Design Optimization of Rain Gardens for Effective Stormwater Management Using the Response Surface Method

Rapid urbanization and climate change have placed increasing pressure on urban stormwater management systems, underscoring the need for sustainable, adaptive drainage solutions, such as low-impact development practices. However, the current optimization process, which involves a broad range of input parameters, carries the risk of evaluating system efficiency under site-specific conditions and lacks applicability to the design process. This study establishes an adaptive design guideline for rain gardens (RGs) in Hue City, Vietnam, by coupling hydrological simulations with a Response Surface Methodology-based optimization framework. The Rosetta3 model was employed to parameterize soil hydraulic properties, and K-means clustering was used to characterize representative rainfall patterns. Sensitivity analysis identified RG area, berm height, and soil thickness as the most influential parameters for runoff volume and water pollution reduction. Adaptive design guidelines for those influential parameters were developed. Results indicated an inflection point in diminishing returns at a 1% area ratio, where runoff-reduction efficiency was maximized relative to spatial cost. The optimization of berm height and soil thickness achieved pollutant removal efficiencies ranging from 34% to 79.5%. Hydrological analysis under extreme events (R3 and R4) revealed the limitations of standalone RG systems in mitigating runoff volume. A management train strategy integrating RGs with detention ponds was proposed, achieving peak flow reductions of up to 50%. This study provides an adaptive decision-support framework based on parameterized and integrated hydrological simulations. This framework enables stakeholders to select design parameters that balance hydrological performance and spatial constraints, enhancing system resilience to increasing climate-related risks.