Integrating social narratives of flood events into a text network analysis-based decision support framework to reduce vulnerability to climate change in Africa

Abstract

In many African countries, the response to climate change is obstructed by a lack of accessible and usable information, such as localised flood maps. Compounding this, current disaster risk management systems often fail to account for context-specific drivers of social vulnerability and environmental risks, crucial for enhancing social resilience to flood impacts. This paper captures the community-based narratives of flood risk in Lusaka, Zambia. Using a well-established network from the Future Resilience for African Cities And Lands (FRACTAL) group, a cross-disciplinary approach of natural and social sciences to support decision-making for flood resilience is presented as the Participatory Climate Information Distillation for Urban Flood Resilience in Lusaka (FRACTAL-PLUS) project. Local flood inundation maps were created using global rainfall and GIS datasets and then analysed across two interactive “Learning Labs” with local stakeholders. Historical observations and lived experiences were distilled from the learning labs into three community-based social narratives of flood risk. These narratives were used to calibrate the flood maps with insights from Lusaka’s stakeholders using Natural Language Processing (NLP) and Text Network Analysis (TNA). The narrative-informed flood maps provide a dynamic entry point for enhancing stakeholder engagement by discussing social vulnerability to floods and climate change, highlighting future challenges and opportunities for resilience planning. The outputs demonstrate the value of convening stakeholders to discuss these topics in a sustainable setting for addressing the interdisciplinary challenges of climate resilience, offering a benchmark for better use of available resources and enabling a swift evaluation of needs and measures for resilience building.