Quantifying the effects of green roofs on stormwater reduction: A review
Urban areas are becoming increasingly expansive and dense as the global population becomes more urban, in accordance we see a decrease in the amount of pervious surfaces in and around urban areas (Berndtsson, 2009). One of the more critical issues concerning urbanization is the effect it has on the watersheds hydrological system. Conventional stormwater infrastructure, designed to convey stormwater from the city to receiving bodies of water, in combination with high amounts of impervious surfaces result in highly fluctuating volumes of surface water in our urban waterways. Further exacerbating these fluctuations are the effects climate change (Mentens et al., 2005).
The increase of impervious land cover associated with urban development has consistently been shown to degrade aquatic ecosystems (Miltner et al.,2004; Wang et al., 2001). The USEPA (2003) reports that runoff from urbanized areas is the leading source of water quality impairments to surveyed estuaries, and the third largest source of impairments to surveyed lakes. Amplified peak flows impact stream ecologies through increase bed and bank erosion (Wolman, 1967), increased frequency of disturbance (Booth and Jackson, 1997) and rapid and efficient pollutant transport (Olguin et al., 2000). Schueler (1994) developed the impervious cover model (ICM) which identifies a threshold of 10% impervious cover, at which point streams become impacted and a threshold above 25% impervious cover where stream quality is considered non-supporting for aquatic life.
Existing land values in urban areas often exceed 25%. Stormwater Best Management Practices are primarily designed to lessen the impact of urban development and drainage by increasing storage areas across a watershed therefore slowing the flow of water into a receiving water body and/or replacing impervious surfaces with permeable ones that allow stormwater infiltration, (Berndtsson, 2009).
The management of stormwater has for many years now been a very important and highly active area of green roof research. In comparison to conventional roofs, many studies have demonstrated how green roofs can significantly reduce runoff volumes. Roofs can account for up to 30% of urban impervious surfaces thus substantially contribute to stormwater management problems (Carter & Jackson, 2007). High land prices make the creation of green spaces in urban areas very expensive, given the huge amount of unused roof area available green roofs are an appealing component of stormwater management. Green roofs (a.) delay the initial time of runoff due to absorption of water in green roof system; (b.) reduce the total volume of runoff by retaining part of the rainfall; (c.) distribute the runoff over a longer time period through a relative slow release of the excess water that is temporary stored in pores of the substrate (Mentnes et al. 2005).
Water retention capabilities of growing media for green roofs
German FLL guidelines indicate that the main source of variation in water retention capabilities comes from the depth of the growing media, (FLL, 2008). Actual values for water retention capabilities of green roofs in published studies varies widely (Graceson et al., 2013). Metens et al., (2006) in a meta-analysis of 18 German studies from 1987-2003 reported that estimated rainfall retention capability ranges from 65-85% for intensive to 21-87% for extensive green roofs. FLL Guidelines state that the maximum annual water retention for extensive and semi-intensive green roofs is about 60%, as growing media should be no deeper than 200mm. Graceson et al., (2013) found that growing media depth did influence water retention on extensive roofs but the effect was relatively small, a 100% increase in depth brought just 20% increase in retention volume. VanWoert et al., (2005) also found the increase in water retention with increase soil media depth to be relatively small, only 3% from an increase of 25 mm to 45 mm depth. The assumption to be drawn from these findings is that increasing media depth from extensive to semi-intensive (20-200mm) does not always result in appreciable changes in water retention capabilities. Runoff reduction values as a % of total annual rainfall are also difficult to compare between different studies due to the different conditions in which these studies were performed and the different number of events which were included in each study. In Mentens study it was found that annual precipitation, type of roof, number of layers, & depth of substrate layers are significantly correlated with yearly runoff (p>0.05), Metens et al., (2006).
Not only is the growing media for intensive roofs deeper but it tends also to have more organic matter and a higher proportion of particles smaller than 1mm (FLL, 2008). Water retention increase with the proportion of fine particles because these particles fill in pore spaces between larger particles thus creating more water holding pore spaces. In macro-pores water cannot be held against the forces of gravity thus coarse soils are less effective at retaining water. Graceson et al., (2013) conducted a study to investigate the effect of growing media composition on water retention capabilities of green roofs using simulated decks composed of various inorganic substrates mixed with composted green waste. The study looked at two different roof types with characteristically different soil drainage requirements. Sedum roofs contained coarse crushed brick, coarse crushed tile, or Lytag (palletized power station fly ash) mixed with 10% or 20% coarse composted green waste. Meadow roofs contained fine crushed brick, fine crushed tile or Lytag with 20% or 30% fine composted green waste for meadow roofs. It was found that growing media treatment significantly affected water retention (sedum decks retained 40%, compared to 48% for meadow decks) the majority of the effect was attributed to the low water retention of the growing media for sedum green roofs. These results indicate that differences in particle size proportion and distribution largely influence water holding capacity.
In another study VanWoert et al., (2005) looked at hydrographs from conventional roofs compared to vegetated roofs with 2.5 cm of growing medium and non-vegetated roofs with equal depth and found that vegetated roofs contained an average of 48% more than conventional roofs but not significantly more than media only roofs. These findings suggest the main factor in water retention relates more to the properties of the media than the presence of vegetation, although vegetation does play a role.
Antecedent Soil Moisture
The performance of green roofs in stormwater management can be linked to both soil properties such as precursory soil saturation levels, water holding capacity of the soil, and age characteristics but also climatic factors such as rainfall event size, frequency, duration and intensity and seasonal fluctuations in solar radiation. The water retaining capacity of green roofs regenerate more quickly in the summer months as lower rainfall depth and higher evapotranspiration losses occur during this time, (Mentens et al., 2005). Schroll et al., (2011) found that 65% of rain falling onto roofs during the dry season was retained compared to only 26% during the wet season. Graceson et al. (2013) findings show runoff was equal to or exceeded rainfall during the winter months, average water retention between Dec. and Feb. being just 6% compared to 76% Mar. to May. Mentens et al. (2005) also concluded that retention of rainwater is lower in winter than in summer. In combining all studies runoff was 30% for the warm season, 51% for the cool season and 67% for the cold season.
Rainfall event size and intensity have also been shown to be highly correlated with retention. Carter and Rasmussen (2006) found an inverse relationship between the depth of rainfall and the percentage of rain that was retained. Small storms retained 88%; medium, 54%; and large storms, only 48%. Villarreal and Bengtsson (2005) found that lower intensity rainstorms lead to higher retention values; for rain intensity .4mm/min and slopes of 2 and 14% the retention was 62% and 39% respectively; for 1.3mm/mm retention was 21% and 10% respectively for those same slopes. Mentens et al., (2006) found the rainfall-runoff relationship is linear for hard roofs but contains a quadratic factor for green roofs due to the fact that higher annual precipitations mean a higher amount of extreme annual events or which retention is lower.
Roof age may also play a role. Getter et al., (2007) suggests roof maturity to be an influencing factor in the results of their study which observed minimal runoff delay in comparison to other studies (VanWoert et al., 2005; Simmons, 2008) which found an average of 10 min. delays for runoff. In a separate study comparing organic matter content and physical properties of soil after 5 years on a green roof Getter et al., (2007) found that organic matter content and pore space doubled in that time from 41-82% and the water holding capacity also increases during that time from 17%-67%.
Water retention capabilities for different vegetation types
Green roof vegetation can affect the amount of water runoff depending on each plant’s capacity for water interception, water retention, and transpiration. Nagase & Dunnett (2011) found there to be significant differences in the amount of water runoff between vegetation types. The above ground plant structure and surface water storage capacity can impact interception which then influences water runoff. Clark (1937) found low growing mat-forming plants don’t intercept as much rain as taller plants do. Nagase & Dunnett (2011) showed there to be a significant negative relationship between plant height and amount of water runoff, indicating that modules of taller species retain more water than shorter species with smaller diameters. In the same study they found that water droplets more effectively adhered to species with fine structures on the leaf surface such as fine hairs or impermeable wax than those that did not contain such features.
Diversity in plant structure may also affect hydrological performance on green roofs. The reduced runoff associated with vegetation composed of many structural types could be ascribed to the sampling effect which explains that a greater inclusion of species increases the probability of selecting a species of high absorption capacity (Rixon, Mudler, 2005). Different architectures also perform differently at different tasks and may be complementary to one another. Lundholm et al., (2010) in a study looking at functional group combinations, found an indication of low diversity canopies preventing evaporation from the soil surface, reducing the amount of water that can be captured in subsequent rain events.
Below ground plant structure was also found to play a role in water retention. Plant species that form extremely dense fibrous roots reduce the porosity in the growing medium and thus the volume of space in which water can be retained (MacIvor and Lundholm, 2011). Nagase & Dunnett (2011) found there to be a significant negative relationship between mean dry weight of roots and mean water runoff indicating that species with more root growth have greater water capture. These results are inconsistent with the above study and the authors offer that these contradictions may be due to differences in physical properties of the growing mediums used between each study. Lamont and Bergl (1991) point out that many species of contrasting growth form behave differently in their use of soil water. Different rooting patterns create below ground niche separation facilitating coexistence by allowing differential exploitation of the soil profile.
Difference in transpiration rates can affect reduction of runoff. Species that transpire at a faster rate will create more space sooner within the growing medium allowing for additional water capture in subsequent rain events, (Lundholm et al., 2010). Sedum and succulent species use water more efficiently than herbaceous species and therefore diminish the ability of a green roof to mitigate stormwater retention however they also exhibit greater survivability during times of water stress (Nagase & Dunnett (2011). the photosynthetic mechanism used by certain species can also affect transpirational water loss and thus affect runoff rates. C3 and C4 plants are more effective at reducing water runoff compared to CAM plants because CAM plants have higher water use efficiency by keeping their stomatas closed during the day (Gravatt & Martin, 1992).
Multi-functionality turns out to be a very important component in the way plants capture and transpire rainfall. Plant life forms such as grasses, shrubs, and forbs represent different life history strategies, resource use patterns and adaptations to the external environment and may act as a coarse surrogate to functional diversity. In a study Lundholm et al. measured how green roof services, such as water capture and roof temperatures were affected by plant type and plant diversity and found that mixtures containing 3 & 5 life-form groups optimized several green roof ecosystem functions and outperformed monocultures and single life-form groups. 3 & 5 life-form groups captured more water, and transpired more water than monoculture treatments on average. No single life-form optimized all functions equally, some mixtures performed better at water capture and evapotranspiration while others performed better at temperature reductions and seasonal stability. The main benefit of including all three groups was not to maximize any single process but to perform a variety of functions well. Nagase & Dunnett (2011) looked at how speices richness influenced retention values and found no consistent relationship between species richness and water runoff. Water runoff from diverse combinations of 12 species was higher than that of the best performing monocultures. This illustrates the point that species richness alone is not enough to improve water retention values and that functional diversity is the actual underlying cause.
Effect of slope on water retention
Water Retention and Slope Angle
Somewhat obviously, Getter et al., (2007) found that retention values decrease with increasing slope angle and that mean retention was least in slopes of 25% (75.3%), and greatest in slopes of 2% (85.2%). VanWoert et al., (2005) similarly found that slopes of a 2% slope retained significantly more rainfall than those of a 6.5% slope. In combining the effect of media depth with slope increase Getter et al., (2005) found that increasing media depth was not enough alone to compensate for increasing slope, no difference occurred in retention between 4cm and 6cm of media on a 6.5% slope.
Widespread green roof application
Green Roofs are a valuable urban stormwater BMP and while many studies have examined the effectiveness in rainfall retention brought at the roof scale an important aspect in determining their viability is an understanding of how they perform at a watershed scale. Widespread green roof implementation can significantly reduce the total impervious area of a watershed and through the provision of additional storage significantly reduces the peak runoff rates. Changes in hydrology due to roof greening however, are very dependent upon the size of the storm event. Even with widespread green roof installation, change in hydrology across the watershed will be minimal for large storm events. Carter and Jackson, (2006) found that, for a small storm, if all roofs across a watershed were greened this could amount to a 36.9% reduction in runoff volume, and when only flat roofs are considered in this scenario a 18.9% reduction. For a large storm a 3.4% reduction in volume could be seen for all roofs and a 1.6% reduction for flat roofs.
Carter and Jackson also looked at how zoning categorization effected sub-watershed retention values and found the areas zoned commercial downtown retained the most. Such areas often contain the largest proportion of flat roofs and greening of all roofs in this area resulted in a 45% reduction in runoff volume and greening only of flat roofs resulted in a 40% reduction. Residential zoning blocks showed a high reduction when all roofs where considered (47%) but when only flat roofs where examined this number dropped to zero. Residential zoned areas have a high proportion of rooftop area to land area but single family residences typically have pitched roofs which have significantly reduced effectiveness in capturing stormwater and therefore are not as suitable to roof greening.
Green roofs make use of previously unused space within the urban environment and thus do not limit the demands of people for “open space” on the ground. They can play a significant role in stormwater management but the magnitude of their retention depends on the structure of the green roof and the amount and intensity of rain falling on them, and for this reason they should be used in conjunction with other BMP measures to provide a more complete watershed management plan. The reduction of runoff accomplished by green roofs is higher in the city center than in suburbanized areas because urban areas have a higher proportion of impervious surfaces to permeable land cover. There is little opportunity for practical green roof implementation in single family residential areas due to the high proportion of slopes roofs in these areas, thus flat green roofing is considered most feasible in areas zoned commercial, industrial or institutional, places known to contain large flat-roofed buildings. Roofs with deeper substrate further enhance the benefit of rainfall retention, however intensive roofs also require more maintenance and extra watering during the dry season. During normal periods deeper substrate layer can support vegetation with greater water requirements but this vegetation is less resilient to water shortages and furthermore the weight of the additional soil requires more structural support that can be costly. In terms of the benefits brought by vegetation, while still playing a smaller role than that of the actual growing media, functional plant diversity brings greater benefit in comparison to monocultures as it allow for the optimization of multiple ecosystem services to be met simultaneously.
Retention on any roof depends on rainfall distribution throughout the year, the intensity of each event, ambient air temperate, plant selection and the influence of local environmental conditions on evapotranspiration. Often the total amount of water and rain event duration are not the problem in stormwater management, it is the rate of that incoming water that needs to be treated. Findings then that support green roofs capacity to reduce peak flow and extend runoff over longer periods are important in regard to this matter. When applied on a wide scale the additional storage provided by green roofs can significantly reduce total peak flow volume in the watershed. This decline in peak flow volume reduces the frequency of the bank full flow events which may result in changes in bank full cross-section dimensions. Green roofs, when designed to maximize their water retention capabilities can serve as a valuable tool towards the goal of hydrologic remediation of watersheds and should be incorporated in larger watershed plans as a measure to mitigate the effects of urban density on loss of permeability.