Spatio-temporal transitions of disaster vulnerability in Nepal

Climate change, in combination with evolving development pathways, is contributing to increasing disaster risks globally. Understanding these risks requires the assessment of risk components, i.e., hazard, exposure, and vulnerability. Among them, social vulnerability is particularly challenging to assess due to its dynamic nature and the limited data availability in resource-constrained, high-risk countries for instance, Nepal. Existing studies in such regions often utilize open-source census data to assess vulnerability using a composite vulnerability index, but overlook spatio-temporal shifts in vulnerability and its components.

To address this gap, our study explores spatio-temporal disaster vulnerability in Nepal by applying Principal Component Analysis (PCA) to municipal-level population census data of 2011 and 2021. We applied PCA separately to individual vulnerability components of both years to identify changes in explanatory indicators. Then, we illustrate disaster vulnerability across Nepal for 2011 and 2021 and assess how it has changed over the decade. Finally, we investigate changes in central vulnerability components, namely, sensitivity and adaptive capacity.

The PCA reveals both continuity and transformation of drivers of sensitivity and adaptive capacity. Migration and literacy newly emerged in 2021 as principal components in sensitivity, while housing ownership and quality, as well as access to electricity, emerged in adaptive capacity. Overall, we observe a slight increase in the aggregated national vulnerability score, with approximately 45% of municipalities exhibiting high vulnerability classes in 2021. Most urban metropolitan cities and lowland regions (Terai) exhibit increased vulnerability, whereas Far Western regions witnessed a slight decrease in their vulnerability levels. A closer look at the shifts in sensitivity and adaptive capacity reveals that the increase in overall vulnerability was largely driven by a strong decrease in adaptive capacity in metropolitan cities and increased sensitivity in Terai regions. These findings suggest that focusing solely on composite vulnerability might lead to misguided mitigation strategies and that dissecting vulnerability into sensitivity and adaptive capacity offers actionable insights for decision-making. Furthermore, our approach supports multi-hazard risk and impact assessments in data-limited settings.

By investigating the temporal and spatial changes in vulnerability components, our study enhances the understanding of vulnerability dynamics in Nepal over the past decade, developing a refined approach for spatio-temporal index-based vulnerability assessments. To illustrate the potential applications of the findings in disaster risk management, we explored sectoral vulnerability interventions through key informant interviews with relevant authorities. Furthermore, our vulnerability assessment is being employed in a flood impact model that aims to identify the main drivers for reported flood fatalities in Nepal.