 7.22% 7.26% 7.28% 7.51% 30+ 13.14% 11.84% 12.03% 11.67% 11.11% 10.98% 12.55% 12.65% 12.53% 12.34% 12.53% Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%

Table C. 2: Commuting Distance in Different Scenarios

Region Baseline A B C D E 2A 2B 2C 2D 2E Main City 78,161,197 60,348,699 58,376,792 55,878,775 54,196,162 48,798,350 59,526,420 57,762,854 55,628,495 54,221,963 49,525,082 Extended City 25,361,372 16,500,919 16,320,517 15,973,596 15,652,058 14,969,453 18,304,838 17,775,679 17,123,868 16,682,974 15,247,430 Inner Ring 47,963,075 66,723,753 70,247,349 69,343,588 64,599,086 62,825,745 66,272,178 69,459,895 68,847,258 64,856,349 63,653,126 Outer Ring 24,973,194 31,833,337 31,927,012 35,094,756 39,957,229 48,849,778 33,534,762 33,696,096 37,190,970 42,511,931 52,162,278 Total 176,458,838 175,406,708 176,871,671 176,290,715 174,404,536 175,443,327 177,638,199 178,694,524 178,790,591 178,273,216 180,587,916

Appendix C: Maps for Different growth Scenarios 277