Liz Denman, Burnley College, University of Melbourne


In recent times, a greater emphasis has been placed on the environmental aspects of urban stormwater management. An experiment has been designed to investigate the potential of using street tree and tree soil systems to treat stormwater. The proposed treatment system fits within the concept of Water Sensitive Urban Design, with the aim of achieving water quality improvement, flow attenuation and integration of stormwater into the landscape. This paper describes the experimental design and methodology developed to assess the feasibility of using street tree and designed soil systems as a stormwater best management practice.


The health of aquatic ecosystems in urban areas has received much attention in recent times.  Stormwater generated in urban areas is recognized as a significant source of non-point pollution in waterways.  Non-point pollution comes from a wide range of diffuse sources in contrast to point-pollution emitted from sources such as industrial or sewage treatment plants.  Marsalek and Chocolat (2002) define stormwater as “the water running off urban surfaces, as a consequence of rainfall over urban catchments”.   Pollutants commonly found in stormwater include: heavy metals, oils and surfactants, toxic organics and nutrients (Victorian Stormwater Committee 1999).  Urbanisation and its associated high proportion of impervious surfaces also leads to increased quantities of stormwater.  The greater frequency, volume and velocity of runoff events also harm aquatic ecosystems (Wong 2000). Conventional urban drainage systems have been designed to rapidly transport stormwater to the point of discharge, the sole aim being flood prevention.

The concept of water sensitive urban design (WSUD) is a sustainable approach to urban development.  It emphasizes an holistic approach to water management with the aim of achieving multiple objectives, including:

  • protection of natural aquatic ecosystems
  • integration of stormwater treatment into the landscape
  • protection of water quality
  • reduction of run-off and peak flows (Victorian Stormwater Committee 1999)

Stormwater management is a component of water sensitive urban design (Lloyd et al. 2002).  This alternative approach views water as a resource rather than a waste product (Lloyd et al. 2002).  Examples of best management practices being incorporated into new urban developments, such as Lynbrook Estate in Melbourne, include grass swales, vegetated infiltration systems (biofilters) and constructed treatment wetlands (Lloyd et al. 2002).  In addition to flood prevention, these control measures seek to achieve water quality improvement and reduced peak flows (Wong 2000).

There is a potential to incorporate other landscape elements into urban stormwater treatment systems. These are particularly important in high-density urban zones, such as central business districts, where grassed waterways, biofilters and constructed wetlands are less easily incorporated.  The possibility of using street trees and tree soil as another stormwater management control measure exists.  Soil properties required to achieve successful tree growth and water treatment appear to overlap.

It is proposed that runoff be directed from the street through an inlet in the gutter and collected above the tree pit (Figure 1).  The stormwater would filter down through the soil profile and be collected by a drain installed at the bottom of the tree pit.  After treatment, the stormwater could be discharged into nearby waterways or potentially re-used in the landscape.  This research project will investigate the effects of street tree and soil systems on stormwater.

Figure 1. Diagram of proposed street tree and designed soil system

Figure 1.   Diagram of proposed street tree and designed soil system

Aim of experiment

An experiment has been designed to analyse the performance of street tree and tree soils systems. The focus has been limited to nitrogen and phosphorus.  The hypothesis is that certain soil and street tree combinations will achieve both adequate tree growth and stormwater treatment.  It is also hypothesised that the presence of a healthy tree within the system will improve treatment of the stormwater in comparison to unplanted systems.

Background to experiment

Target pollution

The nutrients of principal concern in stormwater are nitrogen and phosphorus.  The sources of nitrogen and phosphorus in urban areas include fertilisers, organic matter, car washing detergents, animal waste and sewer overflows (Chiew et al. 1997).  Excess levels of these nutrients contribute to blue-green algal blooms in receiving waters, although the processes are not fully understood (Poplawski & Jones 1999). Stormwater discharge into rivers may also be a concern for communities downstream if the waterway is used as a source of drinking water.  The Draft Australian Drinking Water Guidelines suggest a limit of 100mg/L for nitrate in drinking water for adults, and children over the age of 3 months, as it can cause the blood disorder methemoglobinemia (National Health and Medical Research Council 2002).  Infants are more susceptible to this condition which reduces the bloods capacity to carry oxygen.

In compiling results from various studies, Kadlec & Knight (1996) reported that the average concentration and range of total phosphorus and total nitrogen measured in urban runoff is 0.36 (0.02-4.3) and 2.0 (0.7-20) mg/L respectively.  The receiving water objective for total nitrogen and phosphorus set out in Urban stormwater: Best practice environmental management guidelines (Victorian Stormwater Committee 1999) is for the base flow concentration to comply with the relevant State Environmental Protection Policy that is to not exceed 0.9 mg/L and 0.08 mg/L respectively.

Treatment of nutrient rich water

The treatment of stormwater and wastewater using a combination of vegetation, soil and associated microbiota has been carried out in a range of different scenarios.   The use of aquatic plants in treatment wetlands is well documented (Kadlec & Knight 1996).  Irrigation of forestry plantations with treated effluent is practiced (Myers et al. 1999) (Falkiner & Polglase 1997).  Stormwater treatment has also been achieved using bio-filtration systems (Davis et al. 2001) (Lloyd et al. 2002).  These treatment systems consist of vegetated swales overlying gravel or soil filled trenches.  Lloyd, Wong and Chesterfield, (2002) reported that total phosphorus and total nitrogen loads were reduced by 77% and 70% respectively in bio-filtration systems as compared with conventional piped systems. Similarly, Davis et al. (2001) reported moderate reductions of phosphorus (~80%), total kjeldahl nitrogen (65 to 75%) and ammonium (60 to 80%) in column experiments and pilot scale plant-soil-mulch bio-retention systems.  The removal of nitrate however, was variable with both increases and decreases recorded in the effluent concentration sampled from different depths (Davis et al. 2001).  Treatment of stormwater by infiltration through soil systems planted with trees has however not been reported.

The possible fate of nitrogen and phosphorus in infiltration systems are listed below in Table 1.

Table 1 Possible fate of nitrogen and phosphorus infiltrating through planted soil systems

Pollutant Fate
  • Biological absorption by tree roots or microbiota
  • Soil adsorption
  • Transformation to nitrate (nitrification)
  • Leached from the system
  • Biological absorption by tree roots or microbiota
  • Leached from the system
  • Transformation to gaseous nitrogen (denitrification)
Organic N
  • Transformation to inorganic N (mineralization)
  • Leached from system
  • Biological absorption by tree roots or microbiota
  • Soil adsorption
  • Leached from the system

Experimental design

Plant Material

Four urban tree species, commonly grown in south eastern Australia were selected for this experiment; Callistemon salignus (Willow Bottlebrush), Eucalyptus polyanthemos (Red Box), Lophostemon confertus (Brush Box) and Platanus orientalis (Oriental Plane).  The four species chosen display a range of waterlogging tolerances (Table 2).  Both deciduous and evergreen species have been included in this experiment.  An unplanted treatment was also included to act as a control.

Table 2 Tree species selected for the experiment

Species Type Container size Tolerance to waterlogging
Callistemon salignus
Evergreen, native 140mm High
Eucalyptus polyanthemos Evergreen, native 140mm Low
Lophostemon confertus Evergreen, native 140mm Medium
Platanus orientalis
Deciduous, exotic bareroot High


The combined objectives of healthy tree growth and adequate stormwater treatment influenced soil selection.  Healthy tree growth requires drainage rates fast enough to avoid prolonged waterlogging and stormwater treatment benefits from greater retention times within the soil profile.  It was predicted that soils with drainage rates of between 40 and 180 mm/hr would achieve both goals.

Three soil blends were sourced with saturated hydraulic conductivities of 4, 95 and 170 mm/hr (
Table 3).  The hydraulic conductivity of the slowest draining soil is below the range (20-1000 mm/hr) stipulated in the Australian Standard ‘Soils for landscaping and garden use’ AS 4419-1998 (Standards Australia 1998).  Difficulty was experienced in sourcing a soil blend with saturated hydraulic of approximately 40mm/hr, hence the inclusion of this very slow draining soil.

Table 3 Soil drainage rates and supplier details

Soil Saturated hydraulic conductivity (mm/hr)

AS 4419 1988

Low 4 TGS Sand & Soil

1550 Thompsons Rd, Cranbourne

Medium 95 Rocla Sands

150 Quarry Rd, Langwarrin

High 170 Rocla Sands

150 Quarry Rd, Langwarrin

The soil profile was approximately 500mm deep, with composted green waste added (10% v:v) to the surface 200mm.  Composted green waste ‘Premium organic compost’, certified to Australian Standards 4454 Lic. 2718 was supplied by Soil Power Pty Ltd .


The irrigation treatments will be applications of either tap water (control) or a stormwater simulation.  The trees will be irrigated weekly with a volume of four litres. A regular interval between irrigation events, although not reflective of “real life” runoff events, was selected as it is more manageable.  In Melbourne the long-term mean number of rain days (147) per annum indicates that rainfall events on average occur every 2.5 days (Bureau of Meteorology 2003).  However, not all of these events would be sufficient to generate runoff, so a weekly irrigation frequency simulates close to a worst-case scenario.  The irrigation volume  (4L) corresponds to an irrigation depth of approximately 90mm.  This value approximates a maximum acceptable depth to lower the tree pit below the road surface.  The chemical makeup of the stormwater solution to be applied is listed in Table 4.   The chemical makeup was adapted from a solution devised by Davis et al. (2001) for a similar study.  Calcium chloride was replaced with sodium chloride and magnesium chloride to more closely reflect the composition of stormwater in Melbourne (P. F. Breen pers. comm.).

Table 4 Chemical makeup of stormwater to be applied

Concentration (mg/L) Concentration of target nutrient (mg/L)
   Nitrate Sodium nitrate NaNO3 12.1 2 (as N)
  Organic       nitrogen Glycine NH2CH2COOH 21.4 4 (as N)
   Phosphorus Dibasic sodium phosphate Na2HPO4 2.75 0.6 (as P)
 Heavy Metals
   Copper Cupric sulfate CuSO4 0.08
   Lead Lead chloride PbCl2 0.08
   Zinc Zinc chloride ZnCl2 0.6
 Dissolved solids
Sodium chloride NaCl 96
Magnesium chloride MgCl2 24


In total the experiment contains 30 treatments with 8 replicates of each making a total of 240 experimental units (see Table 5). The experiment has been set out as a randomised complete block design.

Table 5 Summary of experimental treatments

Treatment Number Description
Tree species 5 Callistemon salignus, Eucalyptus polyanthemos, Lophostemon confertus, Platanus orientalis and, unplanted (control)
Soil 3 Low hydraulic conductivity (4 mm/hr)

Medium hydraulic conductivity (95 mm/hr)

High hydraulic conductivity (170 mm/hr)

Irrigation 2 Stormwater

Tapwater (control)

Replication 8 Randomised complete block design

Sampling tubes throughout soil profile incorporated in 5 of the 8 blocks.

Experiment setup

Model soil profiles were built in above-ground containers.  These were built from Polyvinyl chloride (PVC) stormwater pipe, internal diameter 240mm, cut into 600mm lengths (Figure 2. Model Soil Profile). Each container stands on a plastic saucer with the centre drilled out.  These were raised off the ground to allow the collection of leachate.  Collection and analysis of leachate will occur on a fortnightly basis immediately following an irrigation event.


The irrigation system is set up in two sections (Blocks 1-3) and (Blocks 4-8) with a loop design to minimise the variation in water pressure.  Each system has a microspray inserted into a 500ml plastic takeaway food container with two holes drilled around the outside (Figure 3).  Concentrated stormwater solution (10ml) will be pipetted into the containers prior to the commencement of irrigation.  The concentrate and tap water will mix inside the plastic container and then pour out through the drilled holes. This irrigation system was devised to minimise contact between the phosphorus in the stormwater simulation and the plastic irrigation piping and therefore avoid phosphorus adsorption.


The evergreen tree species (C. salignus, E. polyanthemos and L. confertus) were planted between 29th March and 1st April, 2003 in the field station at Burnley College.  The potting mix was washed off the rootballs before planting to maintain homogenous soil properties within the system.  At this time the unplanted and Platanus orientalis treatment cylinders were filled with soil.  Seasonal availability constraints meant that bareroot Platanus orientalis were sourced and planted on 11th June.

Establishment period

An establishment period following planting is being undertaken to allow the system to stabilise.  During this period mains water irrigation has been applied as required.   Some initial applications of liquid fertiliser were required as the evergreen trees, sourced in the Spring of 2002, were displaying visual signs of nutrient deficiency.  Healthy foliage colour is now evident.  The removal of lateral shoots from the lower trunk of each seedling has been initiated to reflect the management of street trees in the nursery and also improve the reliability of height increases as a measurement of tree growth.


Data collection will focus on tree growth and stormwater treatment and is scheduled to begin in September 2003.  The measurements are listed below (Table 6).

Table 6 Summary of data collection

Goal Measurements Frequency
Tree health / growth
Increase in tree height Monthly
Caliper Experiment commencement and conclusion
Visual inspection of foliage As evident
Stormwater treatment
Oxidised nitrogen (nitrate and nitrite) Fortnightly
Ammonium Fortnightly
Total kjeldahl nitrogen Fortnightly
Reactive filterable phosphorus Fortnightly
Total phosphorus Fortnightly


Analysis of the results will be through the use of the general linear model ANOVA (a 5 x 3 x 2 factorial).  The following questions will be addressed in the analysis of the results.

  • Is tree growth adequate with the imposed irrigation regime (4L applied weekly)?
  • Do street tree and tree soil systems reduce nitrogen and phosphorus levels in stormwater?
  • Does the presence of a tree have a significant effect?
  • Are there differences between the four species selected?
  • Are there differences between the three soil blends?


An experiment has been designed to investigate the possible use of street tree and designed soil systems in treating stormwater.  The experimental treatments include, four street trees species with a range of waterlogging tolerances, and three soils with differing saturated hydraulic conductivities.  It is hypothesised that certain street tree and soil combinations will achieve both adequate tree growth and stormwater treatment.  Data collection is scheduled to start in September 2003.


  • Bureau of Meteorology 2003, Averages for Melbourne Regional Office,
  • Chiew, F.H.S., Mudgway, L.B., Duncan, H.P. & McMahon, T.A. 1997, Urban stormwater pollution, Cooperative Research Centre for Catchment Hydrology, Melbourne, Australia.
  • Davis, A.P., Shokouhian, M., Sharma, H. & Minami, C. 2001, Laboratory study of biological retention for urban stormwater management, Water Environment Research, 73(1), pp. 5-14.
  • Falkiner, R.A. & Polglase, P.J. 1997, Transport of phosphorus through soil in an effluent-irrigated tree plantation, Australian Journal of Soil Research, 35 pp. 385-397.
  • Kadlec, R.H. & Knight, R.L. 1996, Treatment wetlands, CRC Lewis Press, Boca Raton.
  • Lloyd, S., Wong, T. & Chesterfield, C. 2002, Water sensitive urban design – a stormwater management perspective, Cooperative Research Centre for Catchment Hydrology, Melbourne.
  • Marsalek, J. & Chocat, B. 2002, International report: stormwater management, Water Science and Technology, 46(6-7), pp. 1-17.
  • Myers, B.J., Bond, W.J., Benyon, R.G., Falkiner, R.A., Polglase, P.J., Smith, C.J., Snow, V.O. & Theiveyanathan, S. 1999, Sustainable effluent-irrigated plantations: an Australian guideline, CSIRO Forestry and Forest Products, Canberra.
  • National Health and Medical Research Council 2002, Australian drinking water guidelines – Draft for consultation,
  • Poplawski, W.A. & Jones, G.J. 1999, The blue-green algal scourge – emerging developments in prevention and management, in Water 99: Joint Congress: 25th Hydrology & Water Resources Symposium, 2nd International Conference on Water Resources & Environment Research, vol.2, ed. W. Boughton, Barton A.C.T., Institution of Engineers, Brisbane, Australia, pp. 836-840.
  • Standards Australia 1998, Australian Standard: Soils for landscaping and garden use AS 4419-1998, Homebush, NSW.
  • Victorian Stormwater Committee 1999, Urban stormwater – best practice environmental management guidelines, CSIRO Publishing, Melbourne.
  • Wong, T.H.F. 2000, Improving urban stormwater quality from theory to implementation, Water – Journal of the Australian Water Association, 27(6), pp. 28-31.