Green infrastructure layout optimization and screening for stormwater management in dense urban catchments under short-return-period rainfall scenarios

ABSTRACT

Green infrastructure (GI) optimization typically focuses on extreme storms and yields a single compromise layout, overlooking frequent rainfall conditions and diverse stakeholder preferences. This study proposes a GI layout optimization and screening framework coupling SWMM with NSGA-II and a two-stage clustering-decision approach. Taking a dense urban catchment in Ningbo's Sponge City pilot zone as a case study, high-dimensional GI optimization targeting short-return-period rainfall was implemented with 520 configuration variables. Synthetic design rainfall return period greatly affected cost–runoff–pollutant trade-offs. As return period increased from 0.25 to 1 year, the cost range narrowed from 59.59–152.24 to 61.29–140.60 million CNY, and runoff reduction rates declined from 31.04–48.92% to 26.32–45.77%. Integrated pollutant removal remained relatively stable, varying from 35.49–51.80% to 33.30–51.05%. Three representative schemes were identified: environment-prioritized, cost-driven, and balanced. Their hydrological responses were rainfall-sensitive, whereas pollutant removal remained stable among the investigated scenarios, with averages of 47.3% for total suspended solids, 47.0% for chemical oxygen demand, 49.9% for total nitrogen, and 44.6% for total phosphorus. Sensitivity analysis confirmed their rankings remained stable under ±20% unit cost fluctuations. This framework offered a generalized methodology to generate practical GI layouts for dense urban catchments.