Stormwater retention and detention performance of green roofs with different substrates: Observational data and hydrological simulations.

Green roofs are widely considered as a promising nature-based solution for urban stormwater management. In this study, the stormwater retention and detention performance of 6 green roof modules with different types and depth of substrates at Beijing, China was investigated through 3-year continuous monitoring. The Hydrus-1D was applied to further explore the stormwater management performance of green roofs under extreme storms. The average event-based stormwater retention and detention rates of the green roofs with 10 cm substrates ranged between 81% and 87%, and 83%-87%, respectively; and the average time delays in runoff generation and peak discharge ranged between 82 and 210 min, and 63-131 min, respectively. Green roofs with 15 cm depth of substrates offered higher stormwater retention and peak runoff attenuation rates than those with 10 cm substrates. However, due to the high frequency (55 out of total 92) of light rainfall events (<10 mm) and short antecedent dry weather periods (3.8 days in average), no significant difference was found on stormwater control performance of those green roofs. The Hydrus-1D simulations revealed that green roof stormwater retention rate decreases exponentially with return periods of extreme storms but increases with substrate depth. There exists a critical depth of substrates and further increases in substrate depth beyond this critical value could not bring much improvement in stormwater retention performance of green roofs. The application of extensive green roofs with 10-15 cm substrates provides promising stormwater retention and detention performance in highly urbanized area of Beijing.