Human mobility dynamics for cost-effective flood emergency resource allocation

Abstract

Extreme flooding displaces millions of people annually worldwide. However, emergency resources are often inefficiently allocated, as conventional strategies usually overlook mobility dynamics and thereby exacerbate vulnerability. Here, by capturing flood-induced human dynamics using agent-based modelling, we investigate the emergency resource allocation issue (official emergency shelters and dual-use public service facilities) across China’s 21 large cities. We find that 43% of these cities face notable resource demand‒supply mismatches (>2,000 km2 per city) during 100-year floods, exposing 16.76 million residents to preventable risks. Our SSP5-8.5 projections indicate that substantial mismatch will persist by 2050 despite the declining population in China, while flooded areas are expected to expand fivefold owing to climate change. As a solution, strategically retrofitting 30% of high-deficit zones yields benefit-to-cost ratios exceeding 28 via reduced mortality, societal co-benefits, carbon savings and economic continuity (US$0.20–3.05 billion savings per city with a per capita investment of merely US$1.39). These findings underscore the critical links between flood-induced human behaviour and emergency resource optimization in climate-adaptive urban development. Flood emergency resources are often misallocated as planning relies on static population data rather than dynamic human mobility during disasters. A study now models flood-induced mobility across 21 Chinese cities, revealing widespread resource mismatches and cost-effective retrofitting strategies.