Flooding in mega-cities: using structural equation modeling to assess flood impact in Dhaka


Purpose: This paper aims to bring the more recent discourse on the multilayered and interconnected dimensions of flood vulnerability, damage and risk reduction at the microlevel of global south cities to Dhaka, by looking at multiple factors and their relationships. Design/methodology/approach: A cross-sectional research design was used to generate data from 315 respondents in five neighborhoods in Eastern Dhaka, located in high flood damage zones with previous flood experience, using a structural equation model to test nine hypothetical relationships. Findings: The model confirms that low socioeconomic conditions often lead households to use social capital to traverse flood vulnerabilities in cities. It also advances this notion to show that flood impact unleashes social capital through collective activities in responding to flooding. Further, it reveals that while socioeconomic conditions influence flood impacts, these also engender the necessary mechanisms to unleash collective responses to flooding. Practical implications: This paper suggests the need for context-specific interventions that transcend physical and infrastructural responses to integrate socioeconomic conditions as a basis of understanding and addressing flood vulnerabilities. To achieve this requires transcending generic participatory mechanisms to use frameworks that encourage genuine participation and partnerships using coproduction. Originality/value: This paper engages both the inner city and peri-urban areas of Dhaka to extend current conversations on the various conditions underlying flood impact to offer entry points for integrated flood management interventions at the microlevel. This paper contributes to fill the research gap in Dhaka where very few studies have examined flood damages to residential buildings and its driving factors at the neighborhood level.

Publication Title

International Journal of Disaster Resilience in the Built Environment