为科学定量评价不同降雨特征对多雨城市海绵改造小区径流控制效果的影响，该文基于在线监测和模型模拟相结合的方法，以某多雨地区海绵城市试点区内一个典型小区为研究对象，布置代表性的监测点对其排口流量和悬浮物（suspended solid，SS）进行连续在线监测。在此基础上构建暴雨洪水管理模型（storm water management model，SWMM），开展海绵小区改造对径流水量控制和水质净化效果评价的实证研究。利用实测在线降雨水量水质数据对模型关键参数进行率定和验证，Nash-Sutcliffe效率系数均大于0.71。在实际场次降雨和典型年降雨情景下，模拟分析了海绵改造对雨水径流水量控制和水质净化的效果。结果表明，海绵改造后，小区范围内的年径流总量控制率为52.9%，年径流削减率为28.0%，污染物负荷削减率为66.3%。整体而言，共计81场降雨事件下的径流控制率均达到70%以上。同时，研究发现极端降雨事件与各降雨事件下的降雨特征均对场降雨径流控制能力有不同程度的影响。研究区2009年内平均场降雨控制率为86.5%，而当分别不考虑极端小降雨/暴雨事件时，平均场降雨控制率为79.9%和88.2%。降雨量与雨前干期对场降雨径流控制能力的影响较大，在小雨（< 15 mm）、中雨（10~25 mm）及大雨（>25 mm）3种不同情况下，研究区的场降雨径流平均控制率分别为98.1%、73.8%和52.9%。在雨前干期大于等于6 d的降雨事件下（共15场），场降雨径流控制率均达到90%以上。而降雨历时与平均雨强对场降雨径流控制率的影响并不显著。

Runoff control and non-point source pollution reduction in a sponge reconstruction community in a rainy city were studied using online monitoring and model simulations of a typical community in the sponge city pilot area. The rainfall-runoff volume and the suspended solid (SS) concentration were monitored continually at the two outlets from the community. A storm water management model (SWMM) was used to simulate the runoff control and non-point source pollution reduction in the sponge city design. Then, the model parameters were calibrated and validated against a natural rainfall event during the monitoring period. The results showed that all of the Nash-Sutcliffe efficiency coefficients were greater than 0.71, which ensured the model accuracy. Finally, the model was used to evaluate the effect of total runoff control and non-point source pollution reduction for 2 scenarios "before" and "after" the sponge city construction. The results showed that the sponge city construction helped control the runoff volume and reduce the non-point source pollution. After the sponge city construction, the total annual runoff rate was 52.9% of the total rainfall, the annual runoff was reduced by 28.0%, and the pollutant load was reduced by 66.3%. Overall, the runoff rate exceeded 70% for 81 rainfall events. Extreme rainfall events and the rainfall characteristics for each rainfall event have varying degrees of influence on the rainfall control rates. The average rainfall control rate in this study area in 2009 was 86.5%. The average rainfall control rates were 79.9% for light storms and 88.2% for severe storms. The rainfall amount and the preceding dry period strongly influenced the runoff control capacity. The average rainfall control rates were 98.1% for a light rainfall (< 15 mm), 73.8% for a moderate rainfall (10~25 mm) and 52.9% for a heavy rainfall (> 25 mm). For rainfall events after dry periods of 6 days or more (15 events in total), the rainfall runoff control rate exceeded 90%. The rainfall duration and average rainfall intensity had little effect on the rainfall runoff control rate.