Hydrological Performance Assessment for Green Roof with Various Substrate Depths and Compositions

Low Impact Development (LID) techniques have been drawing an increasing attention in recent years as efficient ways of stormwater management. These methods try to mimic nature by providing pervious area which has been decreased due to urbanization. One of the most efficient practices of LID is the use of Green Roof (GR) systems. Although GRs have been traditionally used in the past, their engineering design and construction is a recent trend. Since GR systems, similar to other LID techniques, are supposed to endure local climate and environmental conditions, their design should be accomplished regarding the conditions of the area where they will be installed. In order to systematically examine, design, and construct various LID techniques for Korean climate and environment, the LID & GI Center of Korea is built in Pusan National University. This center tries to hire hardware and software facilities to study and design LID practices; Hydrological Performance Tester for Green Roof (HPT-GR) and Korea Low Impact Development Model (K-LIDM) are made for this purpose. In order to examine the performance of a Green Roof (GR) system composed of layers with different material and size, five trays were made at LID & GI Center of Korea. The size and composition of these trays were: (A) plant + soil + drainage mat with total depth of 15 cm, (B) plant + soil + sand with total depth of 15 cm, (C) plant + soil + sand with total depth of 9 cm, (D) a bare soil with total depth of 15 cm, and (E) a concrete with total depth of 15 cm. All the trays were examined under different uniform precipitations with rates of 50, 80, 100 and 150 mm/h using HPT-GR. The results showed that the use of drainage mat in tray A could significantly improve the performance of the GR by increasing its storage capacity up to 40% of the rainwater, decreasing its saturation rate, and reducing its discharge peak value. Then, as the first use of K-LIDM for simulating a green roof, all the experimental GRs were numerically simulated using this model; the numerical hydrographs were compared to those achieved from the experiment. The results showed that K-LIDM can be used for performance prediction of GR systems with slightly conservative but acceptable results. With an increase in the rainfall intensity, the discharge peak values increased and their times decreased in both numerical and experimental scenarios; however, the peak values were slightly higher in simulation results than those of the corresponding experimental ones.