Exploring the response of urban storm sewer system to the implementation of green roofs

With the increasing popularity of Best Management Practices (BMPs), research about the effect of BMP implementation on the hydrologic response of urban watersheds has become a hot topic. Green roofs, which have been reported as an effective means of reducing urban storm water runoff, have been widely used in Europe and are gaining increasing attention in the US. This thesis presents a method to study the changes of urban catchment runoff behavior from green roof implementation. The proposed method takes into account the thresholds in green roof response and abrupt changes in areal characteristics for better understanding of urban catchments. This allows the method to correctly scale various types of green roofs, and ultimately other BMPs, for a range of storm characteristics and BMP implementation. This study used the Illinois Urban Hydrologic model (IUHM) as a base model for urban sewer system, except that the IUHM model was modified to allow the overland flow to be routed from impervious to pervious areas. A mathematical model was developed to simulate the major processes affecting the hydrologic response of green roofs (ponding, infiltration, drainage, and recession). The green roof model was calibrated to match the runoff observed from the green roof on the Business Instructional Facility (BIF) at the University of Illinois campus. The green roof model was incorporated into the probabilistic IUHM model by coupling the hydrologic response of the green roof from the green roof model with the probability of rain falling on a green roof and the probability of the green-roof runoff following different possible paths to the watershed outlet. The modified IUHM model, herein named Combined IUHM-Green Roof Model (CIGM) was used to test green roof impact on hydrologic response at the catchment scale, with the green roof coverage ranging from 10% to 40% of the impervious area of the whole watershed. Two hypothetical storms (a uniform storm with an intensity of 7.62 mm/hr for 2 hours, and a 5yr Average Return Interval storm for 2 hours,) and two real storms (a January 2008 storm, and a July 2007 storm) were used in the model to test green roof performance under different rainfall types. A SWMM model was also developed and used for comparison purpose. Results indicate that the implementation of green roofs has a distinct effect on decreasing the volume and peak of storm water runoff and postponing the time to peak runoff. The volume and peak discharge reduction depends on the temporal distribution of rainfall as well as the volume of rainfall and green roof coverage. The model also shows that routing overland flow from impervious to pervious areas can significantly reduce the catchment outflow and postpone the time to peak discharge. This overland flow routing alteration shows an even bigger effect on catchment outflow response than the change of green roof coverage does. In addition, catchment outflow can be further decreased by routing green roof runoff to pervious areas rather than impervious areas