Abstract
Urbanisation has delivered society many benefits but has also created problems. Low albedo and high thermal mass create urban heat islands. Altered hydrological cycles can lead to downstream flooding or desertification, pollution, and erosion. Loss of vegetation and the resulting nature deficit increases human physical and mental health disorders and diseases. With traditional engineering and urban design approaches problems have increased as city size has increased. More of the same can’t fix this. The solutions to these issues will be holistic, multidisciplinary, and focused on life. Traditional urban engineering is evolving to support urban green life that manages stormwater, moderates microclimates, remediates soils and sustains the micro and macro biodiversity we can’t live without. This paper reports research and case studies that show green engineering can support life and engineering outcomes simultaneously. Research and working demonstrations are detailed that show infiltration devices and porous surfaces can provide effective stormwater management solutions in roads, on verges, in car parks, on parks and reserves and in private gardens. Green engineering solutions support urban forests and the vegetation enhances the efficiency and effectiveness of the engineering, which combined help to restore nature’s circular systems in urban areas.
Introduction
Urbanisation’s increasing demands on infrastructure provision and impacts on environmental sustainability have been serious concerns for decades. Highly developed impervious urban catchments raise stormwater discharge volumes and reduce soil water recharge, create urban heat islands and decrease water availability to urban vegetation. Diminished creek and river base flows and poorer water quality in creeks and rivers lead to local extinctions of aquatic and terrestrial species, degraded marine environments, and reduced human health and wellbeing.
Green engineering approaches based on or incorporating water-sensitive urban design (WSUD) can help to address these problems. WSUD is a major component of green engineering, it’s aim is to detain and manage rain where it falls. WSUD can improve the efficiency of stormwater management by reducing costs associated with traditional pipe networks, instead using the water to deliver human and environmental benefits locally. Emerging from the last decade of the last millennium, the adoption of WSUD was helped by policy introduced nearly 20 years ago (National Water Commission, 2004) which set capacity-building targets toward creating water sensitive Australian cities.
Following at state level, in South Australia the functions of the Stormwater Management Authority (SMA) were updated in the Local Government Act 1999. The SMA’s updated goals guided progress to more effective use of urban stormwater for environmental and sustainability outcomes: ‘…for the maintenance of biodiversity, …for human consumption and …other appropriate purposes.’ Another function of the SMA is to facilitate programs by councils to address such matters.
Putting new policy into practice usually requires innovative approaches for which standards do not yet exist. The absence of WSUD and green engineering standards in the standards-driven arenas of local government and civil engineering constrained it to trials and research projects; it is still not widespread and mainstream after two decades of policy direction but its use is increasing. Resistance to a new approach or change in practise is common, and this was the case when policy proposed the merger of green and grey infrastructure goals.
Trees have long been viewed as problematic or even destructive to utilities, pavements and other built assets above and below ground, and efforts to increase tree canopy cover continue to meet resistance as a result. However, trees have been combined with WSUD devices and other engineering approaches with good effect. Devices and systems like rain gardens, green roofs and walls, infiltration trenches, leaky wells, pervious pavements, bioretention systems, swales and buffer strips, sedimentation basins and constructed wetlands can all be designed to incorporate trees and other vegetation. These systems can infiltrate stormwater more efficiently than parkland surfaces such as grassed areas, and so can help to restore pre-development runoff regimes with minimal land requirement. Through appropriate design any problematic interaction between natural and built assets can be greatly reduced, and research has begun to identify and quantify benefits of combining trees with WSUD devices.
Trees can help to manage stormwater quality and quantity; they intercept rainfall, increase infiltration, and convey much of what is detained in the soil back to the atmosphere (Thom et al. 2022; Grey et al. 2018; Szota et al. 2019). Pollutants that can be toxic to aquatic and marine life can be harvested to nourish trees and aid their growth (Denman et al. 2016). Soaking harvested stormwater into tree root zones can increase tree hydration and increase photosynthesis, transpiration and growth of saplings and mature trees (Gleeson et al. 2022), and it can reduce ground movement effects of trees in reactive soils (Johnson et al. 2020). Opportunities exist to deliver these benefits by combining trees with appropriate, compatible engineering.
Integrating trees and other vegetation into urban engineering is becoming easier as knowledge and expertise increase at the boundaries between the engineering, horticulture, arboriculture and landscape architecture disciplines. Interdisciplinary knowledge and collaboration have increased markedly in the two decades since the National Water Initiative and industry and Australian Standards now exist to inform planning and design for construction and protection of urban green infrastructure (UGI). The new Handbook, Urban Green Infrastructure – Planning and decision framework (Standards Australia, 2023) provides guidance for planning and decision making and is relevant to works on public and private land. Application of the approach described in this new handbook and compliance with AS4970-2009 Protection of trees on development sites and AS2303-Tree stock for landscape use should be considered as the mandatory minimum standards required for protecting existing trees and providing new saplings. To guide WSUD planning, best practice manuals exist in some jurisdictions and preparation of standard drawings is well progressed in South Australia. Applying Arboriculture Australia’s MIS506 Tree Valuation, the industry’s accepted minimum industry standard on tree valuation, to reliably attribute financial values to tree assets (Wilson, 2022), is also appropriate to inform project scoping and planning decisions.
With increasing knowledge, tools and standards, tree protection and provision are becoming more mainstream, standardised and effective. Research is ongoing and mainstreaming will help to deliver improved designs, efficiency and effectiveness, but much is now known from many early projects. Projects built over the past decades that continue to provide problem-free service with little or no additional maintenance requirement indicate that more widespread application is warranted. This paper summarises the authors’ experience with some UGI devices and approaches over nearly two decades. The use of soakage trenches, kerb inlets and leaky wells, and pervious pavements is described, including some updates on projects that have been introduced at Treenet symposia previously.
Soakage Trenches
Soakage trenches detain water and allow it to infiltrate the soil where it can be accessed by trees and other vegetation. Soakage systems were common in South Australia for much of the early and mid-20th century, for disposal of stormwater and effluent, but more recently ‘waste water’ is conveyed through pipe networks to treatment works or to receiving waters. Stormwater soakage system trials built in the City of Mitcham over the last two decades were combined with vegetation as trees are known to increase infiltration rates in fine soils (Day and Dickinson, 2008), thereby increasing the capacity and efficiency of the systems. Road runoff now irrigates many reserves and sustains healthy trees and vegetation in the City of Mitcham.
Stormwater from surrounding streets has recharged subsoil moisture at Burbank Reserve at Bedford Park since 2009. All runoff from three tennis courts (1,500 m2) in Naomi Terrace, Pasadena has soaked into the adjacent reserve since 2014, likely almost doubling the soil moisture that was previously available to the reserve’s trees. In 2015 a series of leaky wells was built to soak stormwater harvested from Freeling Cresent into the root zones of heritage-listed River red gum trees adjacent to Colonel Light Gardens Primary School. In 2017 over 180 linear metres of soakage trench was built in reactive soil at Thurles Reserve, St Marys, which harvested stormwater from three of the four adjacent roads to support a diverse planting of ornamental, fruit and nut trees. These systems and others were based on the innovative work of The Late Prof. John Argue, in particular his working example built in 1991 that continues to harvest runoff from 18 townhouses to provide passive irrigation to Rowley Reserve, Brompton (Argue 1999, 2004; Water Sensitive SA 2020). Details of soakage systems in Dorset Avenue, in Netherby Reserve and in Skitch Reserve that were based on Prof. Argue’s work are provided below.
Where to use soakage trenches: the ongoing, trouble-free operation of the stormwater soakage trenches reported in this paper demonstrate their suitability for use on large and small public reserves including in reactive soil of low hydraulic conductivity. Combining soakage trenches with tree cover will improve their efficiency and effectiveness.
Benefits of soakage trenches: strategically located soakage trenches across urban catchments will reduce stormwater nutrient loads and discharge volumes, improving downstream aquatic and marine environments and potentially generating substantial savings by avoiding the need for pit and pipe infrastructure upgrades. Increased soil water recharge in semi-arid and arid regions will enhance tree transpiration and growth, increase shading, and mitigate urban heat island effects. Soakage trenches will support biodiversity conservation and urban forest resilience in the changing climate by increasing the range of species that may be planted.
Dorset Avenue
A soakage system built in a verge in Dorset Avenue, Colonel Light Gardens in 2008 (Johnson, 2009), as an addition to a kerb, gutter and road renewal project, continues to demonstrate the effectiveness of this approach after 15 years of trouble-free service. Water harvested from the road by a 0.6 x 0.9 m side entry pit flows into the trench through a coiled agricultural drainage pipe which serves as a filter (Figure 1). The base of the pit was not sealed; the soil at its base and part of the agricultural pipe were covered with a layer of 20 mm gravel. The 9 m long, 0.9 m wide soakage trench was filled between 0.75 m and 1.5 m deep with 60 mm stone. Stone was wrapped in geotextile then covered with soil excavated during trench construction. Note that the site’s topsoil with its high organic content should be reinstated at the surface following construction.
Figure 1. Adapted side entry pit design used to harvest, filter and convey stormwater received from a road to a soakage trench.
Built in desiccated moderately reactive clay soil near the end of the 1997 – 2009 Millennium Drought the potential for reactive ground movement at this site upon re-wetting was considered to be substantial. The catchment area includes road and footpath surfaces along 90 m of road, the roof of a local hall, and part of the rooves of two homes. Adelaide receives 80% of its 560 mm average annual rainfall at intensities not exceeding 4 mm/hr, so flows bypassing the pit are uncommon. In a year of average rainfall it is likely that this system harvests over 100,000 litres of stormwater.
15 years after construction the kerb and gutter along the street has developed minor hairline cracks in some places but generally it remains in very good condition. No cracks in infrastructure have been observed near the soakage trench but they have been observed elsewhere in the street. No issues or problems have been observed in the road surface in the vicinity of the pit and trench since this infrastructure was built. Apart from the presence of the drainage pit, the only other obvious difference between the area near the soakage trench and elsewhere in the street more distant from it is that beneath the trees near the soakage trench the shade appears much denser than elsewhere (Figure 2).
Figure 2. Shade density beneath the mature tree nearest the soakage trench (right hand side of the road, pit is visible near centre of image) was greater than beneath other trees along the street.
Litter and clogging-related maintenance requirements were investigated at the Dorset Avenue soakage trench. Litter was not managed in the pit for the first 9 years; the usual practise of annual vacuuming was not conducted at all in the first 9 years of operation yet the system continued to function. The deciduous native White cedar (Melia azedarach) trees that line the street shed flowers, fruits, leaves and twigs so litter built up in the pit and at times it covered the filter, but the litter was not removed. After nine years without maintenance stormwater inflow was consistent at ~70 litres/minute. Litter that collected in the pit soon decomposed, i.e. in-situ composting, which may have been enabled or enhanced by biological activity through the unsealed base of the pit.
Netherby and Skitch Reserves
In March 2016 a soakage trench system was built in Netherby Reserve to manage localised flooding in Bartley Avenue, Netherby. Surface flows along the road and verges had caused localised flooding at times, so a system was designed to divert surface runoff into a drainage pit on the road frontage of Netherby Reserve, from which soakage trenches extended into the reserve. The capacity of the system and the infiltration into the reserve soils reduced surface flows and alleviated the frequency and extent of localised flooding. Construction of the reserve soakage system and associated drainage beneath the road was calculated to have saved $150,000 (AU$ in 2016) compared with extending deep drainage between the reserve and the nearest pipe network downstream. A summary of this project provided at a Treenet symposium workshop is available as a video (Johnson & King, 2019).
Trees planted in Netherby Reserve following the Millennium Drought but prior to construction of the soakage system had failed to thrive due to insufficient subsoil moisture. Although the reserve’s irrigation had been recommissioned following the end of water restrictions when the drought broke, turf irrigation is typically shallow, rarely supports tree growth, and can result in shallow tree root development. However, the growth rates of trees planted following construction of the soakage system exceeded expectations. English oak (Quercus robur) is commonly understood to grow slowly, but with subsurface irrigation supplied by the stormwater soakage trenches this species achieved apical growth exceeding 1.4 m per year in the first few years (Figure 3).
Trouble-free function of the early soakage trenches led to their wider use, including in small pocket parks, to enhance amenity, biodiversity conservation, cooling, and to manage stormwater. In 2019 a soakage system was built in Skitch Reserve, Melrose Park; a housing allotment-scale pocket park. The concept was simple: a side entry pit was built immediately upstream from an existing drainage pit that connected into the pipe network. The new pit diverted storm flows into the reserve soakage system until full, after which the gutter flow then bypassed the soakage system pit and continued down the gutter and into the downstream pipe network (Figure 4).
Figure 3. English oak saplings planted after the soakage system was installed have thrived.
The soakage trenches used in Netherby and Skitch Reserves were smaller than the original trial in Dorset Avenue; at 0.45 m wide they were half the width and they were less than 1 m deep (Figure 5). The shallower, narrower trenches were built with a level base and the slotted pipes were installed level to promote even water distribution and infiltration along their length. The smaller soakage trenches were quicker and cheaper to build and are probably more efficient, given that their greater surface area to volume ratio is likely to enhance infiltration. Trenches were fitted with a vented riser positioned beneath a vented valve box (i.e. 2 x 25 mm diameter holes were drilled in the lids) at the end of each pipe run to allow air to escape and so speed stormwater ingress.
Figure 4. Plan view of Skitch Reserve soakage system to intercept and infiltrate stormwater from Winston Avenue.
Figure 5. Cross-sections (top) and longitudinal section (bottom) showing the small, level design used to distribute water evenly throughout Skitch and Netherby Reserves.
Kerb Inlets
Early stormwater harvesting trials conducted adjacent to the Waite Arboretum in Claremont Avenue were reported at Treenet’s 2003 symposium (Porch, Zanker & Pezzaniti, 2003). Operational issues with early infrastructure were examined during Treenet’s symposium field day five years later (Lawry, 2008). These early investigations sparked ongoing research that led to the development of the kerb inlet and ‘leaky well’ (Figure 6) as was used in initial research projects. A beauty of the kerb inlet design is that it can be connected to soakage devices of any size and shape for any street verge or other application. A working kerb inlet in a street in the City of Unley was demonstrated at the 2010 Treenet symposium’s field day (Lawry & Smith, 2010).
An experiment to evaluate the performance of the kerb inlet began when road and drainage asset renewal in the City of Mitcham provided an opportunity to install 28 inlets in Eynesbury Avenue, Kingswood, in 2014. Infiltration performance and water quality benefits of the inlets were investigated, and the research installation also served as a working demonstration. The backfill media in the Eynesbury Avenue experiment included washed gravel (14 mm Stonyfell quartz screenings) but it also tested the site’s silty clay loam, a commercially sourced sandy loam, and a self-granulating, clay-based waste product of water filtration (‘SPACE’, Space Down Under, Adelaide, South Australia) (Sapdhare et al., 2018; Sapdhare et al., 2019). Gravel supported the highest infiltration rate in the local soil, followed closely by the SPACE which had the added benefit of a very high cation exchange capacity which gives it greater potential to contribute to pollution management including heavy metals.
Figure 6. Kerb inlets divert stormwater into soakage wells located in Eynesbury Avenue’s verge.
Concerns are routinely raised regarding the potential for problems to arise from infiltration into reactive soil but very little research has been conducted to inform this. Shrinkage of reactive soil upon drying is well understood, and the contribution of trees to this effect through water extraction is known, so logic suggests that offsetting the water loss due to trees through infiltration might reduce ground movement (Goldfinch 1995). To help to inform this with regard to infiltration through kerb inlets, kerb elevations were surveyed along Eynesbury Avenue to measure ground movement. Survey points were established at the locations of the inlets and wells, and midway between them. The experiment revealed that there was no difference between ground movement at the infiltration points with inlets and at points without infiltration midway between the inlets. Infiltration through inlets into moderately reactive soil did not induce ground movement during the study period from 2014 to 2016.
To investigate the benefit of these dispersed, small-scale stormwater harvesting devices at the catchment scale a larger study was established in suburban Hawthorn (34°58’30.7″S 138°36’10.0″E). This study measured the cumulative benefit of 183 inlets (model R-750, Space Down Under, Adelaide, Australia). Catchment outflow and rainfall data were collected for 12 months before the inlets were installed across the 17.5 hectare catchment; these data were used to calibrate a model to predict outflows under the catchment’s conditions prior to installation of the inlets (Shahzad et al., 2021). Comparison of the modelled flows with data measured following installation of the inlets revealed the effect of the inlets on discharge volume and concentration time.
The study showed that under field conditions in an established residential area (i.e. with minimal sediment but abundant tree litter build-up between routine road sweeping) the average harvest volume per inlet was 1.6 kilolitres in 2017 and over 2018 the average harvest was 1.4 kilolitres per inlet (Shahzad et al. 2022). The difference in harvest quantity over these two years was due to rainfall variability. Distributed kerbside stormwater storage and infiltration devices were shown to be effective in managing flows from small to medium storms, and modelling showed that with increased storage volume inlets and wells ‘…could provide a significant reduction in the runoff volume and flow rate at the catchment scale (Shahzad et al. 2022). The inlets continue to provide maintenance-free service after nine years in-situ.
In an ongoing study in the Hawthorn sub-catchment the benefits of the harvested stormwater to street tree growth and urban cooling are under investigation. Research by Flinders University staff and students has shown that mature White cedar (Melia azedarach) trees with inlets in their root zones transpired 29% more water during summer than similar trees without inlets (Gleeson 2022). Additionally, photosynthesis was 94% higher, trunk diameter increased 25% more and height increased 50% more for White cedar saplings with inlets than for saplings without inlets.
Where to use kerb inlets: anywhere where there is a kerb and gutter near vegetation. Inlets will benefit trees and communities particularly in urban streets with impermeable surfaces and high levels of urban heat, even in narrow streets with narrow verges.
Benefits: Increased transpiration in mature trees and faster growth of saplings have been demonstrated. Stormwater quality and quantity benefits result from widespread distribution of inlets across urban catchments.
Permeable paving
Permeable paving delivers trafficable surfaces, stormwater quality and quantity benefits, enhanced ecosystem services, and reduced tree root impacts but its use is still the exception in Australia despite decades of application internationally. Updated pavement base design software (DesignPave v2.0) is available for free download to aid structural and hydrologic design of permeable pavements (CMAA, undated), including on low strength subgrades that can support tree root growth. The authors’ experience with permeable paving began when solutions were needed to manage stormwater and prevent pavement damage associated with tree root growth.
Pavement surfaces are damaged when forces exerted by radially expanding tree roots exceed the shear strength of subgrades, pavement base materials, or the wearing surface. A layer of gravel has been demonstrated to deter tree root growth (Cline et al., 1980; Hakonson, 1986; Reynolds, 1990, Gilman 2006), and gravel used for structural and stormwater detention purposes beneath permeable paving has shown a similar root-deterrent benefit in experiments conducted in Adelaide (Johnson et al. 2019) and on the Sunshine Coast (Lucke and Beecham 2019). The highly porous gravel base protects the pavement assets and the trees; it provides stormwater detention volume to increase infiltration to nurture trees, and it supports seasonal drying of the gravel base and the subgrade surface which suppresses shallow root growth.
Examination of Callery pear (Pyrus calleryana Chanticleer) roots after 5 years of growth in the Adelaide experiment (Johnson et al. 2019), and of Broad-leaved paperbark (Melaleuca quinquenervia) roots 4.5 years after planting on the Sunshine Coast (Lucke and Beecham 2019), revealed significant differences in root development beneath permeable and impermeable pavements. Beneath permeable paving there were more fine, non-woody roots but fewer coarse, woody roots in the subgrade than beneath impermeable paving. Beneath permeable paving the larger diameter roots were also deeper in the subgrade than was the case beneath impermeable paving. Fine roots in the gravel layer beneath the permeable paving were of a seasonal nature and did not exceed 2 mm in diameter; evidence of seasonal fine root turnover was abundant. The study concluded that the gravel layer did function as a root deterrent; the gravel served as a buffer between the pavement surface and thickening roots in the subgrade. Growth of thicker, woody roots deeper in the subgrade beneath permeable paving further reduces the likelihood of pavement damage.
Updating the previously published results, after ten years of tree growth the Adelaide experiment’s footpath pavement sections were lifted to examine the Callery pear trees’ structural roots. The trees at this time had attained trunk diameters of over 200 mm in the 600 mm wide tree pits. Structural roots had not grown into permeable pavement gravel base layers immediately adjacent to the planting pit (Figure 7).
With the geotextile lifted and the soil washed away it was clear that all of the structural roots had developed outside the gravel and grown vertically downward to at least the depth of the pavement’s subgrade. Some roots had grown horizontally beneath the gravel without having entered the base layer (Figure 8). This growth habit was observed consistently at all of the six permeable pavement sections that were excavated for root examination. Pavement repair and root removal were required at impermeable concrete block paved sites after 10 years of street tree growth, but repairs were not required at sites with permeable paving.
The gravel base layer beneath permeable paving consistently prevented the shallow growth of large roots which is often observed to damage paving and create tripping hazards. Not compacting the subgrade, as suggested by Eisenberg et al. (2015), Drake et al. (2013) and others for hydrologic benefit, is likely to support this root growth habit. Uncompacted low-strength subgrade allows root penetration, elongation and radial expansion at depth. Compacted subgrade restricts root penetration and growth and, because compaction increases subgrade strength, if a root did develop along a line of weakness in or beneath compacted subgrade then it might impact the subgrade as a structural unit and thereby cause greater damage.
Figure 7. Structural roots did not develop in the permeable pavement gravel base layer.
Figure 8. Structural roots descended outside the gravel base before developing horizontally in the subgrade. The string marked the edge of the header course 600 mm behind the kerb. Scale shows centimeters and inches.
In a garden adjoining one of the permeable pavement sections in the Adelaide experiment a Hill’s weeping fig (Ficus microcarpa var hillii) had been planted in 2009. Separate roots from this tree grew beneath impervious paving (Figure 9) and beneath permeable paving (Figure 10). Beneath impervious paving the Ficus root growth habit was similar to the shallow growth of Chanticleer callery pear beneath impervious paving: it required roots to be cut and removed to enable footpath repair. A larger root from the same tree grew beneath a pervious pavement section 3.5 m to the north of the root shown in Figure 9. Beneath the permeable paving this larger root grew at the interface of the gravel and subgrade (Figure 10).
Following extensive research and ongoing development the City of Mitcham revised its standard design for permeable paved footpaths with the view to reducing asset life cycle costs and achieving environmental benefits. Damage to footpath assets will be avoided or reduced and injury to trees will be eliminated or reduced through the use of permeable paving, which will reduce expenditure on pavement repairs and tree replacement while enhancing ecosystem services and benefits to residents. The cost of permeable paving remains slightly higher than impermeable paving due to supply and demand, so the City of Mitcham has sought ways to reduce construction costs. As in all areas of council business the search for cost effectiveness and efficiency in asset management is ongoing, and WSUD and other innovations should continue to become more cost effective as new materials and processes are mainstreamed.
Figure 9. Shallow Ficus root growth beneath impervious concrete block paving
Figure 10. Deep Ficus root growth beneath pervious paving with a gravel base layer (the tree from which this root grew is visible in the top right of the picture).
Mitcham’s revised footpath pavement base design uses a single layer of 10 mm gravel in its base, above uncompacted subgrade (Figure 11). The 10mm gravel base replaced the previous two-layer base with its 5-7 mm gravel bedding layer above 14 – 16 mm gravel. The standard design for impermeable paved footpaths uses a sand bedding layer above a base layer of compacted quarry rubble. The new permeable design with its single material and single operation for base construction will generate savings compared with cartage, storage and handling of two bulky quarry products and construction of two separate base layers. This should more than offset the currently marginally higher cost of permeable paving to make permeable paving less costly to construct than impermeable paving. Elimination of subgrade compaction reduces time and construction equipment requirements to generate further saving. Such savings are unlikely to be delivered in the short term however, until enough contractors become familiar with the work to be confident enough to begin to compete on pricing.
Figure 11. The City of Mitcham’s adopted design for some permeable paved footpaths utilizes a single 10mm gravel base layer on uncompacted subgrade.
The City of Mitcham’s experience with permeable paving in footpath situations led to trials of its suitability for use in road surfaces. A trial in Kegworth Road, Melrose Park was designed to resolve regular local flooding issues (Figure 12). The use of permeable paving also supported the good health of significant large trees in the reserve adjacent and additional plantings in the streetscape. To maximize exfiltration the subgrade was not compacted; in its saturated state it had a Californian Bearing Ratio of 3%. Drainage was installed in the gravel base (Megaflo 170, Geofabrics Australasia) but was perched in the gravel above the subgrade to support infiltration prior to excess water during storm events being drained to leaky wells in road verges and soakage trenches on a nearby reserve. 80 mm thick Ecotrihex pavers were used above a 50mm thick bedding layer of 5-7 mm screenings and 175 mm thick base layer of 20 mm screenings. 2-3 mm sand was used to fill the joints and voids between the pavers.
Figure 12. Permeable paving resolved a localized flooding issue at Kegworth Road, Melrose Park, saving $1M compared with the cost of upgrading existing pit and pipe drainage.
The area of permeable paving at Kegworth Road was less than 600 m2 and it received stormwater runoff from a large suburban catchment. Dispersing WSUD infiltration devices throughout catchments would be a better way of preventing the localized flooding downstream, i.e. to prevent the downstream problem rather than to treat its symptoms, so this example may present a ‘worst case’ scenario in terms of repeated saturation. The Kegworth Road solution has proved successful, however some settling of the surface occurred over a trench alignment in its first few years. Settling of utilities trenches is a common problem in streetscapes with sealed roads too, and the ‘worst case’ aspect of drainage from an expansive catchment may have contributed to it happening more rapidly in this instance.
Early Australian experience with permeable paving includes Olympic Park and Smith Street, Manly, in New South Wales, Kirkaldy Avenue, Henly Beach in South Australia and other examples that date from the 1990s. These examples remain fully functional. Some other SA examples include Holland Street in Thebarton (City of West Torrens), Alfred Street in Walkerville (Town of Walkerville), Enterprise Lane in Hyde Park (City of Unley), and Druid Avenue in Stirling (Adelaide Hills Council).
Where to use permeable paving: Urban footpaths with adjacent street trees (particularly prior to planting), plazas, patios, residential roads, carparks.
Benefits: Avoid or delay root-related pavement damage, hazard, risk, expense and liability; reduced injury or loss of trees during maintenance.
Porous asphalt
Open-graded or porous asphalt has been widely used as a wearing course over impervious asphalt for many years; it reduces traffic noise, reduces or prevents plumes of water spray that can obscure visibility and it can help to reduce aquaplaning. In these applications the surface water flows into and through the voids in the asphalt to the road edge where it can drain away into conventional drainage systems. Major roads carrying vehicles traveling at higher speeds benefit most from the noise attenuation, splash, spray and aquaplaning reduction (Isenring, Koster & Scazziga, 1990). There are also substantial pollution reduction benefits, most likely due to the double effect of reduced splash removing less material from the undersides of vehicles and the trapping of contaminants within the pavement surface (Barrett, Kearfott & Malina 2006). Porous asphalt systems have also been developed using waste tyre rubber content to add flexibility, maximise fatigue resistance, and to reduce reliance on virgin materials and synthetic binders.
Recent innovations in techniques developed to create a structural underlying pavement base for permeable block paving that is suitable for roads have been applied beneath porous asphalt to capture and infiltrate the water to the underlying soils and street tree roots. The underlying base is constructed using the same method as permeable paving with an embedded geogrid system to distribute the load, however with a 75-100mm thick layer of porous asphalt over the top of the screenings made with recycled waste tyre aggregate that provides a resilient surface layer. The City of Mitcham collaborated with researchers based at the University of Melbourne to conduct an experiment in a full-scale working carpark in St Marys in South Australia. The experiment tested the porous asphalt and pavement structure under field conditions and demonstrated that these systems are viable for sustaining low to medium traffic loads (Raeesi et al., 2021).
Porous asphalt has a higher infiltration rate to most permeable pavers (Raeesi et al., 2022), and, while it is expected to reduce over time due to clogging, relative to clogging in permeable paving and even without regular maintenance the asphalt’s infiltration rate remains well above local South Australian rainfall intensities. Water that soaks into porous asphalt base layers then infiltrates into the underlying soils where it can provide moisture to trees and reduce pollution impacts on local environments (Raeesi et al., 2022). Following the successful trial at St Marys, porous asphalt has been used in many applications by the City of Unley to support the health of newly planted trees and as part of streetscape upgrades to sustain mature trees.
Unlike permeable paving, porous asphalt visually blends with conventional asphalt road surfaces which reduces the potential for distraction and unnecessary delay of motorists. Porous asphalt may also more readily fit the required dimensions and site constraints of the Manual of Uniform Traffic Control Devices (Standards Australia, 2021) than block paving. As an example, porous asphalt was used around significant trees in Dunrobin Street, Black Forest (Figure 13) where the kerb alignment could not be altered and where the previous kerb had been overgrown by the trunks of trees. The porous asphalt sections now work in harmony with the trees; they ‘future proof’ the site against further kerb lift. While the effects on trees due to the porous asphalt have not yet been investigated, the gravel layer and seasonal drying of the subgrade surface beneath the porous surface are likely to retard shallow root growth similarly to the block paving effect described previously and the infiltration of water is likely to deliver similar benefit to that shown in response to infiltration through inlets in the experiment in Hawthorn.
Figure 13. Porous asphalt construction in Dunrobin Street, Black Forest, used kerbing to create a build-out which was temporarily filled with screenings prior to installing porous asphalt.
Where to use porous asphalt: Carparks and local roads are ideal to convert to infiltration zones, with or without existing trees, as are streetscapes and parking spaces where new trees are to be planted.
Benefits: Reduced water ponding, increased shade and transpiration, larger and denser tree canopies, reduced heat island effects, reduced downstream stormwater quantity and quality issues, reduced risk of trip hazards and related liability and expense.
Discussion
The benefits of soakage systems, kerb inlets, swales and other stormwater harvesting devices have been reported in relation to Australian research and trials over many years. System performance has been quantified in working demonstrations in urban streets including in areas with moderately reactive clay soil. Enhanced tree growth and transpiration have been reported (Gleeson et al. 2022), as have stormwater quality and quantity benefits (Sapdhare et al, 2018, Sapdhare et al. 2019; Shahzad et al. 2021, Shahzad et al. 2022). Devices that have operated over years and decades with little or no maintenance remain fully functional. These well-established Australian examples show that urban green infrastructure and water sensitive urban design work from environmental, engineering and human perspectives.
The soakage trench system at the City of Charles Sturt’s Rowley Reserve in Brompton has been in operation since the surrounding townhouse development was completed over thirty years ago. This site was previously a ‘pughole’ (a quarry where clay was sourced for the local brickmaking industry) before it was partly filled to create the Rowley Park Speedway and subsequently filled more to provide a level site for housing. The continuing operation of Rowley Reserve’s soakage system on this reclaimed fill site after thirty-two years of trouble-free service, with the surrounding townhouses remaining in apparently excellent structural condition and the reserve a cool, green, real estate value-adding oasis, is proof that these high-functioning systems can have long service lives, be cost effective, and present minimal risk.
The installation of some of the City of Mitcham’s devices into desiccated soil during the Millennium Drought preceded massive fluctuation in rainfall. In subsequent years the lowest annual rainfall total of 338 mm (2015) and the highest 820 mm (2016) – the second highest ever recorded – were 38% lower than and 51% higher than Adelaide’s annual mean of 541 mm. In 1992, one year after the Rowley Reserve system was built and when the trees were still saplings, it received and managed 63% more than Adelaide’s average annual rainfall, i.e. 883 mm total annual rainfall. Through these rainfall extremes no problematic shrink-swell effect or built infrastructure impact has been observed in relation to kerb inlets or larger soakage devices like the trenches referred to in this paper.
Although many in the industry still raise concerns that stormwater infiltration into reactive soil may be problematic, published research and local examples working over long periods indicate to the contrary. Research shows that infiltration near trees can marginally offset some of the settlement that is due to trees, which seems logical. Much infrastructure damage occurs due to settlement during drought conditions; it seems logical that infiltration may also offset this seasonal effect to some degree. In contrast, urban drainage and impervious surfaces that prevent soil moisture recharge might contribute to increased settlement during dry periods and so present the greater risk.
Cost is still cited by some as a barrier to adoption of UGI and WSUD. This is surprising because accurate life-cycle costs are unavailable. The actual life-cycle costs of these innovative approaches will become clearer as initial installations approach their service life spans. Costs of materials are typically higher early in a product’s life cycle, with innovators and early adopters paying a premium but prices then becoming more reasonable as demand increases. Contractors unfamiliar with new approaches may charge a premium before mainstreaming increases competition in the market, but novel practices can still save money. The integrated design approach in the Kegworth Road permeable paving job, for example, saved $1 Million compared with business-as-usual pit and pipe extension.
The City of Mitcham’s permeable paved footpath standard with its single base layer design using a single material is likely to be cheaper to construct than the previous double layer bases for permeable and for impermeable paving. Lower life cycle costs have been reported in a UK comparison between permeable and impermeable pavement options (Wilson, 2006). Not only did permeable paving options cost less than impermeable paving in the UK comparison where WSUD has been mainstream for decades, the permeable option avoided substantial additional costs associated with conventional drainage requirements of impermeable pavement.
Risk is sometimes cited as a reason to be cautious of new and ‘unproven’ approaches. Just as life-cycle costs are unknowns with any new innovation, so UGI risk must also be considered unknown. This presents a real dilemma for asset managers who must choose between the unknown that might possibly involve a greater degree of risk than an accepted business as usual approach, but the risk of the unknown might also be less than the known and universally accepted risks of BAU. An alternative perspective might be to accept the realities that current approaches are delivering increasing problems (e.g. UHI, flooding, pollution, erosion, increasing built asset costs) and that more of the same will inevitably increase these problems further, so risk from BAU should not be accepted but considered as an increasingly expensive and problematic certainty. The risk manager’s decision might then be more clearly framed as a choice between an innovative solution that may eventually require some remedial action or an increasingly expensive and unacceptably problematic certainty.
There will always be unknowns with every new process or technology; these diminish with ongoing research. Unknowns regarding the consequences of long-term application of long-held standards and business as usual practices similarly diminish with time and research. Research over past decades has shown that fears surrounding UGI impacts are unfounded and that UGI benefits are generally greater than initially anticipated. Conversely, well-established grey engineering approaches have been shown to contribute to unsustainable and increasingly problematic environmental issues that are expensive to mitigate.
New capital works and asset renewal projects are funded routinely in council budgets and these provide opportunities to tap into the substantial long-term savings and benefits of WSUD and UGI options. Designing green assets into towns and cities, protecting and nurturing existing mature trees, and creating spaces for new trees aren’t difficult tasks provided green engineering thinking begins at the project scoping stage and continues throughout the project. Appropriately scoping UGI outcomes at project commencement will help to ensure the required interdisciplinary expertise will be brought to all project stages. It’s not a difficult task nor should it be considered risky, as since March 2023 it is informed and guided by Australian Standards’ Urban Green Infrastructure – Planning and decision framework. Choosing to not comply with the Australian Standard (as a minimum) must surely present greater professional risk.
Conclusion
South Australian experience with passive irrigation of urban trees through stormwater harvesting and infiltration exceeds thirty years. Treenet’s experience with stormwater harvesting and infiltration into road verges through kerb inlets extends over twenty years with units installed on residential, collector and arterial roads. Local experience with permeable paving began with the earliest installations dating from the 1990s. All of these systems have shown benefit to residents, local and downstream environments. None of these approaches have resulted in problematic outcomes as were predicted by many, either in years of high rainfall or low, yet unfounded concern continues to dissuade many asset managers from adopting urban green infrastructure and water sensitive urban design approaches.
The case studies reported in this paper show that the paradigm is shifting. The change to urban green infrastructure is accelerating. A goal of this paper is to encourage the further uptake and more widespread trial, application and research of urban green infrastructure. It is hoped that this paper will encourage asset managers to be open to opportunities provided by annual capital works programs to test such devices as kerb inlets and soakage wells, soakage trenches, permeable block paving and porous asphalt. Engaging in discussion with research institutions during project planning will inevitably identify opportunities to collaborate and to increase knowledge and understanding based on the performance of working infrastructure. The alternative is to perpetuate the increasingly expensive, problematic and unsustainable status quo.
