ASSESSING DYNAMICS OF RURAL-URBAN LINKAGES AND THEIR INFLUENCE ON RURAL VULNERABILITY TO EXTREME FLOOD EVENTS CASE STUDY OF THREE RURAL FARMING COMMUNITIES IN PUNJAB, PAKISTAN Von der Fakultät für Bau- und Umweltingenieurwissenschaften der Universität Stuttgart zur Erlangung der Würde eines Doktor-Ingenieurs (Dr.-Ing.) genehmigte Abhandlung vorgelegt von Ali Jamshed geboren in Doha, Qatar aufgewachsen in Lahore, Pakistan Hauptberichter: Prof. Dr.-Ing. habil. Jörn Birkmann Mitberichter: Prof. Dr.-Ing. Stefan Greiving Tag der mündlichen Prüfung: 15.01.2021 Institut für Raumordnung und Entwicklungsplanung der Universität Stuttgart 2021 Dedicated to my parents, wife, daughter and teachers I DECLARATION I hereby declare that this doctoral dissertation is composed independently, and all the sources of information and material have properly acknowledged. Ali Jamshed Stuttgart, 15-03-2021 II PREFACE This dissertation is submitted in partial fulfilment of the requirements for obtaining the Ph.D. degree in Engineering (Doctor Ing.) at the University of Stuttgart, Germany. The research work described in this document was conducted and defended at the Institute of Spatial and Regional Planning (IREUS), Faculty of Civil and Environmental Engineering between April 2016 and January 2021 (see the timeline in Annex A) under the supervision of Prof. Dr.-Ing (habil) Joern Birkmann and Prof. Dr.-Ing Stefan Greiving. The research was funded by the Higher Education Commission of Pakistan (HEC) in cooperation with German Academic Exchange Services (DAAD) under “Faculty Development of MSc leading to PhD for Universities of Engineering, Science and Technology (UESTP)”. This dissertation is mainly based on the work published/accepted/in preparation to be published as research articles and a book chapter which include: Publication 1 Journal Article: Jamshed, A., Birkmann, J., Feldmeyer, D., Rana, I.A., (2020) “A Conceptual Framework to Understand the Dynamics of Rural-Urban Linkages for Rural Flood Vulnerability”. Sustainability 12 (7), 2894. DOI: https://doi .org/10.3390 /su1207289. Publisher: MDPI. (see Chapter 2) Publication 2 Book Chapter: Jamshed, A., Birkmann, J., McMillan, J., Rana, I.A. and Lauer, H. (2020) “The Impact of Extreme Floods on Rural Communities: Evidence from Pakistan” in Leal, W.F., Nagy, G., Borga, M., Chavez, D. and Magnuszewski, A. (Eds.), Climate Change, Hazards and Adaptation Options: Handling the impacts of a changing climate., Climate Change Management, 1st ed., Springer, Cham, 585-613. DOI: https://doi.org/10.1007/978-3-030-37425-9_30 Publisher: Springer. (see Chapter 4) Publication 3 Journal Article: Jamshed, A., Birkmann, J., Joanna, M.M., Rana, I.A., Feldmeyer, D., Sauter, H (2021) “How do Rural-urban Linkages Change After an Extreme Flood Event? Empirical Evidence from Rural Communities in Pakistan” Science of the Total Environment, Vol 750C, 141462. DOI: https://doi.org/10.1016/j.scitotenv.2020.141462. Publisher: Elsevier. (see Chapter 4 and 5) Publication 4 Journal Article: Jamshed, A., Birkmann, J., Rana, I.A., Joanna, M.M. (2020) “Relevance of City Size to the Vulnerability of Surrounding Rural Areas: An Empirical Study of Flooding in Pakistan” International Journal of Disaster Risk Reduction, Vol 48, 101601. DOI: https://doi.org/10.1016/j.ijdrr.2020.101601. Publisher: Elsevier. (see Chapter 6) Publication 5 Journal Article: Jamshed, A., Birkmann, J., Rana, I.A., Feldmeyer, D. (2020) “The Effect of Spatial Proximity to Cities on Rural Vulnerability against Flooding: An Indicator Based Approach” Ecological Indicators, Vol 118, 106704. III DOI: https://doi.org/10.1016/j.ecolind.2020.106704. Publisher: Elsevier. (see Chapter 6) Publication 6 Journal Article (in preparation): Jamshed, A., Greiving, S. Birkmann, J., Rana, I.A., “Dynamics of Rural-Urban Linkages due to Flooding and Their Influence on Vulnerability: Case of Pakistan” to be submitted in Science of the Total Environment. Publisher: Elsevier. (see Chapter 7) Research work already published/ in preparation to be published contribute to a different chapter of the thesis (see figure below). The introduction is based on the introduction section of all publications. Chapter 2 is based on publication 1, chapter 3 based on the methodology sections of all publications. Chapter 4 is based on publication 2 and 3. Chapter 5 is based on publication 3. Chapter 6 is based on publication 4 and 5. Chapter 7 is based on publication 6. Chapter 8 is based on the conclusion section of all the publications. Ali Jamshed Stuttgart, 15-03-2021 IV ABSTRACT Although rural areas and cities are intrinsically linked, the vulnerability of rural households and communities to hazards or extreme weather and climatic events is often assessed without considering their relationships to cities. These linkages are important due to interdependencies between rural and urban areas for socio-economic and physical growth. Moreover, extreme events can lead to dramatic shifts in societal processes, disrupt rural-urban linkages, and affect rural vulnerability; these matters need to be investigated. Considering these gaps in knowledge, this study aims to conceptualise and understand rural vulnerability with respect to the dynamics of rural-urban linkages in the case of flooding, with a special focus on spatial factors like city size and proximity to the city. To do so, a mixed methods approach was adopted in this research. Still, the present study is largely based on quantitative techniques. First, the current literature on rural-urban linkages, vulnerability and factors that influence them was critically reviewed, and a unified framework was proposed to connect the elements of rural-urban linkages and flood vulnerability. The framework was designed to examine changes in rural-urban linkages and the subsequent impact on rural vulnerability to flooding. For empirical research, three case studies (Darya Khan, Muzaffargarh, and Multan) were selected in the Punjab province of Pakistan. A multistage, mixed methods sampling approach was applied to derive 325 samples. Secondary data, observations and a focus group discussion deepened understanding of the topic. The household survey, using a structured questionnaire, was administered to collect information from the required sample, comprised of a flood-affected rural population surrounding three different-sized cities and at varied proximity. The data were analysed using descriptive statistics (frequency analysis, cross-tabulation) and inferential statistics (correlation, regression, chi-square, the Mann-Witney U test). Moreover, an index-based approach was developed to obtain the composite values of the three components of vulnerability: (1) exposure, (2) susceptibility and (3) capacity. V The findings show that flooding severely affects rural households both directly and indirectly. The ramifications have led to several changes among rural households; most notably, they have modified how they earn a living and their relationship with the nearest major city. Floods have shifted the flow of people, information, finances, goods, and services between rural and urban areas. The research indicates that rural-urban linkages are altered in that flooding both increases and decreases rural households’ dependence on cities in different ways. These outcomes are largely driven by socio-economic, spatial, and flood-related factors. In terms of vulnerability, first, the findings signal that rural populations surrounding smaller cities are less exposed, but more vulnerable, as compared to rural households that surround larger cities. This is because rural populations adjoining larger cities are better able to deal with flood hazards due to stronger linkages. Secondly, the results confirmed that distance to the city influences the vulnerability of surrounding farming households. Rural farming households located close to cities are less vulnerable, mainly due to a better transfer of services and facilities from cities, which has made such households more educated, informed, financially strong and more closely connected, with easier access to public and private institutions. Thus, city size and proximity to the city modify linkages that further impact the flood vulnerability of the rural population. Lastly, changes in linkages made by rural households following a flood influence their overall vulnerability differently; increasing linkages with the city after a flood reduce their vulnerability, while decreasing linkages with the city exacerbate it. These changes in linkages are used to adapt to future floods and affect rural households’ vulnerability both positively and negatively. Hence, the dynamics of linkages and rural households’ exchanges with cities are crucial to reducing their vulnerability to future flood hazards. This study paves the way for regional planners and disaster managers to establish synergies between them for devising integrated flood management and development strategies that strengthen linkages, mitigate disparities and curtail vulnerability. VI ZUSAMMENFASSUNG Obwohl ländliche Gebiete und Städte eng miteinander verbunden sind, wird die Vulnerabilität ländlicher Haushalte und Gemeinden für Gefahren oder extreme Wetter- und Klimaereignisse häufig ohne Berücksichtigung ihrer Beziehung zu Städten bewertet. Diese Verflechtungen sind aufgrund der gegenseitigen Abhängigkeiten zwischen ländlichen und städtischen Gebieten für die sozioökonomische und physische Entwicklung wichtig. Darüber hinaus können Extremereignissezu dramatischen Veränderungen in gesellschaftlichen Prozessen führen, die Verbindungen zwischen Land und Stadt beeinträchtigen und ländliche Verwundbarkeiten betreffen, die untersucht werden müssen. In Anbetracht dieser Wissenslücken zielt diese Studie darauf ab, die ländliche Verwundbarkeit in Bezug auf die Dynamik der Verflechtungen zwischen Land und Stadt bei Überschwemmungen zu konzipieren und zu verstehen. Hierbei stehen räumliche Faktoren, wie die Größe der Stadt und die Nähe zur Stadt, im Vordergrund. Zu diesem Zweck wurde ein Mixed-Method-Ansatz, mit einem starken Fokus auf quantitativen Methoden gewählt. Zunächst wurde die aktuelle Literatur zu urban-ruralen Verflechtungen, zur Vulnerabilität sowie zu Faktoren, die sie beeinflussen, kritisch überprüft. Es wird ein einheitlicher Rahmen vorgeschlagen, der die Elemente der Verflechtungen zwischen Land und Stadt und die der Anfälligkeit gegenüber Überschwemmungen miteinander verbindet. Der Rahmen dient dazu, Änderungen der urban-ruralen Verflechtungen und deren Einfluss auf die Vulnerabilität des ländlichen Raums bezüglich Überschwemmungen zu untersuchen. Für die empirischen Untersuchungen wurden drei Fallstudien (Darya Khan, Muzaffargarh und Multan) in der pakistanischen Provinz Punjab ausgewählt. Ein mehrstufiger Mixed-Method Stichprobenansatz wurde angewendet, um 325 Proben zu erhalten. Anhand von Sekundärdaten, Beobachtungen und Gruppendiskussionen aus der Haushaltsumfrage wird das Thema erfasst. Es wurde eine Haushaltsumfrage basierend auf einem strukturierten Fragebogen durchgeführt, um die erforderliche Stichprobe von der von Überschwemmungen betroffenen ländlichen Bevölkerung, die im Umfeld von drei VII verschiedenen Städten unterschiedlicher Größe und Lage leben, zu sammeln. Die Daten wurden unter Anwendung deskriptiver Statistiken (Frequenzanalyse, Kreuztabelle) und Inferenzstatistiken (Korrelation, Regression, Chi-Quadrat, Mann-Witney-U-Test) analysiert. Darüber hinaus wurde ein indexbasierter Ansatz entwickelt, um die zusammengesetzten Werte der drei Komponenten der Vulnerabilität zu ermitteln, nämlich Exposition, Anfälligkeit und Kapazität. Die Ergebnisse zeigen, dass ländliche Haushalte sowohl direkt als auch indirekt besonders schwer von Überschwemmungen betroffen sind. In Folge dieser Auswirkungen ist eine Reihe von Verhaltensänderungen bei ländlichen Haushalten zu beobachten – beispielsweise im Hinblick auf Einkommensquellen oder die alltäglichen Stadt-Land-Verflechtungen. Überschwemmungen führten zu einer grundlegenden Veränderung des Stadt-Land-Gefüges und den Beziehungen zwischen ländlichen und städtischen Gebieten im Hinblick auf den Strom von Menschen, Informationen, Finanzen, Waren und Dienstleistungen. Die Forschungsergebnisse verdeutlichen, dass sich die Verflechtungen zwischen städtischem und ländlichen Räum insofern verändert haben, als dass Überschwemmungen die Abhängigkeit der ländlichen Haushalte vom städtischen Gebiet auf unterschiedliche Weise sowohl erhöht als auch verringert hat. Diese Veränderungen sind weitestgehend auf sozioökonomische und räumliche Faktoren zurückzuführen sowie auf Faktoren, die in direktem Zusammenhang zur Überschwemmung stehen, wie der Verlust von landwirtschaftlichen Flächen. Im Hinblick auf die Verwundbarkeit lassen sich zwei zentrale Ergebnisse festhalten. Erstens sind ländliche Haushalte in der Umgebung kleinerer Städte im Vergleich zu ländlichen Haushalten, die in der Umgebung größerer Städte leben, weniger exponiert, aber verwundbarer. Dies lässt sich dadurch erklären, dass ländliche Haushalte in unmittelbarer Nähe zu größeren Städten stärkere Stadt-Land Verflechtungen haben und daher besser in der Lage sind, mit Überschwemmungsgefahren umzugehen. Zweitens bestätigten die Ergebnisse, dass die Entfernung zu den Städten die Verwundbarkeit der umliegenden landwirtschaftlichen Haushalte beeinflusst. Ländliche Haushalte, die in der Nähe dieser Städte leben, waren weniger verwundbar, was hauptsächlich auf den VIII besseren Transfer von Dienstleistungen und Einrichtungen aus den Städten zurückzuführen ist, wodurch sie besser ausgebildet, informiert und finanziell unabhängig sind und einen leichteren Zugang zu öffentlichen und privaten Institutionen haben. Abschließend lässt sich festhalten, dass die Stadtgröße und die Distanz ländlicher Haushalte zur Stadt das Beziehungsgeflecht zwischen Stadt und Land formen, welches wiederum die Verwundbarkeit ländlicher Haushalte beeinflusst. Verhaltensänderungen von ländlichen Haushalten aufgrund der Überschwemmung beeinflussen die Vulnerabilität unterschiedlich. Eine Intensivierung der Stadt-Land- Verflechtung verringerte die allgemeine Vulnerabilität, während gleichzeitig eine Abnahme von Verbindungen diese erhöhte. Die Verhaltensänderungen dienen zur Bewältigung von negative Folgen und Anpassung an zukünftige Überschwemmungen und können die Vulnerabilität von ländlichen Haushalten sowohl positiv wie negativ beeinflussen. Daher ist die Veränderung der Stadt-Land-Verflechtung von entscheidender Bedeutung für die zukünftige Verletzlichkeit von ländlichen Haushalten. Diese Studie ebnet den Weg zur Schaffung von Synergien zwischen Regionalplanern und Katastrophenmanagern für das Hochwassermanagements und allgemeine Entwicklungsstrategien, die Verbindungen stärken, Ungleichheiten verringern und die Verwundbarkeit reduzieren. IX ACKNOWLEDGEMENTS This research work has been made possible due to the contribution, coordination and assistance of many individuals and organizations. First, I would like to thank Allah (SWT) for his endless blessings, always being there for me and who granted me wisdom, strength and insistence for completing this research work. My heartiest gratitude and respect to my supervisor Prof. Dr. (Ing) habil Joern Birkmann for his invaluable guidance and continuous assistance during the Ph.D. process. His kind support, encouragement and appreciation during this time are unforgettable. I am thankful to him for trusting me and granting me the freedom to pursue my ideas and interests. I would also like to extend my sincere gratitude to my second supervisor, Prof. Dr.-Ing Stefan Greiving, for his constructive criticism and suggestions on my research. It was an honour to be part of his research group. Especially, I would also like to thank and mention the contributions of Dr. Irfan Ahmad Rana in my research, without his opinionated views and critical remarks this research could not have been shaped in its final form. Thanks for being my mentor and motivator throughout my Ph.D. I am extremely grateful to my institute colleagues for their enormous help, moral support and good company. I would like to mention Joanna McMillan for proofreading, editing and commenting on my papers and dissertation; Daniel Feldmeyer for providing help and insights into the analytical part of my papers and dissertation as well giving valuable comments on them; Hannes Lauer for his contribution in writing book chapter and proofreading dissertation; Britta Weisser for proofreading and providing the opportunity to work with her on RESI-Extreme project; Holger Sauter for his valuable ideas and input in one of the papers and dissertation chapters; Dr. Mohammad Ravankhah and Angela Wendnagel-Beck for proofreading. Thanks to the colleagues for providing German translation of my dissertation summary. I am also indebted to Iris Petersen and Dr. Richard Junsech for helping with administrative issues, as well as Peter Kindl for handling technical issues. I would also like to mention Linda Sorg for her support in several matters. I gratefully acknowledge the experts and officials of various international and local institutions in Pakistan for providing useful information to shape the research and questionnaire as well as coordination and support in field surveys. I would like to X mention persons from academics (Dr. Obaid Ullah Nadeem, Dr. Atif Bilal Aslam, Dr. Shaker Mehmood, Dr. Muhammad Asim, Dr. Shahina Tariq and Dr. Seeme Mallick); FAO (Muhammad Qadeer), Leads Pakistan (Muhammad Arif Goheer), UNOCHA (Muhammad Amjad), ADB (Saad Malik), SPDI (Shareef Hussain, Shafqat Muneer), PDMA, Punjab (Khurram Shahzad, Sidra Ikram), NRSP (Khurram Shahzad), PRSP (Shahid Nadeem), The Urban Unit (Nizam ud Din), and District Governments (Khizar Abbas, Qaswer Abbas, Muhammad Bilal), Punjab Rescue Services (Aqeel Suleman). I gratefully acknowledge the rural communities of Darya Khan, Muzaffargarh and Multan, who showed immense hospitality by opening their doors and extending their help for me in acquiring the ground information. I would also like to appreciate the efforts of Shafaat Nawaz and Dr. Atif Bilal Aslam in arranging field surveyors. I would like to thank Amir Faridi, Abdul Razzaq, Muhmmad Jahanzaib, Muhammad Sajid, Muhammad Inayat, and Uzair Sandela for their effort in data collection. I would also like to thank Muhammad Bilal Paracha (Transport specialist, World Bank) for helping me in conducting interviews from the various public sector and private sector institutes in Islamabad. Besides, I would like to thank Nimra Iqbal for providing relevant data on the socio-economic profile of districts, Usman Maqsood Mirza for his efforts in proofreading my dissertation and Hafiz Syed Rizwan and, Nikhil Agarwal for patronage. Moreover, it is a pleasure to pay tribute to my friends for their assistance and encouragement. I would also like to thank the Higher Education Commission (HEC) of Pakistan and German Academic Exchange Services (DAAD) for trusting on my abilities and selecting me for the scholarship. I would like to thank the International Office of the University of Stuttgart and the Institute of Spatial and Regional Planning for providing support in my difficult times. I owe a great debt of gratitude to my parents, wife, brothers and sisters, for their motivation and encouragement. Despite the geographical distance, they have always been with me as a source of light and inspiration. This major accomplishment of my life is made possible only because of their continuous support XI LIST OF ABBREVIATIONS BBC Bogardi, Birkmann and Cardona framework BSP Bureau of Statistics Punjab CRI Climate Risk Index DFID Department for International Development DoIP Directorate of Industries Punjab EM-DAT Emergency Events Database FFC Federal Flood Commission FPS Far Proximity Settlements GDP Gross Domestic Product GoP Government of Pakistan IFAD International Fund for Agricultural Development IPCC Intergovernmental Panel on Climate Change Km Kilometres KPK Khyber Pakhtunkhwa MC Municipal Corporation/Municipal Committees MoF Ministry of Finance MOVE Methods for the Improvement of Vulnerability Assessment in Europe NDMA National Disaster Management Authority NPS Near Proximity Settlements NRSP National Rural Support Program PBS Pakistan Bureau of Statistics PDMA Provincial Disaster Management Authority PDS Punjab Development Statistics PKR Pakistani Rupees PMD Pakistan Metrology Department PRSP Punjab Rural Support Program SCARV City Size and Rural Vulnerability framework SDGs Sustainable Development Goals SFDRR Sendai Framework for Disaster Risk Reduction SPDI Sustainable Policy Development Institute UC Union Council UN-DESA United Nations Department of Economic and Social Affairs UNDRR United Nation Office for Disaster Risk Reduction UNFCCC United Nation Framework Convention on Climate Change UNHABITAT United Nations Human Settlements Programme VPN Vulnerability Proximity Nexus framework WASH Water and Sanitation XII TABLE OF CONTENTS DECLARATION ............................................................................................................................I PREFACE ........................................................................................................................... II ABSTRACT ......................................................................................................................... IV ZUSAMMENFASSUNG ......................................................................................................... VI ACKNOWLEDGEMENTS ....................................................................................................... IX LIST OF ABBREVIATIONS .................................................................................................... XI TABLE OF CONTENTS .......................................................................................................... XII LIST OF FIGURES .................................................................................................................. XVI LIST OF TABLES ................................................................................................................. XVIII CHAPTER 1 INTRODUCTION .......................................................................................... 1 1.1 Background: Flood hazards and vulnerability .................................................................. 1 1.2 Motivation: Rural-urban linkages and rural vulnerability ............................................... 3 1.3 Pakistan and flooding ...................................................................................................... 7 1.4 Objectives and research questions ............................................................................... 10 1.5 Research approach.......................................................................................................... 12 1.6 Scope of the research ...................................................................................................... 14 1.7 Organisation of the dissertation ................................................................................... 14 CHAPTER 2 THEORETICAL AND CONCEPTUAL PERSPECTIVES ON RURAL- URBAN LINKAGES AND FLOOD VULNERABILITY AND THEIR FRAMING ....... 16 2.1 Rural-urban linkages ...................................................................................................... 17 2.1.1 Defining ‘rural’ and ‘urban’ .................................................................................. 17 2.1.2 Theories and concepts that correspond to rural-urban linkages ..................... 18 2.1.3 Typologies of rural-urban linkages ..................................................................... 19 2.2 Vulnerability.................................................................................................................... 21 2.2.1 Definitions ............................................................................................................... 22 2.2.2 Components ............................................................................................................ 23 2.2.3 Assessment frameworks ....................................................................................... 26 2.3 Flood hazards, impacts, rural-urban linkages and rural vulnerability .................. 29 2.3.1 Flood, impacts, losses/damages and changes .................................................... 29 2.3.2 Rural-urban linkages in the context of floods events ........................................ 30 2.3.2.1 People .............................................................................................................. 30 2.3.2.2 Information ..................................................................................................... 31 2.3.2.3 Finance ............................................................................................................. 31 2.3.2.4 Goods and services ........................................................................................ 32 2.3.3 The driving factors of rural-urban linkages and flood vulnerability ............. 33 2.3.3.1 Social factors ................................................................................................... 33 2.3.3.2 Economic factors ............................................................................................ 34 2.3.3.3 Environmental factors ................................................................................... 35 XIII 2.3.3.4 Institutional factors ........................................................................................ 36 2.3.3.5 Infrastructure factors ..................................................................................... 37 2.3.3.6 Spatial factors ................................................................................................. 38 2.4 The conceptual framework ........................................................................................... 39 2.5 Summary .......................................................................................................................... 43 CHAPTER 3 THE METHODOLOGY FOR EMPIRICAL RESEARCH ...................... 45 3.1 The selection of the case studies ................................................................................... 47 3.1.1 Punjab: A brief overview....................................................................................... 47 3.1.2 Specific study areas for empirical research ........................................................ 49 3.1.2.1 Spatial proximity criteria .............................................................................. 51 3.1.3 The profile of the selected study areas ................................................................ 52 3.1.3.1 Darya Khan: Representative of a small city ............................................... 52 3.1.3.2 Muzaffargarh: Representative of a medium-sized city ............................ 55 3.1.3.3 Multan: Representative of a large city ........................................................ 56 3.2 Sampling process ............................................................................................................ 57 3.2.1 Sample size calculation .......................................................................................... 58 3.2.2 Sample distribution ................................................................................................ 59 3.2.3 Sorting of the samples ........................................................................................... 60 3.3 Data collection ................................................................................................................. 60 3.3.1 Secondary data ....................................................................................................... 60 3.3.2 Primary data ........................................................................................................... 60 3.3.2.1 Observations ................................................................................................... 61 3.3.2.2 Individual and group discussions ............................................................... 61 3.3.2.3 Questionnaire-based household survey ..................................................... 62 3.4 Data processing and analysis ........................................................................................ 64 3.4.1 Qualitative analysis ................................................................................................ 64 3.4.2 Quantitative analysis ............................................................................................. 65 3.4.2.1 Descriptive statistics ...................................................................................... 65 3.4.2.2 Regression analysis ........................................................................................ 66 3.4.2.3 The composite index development ............................................................. 68 3.4.2.4 Other inferential statistics ............................................................................. 78 3.5 Challenges and limitations ............................................................................................ 79 CHAPTER 4 THE RESPONDENTS’ SOCIO-ECONOMIC PROFILES, FLOOD IMPACTS AND RESPONSES ................................................................................................. 81 4.1 The socio-economic profiles of the surveyed households ........................................ 82 4.1.1 Age of the household head and gender ratio ..................................................... 82 4.1.2 Household size and family type .......................................................................... 83 4.1.3 Education attainment............................................................................................. 84 4.1.4 Primary income source and household income ................................................. 85 4.1.5 House type and ownership of farmland ............................................................. 86 4.2 The impacts of flooding on the surveyed households .............................................. 86 4.2.1 The impacts of flooding: 2007 – 2017 ................................................................... 87 XIV 4.2.2 The impacts of the most recent, worst flood event............................................ 88 4.3 Response to flooding in rural areas of Pakistan ......................................................... 91 4.3.1 Immediate responses from the government and donors ................................. 91 4.3.2 Changes to structures and livelihood practices ................................................. 93 4.4 Summary .......................................................................................................................... 95 CHAPTER 5 CHANGES IN RURAL LINKAGES WITH CITIES AND THEIR DETERMINANTS ..................................................................................................................... 97 5.1 Conceptual framework .................................................................................................. 98 5.2 Analytical approach ..................................................................................................... 100 5.3 Results ............................................................................................................................ 100 5.3.1 Changes in rural-urban linkages ........................................................................ 100 5.3.1.1 Increased dependence on cities: More frequent flows ........................... 100 5.3.1.2 Decreased dependence on cities: Change in the location of flows ....... 102 5.3.1.3 Comparing more frequent flows and changes in the location of flows104 5.3.2 Factors influencing changes in rural-urban linkages ...................................... 104 5.3.2.1 Descriptive statistics of the explanatory variables .................................. 104 5.3.2.2 The evaluation of binary logistic regression models .............................. 105 5.3.2.3 Increased dependence on cities: More frequent flows ........................... 107 5.3.2.4 Decreased dependence on cities: Change in the location of flows ....... 110 5.4 Discussion ...................................................................................................................... 112 5.4.1 The flow of people: Changing mobility patterns ............................................. 114 5.4.2 The flow of finances: Credit and remittances................................................... 114 5.4.3 The flow of information: Access to support ..................................................... 115 5.4.4 The flow of goods and services: Farm input and markets ............................. 117 5.4.5 The main differences between the case studies ............................................... 118 5.4.6 The impact of changes in rural-urban linkages on vulnerability .................. 119 5.5 Summary ........................................................................................................................ 119 CHAPTER 6 INDEX-BASED ASSESSMENT OF THE LINKAGES BETWEEN THE SIZE OF THE CITY, PROXIMITY TO THE CITY AND RURAL FLOOD VULNERABILITY .................................................................................................................... 122 6.1 Conceptual frameworks .............................................................................................. 123 6.1.1 City size and rural vulnerability (SCARV) ....................................................... 124 6.1.2 The vulnerability proximity nexus (VPN) framework ................................... 125 6.2 Analytical approach ..................................................................................................... 127 6.3 Results and discussion ................................................................................................. 127 6.3.1 Index validation .................................................................................................... 128 6.3.2 City size and rural vulnerability ........................................................................ 129 6.3.2.1 Exposure ........................................................................................................ 129 6.3.2.2 Susceptibility ................................................................................................ 130 6.3.2.3 Capacity ......................................................................................................... 134 6.3.2.4 Vulnerability ................................................................................................. 136 6.3.3 Proximity to the city and rural vulnerability ................................................... 136 XV 6.3.3.1 Exposure ........................................................................................................ 136 6.3.3.2 Susceptibility ................................................................................................ 138 6.3.3.3 Capacity ......................................................................................................... 139 6.3.3.4 Vulnerability ................................................................................................. 141 6.3.3.5 Rural vulnerability vis-à-vis distance to the city ..................................... 143 6.3.4 The relevance of city size, proximity to cities and rural-urban linkages for rural flood vulnerability .................................................................................................. 144 6.4 Summary ........................................................................................................................ 145 CHAPTER 7 CHANGING RURAL LINKAGES WITH CITIES AND THEIR INFLUENCE ON FLOOD VULNERABILITY .................................................................... 147 7.1 Analytical approach ..................................................................................................... 147 7.2 Results and discussion ................................................................................................. 148 7.2.1 Increased dependence on cities and flood vulnerability ................................ 148 7.2.2 Decreased dependence on cities and flood vulnerability ............................... 152 7.3 Summary ........................................................................................................................ 155 CHAPTER 8 CONCLUSIONS AND RECOMMENDATIONS ................................. 157 8.1 Conclusions ................................................................................................................... 157 8.2 Implications and recommendations for policy and practice .................................. 162 8.3 Research contributions ................................................................................................. 165 8.4 Limitations of the research .......................................................................................... 168 8.5 Future research ............................................................................................................. 169 REFERENCES ........................................................................................................................ 171 ANNEXURES ........................................................................................................................ 193 Annex A ...................................................................................................................................... 193 Annex B ...................................................................................................................................... 194 Annex C ...................................................................................................................................... 198 Annex D ...................................................................................................................................... 204 Annex E ...................................................................................................................................... 206 Annex F ....................................................................................................................................... 208 Annex G ...................................................................................................................................... 209 XVI LIST OF FIGURES Figure 1.1. Gaps in examining rural flood vulnerability ......................................................... 7 Figure 1.2. Hazard impacts between 1970-2017. (A) the proportion of economic damages; (B) the proportion of the population affected; (C) the proportion of flood events by type; (D) the proportion of the population affected by flood type. ................................................. 8 Figure 1.3. Overview of the research approach and research tasks .................................... 13 Figure 2.1. Linkages between rural and urban areas ............................................................. 20 Figure 2.2. Components of vulnerability ................................................................................. 24 Figure 2.3. Role of the capacity component in influencing vulnerability. .......................... 25 Figure 2.4. Interconnections of driving factors, rural-urban linkages, and flood vulnerability. ................................................................................................................................ 33 Figure 2.5. Conceptual framework demonstrating the dynamics of rural-urban linkages and its influence on rural flood vulnerability ......................................................................... 40 Figure 3.1. Research process ...................................................................................................... 46 Figure 3.2. The figure showing the location of the case study areas. .................................. 50 Figure 3.3. Location of surveyed villages in Darya Khan. The boundary of the city (in green) is defined by combining the boundary of all urban-UCs. The labels on green dots represent the identity code of the villages which are available in Annex B. ...................... 54 Figure 3.4. Location of surveyed villages in Muzaffargarh. The boundary of the city (in green) is defined by combining the boundary of all urban-UCs. The labels on green dots represent the identity code of the villages which are available in Annex B. ...................... 56 Figure 3.5. Location of surveyed villages in Multan. The boundary of the city (in green) is defined by combining the boundary of all urban-UCs. The labels on green dots represent the identity code of the villages which are available in Annex B. ....................................... 57 Figure 3.6. Sampling framework for the household survey ................................................. 59 Figure 3.7. Individual and group discussion, (A) shows the discussion with an official of the Punjab Rural Support Program (B) shows the discussion with community members in Multan ........................................................................................................................................... 62 Figure 3.8. Enumerators collecting data from respondents (A) Darya Khan; (B) Multan 64 Figure 3.9. Steps to construct a vulnerability index for rural households affected by floods in Pakistan .................................................................................................................................... 69 Figure 3.10. Farmers harvesting crops during the time of the survey ................................ 80 Figure 4.1. The average number of floods experienced by households between 2007 and 2017, the number of times flood reached inside the house and the number of times livelihood destroyed by flooding between 2007 and 2017. ................................................... 87 Figure 4.2. Comparison of direct impacts of the extreme flood event in case study areas ........................................................................................................................................................ 88 Figure 4.3. Comparison of indirect impacts of the extreme flood event in case study areas ........................................................................................................................................................ 90 Figure 4.4. Percentage of households that got multiple relief and recovery items during and after the flood ....................................................................................................................... 92 XVII Figure 4.5. Structural and livelihood changes following the flood event in case study areas ........................................................................................................................................................ 93 Figure 5.1. The conceptual framework of the study .............................................................. 99 Figure 5.2. Percentage of responses that indicated an increased frequency of the different types of flows after the flood, which shows a strengthening of rural-urban linkages ... 102 Figure 5.3. Percentage of households acquiring facilities and services from the rural village after the flood event that were previously acquiring them from the city ............ 104 Figure 5.4. Factors that determine changes in rural linkages with cities .......................... 109 Figure 6.1. City Size and Rural Vulnerability (SCARV) framework ................................. 125 Figure 6.2. Vulnerability-Proximity Nexus Framework (VPN) ......................................... 126 Figure 6.3. Sensitivity analysis of exposure, susceptibility, and capacity indicators ...... 128 Figure 6.4. The boxplot shows A) Exposure, B) Susceptibility, C) Capacity, and D) Vulnerability of rural households surrounding cities of various sizes. The number of circles (°) over whiskers represents the number of mild outliers and asterisks (*) show the extreme outliers. ........................................................................................................................ 131 Figure 6.5. A-D showing exposure, susceptibility, capacity, and vulnerability of each case study area. 0 represents the lowest value and 1 represents the highest value in each case. ...................................................................................................................................................... 139 Figure 6.6. Comparison of flood vulnerability level of villages near (represented by ‘N’) and far (represented by ‘F’) from the Darya Khan city. ...................................................... 141 Figure 6.7. Comparison of flood vulnerability level of villages near (represented by ‘N’) and far (represented by ‘F’) from the Muzaffargarh city. ................................................... 142 Figure 6.8. Comparison of flood vulnerability level of villages near (represented by ‘N’) and far (represented by ‘F’) from the Multan city. ............................................................... 142 Figure 6.9. Figure showing the scatterplots between distance and vulnerability ........... 143 Figure 7.1. Box plot showing the difference between the vulnerability of households with and without increased dependence on cities after the flood event A) Flow of people, B) Flow of finance, C) Flow of information, D) Flow of goods. The number of circles (°) over whiskers represents the number of mild outliers and asterisks (*) show the extreme outliers. ....................................................................................................................................... 149 Figure 7.2. Box plot showing the difference between the vulnerability of households with and without reduced dependence on cities after the flood event A) Flow of finance, B) Flow of information, C) Flow of goods, D) Market and trading services. The number of circles (°) over whiskers represents the number of mild outliers and asterisks (*) show the extreme outliers. ........................................................................................................................ 153 Figure A.1: Flood risk map of Pakistan. ................................................................................ 194 Figure A.2. Locations where households lived during the duration of the flood ........... 206 Figure A.3. Changes in farming practices in order to adapt to future flood events ....... 206 XVIII LIST OF TABLES Table 2.1. Development of vulnerability definitions from the context of hazard, disaster risk, environment, and climate change .................................................................................... 22 Table 2.2. Conceptual frameworks with respect to assessment approaches, conceptual understanding of vulnerability, applicability as per spatial scale and reflection on the interaction between spatial units .............................................................................................. 27 Table 3.1. Characteristic of flood events between 2010 and 2014 in Punjab ...................... 48 Table 3.2. Case study areas and key characteristics .............................................................. 53 Table 3.3. Classification, standardized scores and explanation of vulnerability indicators with respective components (exposure, susceptibility and capacity) ................................. 71 Table 4.1. The table shows the age of household head and male to female ratio in surveyed villages .......................................................................................................................................... 83 Table 4.2. Differences among case studies concerning household size and family type in surveyed villages......................................................................................................................... 84 Table 4.3. Education attainment status of households in surveyed villages ..................... 84 Table 4.4. Primary income sources and family income of households in surveyed villages ........................................................................................................................................................ 85 Table 4.5. Housing conditions and agriculture land ownership status of households .... 86 Table 4.6. Duration of households’ displacement due to the last worst flood event........ 89 Table 5.1. Description of explanatory variables selected for different linkages .............. 106 Table 5.2. Testing hypothesis for accuracy and significance of the model ...................... 107 Table 5.3. Coefficient estimates for binary logistic regression, marginal effect and elasticity for changes in a different type of linkages after the extreme flood event ........ 113 Table 6.1. Degree of exposure, susceptibility, capacity, and vulnerability rural households living around small, medium and large cities. ..................................................................... 133 Table 6.2. Mann-Whitney U test for proximity groups of holistic vulnerability............. 143 Table 6.3. Correlation between rural households’ distance to cities and their vulnerability ...................................................................................................................................................... 144 Table 7.1. Comparison of vulnerability level of households that intensify their linkages (flow of people, finance, information, goods) with cities after the flood .......................... 150 Table 7.2. Comparison of vulnerability level of households that reduce their dependence on cities (in terms of finance, information, goods, services) after the flood ..................... 154 Table A.1. The distance of different villages from their respective cities. ........................ 194 Table A.2. Sample size calculation ......................................................................................... 197 Table A.3. Explanatory variables used in different models ............................................... 204 Table A.4. Changes in mobility pattern, access to information, finance/income and flow of goods and services after the extreme flood event ............................................................ 206 Table A.5. Cross-tabulation of households that were frequently accessing facilities and services with those households getting these facilities and services from the village after the floods .................................................................................................................................... 207 Table A.6. Correlation Matrix ................................................................................................. 208 XIX Table A.7. Correlation between structural-livelihood modifications and changes in linkages ....................................................................................................................................... 209 1 Chapter 1 Introduction 1.1 Background: Flood hazards and vulnerability Extreme weather and climatic events, especially in the form of hydro-meteorological hazards, have wreaked havoc around the world. These events cause severe impacts, including fatalities, injuries and damage to homes and infrastructure, which disrupt the economy, infrastructure systems, people’s sources of livelihood and societal services (Kirsch et al., 2012). A flood is one of the most damaging types of events and is likely to become more frequent (Willner et al., 2018). It is expected that climate change will further increase the intensity and magnitude of such events (Jongman et al., 2018). Different regions around the world are exposed to different kinds of floods, including fluvial or river-related floods, pluvial and flash floods, and coastal (i.e. due to storm surges) floods. Their destructive effects, both in rural and urban areas, have become a global issue. Floods have resulted in widespread human, economic and physical losses. According to an estimate, flood events in the last two decades have caused economic losses of more than US$537 billion, affected 1.6 billion people, and taken more than 100,000 lives (EM- DAT, 2020). The aftermath is worse in low-income and developing countries in Asia, where the majority of people live in rural areas. In these regions, the flood impacts and resulting losses are projected to increase in the future (Winsemius et al., 2016; Willner et al., 2018). These increasing impacts and losses are not only due to an increase in the intensity and magnitude of flood hazards, but also the vulnerability of the socio-ecological system. These systems’ vulnerability1 depends on location-specific social, economic, demographic, political, physical and environmental attributes (Greiving et al., 2006a; 1 Vulnerability refers to characteristics and circumstances of people, their surroundings and systems, which makes them likely to be adversely affected by a hazard or climatic event. Vulnerability has several components and dimensions. In this research, vulnerability encompasses the aspects of exposure, susceptibility and capacity. 2 Birkmann et al., 2013; Jamshed et al., 2017) that make vulnerability spatially dynamic (i.e. vulnerability varies between regions, countries and counties, as well as between urban and rural areas) (Vogel et al., 2004; Cutter et al., 2008b; Greiving, 2013; Cutter et al., 2016; Feldmeyer et al., 2017). When comparing the vulnerability of rural and urban regions, rural regions are considered more vulnerable (Bird et al., 2002; Turpie et al., 2012). This higher vulnerability in rural areas is caused by lower human development, poor physical infrastructure, high dependence on agriculture with impoverished conditions, and little attention from public institutions (Kapucu et al., 2013; Dasgupta et al., 2014; Lazarte, 2017). Rural locales are also less prepared for (and have fewer mechanisms to adapt to) weather or climate-related events like floods (Steinberg, 2014; Cutter et al., 2016). Hence, rural areas are hotspots of high flood impacts, and are likely to continue to be in the future unless vulnerability is significantly reduced. In this regard, understanding vulnerability is imperative for informed decision-making that aims to reduce flood impacts. A vulnerability assessment is important for designing flood risk reduction and adaptation measures, and is a vital step toward alleviating and managing flood risk over space and time. There is agreement on this in the different—but related—international discourses on disaster management, sustainable development and climate change adaptation (Birkmann, 2013b; IPCC, 2014b; Birkmann et al., 2020a). This can be seen in the Sendai Framework for Disaster Risk Reduction (SFDRR), in which Priority 1 calls for understanding disaster risk in all its dimensions, including vulnerability (UNDRR, 2015). In the United Nations’ (UN) Sustainable Development Goals (SDGs), goals 11 and 13 call for safe and resilient human settlements by strengthening resilience and capacity (United Nations, 2018). The Paris Agreement on combatting climate change calls for climate change impact and vulnerability assessments of people, places and ecosystems to prioritise actions for reducing vulnerability (UNFCCC, 2015, pp. 9–10). Much of the research on flood vulnerability assessment focuses either on the vulnerability of urban hubs/cities (Douglas et al., 2008; Adelekan, 2011; Balica et al., 2012; Welle et al., 2015b; Rana et al., 2018b) or rural areas (Mustafa, 1998; Brouwer et al., 2007; Armah et al., 2010; Schütte et al., 2012; Jamshed et al., 2019b; Sarker et al., 2019), but not on the 3 relationship between the two. Although such studies have investigated different components (exposure, susceptibility and capacity to adapt) and dimensions (social, economic, physical, institutional, environmental and others) of vulnerability to flood events, they have not paid sufficient attention to the relationship between rural areas and cities/urban areas that can be affected due to flood events, and how it can influence vulnerability. In short, assessment of the vulnerability of rural areas and their residents in relation to the nearest city (on which rural households depend for various services and facilities) has been widely neglected. 1.2 Motivation: Rural-urban linkages and rural vulnerability Over 90% of the world’s rural population lives in less developed regions2 (UN-DESA, 2018b). Rural settlements in these regions are usually undersupplied and underdeveloped (Srivastava et al., 2016). These settlements are going through socio-economic, demographic, environmental and governance changes due to interdependencies between rural and urban areas (Dasgupta et al., 2014). Rural populations depend on nearby cities as they provide employment, health, education, emergency services, markets and information to the rural population. Similarly, cities and their inhabitants depend on rural areas for labour, food and other ecological services (Gebre et al., 2019). These dependencies are often referred to as rural-urban linkages in the literature, and are explained in terms of the flow of people, finances/income, information, goods and services (Tacoli, 1998; Lynch, 2005). Spatial and regional development studies underscore that these linkages result in the social, economic and physical growth of rural regions (see for example Douglass, 1998, Bah et al., 2003, Tacoli, 2007). There is a need to consider rural- urban linkages not only from a spatial development perspective but also how they can be affected by floods or other hazards (Srivastava et al., 2012). There is a consensus—but limited evidence—that the dependencies and linkages between rural and urban places may be affected by the repercussions of weather or climatic events 2 Less developed regions comprise all regions of Africa, Asia (except Japan), Latin America and the Caribbean plus Melanesia, Micronesia and Polynesia. 4 like flooding (Dasgupta et al., 2014), which can considerably influence the vulnerability of rural communities. The Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) highlighted the need for improved understanding of rural-urban linkages and their management in the face of climate change and its impacts (Dasgupta et al., 2014, p. 645). UN-Habitat’s ‘New Urban Agenda’ has also called for integrated initiatives where disaster risk, climate change, and spatial and territorial planning should be harmonised. It has called for strengthening rural-urban linkages, especially in the face of hazards and disasters (UNHABITAT, 2016, 2017). Hence, examining changes in rural- urban linkages due to the impact of extreme events is a relevant area of research for understanding and reducing vulnerability and risk. Research has shown that the impacts of extreme events, such as flooding, can cause both formal (by government agencies) and informal (by households, communities) changes in society (Birkmann et al., 2010), including organisational and legislative shifts, as well as modifications of housing structures and livelihood/farming practices or linkages between rural and urban areas (Douglass, 1998; Birkmann et al., 2010; Birkmann, 2011). These changes can influence vulnerability and hence communities’ risk level (Jamshed et al., 2017). As such, it is important to explore how extreme events in rural locales alter livelihood practices, as well as rural linkages with cities, and how these changes influence rural households’ vulnerability (see Figure 1.1). In addition to the impacts of an extreme event, several other factors influence rural-urban linkages and vulnerability, especially in post-disaster situations (Douglass, 1998; Birkmann et al., 2013); these include the socio-economic, physical, environmental and spatial characteristics of society. Although the nature of the rural-urban linkages between a rural community and its nearest city—on which the rural community depends for multiple city-based services (e.g. markets, health and emergency services, institutions)— is determined by several factors, two spatial features are important in this regard: (1) the nearest city’s proximity to the rural community and (2) city size (Jamshed et al., 2020b). These two elements can significantly influence the social, economic, physical and institutional attributes of rural-urban linkages, and thus the rural landscape surrounding 5 the city (Berdegué et al., 2015; Abbay et al., 2016; Sharma, 2016; Rana et al., 2020c). Neither of these two factors (as reviewed by Jamshed et al., 2020b) has been given enough attention in previous studies on rural vulnerability, particularly in Pakistan (see Figure 1.1). A large body of literature indicates that larger cities are less at-risk than smaller cities (Cross, 2001; Birkmann et al., 2016b; Handayani et al., 2017). Cross (2001) and Handayani et al. (2017) underscored that larger cities can have higher exposure to hazards or climatic events, and some parts of such cities can be more susceptible. However, large cities generally have greater social, economic, physical and institutional resources compared to smaller cities, which results in the varied overall vulnerability between these city types (Cross, 2001; Birkmann et al., 2016b; Handayani et al., 2017). Despite the substantial body of research on the vulnerability of cities of different sizes, the investigation of how city size influences the vulnerability of households in surrounding rural settlements has been largely neglected. There is a dearth of research on the relationship between the proximity of rural households and their settlements to a district’s main city, as well as their vulnerability. Some research has shown that different human, natural, physical and financial factors differ with a community’s proximity to the city; this affects the community’s vulnerability to climate change (Pandey et al., 2012; Pandey et al., 2017). However, these studies only categorise rural settlements as near or far from the city, and do not provide any information on proximity criteria (e.g. distance). Moreover, these studies do not present information on the characteristics of cities and types of linkages. Socio-economic and infrastructure conditions, as well as institutional coverage, can be significantly influenced by a household’s proximity to the city and city size. In this way, rural households’ vulnerability to floods can be affected by city size and proximity. These two factors can serve as baseline parameters in investigations of flood vulnerability, as they influence all other social, economic, institutional and infrastructure factors, as well as rural-urban linkages (Ferré et al., 2012; Abbay et al., 2016; Sharma, 2016; Rana et al., 2020c). Further, changes in rural linkages with cities following a flood event can impact rural households’ vulnerability, as mentioned earlier. In this context, changes in linkages 6 need to be connected with rural households’ overall vulnerability in order to examine how the modification of various linkages could influence rural households’ vulnerability to flooding (see Figure 1.1). Despite the importance of rural-urban linkages for rural vulnerability, linkages and vulnerability have not yet been theoretically framed together. Several theoretical frameworks for vulnerability assessment have been developed for different scales (local, national, global), components (exposure, susceptibility/sensitivity, capacity), and dimensions (social, economic, physical, environmental, institutional) depending on the research field (Cutter, 1996; DFID, 1999; Cardona et al., 2000; Bohle, 2001; Turner et al., 2003; Wisner et al., 2004; Birkmann, 2006b; Füssel et al., 2006; Birkmann et al., 2013; IPCC, 2014b). However, none of the frameworks or assessment studies has adequately given attention to (and revealed) the effects of rural flooding or other rural hazards on the interactions/linkages between spatial units (rural and urban), or their influence on vulnerability. The factors and issues that characterise rural-urban linkages and vulnerability need to be outlined together and assessed jointly. To address the above research gaps, rural vulnerability to flooding needs to be evaluated and filtered from the perspective of rural-urban linkages and spatial factors, which are closely related; for example, flood impacts can influence rural-urban linkages, but rural- urban linkages also affect impact typologies and other socio-economic and spatial conditions (see Figure 1.1). The key proposition of this dissertation is: Extreme flood events modify the linkages between rural and urban areas and hence the vulnerability of rural households. A rural settlement’s proximity to a city and the size of that city are deciding factors in how these rural-urban linkages and rural households’ vulnerability are altered. This thesis conceptualises the rural-urban linkages, vulnerability and associated factors into a framework that is operationalised using three case studies in Punjab, Pakistan. 7 Figure 1.1. Gaps in examining rural flood vulnerability (Own figure, 2020) 1.3 Pakistan and flooding Pakistan is a nation in South Asia with an agro-industrial economy, a population of over 207 million, an average annual gross domestic product (GDP) growth rate of 5.5%, and is ranked at 150 on the Human Development Index (PBS, 2017; Jahan et al., 2018). The rural areas are crucial as they house more than 65% of the population and form the backbone of the country’s agricultural economy (PBS, 2017; Mughal, 2019). Agriculture alone employs 43% of the total workforce and contributes about to 20% of national GDP (MoF, 2016) There is a high development disparity between the rural and urban regions (Rana et al., 2020c). Areas further away from primary cities are less developed, and conditions in rural areas are far more inferior than those in urban areas (Rana et al., 2017; Rana et al., 2020b). This situation has caused the rural population to depend heavily on cities for many facilities and services (e.g. markets to buy farm inputs and tools, basic health facilities, information services). 8 Figure 1.2. Hazard impacts between 1970-2017. (A) the proportion of economic damages; (B) the proportion of the population affected; (C) the proportion of flood events by type; (D) the proportion of the population affected by flood type. (EM-DAT, 2017) Pakistan is prone to numerous natural hazards, but flooding is the most common and devastating kind due to diverse climatic and hydrological conditions. Among all natural hazards between 1970 and 2017, floods caused 74% of economic damage (worth US$20 billion) and affected 87% of the population, as shown in Figure 1.2A&B (EM-DAT, 2017). According to Pakistan’s National Disaster Management Authority (NDMA), 67 out of the country’s 156 districts are at high flood risk. The frequency and intensity of floods have increased, and climate change is considered one of the main reasons (GoP, 2012). 0.9% 18.8% 0.1% 74.1% 0.1% 6.0% A) Proportion of economic damages Drought Earthquake Extreme temperature Flood Landslide Storms/cyclones 2.4% 8.0% 0.1% 87.0% 0.0% 2.4% B) Proportion of affected population Drought Earthquake Extreme temperature Flood Landslide Storms/cyclones 23.5% 49.4% 27.1% C) Flood types Flash flood Riverine flood Unspecified 27.9% 44.2% 27.9% D) Population affected by flood type Flash flood Riverine flood Unspecified 9 According to the Climate Risk Index (CRI), Pakistan was the fifth most affected nation by climate change in 2020. The country’s rank was first in 2010, third in 2011 and 2012, and sixth in 2013 (Kreft et al., 2016). Willner et al. (2018) found that Pakistan is already highly affected and will observe almost a doubling in high-end flood risk at the sub-national level within the next 25 years. Pakistan can be divided into three major hydrological landscapes—the Indus Basin, the Kharan Basin, and the Makran coastal drainage area—with each one being prone to a different major flood type, specifically riverine, flash and coastal flooding, respectively (Tariq et al., 2012). Riverine floods account for almost 50% of all floods in Pakistan (see Figure 1.2C). Flooding in the Indus Basin has caused major economic and human losses; the proportion of affected people is also highest in this area (see Figure 1.2D). The Indus Basin is very fertile for agriculture and plays a major role in the country’s economy; it is highly populated, and population pressure has led to development in floodplains, with no regard to the resulting increase in flood risk due to higher exposure. The government has spent a vast amount of money on structural engineering to counter flood risk, but these measures have been unable to prevent large-scale disastrous flood impacts in recent years (Tariq and van de Giesen, 2012). Between 2010 and 2015, Pakistan faced five consecutive large-scale floods. The flooding of 2010 was extreme and left half of the nation paralyzed due to devastating ramifications for people, the economy and infrastructure (Solberg, 2010; Shah et al., 2020). Disastrous floods mainly affected rural areas and their inhabitants, while cities and industrial hubs remained relatively safe (Asgary et al., 2012). Aslam et al. (2017) and van der Schrier et al. (2018) revealed that a flood event like that of 2010 could occur again and may have even more severe socio-economic impacts, especially on the rural population. With the current situation and impending climate change, rural locales and the livelihoods of their residents are highly threatened (Abid et al., 2016a). Moreover, effects on infrastructure and rural households may have influenced linkages between rural and urban areas and hence vulnerability. It is important to note that existing disaster risk management and climate change adaptation policies and strategy documents (Government of Pakistan 10 GoP, 2012b, 2012a; Shahzad et al., 2018)—both at the national and provincial levels—do not acknowledge the impacts of climatic events or hazards on rural-urban linkages and their influence on people. Thus, a rigorous assessment of rural households’ vulnerability, considering their linkages with cities, is critical to guide future policies, as well as for risk- informed decision-making. 1.4 Objectives and research questions The limited consideration of rural-urban linkages in the evaluation of rural vulnerability to floods has been identified as a vital research gap. The overall goal of this research is to examine the dynamics of rural linkages with cities due to flood impacts and their effects on rural households’ vulnerability. Based on multiple research gaps, the study’s sub-objectives are: Objective 1. To develop a framework to scrutinise households’ changes and modifications of rural-urban linkages following a flood event in rural areas, as well as their effects on residents’ vulnerability. Objective 2. To investigate the impacts of flooding, as well as other socio- economic and physical factors, in shifting rural-urban linkages. Objective 3. To determine the impact of city size and proximity on rural households’ flood vulnerability, with a focus on rural-urban linkages. Objective 4. To explore the influence of changing rural linkages with cities on rural households’ flood vulnerability. To achieve these goals, the following research questions (RQs) were developed. RQ1. How can the theoretical perspectives on rural-urban linkages and vulnerability be framed together? o How are rural-urban linkages defined, and how is vulnerability understood in different schools of thought? o What are the driving factors of rural-urban linkages and flood vulnerability? 11 o In what ways can floods affect rural-urban linkages, and how can the impacts of flooding on these linkages be connected to rural households’ vulnerability? RQ2. How do the effects of flooding lead to changes in rural households and influence the linkages between these households and cities? o What are the direct and indirect impacts of flood events on rural farming households in Pakistan? o How do rural households change their housing structure, livelihood patterns and linkages with cities following an extreme flood event? o Which drivers modify rural-urban linkages after flooding? o How are changes in rural-urban linkages related to rural households’ vulnerability? RQ3. To what extent can city size and proximity to the city influence the vulnerability of surrounding rural settlements, keeping in mind the rural-urban linkages? o How can rural vulnerability be operationalised in the context of city size and proximity to the city? o How different are the vulnerability levels of rural households around cities of different sizes and at different proximities? o Are rural-urban linkages relevant to rural households’ vulnerability? RQ4. How do changes in rural-urban linkages after a flood influence rural households’ vulnerability to flooding? o Are there any differences in the vulnerability of households that changed their linkages with cities compared to those that did not? o To what extent do changes in linkages influence rural households’ vulnerability? 12 1.5 Research approach This study is primarily based on empirical research and predominantly uses a quantitative approach, while a qualitative approach is also employed for conceptual framing, as well as for empirical research (see Figure 1.3). The triangulation of both approaches is performed to support and test the validity of the findings. To answer the RQs, this study is divided into four parts: (1) a framework for the study; (2) the impacts of floods and resulting changes, including changes to rural-urban linkages; (3) the vulnerability assessment; and (4) the influence of changing linkages on overall flood vulnerability. Case studies were used to perform empirical research; they were carefully selected and encompass differences in economic functions and spatial characteristics. Data were collected by conducting face-to-face interviews using a questionnaire (these aspects are described more precisely in the methods section). RQ1 sought to clarify different theoretical and conceptual perspectives on rural-urban linkages and vulnerability, and how they can be framed together under the umbrella of flood events and their influence. To do so, multiple literature sources (e.g. research articles, books, published reports, dissertations and newspaper articles) were scrutinised, and a conceptual framework was proposed (see Chapter 2). The conceptual framework provides the basis to answer further RQs. RQ2 aimed to investigate the impacts of past flood events and the changes following a flood (Chapter 4 and Chapter 5), including shifts in livelihood and in rural-urban linkages. These were analysed using descriptive statistics. Further, inferential statistics were employed to examine the relationship between changes in livelihood and in rural-urban linkages. Factors that determine the modification of rural-urban linkages were also explored using inferential statistics. The findings were supported and validated by excerpts from responses to the household survey. RQ3 intended to assess rural households’ vulnerability considering city size and proximity parameters. This built on the findings from RQ2, which identified the importance of changes in rural-urban linkages and other relevant factors for investigating 13 vulnerability. The findings highlight the importance of city size and distance to the city as parameters that influence rural-urban linkages and can be pertinent to assessing vulnerability. The research applied a composite indicator method for evaluating vulnerability, followed by inferential statistics to discern differences in the vulnerability of rural households living on the peripheries of different cities (using the criteria of city size and proximity). Figure 1.3. Overview of the research approach and research tasks (Own figure, 2020) RQ4 aimed to combine the results of RQ2 and RQ3, and to study the influence of changing linkages on rural households’ vulnerability. The findings suggest that shifting linkages 14 can both increase and decrease rural households’ vulnerability. This section is based on descriptive and inferential statistics. 1.6 Scope of the research This research focuses on the local community and household levels, since investigations at the local level are often viewed as effective in capturing the complexity of interactions in coupled human-environmental systems (O’Brien et al., 2004). Rural households are the key focus of this research, as floods particularly affect rural communities in Pakistan. Three case studies were selected from the province of Punjab (i.e. Darya Khan, Muzaffargarh, and Multan). These case studies represent important research sites because the study areas have a history of flooding and different flood profiles (in terms of the number of flood events, risk level, etc.); in addition, the study areas are around cities of different sizes, with different demographic and socio-economic profiles. Flood impacts can lead to both formal (carried by the public sector) and informal (carried by private sector; e.g. households, non-profit organisations [NGOs]) changes (Birkmann et al., 2010). This thesis focuses mainly on informal changes (adopted by households), and also refers to changes in linkages. Four central types of rural-urban linkages are studied, specifically the flow of (1) people, (2) information, (3) finances, and (4) goods and services. Vulnerability is explained as the function of exposure, susceptibility and the capacity to cope with and adapt to flood events (Birkmann et al., 2013). Although the flood impacts on rural households and the resulting changes in rural-urban linkages can also affect urban residents, the scope of this research is on affected rural households. 1.7 Organisation of the dissertation The dissertation is divided into eight chapters. Chapter 1 outlines the background of flood events and their trend in low-income and developing countries; it describes the research gaps, objectives, questions and research approaches. Chapter 2 explains the theoretical context of rural-urban linkages and vulnerability, and develops a conceptual framework that guides the empirical part of the thesis. Chapter 3 provides the methodology; it elaborates on the selection of the case study areas, the sampling and data collection 15 methods, and the different analytical approaches used to answer the various RQs. Chapters 4 through 7 present the findings. Chapter 4 details the respondents’ socio- economic profiles, the impacts of flooding on the surveyed households, and their responses. Chapter 5 uses descriptive statistics to identify changes in linkages following a flood event; regression analysis was applied to determine the factors that result in changes in rural-urban linkages, and correlation analysis was performed to reveal the relationships between changes. Chapter 6 employs an index-based approach to assess vulnerability considering city size and proximity. The indicators were chosen and validated using selected statistical approaches. Chapter 7 uses the results of all the empirical chapters and analyses the influence of changes in linkages on rural households’ vulnerability. Chapter 8 concludes the thesis and provides policy recommendations, as well as possible directions for future research. 16 Chapter 2 Theoretical and conceptual perspectives on rural-urban linkages and flood vulnerability and their framing The parts of this chapter (including figures and tables) have already been published in a peer-reviewed journal ‘Sustainability’ (see the citation below) Jamshed, A., Birkmann, J., Feldmeyer, D., Rana, I.A., (2020) “A Conceptual Framework to Understand the Dynamics of Rural-Urban Linkages for Rural Flood Vulnerability”. Sustainability 12 (7), 2894. DOI: https://doi .org/10.3390 /su1207289. Publisher: MDPI. One objective of this research was to develop a conceptual framework that could connect the elements of rural-urban linkages and vulnerability. The conceptual framework was crucial for the development of the methods, and highlighted relevant aspects to help design the empirical research and answer the research questions (RQs) (see Chapter 1, Section 1.5). This chapter presents a unified conceptual framework to assess how rural- urban linkages evolve in the aftermath of a flood event in rural areas, and how these influences rural residents’ vulnerability. The RQs are: (a) How are rural-urban linkages defined, and how is vulnerability understood in different schools of thought? (b) What are the driving factors of rural-urban linkages and flood vulnerability? (c) In what ways can floods affect rural-urban linkages, and how can the impacts of flooding on these linkages be connected to rural households’ vulnerability? This chapter is based on an extensive literature review. A systematic scrutiny of peer- reviewed papers and books on the topic of vulnerability, regional development and rural- urban linkages was carried out in three steps. In the first step, theoretical studies on rural- urban linkages in the context of regional/rural planning and development were assessed. In the second step, conceptual frameworks of vulnerability and its assessment from the perspective of climate change and disaster risk research were shortlisted. Lastly, screening of empirical studies on flood hazards, vulnerability assessment, rural-urban linkages and rural development was performed. This literature analysis enabled connections to be 17 made among various components and factors of rural-urban linkages and vulnerability in the context of flooding. The framework incorporates theoretical views on vulnerability and rural-urban linkages. Moreover, factors that determine variations in vulnerability, as well as linkages, are discussed. Vulnerability is explained in terms of exposure, susceptibility and response capacity to flood events (Birkmann, 2013b; Greiving, 2013), whereas rural-urban linkages are defined as the flow of people, information, finances, and goods and services (Tacoli, 1998). Factors that influence vulnerability and linkages include social, economic, institutional, infrastructure, spatial and environmental aspects (Cutter, 1996; Douglass, 1998; Tacoli, 2003; Birkmann et al., 2013). In this thesis, rural-urban linkages, vulnerability components and their associated factors collectively define rural vulnerability to flooding. Moreover, the present study ascertained the dynamics of rural-urban linkages in flooding conditions and their role in influencing rural vulnerability. The framework and underlying theoretical discussion serve as a parent framework for conceptualising and empirically analysing specific questions within chapters 4, 5 and 6. The following sections provide definitions, concepts (as well as key elements) and components of rural-urban linkages and vulnerability. 2.1 Rural-urban linkages 2.1.1 Defining ‘rural’ and ‘urban’ The distinction between rural and urban areas can be described by many criteria, but there is no general definition available. Rural and urban areas are defined based on one or a combination of demographic, economic, infrastructure and administrative factors (Rana et al., 2020c). Within these elements, population size and density, the predominant type of economic activity, conformity with legal and administrative status, as well as specific services and facilities, are prominent criteria to define rural and urban regions (UN-DESA, 2018a). However, definitions vary even more often when considering different parts of the world. In South Asian countries, areas within the boundaries of local administrative agencies (e.g. municipalities, councils, committees) are declared as urban (UN-DESA, 2018a). For example, in Pakistan, “all localities which are either a metropolitan corporation, a 18 municipal corporation, a municipal committee or a cantonment at the time of the census are treated as urban” (Reza, 2013). Based on this, areas are divided into rural and urban Union Councils3 (UCs) in Pakistan. Many African countries define ‘urban’ with respect to administrative, demographic and infrastructure characteristics, while in several Latin American nations, a population threshold of 2,000 or 2,500 is used to define urban areas (Tacoli, 1998). The term ‘urban areas’ is also frequently used to describe a city or a town. In this context, rural areas are considered the inverse of urban areas or cities (i.e. a residual or relative category) (Lerner et al., 2011). Consequently, low population and housing density, a high dependence on natural resources for one’s livelihood, limited infrastructure, and social services can characterise rural places in low- and middle-income countries (Lerner et al., 2011). Although a dichotomy exists between the features of rural and urban areas, they are essentially related due to the dependence of resources and facilities that both areas provide to each other. 2.1.2 Theories and concepts that correspond to rural-urban linkages Many concepts describe these linkages; most of them have their roots in spatial and regional development theories that indicate how a city can influence the development of its surrounding rural settlements. Von Thünen’s model of agricultural land uses suggests that land rent and transportation costs define the economic activities in a city’s hinterland. Thus, specialisation of economic activities in rural areas is based on the distance to the central city (Thunen, 1826). Christaller’s central place theory explains the distribution of central places (cities) of different sizes, on which their hinterland depends for various (central) services and facilities (Christaller, 1933). This theory implies that cities of different sizes can have varied linkages with the rural hinterland, depending on the types of goods and services, as well as spatial proximity to these services. Other models, like the growth pole model and the core-periphery model, infers that core areas/cities are the heart of economic activities, whereas periphery/rural locales deliver resources in the form of 3 The Union Council (UC) is the lowest administrative unit in Pakistan’s local administrative setup. Areas with administrative units (i.e. metropolitan/municipal or corporation/town committees in a given area) are termed ‘urban’. Accordingly, the Bureau of Statistics in Punjab (BSP) designated UCs as both urban and rural. 19 labour and goods, among others (Perroux, 1955; Friedmann, 1966). Both theories depict a dominating core, whereas the periphery is dependent. This dependence is structured through the relations of exchange between the core and the periphery. The virtuous circle model presents how the flow of people, goods, information, and finances between rural and urban regions leads to the development of rural areas and their residents (Evans, 1992). The model includes the spatial dimension by stressing proximity to cities in providing income opportunities and services for the rural hinterland. The environmental notion of the urban ecological footprint suggests that cities need a larger space than their actual size, on which their inhabitants rely for food, natural resources and the absorption of carbon (Rees, 1992). This dependence on food and other natural resources is primarily fulfilled by rural areas, hence demonstrating another facet of rural-urban linkages. Despite their differences, all of these theories indicate that rural and urban areas are (intrinsically) linked. The dominant features in these theories are the size of and proximity to cores/cities for the linkages between rural settlements and cities. However, several other socio-economic, institutional, infrastructure, spatial and environmental factors affect these linkages. 2.1.3 Typologies of rural-urban linkages The linkages between urban and rural areas can be categorised into four types, specifically the flow of (1) people, (2) information, (3) finances, and (4) goods and services (Lynch, 2005). There is a trade-off between rural and urban areas, as urban population depend on rural resources (food, labour and others), and urban services are vital for rural communities (Tacoli, 1998). These linkages are depicted in Figure 2.1. The flow of people indicates human mobility between rural and urban locales. Mobility takes several forms, such as temporary, permanent, circular migration and commuting (Tacoli, 2003). The flow of information represents information exchanges between rural and urban areas regarding population needs, job opportunities, market status, innovations, and new technologies for increased agricultural production, lifestyles and others (Srivastava et al., 2016). Financial flows can be classified under three types (1) formal, institutional; (2) informal and; (3) investments made by the government and aid agencies (Lynch, 2005, p. 20 164). First, formal flows include micro-credit schemes for economically active poor households from formal financial institutes, such as banks. Second, the informal exchange of finances involves remittance, taking loans from moneylenders, and landlords or relatives. Third, investments are made by urban-based government and aid agencies for the human and socio-economic growth of the rural population and the physical growth of rural areas (Lynch, 2005). Lastly, the transactions of goods and services are one of the most critical features of rural-urban linkages. Cities depend on rural resources; for example, agricultural products, water, and others (Gebre et al., 2019). Rural inhabitants purchase durable and non-durable goods for household use, as well as for expanding their livelihoods. Agriculture input, tools, building materials and household items are a few examples of goods needed by rural settlements (Douglass, 1998). In terms of services, urban areas often provide higher education, health and emergency services. Additionally, they offer rural households several off-farm opportunities for livelihood diversification (Tacoli, 2003). Figure 2.1. Linkages between rural and urban areas (Own figure, 2020) 21 Such flows modify the social, economic and physical landscape in rural areas. However, these linkages differ from location to location, depending on the rural development efforts made to upgrade them. Rural-urban linkages are not autonomous, but rather overlap, and are closely interlinked. For example, migration flows toward cities are amplified by the increased flow of information to rural settlements based on employment prospects that augment the financial flows from cities. Similarly, off-farm livelihoods in cities result in remittances, which tend to be used to increase agricultural production, improve one’s lifestyle, and send additional household members to urban hubs (Douglass, 1998; Tacoli, 2007). Moreover, better information services on market demand and trends increase the wellbeing of rural entrepreneurs (Mayer et al., 2016). These linkages, if properly understood, are important for framing development policies, and are vital for reducing poverty and social vulnerability. 2.2 Vulnerability The term ‘vulnerability’ comes from the Latin word vulnerare, which means ‘to wound’. In its most basic form, it indicates the fragility of living and non-living things (Luna, 2018). The concept has been widely used in the discourse of geographic development, poverty, human ecology, hazard and disaster risk reduction research, as well as climate change adaptation research (Birkmann, 2013b). Various fields have used this notion according to their applicability; therefore, the conceptual understanding of vulnerability has multiple schools of thought, contexts, dimensions and professions, which has resulted in numerous definitions and interpretations (Birkmann, 2013b). Human geography and human ecology have, in particular, theorised vulnerability to environmental change (Adger, 2006). In the field of disaster risk science, early perspectives on vulnerability were placed in the context of the physical resistance of engineering structures. Later, it was viewed as a characteristic of social and environmental systems (Cardona et al., 2012). The extensive use of these interpretations shows that vulnerability has become a critical concept in both disaster risk reduction and climate change adaptation research. 22 2.2.1 Definitions Diverse definitions of vulnerability suggest different views of the concept, which may lead to specific priorities in assessments (Birkmann, 2013b). Table 2.1 presents a range of definitions from the angle of research on hazards and disasters, as well as on climate change. Table 2.1. Development of vulnerability definitions from the context of hazard, disaster risk, environment, and climate change Source Definitions Context Vulnerability is defined as Mitchell (1989) “… the potential of loss” Hazard and disaster risk Cutter (1993) “… the likelihood that an individual or group will be exposed to and adversely affected by a hazard” Hazard and disaster risk Blaikie et al. (1994) “… the characteristic of person or group and their situation that influences their capacity to anticipate, cope with, resist and recover from the impact of a natural hazard” Hazard and disaster risk Adger (1999) “… the exposure of individuals or collective groups to livelihood stress as a result of the impacts of climate change and related climatic extremes” Climate Change Turner et al. (2003) “… the degree to which a system, subsystem, or system component is likely to experience harm due to exposure to a hazard, either a perturbation or stress stressor” Both hazard/disaster risk and global environmental change Adger (2006) “… the state of susceptibility to harm from exposure to stresses associated with environmental and social change and from the absence of capacity to adapt” Both hazard/disaster risk and climate change Intergovernmental Panel on Climate Change IPCC (2007) “… the degree to which a system is susceptible to, or unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the character, magnitude, and rate of climate change and variation to which a system is exposed, its sensitivity, and its adaptive capacity”. Climate Change United Nations Office of Disaster Risk Reduction UNDRR (2009) “… the characteristics and circumstances of a community, system or asset that make it susceptible to the damaging effects of a hazard” Hazard and disaster risk Intergovernmental Panel on Climate Change IPCC (2014a) “… the propensity and predisposition to be adversely affected. Vulnerability encompasses a variety of concepts and elements including sensitivity or susceptibility to harm and lack of capacity to cope and adapt” Climate Change There are several advancements in defining vulnerability over time. Vulnerability was once seen as possible losses from a hazard event and mainly focused on individuals. Later, it was defined in terms of the socio-economic characteristics of individuals, households and communities exposed to (and affected by) hazards or climate change. In the last two 23 decades, the definition of vulnerability has become more system-oriented in that it considers human, social, economic, physical and environmental systems, as well as their characteristics, that make people or ecological systems likely to be adversely affected. Thus, coupled socio-ecological systems were placed at the centre of vulnerability analysis (Turner et al., 2003; Birkmann et al., 2013). These social, economic, physical and environmental systems refer to the thematic dimensions of vulnerability (Birkmann et al., 2013). Overall, several studies have defined vulnerability as a function of exposure, susceptibility/sensitivity/fragility, and a lack of a system’s capacity (see Table 2.1). Climate change research (Füssel et al., 2006; IPCC, 2007) has mostly centred on the impact-oriented perspective of vulnerability—regarding probability character, magnitude, and the rate of climate change and variation—which differs from the definition of the disaster risk reduction community (Birkmann, 2013b). IPCC Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX) and Fifth Assessment Report (AR5) defined vulnerability in terms of susceptibility and a lack of capacity while considering exposure as a separate factor. No