Evaluation of the Reduction of Rainwater Runoff from a Green Roof Jung Soo Mun 1, a, Hyoung Jun Kim 1, b, Young Jin Kim 1, c and Moo Young Han 1, d 1 Dept. of Civil and Envir. Engrg., Seoul National University, Seoul, Korea, 151-74 a jsmun1@snu.ac.kr, b kimhj9415@hanmail.net, c mukta7@snu.ac.kr, d myhan@snu.ac.kr ABSTRACT According to fast increasing of the area of impermeable regions, the change in rainwater runoff patterns in cities has increased sharply. Therefore, the risk of urban flooding has increased, a distortion of the water cycle has occurred, and the heat island effect becomes serious. A green roof is an appropriate tool for supplying a green area and for managing rainwater to reduce rainwater runoff. However, the reduction in rainwater runoff from green roofs has been reported based on small-scale experiments or simulation studies. Therefore, a study is needed based on a real life scale under various rainfall conditions. In this work, we measured the reduction of rainwater runoff on a green roof that was constructed at a university. The ratio between the measured runoff and the calculated runoff was about.9 for a concrete rooftop. In contrast, the ratio for the green roof was.5.45. The rainwater runoff from the green roof was 8~5% of the runoff from a concrete roof. Moreover, a retardation in the runoff peak flow was also observed for the green roof. This effect was also observed using rainwater storage tanks to reduce the rainwater runoff. Therefore, green roofs are considered to be an alternative to rainwater storage tanks in places where there are spatial limitations in a city. Future research into maximizing the reduction in runoff and on the runoff water quality for rainwater usage is required. KEYWORDS: green roof, runoff reduction, concrete roof, rainfall. 1. INTRODUCTION In most cities, due to ongoing urbanization, green areas are being displaced by impermeable surfaces that are covered with concrete or asphalt. Hence, rainwater runoff has increased sharply, a distortion of the water cycle has occurred, and the heat island effect has become serious. In addition, the safety of the infrastructure of cities is also threatened by climate change. Recently, proper on-site rainwater management has been recognized as a potential solution to solve rainwater runoff problems (Han, ). Rainwater storage and infiltration is usually considered appropriate for on-site rainwater management. However, the high density of land use and the high land prices in a city demand alternative methods for rainwater management be found (Mentens et al., ; Na and Byun, ). A green roof is a rooftop area covered with grass or bushes, and has been proposed as a method for rainwater management in a city to overcome spatial and environmental limits (Dunnett and Kingsbury, 4). A green roof provides a green place in the city (Wong et al., ; Kim and Cho, 4), mitigates the heat island effect (Bass et al., ), and reduces rainwater runoff from impermeable surfaces (Han et al., b). Han et al. (a) showed that rainwater runoff was reduced by % 95% of the total volume of rainfall from extensive green roofs, the maximum runoff was decreased to 8~%, and the peak runoff time was decreased from two to 14 h. In addition, the reduction in rainwater runoff depends on the planting depth and type of plant used (Moran et al., 4). However, reduction in rainwater runoff from green roofs has mostly been reported based on small-scale experiments or simulation studies. Moreover, research into extensive green roofs, intensive green roofs, and the reduction in rainwater runoff from gravel roofs and concrete roofs has been studied in Europe (Mentens et al., ). However, research into the reduction in rainwater runoff is insufficiency in various climate condition, for example a monsoon climate, on a realistic scale. This paper provides the following: (1) a survey of rainwater runoff from concrete roofs, () an analysis of 59
rainwater runoff from a mixed catchment that consists of a concrete roof, extensive green roofs, and a marble terrace to understand the reduction in rainwater runoff and () a comparison of the rainfall runoff from concrete roofs in a mixed catchment.. METHODS AND MATERIALS.1 The Building #9 at SNU The Education Research Building (Building 9) at Seoul National University (SNU) in Korea was constructed in October 5 and used for this study. A schematic diagram of this building is shown in Figure 1. Green roofs and a rainwater utilization system have been installed in Building #9. The catchment area of the rainwater utilization system is,55 m, and the capacity of the main rainwater tank, extra rainwater tank, and supply tank are 5, 7, and 4 m, respectively. Rainwater collected from the concrete roof area of 9 m flows into the main rainwater tank. Similarly, rainwater collected from the concrete rooftop area of 9 m, the green roof area of 95 m, and the terrace area of 84 m flows into a extra rainwater tank, and then, when the water level in the extra rainwater tank reaches a depth of 1. m, the water is transported to the main rainwater tank via a pump. The supply tank has both a rainwater pipe and a city water pipe, with rainwater supplied in preference to city water. However, if there is no rainwater in the main rainwater tank, then the city water is supplied instead of rainwater. Figure 1. A schematic diagram of the rainwater utilization system of Building #9 at the SNU.. The green roof system of Building 9 at the SNU Green roofs generally consist of a vegetation layer, a substrate layer, and a drainage layer (Mentens et al., ), and are usually divided into two main types: extensive and intensive green roofs, based on the depth of the substrate layer (Krupka, 199; Kolb and Schwarz., 1999). Extensive green roofs have a substrate layer with a depth of <15 mm, and mainly have Sedum species planted in them. Intensive green roofs have a substrate layer with a depth >15 mm, and grasses and shrubs are the major constituents of the vegetation planted in them. Figure shows the green roof system in Building #. In this study, an extensive green roof with a substrate layer with a depth of 1 mm was used. The depth of the drainage layer in this system was 5 mm, and Sedum species were used as the vegetation. The area of the green roof was 95 m. The rainwater runoff from the green roof flowed into an extra rainwater tank through a series of pipelines.. Analysis of the rainwater runoff We compared the calculated rainwater runoff (CR) and the monitored rainwater runoff (MR) of a concrete roof and a green roof. The CR value was calculated from the product of a catchment and the rainfall intensity with time. We assumed that there was no loss from the roof runoff in the CR value. 54
Sedum Figure. The green roof system of Building #9 at the SNU. (unit : mm) T natural soil over T8 artificial soil Nonwoven Drainage plate ( ⅹ T5) Concrete rooftop An Automatic Weathering Station (AWS) located at the SNU was used to measure the rainfall volume. The value of MR was estimated from the product of the water level in the rainwater storage tanks and the area of the bottom of the tank (see Table 1). The volume of rainwater runoff from the concrete rooftop (area = 9 m ) was estimated from the change in water level in the main rainwater tank. The volume of rainwater runoff from a mixed catchment area that consisted of a green roof (area = 95 m ), a concrete rooftop (area = 9 m ), and a terrace (area = 84 m ) was estimated from the change in water level in the extra rainwater tank. The water level in the rainwater storage tank was monitored using a float-type level transmitter (HT-1R, Hitrol Co. Ltd., Paju, Rep. of Korea), which measured the water level, and was controlled using a Wonderware-In Touch 9. system (Maha Net Co. Ltd., Korea). Experiments were performed four times on the runoff from a concrete rooftop, and three times on the runoff from a mixed catchment for five rainfall events between June and January 7 (see Table ). Table 1. The calculated and monitored runoff values. Runoff Method Remarks Calculated runoff Catchment area rainfall intensity with time Rainfall data from AWS at SNU Monitored runoff Bottom area of the rainwater storage tank the water level with time Water level from a float-type level transmitter Table. A summary of the rainfall events and conditions. Total rainfall Preceding number Event Duration (h) (mm) of dry days Date Catchment 1 8.89 d 14 h 9// Concrete rooftop 5. 5 h 1/7/ Concrete rooftop. Mixed area.5 d 4/7/ including green roof 4.81 d 1 h 9/9/ Mixed area including green roof 5 5.59 4 19 d 1//7 Concrete rooftop Range.8 8.9 h 19 d / 1/7. RESULTS AND DISCUSSION.1 Rainwater runoff from a concrete rooftop The accumulated CR and MR values of a concrete rooftop in Events 1 to and 5 were investigated, and the case of event 1 and 5 are shown in Figure. Four rainfall events were used from June to January 7, and the total rainfall time was <1 mm with a range of.9 8.9 mm for each rainfall event. The rainwater runoff from the concrete rooftop occurred after an elapsed time of 1 min during the rainfall in Events 1 and, which showed a relatively high rainfall intensity in the initial period. In Events and 5, which showed an even rainfall distribution, the rainwater runoff began at an elapsed time of minutes during the rainfall event. It is expected that runoffs with higher rainfall, >1 mm, would occur earlier than the results observed in this study. The pattern of accumulated MR values with time was very similar to that of accumulated CR values at regular intervals. The runoff at 541
each rainfall event ceased after a period of 1 minutes from the end of the rainfall event. The ratio of monitored runoff to calculated runoff (MR/CR ratio) was MR/CR =.87.9. 1 8 4 1 4 5 5 4 1 1 4 5 (a) Event 1 (9//) (b) Event 5 (/1/7). Note: CR = calculated runoff. MR = real, monitored runoff. Figure. Accumulated runoff from the rooftop during rainfall. An MR/CR ratio of around.9 is concurrent with the values from a range of rooftop runoff coefficients of.85 to.95 suggested by KWWA (5). However, it is higher than results of Lee et al. (). The reason for this gap is the difference in the number of preceding dry days, total rainfall, and scale of the catchment. In particular, because the experiments of Lee et al. were on a small catchment area of 5 5 mm, the ratio is expected to be sensitive to the catchment saturation during rainfall.. Rainwater runoff from a green roof The accumulated CR and MR values from a mixed catchment including a green roof are shown in Figure 4. Three rainfall events were used from July to September, and the total rainfall was < mm, with a range of. 5. mm for each rainfall event. The rainwater runoff from the mixed catchment had an elapsed time of 1 minutes during the rainfall events. The pattern of the accumulated MR values increased regularly with time, and had a little relationship with the accumulated CR values. The runoff from the mixed catchment began after a period of 1 min due to the runoff being from the terrace and a concrete roof. In the case of the concrete rooftop, the MR value was sensitive to changes in the CR value. Considering that the mixed catchment including the impermeable areas of the concrete rooftop and the terrace, the reason that the MR value was not sensitive to the CR value must be due to the green roof in the mixed catchment. 1 9 1 4 5 7 1 9 1 4 5 7 (a) Event (1/7/). (b) Event 4 (9/9/). Key: CR = calculated runoff. MR = real, monitored runoff. Figure 4. Accumulated runoff from the mixed catchment during rainfall. 54
In the mixed catchment including a green roof, the runoff finished after a period of 4 h from the end of rainfall. The MR/CR ratio was.9.74, and the rainwater runoff in the mixed catchment was reduced by about %. Using the measured MR/CR ratio from the mixed catchment and the concrete rooftop of MR/CR =.9, the MR/CR ratio from the green roof was calculated using Equation (1). The MR/CR ratio from the terrace was assumed to be the same as the MR/CR ratio from the concrete rooftop. From this, an MR/CR ratio from green roof was calculated to be.5.44. Therefore, the green roof reduced the rainwater runoff by around %. (CR A + T A + GR A ) R MR/CR M (CR A + T A ) R.9 MR/CR GR = (1) GR A R where MR/CR M is the ratio of the measured runoff to the calculated runoff from the mixed catchment, MR/CR GR is the ratio of the measured runoff to the calculated runoff from the green roof, CR A is the area of the concrete rooftop, T A is the area of the terrace, GR A is the area of the green roof, and R is the rainfall intensity. The total rainfall, duration time, preceding number of dry days, and the MR/CR ratio in the experiment of green roof are shown in Table for each rainfall condition. Table. Ratio of monitored runoff to calculated runoff of the runoff from the green roof. Total rainfall (mm) Duration (h) Preceding dry time Ratio of monitored runoff to calculated runoff Event 5. 5 h.5 Event.5 d.44 Event 4.81 d h.4 Range.8 5. 5 h d.5.44. Runoff flow peak on the green roof For Events and during July, the ratio of MR to CR with time from the concrete rooftop and the mixed catchment area including the green roof is shown in Figure 5. The MR/CR ratio of the concrete rooftop and the mixed catchment area for Event were.87 and.5, respectively, and Event were 1.5 and.49, respectively. This means that the reduction in the runoff flow peak in the mixed area was due to the green roof. Therefore we can conclude that a green roof can reduce the runoff peak flow. Runoff ( m ) 8 4 Runoff ( m ) 1 5 1 15 5 5 1 15 5 (a) Event (1/7/). (b) Event (4/7/). Figure 5. Runoff from the concrete rooftop and the mixed catchment area. 54
4. CONCLUSIONS Green roofs are suggested as suitable rainwater management methods to overcome spatial and environmental limitations. Green roofs provide green places, mitigate the heat island effect, and allow for the recovery of sound water circulation. In this work, we measured the reduction in rainwater runoff from an extensive green roof. The ratio of monitored runoff to calculated runoff (MR/CR) was.9 from a concrete rooftop and the ratio of MR/CR was.5.45 from a green roof. Therefore, green roofs reduce rainwater runoff by about %. Moreover, a retardation in the runoff peak flow was observed using a green roof. This effect was also observed using rainwater storage tanks to reduce rainwater runoff. Therefore, green roofs are considered to be an alternative to rainwater storage tanks in places that have spatial limitations, such as in a city. This study considered a large-scale green roof system, but the reduction in rainwater runoff was only studied in events with <1 mm of rainfall, and an analysis of the reduction in rainwater runoff under diverse rainfall conditions is needed. In addition, an analysis of the reduction in rainwater runoff from intensive green roofs is needed for large-scale projects. Future research is needed into structures for maximizing the reduction in runoff water and into the quality of runoff water from green roofs. REFERENCES Bass, B., Stull, P., Krayenjoff, S., and Martilli, A.,. Modelling the impact of green roof infrastructure on the urban heat island in Toronto Green Roofs Infrastructure, 4(1), pp.. Dunnett, N., and Kingsbury, N., 4. Planting Green Roofs and Living Walls. Timber Press, Portland. Han, M. Y., Kim, J. K., and Park, S. C., a. The characteristic of the quantity and quality of rainfall through greening roof during rainy season Proceedings of the Autumn Joint Conference & Forum of KSWW and KSWQ, pp. 5 5. Han, M. Y., Park, S. C., and Kim, J. K., b. The modelling of effect of flooding reduction on the greening roof through calibrated parameter of SWMM Proceedings of the Autumn Joint Conference and Forum of KSWW and KSWQ, pp. 57. Han, M. Y.,. Proactive multipurpose rainwater management in Korea Proceeding of RWHM Workshop, IWA Fifth World Water Congress and Exhibition, pp. 55. Kim, G. K., and Cho, D. G., 4. A study on a rooftop biotope creation technique reflecting the UNESCO biosphere reserve concept focusing on the UNESCO building rooftop J. Korean Env. Res. and Reveg. Tech., 7(4), pp. 4. Kolb, W., and Schwarz, T., 1999. Dachbegrunung, intensiv und extensiv Ulmer, Stuttgart. Krupka, B., 199. Dachbegrunung. Pflanzen-und Vegetationsanwendung an Bauwerken Ulmer, Stuttgart. Lee, D. K., Oh, S. H., Yoon, S. W., and Jang, S. W.,. A field study to evaluate green roof runoff reduction and delay J. Korean Env. Res. and Reveg. Tech., 9() pp. 117 1. Mentens, J., Raes, D., and Hermy, M.,. Green roofs as a part of urban water management In: Brebbia, C.A. (Ed) Water Resources Management II. WIT Press, Southampton, UK, pp. 5 44. Mentens, J., Raes, D., and Hermy, M.,. Green roofs as a tool for solving the rainwater run-off problem in the urbanized 1 st century? Landscape and Urban Planning, 77, pp. 17. Moran, A., 4. North Carolina State University, Department of Biological and Agricultural Engineering, MSc thesis. Na, H. Y., and Byun, B. S.,. Research trend of rooftop afforestation The Korean Association of Professional Geographers, 4(1), pp. 95 1 Wong, N. H., Cheong, D. K. W., Yan, H., Soh, J., Ong, C. L., and Sia, A.,. The effects of rooftop garden on energy consumption of a commercial building in Singapore Energy and Building, 5, pp. 5 4. Korea Water and Wastewater Works Association (KWWA), 5. The standard of a sewer system. 544