Multiobjective Optimization of Residential Rain Garden Siting Using ArcGIS Suitability Modeling and NSGA-II Algorithm

Urbanization increases stormwater runoff and places growing pressure on aging gray infrastructure, such as deteriorating culverts and unmanaged regional detention systems. These challenges have increased interest in green stormwater infrastructure (GSI) as decentralized stormwater control measures. Recent geospatial and computational hydrology models, such as Community-Enabled Lifecycle Analysis of Stormwater Infrastructure Costs (CLASIC), integrate multiple aspects of stormwater planning, including hydrologic responses, stormwater quality, life-cycle costs, triple bottom line (TBL) cobenefits, and climate change factors in a single, user-friendly web-based tool. However, limited guidance exists for determining optimal spatial allocation of distributed GSI installations. This study integrates the CLASIC tool with ArcGIS suitability modeling and Nondominated Sorting Genetic Algorithm II (NSGA-II) to explore alternative GSI deployment within a 64.7-ha (160-acre) pilot study area in the Proctor Creek watershed in Atlanta, Georgia. The analysis is conducted for a 95th percentile storm event, 45.7 mm (1.8-in.), 24-h design storm. Two representative GSI configurations (n=21) were identified with trade-offs between land suitability and flow interception. This study advances green stormwater infrastructure planning from land suitability–based screening toward performance-driven spatial allocation. Future research should validate results against calibrated models, incorporate multistorm evaluation, and incorporate explicit cost surfaces with equity and ownership constraints.