A bioretention unit (BRU) or cell is a green infrastructure practice that is widely used as a low impact development (LID) technique for urban stormwater management. Bioretention is considered a good fit for use in China’s sponge city construction projects. However, studies on bioretention design, which incorporates site-specific environmental and social-economic conditions in China are still very much needed. In this study, an experimental BRU, consisted of two cells planted with Turf grass and Buxus sinica,was tested with eighteen synthesized storm events. Three levels (high, median, low) of flows and concentrations of pollutants (TN, TP and COD) were fed to the BRU and the performance of which was examined. The results showed that the BRU not only delayed and lowered the peak flows but also removed TN, TP and COD in various ways and to different extents. Under the high, medium and low inflow rate conditions, the outflow peaks were delayed for at least 13 minutes and lowered at least 52%. The two cells stored a maximum of 231 mm and 265 mm for turf grass and Buxus sinica, respectively. For both cells the total depth available for storage was 1,220 mm, including a maximum 110 mm deep ponding area. The largest infiltrate rate was 206 mm/h for both cells with different plants. For the eighteen events, TP and COD were removed at least 60% and 42% by mean concentration, and 65% and 49% by total load, respectively. In the reservoir layer, the efficiency ratio of removal of TN, TP and COD were 52%, 8% and 38%, respectively, within 5 days after runoff events stopped. Furthermore, the engineering implication of the hydrological and water quality performances in sponge city construction projects is discussed.
