Monitoring and Modeling the Hydrological Performance of Extensive Green Roof Systems

Urban stormwater runoff causes many problems for watersheds located within large metropolitan areas, including such detrimental effects as flooding, erosion, pollution, and the increased risk of combined sewerage overflows. Increased amounts of impervious areas resulting from urban sprawl have also been shown to escalate stormwater flows, which exacerbates water management issues in these metropolitan areas. Water resource engineers have progressively turned toward green infrastructure to solve stormwater problems, and green roof systems represent one type of this green infrastructure. As of current, however, green roof systems are largely underused in as an effective stormwater management tool. The major factor limiting the installation of green roof systems is the unpredictable hydrological response of green roofs to individual storm events. Currently, many municipalities use the Soil Conservation Service model or rational method and associated curve numbers to estimate stormwater flows, with green roofs typically receiving an assigned value ranging from 75-90 within these models. However, these simple models do not accurately predict the hydrological response of green roof systems, where the overall performance is determined by many supplementary factors including geometry, soil media type and depth, initial conditions, and the individual storm hyetograph. The accurate monitoring of green roof stormwater runoff and the use of data to create models are critical to measuring hydrological response, as well as to assess the benefits of the green roof installation to the local watershed. In this study, four 15 m2 test plots were constructed on the roof of the Milwaukee Metropolitan Sewerage District headquarters located in downtown Milwaukee, Wisconsin. An ET 107 weather station manufactured by Campbell Scientific was installed onsite. Stormwater flows were monitored for each plot using a “WeirBox”, a tipping gauge and v-notch weir combination which was developed and calibrated specifically for this project. Three extensive green roof systems manufactured by Vegetal i.D. were tested, including Hydropack, a standard modular extensive green roof, and Hydro Active Smart Roof (HSRP) and Hydro Active Smart Roof Active (HSRA), both of which are extensive green roof systems with additional water storage basins. A control bare roof plot was also monitored to confirm and compare hydrological performance. The WeirBox flow monitoring equipment displayed impressive results with water budget error typically less than 7% for individual storm events when comparing total runoff volume from the control plot to onsite precipitation data. All three of the tested green roof systems exhibited significant hydrological performance in terms of total runoff retention, peak runoff rate reduction, and peak runoff rate delay. However, depending mostly on rainfall characteristics, the responses to individual storm events varied widely. Total runoff retention for the 8 month monitoring period was calculated to be 64%, 87%, and 91% for Hydropack, HSRP and HSRA respectively. In general, both HSR systems with greater water capacities outperformed the standard extensive green roof system, which suggests that optimization through an integrated storage basin can be achieved to improve overall green roof performance. A conceptual simple bucket model was created for the Hydropack and HSRP green roof systems. Data from 6 individual storm events was used to validate the model. While “simple”, the conceptual bucket model successfully reproduced the hydrological response of the green roof systems to individual storm events. Synthesized storm events with return periods ranging from 1 to 100 years were then analyzed using the calibrated models. Hydrological performance diminished with larger storm events, mirroring results from monitoring and literature review. Both monitoring and modeling showed that the integration of extensive green roofs with storage basins greatly improves performance