Simulation and optimization of rain gardens via DRAINMOD model and response surface methodology

Abstract

This study aims to optimize the key parameters of rain gardens by combining the hydrological models (RECARGA and DRAINMOD) with the response surface methodology (RSM). The impact factors and design parameters which are related to the regulation effects of the two rain gardens (one is traditional media, the other is modified media) are studied. Results indicate that the DRAINMOD shows greater ability to optimize the rain gardens comparing with RECARGA. The runoff reduction rates are significantly decreased by 16.0% and 19.9% by increasing rainfall conditions and the ratio of rain garden area to catchment area (Rr/c) respectively. NO3-N is significantly reduced by (22.5%, 22.5% and 17.9%) on different rainfall conditions, concentrations and Rr/c, respectively, and the levels of pollutant reduction are better than NH3-N. With the thickness of aquifer, filler layer and internal aquifer increasing, the runoff reduction rates are increased by 19.1%, 38.7% and 5.0%, respectively. Meanwhile, the load reduction rates for NO3-N are increased by 5.6%, 28.9% and 16.4%, and for NH3-N are 6.2%, 17.0% and 14.0%, respectively. The RSM is utilized to explore the correlation between different scenarios and water quantity & quality reduction rates, which is used in the optimization of the rain gardens. Optimization results show that with the constraint of higher inflow concentration and 80% of the annual runoff reduction rate, the total height of rain garden with modified media is lower than the traditional one. The results provide a theoretical basis and important data support for the rational construction of rain gardens.