Can raingardens produce food and retain stormwater? Effects of substrates and stormwater application method on plant water use, stormwater retention and yield
Introduction
In cities, stormwater runs off impervious surfaces at artificially high rates, threatening the health of local waterways through channel erosion, loss of habitat and pollution (Walsh et al., 2005). Approaches to urban stormwater management include raingardens (also termed biofiltration or bioretention systems) which protect waterways by retaining, treating and using stormwater at its source (Lloyd et al., 2002), restoring flow regimes closer to pre-urbanised levels (Bratieres et al., 2008). Raingardens are engineered garden beds with sandy soils and hardy perennial plants which capture and treat stormwater that runs off roofs and impermeable surfaces such as roads (Davis et al., 2009). As up to 50% of runoff from urban surfaces comes from private properties (Brown et al., 2014), where space for on-site stormwater management is limited, there is a need to ensure that raingardens maximize benefits to both humans and waterways (Burns et al., 2014, Jennings et al., 2012). While many homeowners are aware of the environmental benefits of raingardens, there are negative perceptions associated with ‘loss of garden space’ and ‘lack of tangible benefit to the household’ (Brown et al., 2014).
Expanding the functionality of raingardens to include vegetable production might increase their popularity, leading to wider implementation. Green roofs have been shown to produce vegetables without affecting their ability to retain stormwater runoff (Whittinghill et al., 2014) and raingardens may have similar potential. Using stormwater also increases overall security of water supply during drought, when restrictions on the uses of potable water may be imposed. Such restrictions were in place in Melbourne, Australia for ten years between 2002 and 2012, and included a ban on all outside watering, significantly constraining urban food production (Edwards 2011). Raingardens which utilize stormwater for food production may therefore be a win-win technology, protecting waterways from stormwater, while enabling food production that would otherwise be constrained. Vegetables produced without additional potable water irrigation could also provide a tangible return on money and time invested in constructing a raingarden.
However, vegetables differ greatly from the plant species conventionally used in raingardens (e.g. Carex spp.), which tend to be perennials selected for their capacity to remove pollutants, maintain infiltration through their root systems and survive large fluctuations in water availability (et al.,periods of drought between rainfall events, exacerbated by the typically free-draining substrate used in raingardens (Bratieres et al., 2010, Davis et al., 2009). To adapt raingardens for vegetable production, as vegetable raingardens, there are two major knowledge gaps and design issues that need to be resolved, particularly 1) the choice of substrate and 2) the method for delivering stormwater (harvested roof water).
Typically, loamy sand with low organic content (<5%) is used in raingardens, primarily to improve the quality of urban runoff (Bratieres et al., 2008, FAWB, 2009). Raingarden substrates need to be low in organic matter to reduce phosphate leaching into receiving waters (Bratieres et al., 2008). However, the low organic content means loamy sand has a relatively low water-holding (<20%) and nutrient capacity, which may not be suitable for vegetable production (Yu et al., 2013). Conversely, traditional containerized vegetable gardens use purpose-made substrates with high organic matter (>50%) to maximize water retention and vegetable yield (Hudson 1994). Therefore, in developing vegetable raingardens it is important to determine whether loamy sand raingarden substrates can satisfy the water and nutrient requirements of vegetable crops in addition to water quality outcomes.
Stormwater runoff usually enters a raingarden at the surface and then drains downward through the substrate, filtering pollutants (Davis et al., 2001, Hatt et al., 2007, Read et al., 2008). This might not be the optimal design for a vegetable raingarden, as watering needs to be managed efficiently to minimise plant water stress. In addition, the requirement to apply large quantities of supplementary irrigation water should be avoided so that the raingarden remains effective for stormwater runoff absorption. A potential solution is to invert the typical raingarden design so that, rather than runoff being applied to the surface, the raingarden is sub-irrigated with stormwater, delivering water below the root zone which then moves upwards through capillary rise. These garden beds are commonly known as wicking beds in Australia and as sub-irrigated planters in the USA, and have been shown to offer higher water use efficiency than surface irrigation (Ahmed et al., 2000, Incrocci et al., 2006, Santamaria et al., 2003). Sub-surface watering might also improve food safety in vegetable raingardens, as direct contact with runoff is minimized and pollutants can be filtered from the runoff water as it moves upwards through the substrate, before reaching plants (Tom et al., 2013).
The success of a vegetable raingarden will be determined by both its effectiveness at retaining stormwater and its ability to produce vegetables despite water limitations between rainfall events. To determine whether raingardens could be used to produce food and retain stormwater we undertook a greenhouse pot experiment using different substrates and methods of stormwater application. Two substrates were used; a loamy sand typically used in raingardens and a potting mix used in conventional containerised vegetable gardens but not normally in raingardens (FAWB 2009). There were two stormwater application treatments; ‘sub-surface’ and ‘surface’ application of water with water quantity based on average Melbourne summer rainfall. It was expected that vegetable growth (biomass, leaf area and root allocation) and yield would be greater in the potting mix than in the loamy sand due to greater water holding capacity, and that sub-surface stormwater application would maintain water availability between rainfall events and therefore have greater yield than surface watering due to reduced plant stress.
Section snippets
Materials and methods
The experiment was conducted in a non-temperature controlled greenhouse at the University of Melbourne’s Burnley campus (37°49′44.22″S, 145°1′13.40″E) from September 2012 to February 2013. Melbourne’s climate is characterised by cool, wet winters and dry, hot summers. Melbourne’s mean annual rainfall is 648.4 mm (1855–2015), with 153.9 mm falling in summer months (Bureau of Meteorology, 2015). This was a very warm spring-summer period; mean daily maximum temperature in summer is 25.9 °C (January),
Substrate properties and effects on biomass, leaf area and root allocation
Substrate properties differed dramatically between the two substrates. Water holding capacity (WHC) of the potting mix (54% ± 0.5) was more than double that of the loamy sand (26.3% ± 0.1). Plant available water was 102% greater in potting mix than loamy sand, even though permanent wilting point (soil matric suction of −1.5 MPa) occurred at a higher soil water content in the potting mix (7.9% of the total water holding capacity) than in the loamy sand (3.5%). Loamy sand had almost 3.5 fold greater
Discussion
By comparing the effects of different substrates and methods of stormwater application (watering treatments) we were able to show that vegetable raingardens could be used to successfully produce vegetables and retain stormwater. However, the choice of substrate and watering treatment will require trade-offs among yield, stormwater retention and potential implications for water quality and long-term stability of hydraulic properties.
Conclusions
Our results showed that a vegetable raingarden could retain stormwater and produce vegetables despite potential water limitations between rainfall events. Stormwater retention was similar to values reported for conventional raingardens and although the four vegetable species varied in their response to the substrate and watering treatments, they were able to tolerate the frequency and volume of average summer rainfall in Melbourne. Overall, yield and biomass differences were determined by
Acknowledgments
This research was funded by Melbourne Water, under the direction of Keysha Milenkovic. Technical support in the Burnley nursery was provided by Sascha Marianna Andrusiak and Nicholas Osborne. Fletcher was supported by an ARC Future Fellowship (FT100100144).
References (45)
- et al.
Nutrient and sediment removal by stormwater biofilters: a large-scale design optimisation study
Water Res.
(2008) - et al.
Common bean (Phaseolus vulgaris L.)
Field Crops Res.
(1997) - et al.
Water requirement of drip irrigated tomatoes grown in greenhouse in tropical environment
Agric. Water Manage.
(2005) - et al.
Evaluation of the influence of irrigation methods and water quality on sugar beet yield and water use efficiency
Agric. Water Manage.
(2010) - et al.
Hydrologic and pollutant removal performance of stormwater biofiltration systems at the field scale
J. Hydrol.
(2009) - et al.
The influence of drip irrigation or subirrigation on tomato grown in closed-loop substrate culture with saline water
Sci. Hortic.
(2006) - et al.
Moderate drought and signal transducer induced biosynthesis of relevant secondary metabolites in thyme (Thymus vulgaris), greater celandine (Chelidonium majus) and parsley (Petroselinum crispum)
Ind. Crops Prod.
(2015) - et al.
Improvement of soil properties using compost for optimum parsley production in sandy soils
Sci. Hort.
(2009) - et al.
The effect of water deficit stress on the growth: yield and composition of essential oils of parsley
Sci. Hort.
(2008) - et al.
Variation among plant species in pollutant removal from stormwater in biofiltration systems
Water Res.
(2008)
Vegetable raingardens can produce food and reduce stormwater runoff
Urban For. Urban Greening.
Comparison of sub-irrigation and overhead irrigation of tomato and lettuce seedlings
J. Hortic. Sci. Biotechnol.
Storm-water infiltration and focused recharge modeling with finite-volume two-dimensional Richards equation: application to an experimental rain garden
J. Hydraul. Eng.
Pervious and impervious runoff in urban catchments
Hydrol. Sci. J.
Hydraulic and pollutant treatment performance of sand based biofilters
More than money: how multiple factors influence householder participation in at-source stormwater management
J. Environ. Plann. Manage.
Climate Statistics for Australian Locations (Melbourne Regional Office)
Flow-regime management at the urban land-parcel scale: a test of feasibility
J. Hydrol. Eng.
Climate Change in Australia – Technical Report
Fruit production and quality of tomato plants (Solanum lycopersicum L.) are affected by green compost and arbuscular mycorrhizal fungi
Plant Biosyst.
Laboratory study of biological retention for urban storm water management
Water Environ. Res.
Cited by (23)
Urban surface uses for climate resilient and sustainable cities: A catalogue of solutions
2021, Sustainable Cities and SocietyUrban Green-space Water-consumption characteristics and its driving factors in China
2021, Ecological IndicatorsCitation Excerpt :We investigated 286 cities in China from 2004 to 2017. Candidate driving factors are divided into five categories of social-technology (ST), water-supply (WS), city-development (CD), economic (ECO), and climate (CLI) factors (Richards et al., 2017; Moon et al., 2019). The urban green-space water consumption (m3/ha) (WPA) of 286 cities in China was analyzed in this study to reveal important driving factors among 16 candidates, namely, six ST, three WS, four CD, one ECO, and two CLI variables.
Experimental investigation for impacts of rain storms and terrain slopes on low impact development effect in an idealized urban catchment
2019, Journal of HydrologyCitation Excerpt :Therefore, the rainfall intensity and terrain slope are considered to be two of the most concerned impacting factors for LID effects (Guan et al., 2016; Goh et al., 2017; Sedyowati and Susanti, 2017), and in turn systematically investigation is required. Since the long term data for the effects of the realistic LID measures are normally rare, experimentally evaluating the LID effects are becoming more and more popular (Ahiablame et al., 2012; Andersen et al., 1999; Richards et al., 2017). Many studies on evaluating the effect of LID on surface runoff have engaged in this field (Collins et al., 2007; Aravena and Dussaillant, 2009; Lefevre et al., 2012; Yang et al., 2013; Valinski & Chandler, 2015; Palla et al., 2015).
Influences of stormwater concentration infiltration on soil nitrogen, phosphorus, TOC and their relations with enzyme activity in rain garden
2019, ChemosphereCitation Excerpt :They are scientific issues of great concern to all researchers (Guo et al., 2018). At present, many researchers focus on the structure, running effect, influence mechanism and model simulation of low impact development (LID) facilities themselves (Richards et al., 2017; Trowsdale and Simcock, 2011; Aryal et al., 2016). Meanwhile, some researchers have carried out the influence of stormwater concentration infiltration of rainfall runoff on soil in LID mode.
Reprint of : Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions
2019, Urban Forestry and Urban GreeningCitation Excerpt :Jia et al. (2016) have proposed a tool for RG sizing and hydrological evaluation. To increase the adoption by householders an expansion of the functionality to vegetable production was also suggested (Richards et al., 2015, 2017). A survey conducted of households in Maryland (USA) pointed out that incentive programmes have the potential to increase voluntary adoption of RGs on existing residential properties (Newburn and Alberini, 2016).
Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions
2018, Urban Forestry and Urban GreeningCitation Excerpt :Jia et al. (2016) have proposed a tool for RG sizing and hydrological evaluation. To increase the adoption by householders an expansion of the functionality to vegetable production was also suggested (Richards et al., 2015, 2017). A survey conducted of households in Maryland (USA) pointed out that incentive programmes have the potential to increase voluntary adoption of RGs on existing residential properties (Newburn and Alberini, 2016).