Optimizing the integrated green-gray-blue system helps improve urban flood resilience under a non-stationary climate

Climate change has intensified extreme rainfall events, introducing non-stationarity into climatic patterns and heightening the threat of urban flooding. While green-gray-blue infrastructure is being explored globally to enhance flood resilience, limited understanding remains regarding the optimal scale and spatial configuration of these measures under evolving climatic conditions and growing surface flood risks. This study proposes a novel four-step framework that integrates design rainfall under non-stationary conditions, urban flood resilience assessment, coupled hydrological-hydraulic modeling, and multi-objective optimization. This framework is applied to Shahekou District, Dalian City, China. The results indicate that non-stationary conditions have significantly increased the intensity of short-duration extreme rainfall. Under non-stationary conditions, rainfall intensity during 2-h events increased by 6% for the 50-year return period and 8% for the 100-year return period compared with stationary scenarios. Among the interventions evaluated, pipeline desilting provided the most substantial improvement in flood resilience. An orthogonal design indicates that pipeline dredging was the most significant factor in both the high-resilience and low-investment scenarios, yielding values of 100.54 and 141.77, respectively. Each optimized layout enhanced system resilience in distinct ways. The optimal integrated solution achieved a regional flood resilience value of 0.990 while maintaining a favorable return on investment. The findings suggest that future urban development prioritizes gray and blue infrastructure, while selectively integrating green infrastructure in specific areas to maximize the balance between benefits and flood resilience.