Abstract Despite the benefits of green roofs in managing urban stormwater quantity and quality, a number of studies have demonstrated that green roofs can pose negative impacts on the urban environment due to chemical leaching, in particular in their early age. Besides design variables such as growing media composition and depth, the roof age and hydro-meteorological variables are also expected to affect or govern the temporal evolution of chemical leaching. To characterize the chemical leaching behavior of green roofs and explore possible modeling approaches, a full-scale extensive green roof and a reference roof, which are located in a cold and semi-arid climate region, were monitored in both rain and snowmelt events in 2015–2018. The temporal evolution of chemical leaching was examined at both intra- and inter-annual time scales. The roles of hydro-meteorological variables including the growing media temperature (GMT), moisture condition (antecedent moisture condition (AMC) for rain events, and moisture condition (MC) for snowmelt), event rainfall amount (Ra), and cumulative inflow/precipitation amount on chemical leaching were investigated. The field observations demonstrated the leaching of nutrients (i.e., both nitrogen (N) and phosphorous (P)) and conductivity from the green roof, whereas the roof acted as a sink for metals (Zn, Cu, and Pb). The leaching of N appeared to cease, whereas P leaching was still ongoing at the end of the study period. Although the degree of nutrient leaching was not significantly different between the rain and snowmelt events, the nutrient leaching in rain events appeared to be relatively higher in the spring than in the summer and fall. Furthermore, the leaching of nutrients was found to decline annually, but at different rates for different nutrients. Among the investigated hydro-meteorological variables, the cumulative inflow was identified influencing the temporal evolution of chemical leaching considerably. In addition, the results from two modeling approaches, namely multiple linear regression modeling and semi-physically based modeling adapted from the pollutant wash-off concept for urban stormwater runoff, confirmed the critical role of the cumulative inflow on chemical leaching. Besides the cumulative inflow, the GMT was the other important explanatory variable for P leaching. The results suggested that chemical leaching from green roofs in semi-arid regions, where precipitation and moisture level are low, could persist longer but at a lower degree compared to that in mild and temperate climate regions.

中文翻译：

---

寒冷半干旱气候下全尺寸绿色屋顶的化学浸出行为

摘要 尽管屋顶绿化在管理城市雨水的数量和质量方面有好处，但许多研究表明，屋顶绿化会因化学浸出而对城市环境造成负面影响，尤其是在其早期。除了介质成分和深度等设计变量外，屋顶年龄和水文气象变量也有望影响或控制化学浸出的时间演变。为了表征绿色屋顶的化学浸出行为并探索可能的建模方法，在 2015 年降雨和融雪事件中对位于寒冷和半干旱气候区的全尺寸大面积绿色屋顶和参考屋顶进行了监测–2018。在年内和年际时间尺度上检查了化学浸出的时间演变。水文气象变量的作用，包括不断增长的介质温度 (GMT)、水分条件（降雨事件的前期水分条件 (AMC) 和融雪的水分条件 (MC)）、事件降雨量 (Ra) 和累积流入/考察了化学浸出的沉淀量。现场观察表明，绿色屋顶会浸出养分（即氮 (N) 和磷 (P)）和电导率，而屋顶则充当金属（锌、铜和铅）的汇。在研究期结束时，N 的浸出似乎停止了，而 P 的浸出仍在进行。虽然降雨和融雪事件的养分浸出程度没有显着差异，但降雨事件中的养分浸出程度在春季似乎高于夏秋季。此外，发现养分的浸出逐年下降，但不同养分的浸出率不同。在调查的水文气象变量中，确定了累积流入量显着影响化学浸出的时间演变。此外，两种建模方法的结果，即多元线性回归建模和适用于城市雨水径流的污染物冲刷概念的半物理建模，证实了累积流入对化学浸出的关键作用。除了累积流入量，GMT 是 P 浸出的另一个重要解释变量。结果表明，降水和水分含量低的半干旱地区绿色屋顶的化学浸出，