Evaluating green roof detention with experimental and modeling approaches

Each year more urban areas mandate green roofs as an effective solution for mitigating stormwater runoff in densely developed areas where space is limited.  Green roofs can retain much of annual rainfall, however, relying on retention alone can have little impact on the outflow rate in larger storms.  Similarly in ‘2nd day storms’ or multiple day rain events, the retention capacity is reduced from the previous event significantly impacting the performance during the 2nd event.  These factors combined limit the ability of green roofs to meet increasingly strict municipal stormwater requirements.  Modern approaches in green roof design have begun including a detention component which allows for more temporary storage during large rain events and has the potential to significantly reduce peak flow rates compared to retention alone.  Adding these detention components to traditional designs can improve performance making it a more practical solution in a stormwater management plan and help increase their overall adoption.  Because this detention based approach is relatively new in green roof design, understanding the proper way to quantify performance is critical to their success.  

Performance of various green roof profiles and configurations were evaluated using Green Roof Diagnostics’s (GRD) rainfall simulator under a number of different conditions including rainfall intensity, storm type, initial soil moisture, and storm duration.  Using this extensive dataset, the rainfall-runoff response can be characterized.  Typical modeling approaches were evaluated including the Curve Number method and Rational Method common in municipal guidance manuals.  Because of the limitations of these methods, other models were considered drawing inspiration from traditional aquifer hydraulics and groundwater hydrology.  Further, EPA SWMM models were created in an attempt to characterize performance.  These methods are compared to explore accuracy, applicability, and practically for use by design engineers.