Improvement of simulating rain gardens to advance sustainable stormwater management

Rain gardens can effectively manage stormwater runoff. However, existing hydrological models cannot accurately evaluate rain garden performance, hampering informed decision-making. This study improved the simulation of rain garden's key processes in the Soil and Water Assessment Tool (SWAT), including implementation area, surface runoff distribution, bypass and orifice flows, percolation dynamics, and sub-daily temporal resolution. The improved SWAT was calibrated/validated using field-scale observed rain garden data, which showed good model performance. And then the model was demonstrated in Brentwood watershed (Austin, TX) to assess the long-term impacts of various rain garden designs. Increasing the fraction of runoff from pervious/impervious areas draining to rain garden significantly reduced discharge volume (37.41%-65.02%), peak discharge (e.g., 43.62%-57.32%), and combined sewer overflow (CSO) (e.g., 42.81%-63.53%). Deeper amended soil layers of rain gardens only marginally reduced discharge volume (64.86%-65.08%), peak discharge (e.g., 57.11%-57.39%), and CSO (e.g., 63.35%-63.59%). A larger ratio of rain garden's surface storage area to the pervious area substantially decreased discharge volume (56.81%-74.56%), peak discharge (e.g., 48.56%-67.55%), and CSO (e.g., 54.87%-72.80%). Increasing the depth of rain garden's surface storage and height of the orifice from the bottom of the rain garden's surface storage slightly reduced discharge volume (63.63%-67.75%), moderately decreased peak discharge (e.g., 55.83%-60.32%), and slightly lessened CSO (e.g., 62.35%-65.95%). Among different types of amended soil from sand, loamy sand, sandy loam, to loam exhibited slightly lower effectiveness in reducing discharge volume (65.61%-63.34%), peak discharge (e.g., 57.87%-56.17%), and CSO (e.g., 64.19%-61.86%). Therefore, the improved SWAT can provide valuable decision-support for optimizing rain garden design.