Developing an urban poor-centred (multi-)hazard impact categorisation using multiple data sources: an application to the Kathmandu Valley, Nepal

Here we present a systematic approach to developing an urban poor-centred (multi-)hazard impact classification using multiple data source types, with application to the Kathmandu Valley, Nepal. Marginalised communities, including urban poor communities, are typically neglected from impact data sources, despite these groups often experiencing disproportionate impacts of (multi-)hazard events and having a lower capacity to respond. Gaps in impact data are particularly challenging in regions of data scarcity, where comprehensive evidence bases would support the refinement of existing DRR strategies.

We extracted (multi-)hazard impact exemplars from disaster databases (DesInventar Sendai and the Nepal DRR Portal) and newspaper articles (LexisNexis online newspaper archive) utilising systematic (Boolean) searches. We applied the searches to earthquake, flood, landslide and urban fire events owing to their prevalence in the study area. Following this, we manually reviewed the results for relevancy to specific named informal settlements in the Kathmandu Valley. We supplemented these data with insights from three focus group discussions (FGDs) conducted with residents of informal settlements in the Kathmandu Valley and 11 semi-structured interviews with DRR practitioner stakeholders working with these communities. We co-facilitated the FGDs with members of Nepal Mahila Ekata Samaj (NMES, https://mahilaekata.org/), a network organisation of landless women in Nepal.

We compiled the disaster database, newspaper article, FGD and semi-structured interview results into an Excel database of urban poor-centred (multi-)hazard impacts across the four natural hazard types. Within each row of the database, we included details of the source type, (multi-)hazard event details, and impact information categorised by type. Our results indicated that the disaster databases (45 relevant exemplars) presented an overview of (multi-)hazard event details. However, documentation of impacts was typically restricted to quantitative tangible impacts – including economic losses and fatalities. Newspaper articles (83 relevant exemplars) provided nuance to descriptions of (multi-)hazard impacts, with quotes from affected individuals adding socio-political context. Finally, FGD and semi-structured interview participant perspectives of (multi-)hazard events offered richness through lived experience and qualitative accounts, with an emphasis on disaggregated and intangible impacts.

Applying an iterative approach, we compiled the results into an urban poor-centred (multi-)hazard impact categorisation. This typology summarises the impacts, grouped into categories and subcategories, that affect members of urban poor communities in the (multi-)hazard context of the Kathmandu Valley. In gathering multiple data sources of (multi-)hazard impact, we illustrate the value of supplementing quantitative and qualitative data to evidence a more holistic understanding of impact in data-scarce regions, with the intention of centring urban poor community perspectives. We suggest that our methodology and the development of the urban poor-centred (multi-)hazard impact categorisation could provide a framework for scalability to other data-scarce regions, supplementing existing evidence bases to support more inclusive DRR strategies.