Influence of plant composition and water use strategies on green roof stormwater retention
Graphical abstract
Introduction
Impervious surfaces in urban areas increase the frequency, volume and peak flow of stormwater runoff compared with natural areas which degrades urban aquatic ecosystems (Walsh et al., 2005). As roofs comprise a large proportion of impermeable surfaces, green roofs are being used as a tool to manage urban runoff (Li and Babcock, 2014, Elliott et al., 2016). Although they provide other ecosystem services, such as air purification, noise reduction and urban cooling, reducing stormwater runoff is often considered the major benefit provided by green roofs and is one of the main reasons they are legislated or incentivised in cities worldwide (Oberndorfer et al., 2007, Carter and Fowler, 2008).
Green roofs reduce stormwater runoff by storing water during rainfall events, delaying peak runoff (attenuation through the substrate) and removing water from the substrate to the atmosphere through evapotranspiration (ET) (Stovin et al., 2013). Following rainfall, a portion of the detained water drains from the roof and the amount of rainfall corresponding to substrate storage capacity is retained (Bengtsson et al., 2005). This retained water will then either evaporate or be transpired by plants which dries out the substrate and regenerates retention capacity before the next rainfall event (Berretta et al., 2014, Poë et al., 2015). The rate at which water stored in the substrate is used by plants is an important determinant of how a green roof responds to the next rainfall event (Villarreal and Bengtsson, 2005). The most commonly planted species on green roofs are succulents in the genus Sedum. These plants are ideal for survival in green roof systems, because their shallow root system and conservative water use strategy means that they can withstand frequent periods of drought on green roofs when the shallow substrates dry out (VanWoert et al., 2005, Getter and Rowe, 2009). However, these characteristics also mean that they may not sufficiently dry out substrates via transpiration between rain events to provide significant stormwater retention (Farrell et al., 2013). To optimise rainfall retention and plant survival on green roofs, plants need to have high transpiration when water is available after rainfall, but also be able to survive dry substrates between rainfall events by limiting their transpiration (Wolf and Lundholm, 2008, Farrell et al., 2013). These water use strategies have been found in rock outcrop plants evaluated for green roofs, but their potential for reducing stormwater runoff has not yet been evaluated (Farrell et al., 2013).
It has been suggested that a diverse mixture of plant functional groups will improve the function and resilience of green roofs (Lundholm et al., 2010). Studies in both natural and engineered ecosystems have found positive relationships between plant diversity and the provision of ecosystem services, including: maintaining hydrologic cycles, cleaning air and water and the storing/cycling of nutrients (Christensen et al., 1996, Quijas et al., 2010, Johnson et al., 2016). Two principal mechanisms are thought to drive the positive relationships between species diversity and ecosystem function: (1) transgressive overyielding, where mixtures of plant species produce more biomass than monocultures due to complementarity in resource utilization (Schmid et al., 2008, Prieto et al., 2015); and (2) selection effect, where there is a greater possibility of including the best-performing species in a mixture which enhances resource use efficiency (Chapin III et al., 1997). Species which have similar effects on ecosystem processes (water and nutrient dynamics, trophic interactions, or disturbance regime), but differ in response to environmental conditions, can also provide stability (resistance and resilience) under changing environmental conditions (Chapin III et al., 2000). Consequently, if one species is lost from a diverse ecosystem, there is the potential for other functionally similar species to compensate, making the ecosystem more resilient to environmental changes (Ehrlich and Ehrlich, 1981).
Applying biodiversity and ecosystem theory to improve the performance of green roofs has been gaining traction, with several studies evaluating how plant diversity affects rainfall retention (Dunnett et al., 2008, Lundholm et al., 2010). However, most of these studies have evaluated diversity in terms of mixtures of life-forms or species richness, rather than by mixing species with known functionality, such as the water use strategy. While these studies hypothesised that mixtures would out-perform monocultures in terms of retention, the results have not been straight forward. For example, Lundholm et al. (2010) reported that while mixtures of grasses and succulents had greater retention than monocultures of either life-form, their retention was less than bare substrate. The effects of life-form diversity are likely complicated by differences in biomass and rainfall interception. For example, Dunnett et al. (2008) found there was no significant improvement in retention with life-form diversity, but that the highest retention was from tall plants with high root biomass. This is consistent with Franzaring et al. (2016) which showed that, compared with monocultures of legumes, herbs, small shrubs and succulents, mixtures of these life-forms retained less water than monocultures of grasses. While species richness seems to be positively related to productivity, it may not relate to rainfall retention (Nagase and Dunnett, 2012, Johnson et al., 2016).
These results are consistent with ecological research that shows that not all metrics of biodiversity are related to improved nutrient and water uptake, biomass production and transpiration rate (Díaz and Cabido, 2001, Lavorel and Garnier, 2002). It has been suggested that functional diversity and functional trait diversity may be the most important drivers for specific ecosystem services (Petchey and Gaston, 2002, Cadotte et al., 2011). Consequently, green roof experiments assessing the role of diversity on the basis of life-forms or species richness may not have included the necessary functional or physiological traits to improve water retention. This leads us to consider whether biodiverse green roofs designed on the basis of functional diversity or functional trait diversity could achieve the trade-off between hydrological performance and plant survival. We therefore suggest that green roofs planted with mixtures of species with appropriate water use strategies will improve rainfall retention performance without compromising plant survival (Farrell et al., 2013).
In this study, we tested how mixtures and monocultures of species chosen on the basis of plant water use strategies affect rainfall retention on green roofs, relative to a commonly used succulent species. We evaluated three species (Lomandra longifolia, Dianella admixta and Stypandra glauca) known to have water use plasticity, i.e. they have high evapotranspiration when water is available and reduced water use under drought conditions to avoid drought stress (Farrell et al., 2013). We predicted that: (1) green roof modules planted with monocultures of species with water use plasticity would retain more water via greater evapotranspiration than modules planted with the succulent species with conservative water use and (2) mixtures would retain more rainfall due to complementarity in aboveground biomass and water use, resulting in greater evapotranspiration.
Section snippets
Experimental design
We used a randomized block design with six ‘vegetation’ treatments and three replicates: (1) bare substrate (as a control), (2) monoculture of Sedum pachyphyllum (as a reference for common green roof plantings), (3) monoculture of Lomandra longifolia, (4) monoculture of Dianella admixta, (5) monoculture of Stypandra glauca and (6) mixture of Lomandra longifolia, Dianella admixta and Stypandra glauca. These four species were chosen based on two types of water use strategies, (1) conservative
Total rainfall retention
Rainfall retention differed significantly among treatments (Fig. 2A). Modules planted with L. longifolia had the highest retention, with 16% greater retention than S. glauca and 10% greater retention than S. pachyphyllum. Interestingly, S. glauca, not S. pachyphyllum which has a conservative water use strategy, had the lowest retention. Modules planted with S. glauca and the mixture of species, also had lower rainfall retention than the bare treatment. Modules planted with mixtures retained
Discussion
We aimed to test how mixtures and monocultures of species chosen on the basis of plant water use strategies would affect rainfall retention on green roofs, relative to a commonly used succulent species. We predicted that (1) green roof modules planted with species with water use plasticity would retain more water via greater ET than those planted with the succulent S. pachyphyllum which has a conservative water use strategy and (2) mixtures would retain more rainfall due to complementarity in
Conclusions
Contrary to our hypotheses, green roof modules planted with species with water use plasticity did not always retain more water than the conservative water using succulent, S. pachyphyllum. Also, the mixtures of plant species did not retain more water than monocultures. Although the mixed species treatment had higher biomass and productivity, the performance of the mixture was adversely affected by the key species, S. glauca. The root systems of S. glauca likely created preferential flow
Acknowledgements
Thanks to Ariane Lenhard, Andrea Pianella, Pei-wen Chung and Jessica Kurylo for assisting with data collection. This research was funded by Australian Research Council Linkage grant LP130100731 supported by Melbourne Water and the Inner Melbourne Action Plan (IMAP). Zheng Zhang was supported by the Melbourne International Research Scholarship (MIRS) and Melbourne International Fee Remission Scholarship (MIFRS).
References (56)
- et al.
Moisture content behaviour in extensive green roofs during dry periods: the influence of vegetation and substrate characteristics
J. Hydrol.
(2014) - et al.
A long-term hydrological modelling of an extensive green roof by means of SWMM
Ecol. Eng.
(2016) - et al.
Vive la difference: plant functional diversity matters to ecosystem processes
Trends Ecol. Evol.
(2001) - et al.
Water retention, wash-out, substrate and surface temperatures of extensive green roof mesocosms—results from a two year study in SW-Germany
Ecol. Eng.
(2016) - et al.
Hydrologic and pollutant removal performance of stormwater biofiltration systems at the field scale
J. Hydrol.
(2009) - et al.
Using spatial interpolation to construct a comprehensive archive of Australian climate data
Environ. Model. Softw.
(2001) - et al.
The influence of design parameters on clogging of stormwater biofilters: a large-scale column study
Water Res.
(2012) - et al.
Modelling of green roof hydrological performance for urban drainage applications
J. Hydrol.
(2014) - et al.
Green roofs as a tool for solving the rainwater runoff problem in the urbanized 21st century?
Landsc. Urban Plan.
(2006) - et al.
Amount of water runoff from different vegetation types on extensive green roofs: effects of plant species, diversity and plant structure
Landsc. Urban Plan.
(2012)