Seasonal operation of dual-mode biofilters: The influence of plant species on stormwater and greywater treatment

Highlights

• Biofilter inflows can be switched from stormwater to greywater in dry months.

• Effective plant species maintained low outflow concentrations across both sources.

• Removal of filterable reactive phosphorus linked to larger phosphorus reductions.

• Under stormwater inflows, less total nitrogen outflow linked to oxidised nitrogen.

• Under greywater inflows, reduction in ammonia linked to smaller nitrogen outflows.

Abstract

The use of stormwater biofilters (also known as bioretention systems and raingardens), in tropical and semi-arid areas is hindered by seasonal rainfall patterns which cause extended dry periods. These periods can result in plant die-off, long-term damage to system health and leaching of pollutants when stormwater inflows resume. Using an additional polluted water source during dry periods could minimise system stress and eliminate the need to irrigate biofilters with potable water during dry spells. As such, the presented laboratory study tested the seasonal operation of biofilters, by switching from stormwater treatment in wet months to greywater treatment in dry months. Forty-five single planted biofilter columns, incorporating sedges, grasses, understory ornamentals and vines, were subjected to four months of stormwater inflows, followed by three months of greywater inflows. We also investigated the impact of including a carbon source in the saturated zone on treatment performance. The results showed plant species selection to be critical for nitrogen and phosphorus removal, with consistently effective species such as Carex appressa and Canna x generalis able to maintain low outflow concentrations (e.g. total nitrogen of 0.2–0.3 mg/L and 0.3–0.6 mg/L, respectively) across both water sources. Low outflow phosphorus concentrations were associated with plant species that had high filterable reactive phosphorus removal across both water sources. Similarly, higher removal of ammonia and oxidised nitrogen was associated with lower overall nitrogen concentrations. In contrast, high removal of total suspended sediment (>94%), biochemical oxygen demand (>98%) and some heavy metals (e.g. lead >98% and copper >93%) was reported across all designs. The results suggest that with the careful selection of plant species, biofilters can be operated seasonally as a feasible and practical solution to maintaining system health during extended dry periods.

Introduction

With effective design and careful construction, stormwater biofilters (also known as bioretention systems or raingardens) are a desirable infrastructure for urban areas. While their primary purpose is to treat stormwater, they can also attenuate floods (Batalini de Macedo et al., 2019; Mei et al., 2018), are scalable, passive, low maintenance, provide greening of urban areas, increase biodiversity (Kazemi et al., 2009) and decrease the urban heat island effect (Coutts et al., 2013). However, extended dry periods caused by the irregular characteristic of stormwater diminish the treatment reliability of biofilters, with vegetation and microbe die-off (McDowell et al., 2008) and contaminants released from the systems when stormwater inflows resume (Baldwin and Mitchell, 2000; Barron et al., 2019; Hatt et al., 2007). Biofilters can also be perceived as ugly, due in part to a lack of diversity in vegetation, which can reduce community acceptance of them in the urban landscape (Dobbie, 2016; Funai and Kupec, 2017; Suppakittpaisarn et al., 2019). Unfortunately, these negatives seem to be hampering the adoption of biofilters as legitimate urban water infrastructure.

To maintain system health and treatment performance, alternative water sources to stormwater are required during dry conditions. This is particularly important in tropical and semi-arid regions that have seasonal rainfall patterns, such as the Mediterranean coast (e.g. Spain, Israel), parts of California, Western Australia and the Arabian Peninsula. The successful use of groundwater for this purpose has previously been reported in Israel (Aloni and Brenner, 2017). However, a stormwater-groundwater biofilter is not possible in all situations (e.g. deep aquifers, steep topography). In this regard, light greywater (greywater), which includes wastewater from showers, baths and bathroom sinks (Friedler and Hadari, 2006), could be exploited as it is relatively clean compared to other wastewaters (Eriksson et al., 2002) and is an underutilised and abundant water source in urban areas (Grant et al., 2012). Recent work has reported the successful treatment of greywater using biofiltration technology (Chowdhury, 2015; Fowdar et al., 2017) in a single-mode operation (i.e. using only greywater inflows), highlighting the potential to direct greywater to biofiltration systems when it is not raining. Not only would this sustain biofilter health, it would also maximise the treatment capacity of the system, with greater volumes of water being treated. In turn this would reduce the volume of polluted water requiring conveyance to water treatment plants and provide a fit-for-purpose water source.

While dual-mode systems for combining rainfall/stormwater and greywater are available, they have predominantly used different technology to biofiltration (Leong et al., 2018; Rozos et al., 2013; Sapkota et al., 2015; Stratigea and Makropoulos, 2015). Stormwater and greywater differ significantly in terms of concentrations, volume and frequency of generation (Duncan, 1999; Eriksson et al., 2002) and it is has been hypothesised that dual-mode biofilters may not be able to provide effective treatment due to these stark differences (Barron et al., 2019). In their initial work on dual-mode biofilters, Barron et al. (2019) utilised a mixed-mode operation, where stormwater and greywater inflows were alternated almost daily, and noted significant fluctuations in outflow concentrations across sampling events. The regular switching between water sources seemed to impact the system's ability to provide treatment to a consistent level. Treatment performance could potentially be improved utilising a different operational mode. Perhaps water sources could be swapped less regularly, following seasonal rainfall patterns that are very common (and predictable) in semi-arid regions (e.g. we can expect no rainfall for months in Israel, Southern Spain, and Western Australia). Hence, we propose a seasonal operation of dual-mode biofilters, where stormwater would be treated in wet periods, and greywater in dry periods.

While seasonal operation of dual-mode biofilters may provide more consistent outflow concentrations, the physical design of the system may impact its treatment performance. The effectiveness of biofilters centres on the interactions between surface area, filter media, vegetation and micro-organisms (Payne et al., 2015), and have been optimised for single water source inflows. Successful design of these elements allows biofilters to meet their treatment objectives, however, these elements also present a number of challenges. For instance, while the use of low nutrient media, usually with high percentages of sands, aids both hydraulic and treatment performance, it also limits the diversity of vegetation that can survive in raingardens (Funai and Kupec, 2017). This combined with the intermittent nature of stormwater inflows, sometimes with extended dry periods, has seen the widespread use of a limited number of effective plant species. While the inclusion of a saturated zone (SZ) in more recent designs has minimised the impacts of extended dry periods on plant survival and treatment performance, the dearth in variety of plant species has remained. Therefore, as part of this study we investigated the ability of a range of ornamental plants and vines to treat stormwater and greywater. The optimal characteristics for raingarden vegetation are well understood; high growth rates and dense root systems (Payne et al., 2018; Read et al., 2010). However, the ability of plants to thrive in a dual-mode stormwater-greywater biofilter under seasonal operation and provide effective treatment of both water sources is unknown. We hypothesise that plant species selection may become even more vital under seasonal operation, with some plants effective at treating both water sources, some preferencing a particular water source and others providing no advantage over barren filter media.

The importance of including a carbon source in the SZ of dual-mode stormwater-greywater biofilters also requires investigation. Kim et al. (2003) suggested that a carbon source promotes biological denitrification of stormwater while Fowdar et al. (2015) hypothesised that greywater has sufficient biodegradable organics that a carbon source may not be required for greywater biofilters. Barron et al. (2019) reported minimal practical differences in treatment performance between biofilters with and without a carbon source when employing a mixed operational mode. In contrast, a carbon source may be necessary for dual-mode biofilters under seasonal operation to ensure adequate treatment of stormwater inflows by promoting denitrification.

To optimise biofiltration systems for regions that have distinct wet and dry periods, we investigated how plant species selection influenced treatment efficiency under seasonal operation, mimicking a real-life situation where stormwater would be treated in wet months and greywater in dry months. Research questions for this project were:

• What plant species provide effective removal of key contaminants for biofilters under seasonal operation?

• How does switching between stormwater and greywater impact treatment performance?

• How does the inclusion of a carbon source in the SZ influence treatment performance?

This experiment forms part of a larger study on dual-mode stormwater-greywater biofilters, which included testing of mixed-mode operation (see Barron et al., 2019). To the best of the authors' knowledge, this is the first biofiltration study to test the same plant species under separate stormwater and greywater inflows.