Unveiling vulnerabilities: Integrating socioeconomic and physical dimensions in residential exposure modelling

Natural hazard exposure modelling involves compiling a comprehensive database of elements at risk, such as people, buildings, and infrastructure, within a specified area of interest. Residential building (physical) exposure, in particular, is a critical component of disaster risk modelling and management. However, it cannot be examined in isolation from the socioeconomic characteristics of its residents, especially when estimating advanced, people-centred disaster impact metrics (e.g., population displacement) or disaggregating traditional metrics (e.g., financial losses) to account for the unequal impact of hazards on various societal segments. Household income, alongside other socioeconomic variables, plays a pivotal role in shaping decisions made by homeowners or renters regarding the structural vulnerability of their dwellings. Key decisions - such as home maintenance, the selection and quality of construction materials, compliance with building design codes, and the adoption of private insurance - are profoundly influenced by income levels. These factors, although frequently overlooked in conventional building exposure assessments, can significantly impact risk estimates and recovery planning. The main objective of this study is to develop a residential building exposure model that integrates both physical (i.e., building types) and socioeconomic dimensions (i.e., household income levels) in Saint Lucia, in the eastern Caribbean Sea. The model aims to test the hypothesis that income-poor households are disproportionately likely to inhabit physically vulnerable dwellings. To this end, the study introduces SimLucia, a residential building exposure model to hurricane hazards. Employing spatial microsimulation methodologies, SimLucia combines geographically aggregated data from the 2010 National Population and Household census with a-spatial microdata from the 2016 Living Conditions and Household Budgets Survey. The result is a virtual household population dataset that incorporates both structural and economic attributes at the district level. Outcomes from the SimLucia model reveal that low-income households are over twice as likely to reside in more physically vulnerable building typologies, thereby confirming the initial hypothesis. The SimLucia model can serve as a decision-support tool, enabling targeted, human-focused, and pro-poor policy interventions (e.g., housing recovery financing, income-generating instruments).