Abstract
In recent years, measures proposed to address urban flooding caused by extreme rainfall often demand substantial investment, restricting their broad implementation. This study quantitatively assessed the inundation situations of 138 capital cities under both normal and extreme rainfall conditions. Using machine learning techniques, we found that grey infrastructure-closely commensurate with a city's economic development-dominates flood reduction during normal rainfall events. However, during extreme precipitation, as rainfall intensity rises, the marginal effectiveness of grey infrastructure declines markedly. In contrast, green infrastructure and topography-less commensurate with economic development-play increasingly critical roles in mitigating urban flooding. These findings suggest that economic development has a limited impact on urban flooding during extreme rainfall events. Rationally utilizing topography and enhancing green spaces provides a cost-effective nature-based solution, which is particularly important for urban planning in low- and middle-income countries undergoing rapid urbanization.