The Science of the Total Environment 235 Ž1999. 189]197

The design of vegetative constructed wetlands for the treatment of highway runoff R.B.E. Shutes a,U , D.M. Revitt a , I.M. Lagerberg b, V.C.E. Barraud b a

Urban Pollution Research Centre, Middlesex Uni¨ ersity, Bounds Green Road, London N11 2NQ, UK b Halcrow Water, Burderop Park, Swindon Wiltshire, UK

Abstract The Environment Agency for England and Wales are responsible for assessing the effects of highway runoff and for monitoring the treatment systemsrprocedures which have been introduced for the reduction of deleterious effects. The Agency is looking into the improvement of surface water management in terms of best management practices and plans to work in partnership with the Highways Agency to achieve this aim. Among the treatment options being considered are constructed wetlands. Draft Guidelines have been developed to provide information on their design. This paper describes procedures for carrying out an Environmental Sensitivity Analysis to determine whether treatment by a constructed wetland is appropriate. Information on water quality and quantity is required as well as the sensitivity of the receiving environment. The legislative position, particularly in relation to the discharge quality of the water and the conservation status of the receiving environment, needs also to be considered. The factors that will determine the most appropriate wetland design criteria include traffic loadings, road drainage area, land availability, cost and the sizerextent and type of the receiving water body. The following structures are recommended for incorporation in the overall design; oil separator and silt trap, spillage containment, settlement pond, vegetative wetland and final settlement tank. The operation and maintenance procedures and the monitoring requirements for a functioning wetland are described. Q 1999 Elsevier Science B.V. All rights reserved.

1. Introduction Runoff is an important source of pollution which may be discharged directly, or as treated

U

Corresponding author. Tel.: q 44-181-362-5000; fax: q44441-4672. E-mail address: [email protected] ŽR.B.E. Shutes.

effluent, into receiving environments such as aquifers, watercourses, and wetlands. Careful consideration of the sensitivity of the receiving environment should inform and guide each stage in the selection, construction and maintenance of road runoff treatment facilities. The factors likely to influence the sensitivity of the receiving environment to runoff discharge are: receiving water quality; ecological sensitivity; site protection; and

0048-9697r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 9 9 . 0 0 2 1 2 - 0

190

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

sensitivity of groundwaters. Currently, in the UK, there is no regional or national assessment of the environmental deterioration caused by road runoff discharged to ground and surface waters. Water quality changes resulting from road runoff discharge include the occurrence of surface oil films, deoxygenation, increased concentrations of suspended solids and trace metals such as lead, zinc, copper, cadmium and nickel. Metals tend to precipitate in hard waters and, with hydrocarbons, may accumulate in sediments. Some benthic macroinvertebrate communities are affected by these pollutants; others may rework the sediments releasing toxicants into the overlying water ŽBoxall and Maltby, 1995.. In soft waters or where there are acidic industrial discharges, the concentration of metals in solution may rise, resulting in bioaccumulation or toxicity in aquatic flora and fauna. The sensitivity of the system depends on the inherent buffering capacity of the receiving waters and the extent to which road runoff impacts can be tolerated by the existing flora and fauna. The most sensitive ecosystems are those which cannot tolerate even a minor change in the physical or chemical quality of their environment. Invariably, the most sensitive ecosystems are those which are influenced by limiting factors such as extreme pH, temperature or dissolved oxygen availability. The receiving environment may be sensitive to flooding, scouring, erosion and siltation problems caused by large volumes of runoff discharged to watercourses. Designated areas are identified and protected for their habitats, wildlife species, amenity or cultural assets. Any environmental change Žphysical, chemical or biological. affecting such areas can compromise their value. Designated sites should not be considered to be the only areas worth protecting but their integrity is of paramount importance in the process of decision making for the treatment of road runoff. The development of a constructed wetland system for the treatment of highway runoff, which may lead to infiltration of the effluent into the groundwater, requires prior consultation with the Environment Agency. The policy for water quality objectives and standards for England and Wales

for groundwater is set out in policy and practices for the protection of groundwaters ŽNational Rivers Authority, 1992.. The policy states that discharges to soakaways from major roads will require oil separators. The level of environmental protection required will be determined by existing legislation regarding the conservation status of the receiving environment, and future legislation associated with river quality objectives. Some of the legislation defines standards for water quality which could be used as the basis for deriving discharge quality requirements for road runoff. The results of this study should assist the future establishment of clearer guidelines and regulatory framework. Having obtained data on the runoff characteristics and the ecological sensitivity of the receiving environment, analysis must be carried out to ascertain the overall environmental sensitivity and therefore the requirement, or otherwise, for road runoff to be treated. Environmental sensitivity analysis can be carried out using the results of surveys and hydraulic calculations. However, although the quantity and quality of road runoff can be modelled, the impacts have not yet been described for a sufficient number of different sites or monitored over long enough periods of time to develop a prediction model.

2. Factors influencing constructed wetland design Although relatively new in the UK, the use of constructed wetlands for the treatment of highway runoff has been established in continental Europe and in the US for some time. Attention has been paid to the utilisation of American research data where ambient temperatures are similar to the UK and European averages. The factors that will determine the selection of the most appropriate design criteria include: local climate, topography and geology; traffic loadings Žpresent and future.; road drainage area; land availability; cost; sizerextent and type of receiving water body; water quality classification and objective Žincluding water uses.; and environmental enhancement value. A constructed wetland system to treat highway runoff should ideally in-

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

clude the following structures: oil separator and silt trap; spillage containment; settlement pond and associated control structures; constructed wetland and associated control structures; final settlement tank; outfall into receiving watercourse; and access. An example of an idealised treatment system is shown in Figs. 1 and 2. There are no established design criteria for constructed wetlands for the treatment of highway runoff although the design of urban runoff wetlands has been described by Shutes et al. Ž1997.. This paper describes the development of guidelines for the design, construction, operation and maintenance of these facilities for the Environment Agency and others to use ŽLagerberg et al., 1998.. Performance criteria cannot be set because of the fluctuations in influent flows and pollution loadings. If criteria were adopted for the highest flows or loads the system would be

191

over-engineered, or the water quality standards for certain pollutants would have to be relaxed. In addition, the criteria need to make reference to Water Quality Objectives ŽWQOs. in the UK and take into account the dilution effect of the receiving water. Instead performance criteria may be based upon a selected design storm to be treated and a specified critical flow in the receiving watercourse to be protected. Traditional pollution control measures for urban and highway runoff in the UK have included grit and oil separators for the reduction of sediments and hydrocarbons. They are, however, inefficient in removing the majority of the pollution load and the finer and more mobile sediments Žwhich clog some designs of constructed wetland treating road runoff.. Integrated pollution control systems including a combination of oil separators, spillage containment facilities and wetland-fore-

Fig. 1. Idealised layout of constructed wetland.

192

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

Fig. 2. Section through sub-surface constructed wetland.

bays or lagoons, located prior to constructed wetlands, provide for pretreatment of raw highway runoff and help to prevent siltation in wetland inlet zones. Oil and phytotoxic chemicals in highway runoff can seriously affect the treatment efficiencies of constructed wetlands and the viability or performance of the plants. As constructed wetlands require 1]3 years to mature and become capable of efficient wastewater treatment, bypass oil separators, silt traps and spillage containment facilities must be installed prior to the discharge of highway runoff into the constructed wetland. All three structures must be tamper-proof and easily accessed. The spillage containment facility should have a minimum volume of 25 m3. Where land availability is not limited Žie. rural and semi-rural areas., forebays with additional oil booms on the water surface have been advocated to serve as secondary sedimentation chambers to reduce the initial flush of pollutants into the main wetland ŽCIRIA, 1994.. The incorporation of settlement ponds if space is unlimited is also recommended. Ellis Ž1991. suggested from a review of a number of studies in the US and Europe, that the maximum pollutant removal appears to occur when the settlement pond surface area is 2]3% of the drainage area ŽSD. and the retention volumes exceed 100 m3 per hectare of effective drainage area or 4]6 times the mean total storm

runoff volume. However, a large pond would be required and land space may be limited; therefore, compromises are suggested below. If sufficient land is available, a final settlement tank Žconcrete structure . with a minimum capacity of 50 m3 extending across the width of the wetland can be installed. It will help to prevent fine sediment from the wetland being transferred into the receiving water body especially during the highway construction period. The final settlement tank is an idealised part of the overall system and only needs to be included in the overall design where greatest protection to the receiving water is required. Regular maintenance is recommended to prevent collected sediments being resuspended during high flows. The rate of sediment deposit will vary with each catchment so the frequency of sediment removal cannot be predicted. Annual inspections should be made to determine if sediment removal is required. In principle there are two types of constructed wetlands that could be designed to provide treatment for highway runoff: Surface flow ŽSF. system; and Sub-surface flow ŽSSF. system. The maximum potential treatment will be achieved by the use of a SSF constructed wetland system and the design recommendations that can be made at this stage are based on this system. A SSF system provides very little flood storage and thus, where significant flood storage of runoff is required, it

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

will be necessary to combine the treatment facility with separate storage. The balance between storage and treatment requirements must be made on a case by case basis. For instance: 1. If land space is not limited, the constructed wetland could be sized to receive the design flood at an attenuated flow rate regulated by the settlementrbalancing pond outlet structure. 2. If land space is limited, the constructed wetland could be sized to only treat the first flush. In this case the settlementrbalancing pond outlet control structure will need an overflow incorporated to divert excess flows to the watercourse after detention, bypassing the constructed wetland. If land space is further limited, the next compromise would be to reduce or, if absolutely necessary, remove the spillage containment area and the final settlement tank. 3. Alternatively, if land space is limited the settlement pond and constructed wetland could be incorporated into the balancing pond.

3. Design criteria for constructed wetlands There are no established design criteria for constructed wetlands for the treatment of road runoff. Until further information is available to develop relevant design criteria, the following guidelines are recommended at this stage for the design of a SSF system. Constructed wetlands can be designed to either: retain short duration storms for the maximum retention period, ensuring that the high flows can be accommodated by the constructed wetland without overland flow; or retain longer duration storms ensuring that the first flush containing the heaviest pollution loads receives adequate treatment. It is important that the constructed wetland is large enough to retain the first flush of the longer storms to achieve partial treatment and delay transfers to the watercourse until normal flows have returned. The constructed wetland should be designed to treat

193

storms with a return period of 10 years where the availability of land makes this feasible. If a compromise is necessary requiring a design based on a shorter return period, the system should be capable of treating the polluted first flush of any storm event. Retention time is an important factor in the treatment performance of constructed wetlands. Considerations affecting the retention time include the aspect ratio Žwidth: length., the vegetation, substrate porosity and hence hydraulic conductivity, depth of water, and the slope of the bed. Water level and flow control structures, for example flumes and weirs, are also required to keep the hydraulic regime within desired parameters. Constructed wetlands should ideally have a minimum retention time of 30 min for the design storm event. An ideal design should retain the average annual storm volume for a minimum of 3]5 h and preferably for 10]15 h to achieve a good removal efficiency. A maximum retention time of 24 h has been suggested. For the design storm, the following criteria are recommended. Retention time: Aspect ratio Žwidth:length.: Slope of base of wetland bed: Minimum substrate bed depth: Substrate Žmedia.: Hydraulic conductivity of: substrate

24 h maximum 1:1 to 1:2 1% maximum 0.6 m 0.15 m of soil over 0.45 m pea gravel 10y3 mrs to 10y2 mrs

Once the design storm choice has been made, the size of the conceptual constructed wetland can be calculated using Darcy’s Law and the above criteria. Darcy’s Law assumes laminar uniform and constant flow in the media bed and clean water. In a wetland, flow will be channelled and short-circuited and the media will be covered with biological growths; therefore the equation only has limited usefulness in wetland design. Nevertheless Darcy’s Law does provide a reasonable approximation of flow conditions in beds if moderate sized gravel Že.g. 10 mm pea gravel. is used for the support medium. During storm events, high rates of highway

194

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

runoff may discharge onto constructed wetlands, but optimal hydraulic loading should not exceed 1 m3rm 2 per day in order to achieve a satisfactory treatment. Flow velocity should not exceed 0.3]0.5 mrs at the inlet zone if effective sedimentation is to be achieved. At velocities greater than 0.7 mrs, high flow may damage the plants physically and cause a decline in system efficiency. The inlet pipe should be constructed in such a way that influent flow is evenly distributed across the width of the bed. This may be achieved using slotted inlet pipes where the slots are sufficiently large to prevent clogging by algae. The distribution system must be designed to allow maintenance in case of blockage. Riser pipe distributors are adopted on many wastewater treatment systems ŽCooper et al., 1996.. A stilling structure under the inlet, usually a 1m wide stone trench Žrip-rap or gabion zone., is necessary to either dissipate high water flows or contain the inlet distributor pipe. Rip-rap and gabions are blankets of stones placed to protect erosion zones. The stones for rip-rap are laid directly on the bed, whereas they are packed in cages for gabions. The level at which the outlet is set is determined by the lowest level required. Until further information is available, it is considered that the lowest level in the wetland should be 300 mm below the substrate surface. An additional source of water may be needed to supply the reedbeds during dry periods. Ideally the outlet structure should incorporate control measures which allows the water level in the bed to be varied. The control should at least incorporate a measure to allow periodic flooding. At the outlet zone it is recommended that an additional rip-rap Žor gabion. zone be situated to prevent weed growth and resuspension of reedbed substrates. A grid of slotted plastic pipes Žwith diameter of about 100 mm. should be installed vertically in the substrate Ž100 mm protruding above the surface, and penetrating the full depth of the substrate. at 5-m intervals, to aid aeration of the root zone. Plastic poles should be erected to support lines of bunting to discourage birds from feeding on young plants. The height of the bunting should be about 1.5 m above the substrate surface. Non-

metallic items should be incorporated into the construction of the wetland so that metals in the wetland only come from road runoff. Therefore gabions should be encased with geotextiles and the poles supporting bunting should be plastic. The location of constructed wetlands will depend on: low points in the road drainage system; location of the receiving waters; and local topography. The constructed wetlands will not necessarily be sited beside the road. However, the centres of roundabouts and areas between the arms of slip roads are spaces that may be utilised in wetland construction. Gravel provides the most suitable substrate for constructed wetland emergent plants, supporting adequate root growth, high conductivity and superior permeability. Ideally, prior to use, all components of a substrate mixture should be analysed for hydraulic conductivity, buffering capacity, pH, plant nutrient levels and microbial activity. Hydraulic conductivity is the most important determinant in pollutant removal efficiency, and is especially important in SSF systems where purification processes are largely confined to the root zone. Natural boulder or bentonite clay or geotexile liners may be used as reedbed bases, in instances where prevention of leakage to groundwaters is imperative. An impermeable layer is also necessary to retain water in the wetland during dry periods. The top surface of the substrate must be level. This allows flooding of the reedbed to occur for control of weed growth when the reeds are being established. It is recommended that the constructed wetland is based on two main species: reedmace (Typha latifolia) in the first half and common reed Ž Phragmites australis) in the second half; and a fringe of other plants, such as Iris, to soften the wetland appearance. The use of a range of emergent and floating aquatic plants is recommended to enhance the ecological and visual interest. The use of vegetation is often considered to be a more aesthetically pleasing feature within the landscape than a concreterbrick treatment system with no vegetation. However, constructed wetlands for the treatment of road runoff may well be located in places that are not their natural

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

habitat. Their alien appearance may be accentuated by the design of unnatural shaped beds. Constructed wetlands should be designed to fit in with the natural environment.

4. Retro-fitting of existing treatment systems Retro-fitting means the installation of a system into a structure that already exists. The physical attenuation of storm runoff from highways has been practised for many years. Therefore there are many balancing facilities adjacent to highways throughout the UK. Although, these traditional facilities generally do not include vegetative systems, some have been naturally colonised by reeds and aquatic plants. To provide treatment, as well as flood attenuation, it may be possible to retro-fit a constructed wetland into these existing structures. When considering the retro-fitting of constructed wetlands into existing balancing ponds, the following need to be examined: Does suitable access exist or can it be provided? Can the storage for flood attenuation be reduced so that the 0.6 m deep substrate of a constructed wetland can be incorporated? Is the outlet structure of the balancing pond offset from the inlet structure? If the outlet is offset Žie not directly opposite the inlet. then the flow would short circuit. Short circuiting could be reduced by introducing plastic baffles into the substrate. Does the balancing pond have an impermeable liner? An impermeable lining is necessary to retain a minimum water depth to sustain the plants during periods of no rainfall. It is anticipated that a constructed wetland retro-fitted into a balancing pond will operate such that the initial storm flows pass through the substrate so that subsurface flow treatment occurs. If the storm flows continue until the water level in the pond rises above the surface of the substrate, then the constructed wetland will operate as a surface flow system. An example of a constructed wetland retrofitted into a balancing pond is given in Fig. 3.

195

5. Operation and maintenance requirements The operation and maintenance procedures connected with a constructed wetland are anticipated to include: removal of sediment; maintenance of the substrate and plants; harvesting; maintenance of water levels; maintenance of nutrient levels; general structure maintenance; and control of weed growth. A higher frequency sediment removal will be required during the highway construction period. To carry out the operation and maintenance requirements, good vehicular access is required to all constructed wetlands. Continued hydraulic conductivity of the substrate Žby washing or replacement. must be maintained, with removal of accumulated sludges and replacement of moribund areas of vegetation The accumulation of a litter layer is beneficial for metal removal and plant harvesting is not normally required It is likely that constructed wetlands for highway runoff treatment will only require significant maintenance between 15 and 25 years following commissioning. However, as more information is collected on systems for treating road runoff, the figure for this period may change. Periodic flooding of the constructed wetland may be necessary to control weed growth when the reeds and aquatic plants are initially growing to maturity. However, the density of reeds at maturity would considerably reduce or eliminate the possibility of weed growth.

6. Monitoring of wetland system The duration and frequency of a monitoring programme will be determined by its objectives which may include: monitoring for compliance with a discharge consent or operating agreement with the Regulatory Authority; establishing a long-term management programme to ensure the continued functioning of the constructed wetland; or further understanding of the processes which take place within a wetland. There is a range of water quality parameters

196

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

Fig. 3. Balancing pond after retro-fitting to incorporate a constructed wetland.

which can be monitored to assess wetland performance. A comprehensive monitoring survey would include measurement of: pH; dissolved oxygen; total suspended solids ŽTSS.; Biochemical and

Chemical Oxygen Demands ŽBOD and COD.; nitrates and phosphates; heavy metals; and hydrocarbons. Aquatic sediments are frequently a repository

R.B.E. Shutes et al. r The Science of the Total En¨ ironment 235 (1999) 189]197

of metals, hydrocarbons and pesticide residues, many of which may be toxic to benthic macroinvertebrates ŽBoxall and Maltby, 1995. or may be transported to more sensitive sites. The water quality response to the impact of road runoff discharge may be assessed by a number of methods including: direct measurement of a number of chemical parameters; biotic indices based on the biodiversity of communities of aquatic organisms; toxicity-based criteria; and river habitat surveys. Further evaluation of constructed wetland performance and its effectiveness in protecting the receiving environment is essential. It is recommended that performance data, based on the receiving environment, is collected and used to develop a predictive model. This would then become applicable at the design stage of road building and widening projects, as well as for the selection of sites for retro-fitting.

Acknowledgements Funding to support this work has been provided by Thames Region of the Environment Agency and has resulted in the publication of a report entitled ‘Treatment of Highway Runoff Using Constructed Wetlands}An Interim Manual’ dated February 1998. The project is managed

197

by Maxine Forshaw and Barry Winter. The authors thank Mr Chris Birks, Regional General Manager of Thames Region for the assistance given during the project and for his approval to publish this paper. However, the views expressed are those of the authors and may not necessarily be shared by the Environment Agency. References Boxall ABA, Maltby L. The characterization and toxicity of sediment contaminated with road runoff. Water Res. 1995;29Ž9.:2043]2050. CIRIA. Control of pollution from highway drainage discharges. Luker M, Montague K, editors, CIRIA Report 142, London: 1994. Cooper PF, Job GD, Green MB, Shutes RBE. Reedbeds and constructed wetlands for wastewater treatment. WRc plc, Swindon: 1996. Ellis JB. The design and operation of vegetation systems for urban source runoff quality control. In: Proc.3rd Standing Conf. Stormwater Science Control. Coventry Polytechnic, 1991. Lagerberg I, Barraud V, Shutes B, Revitt M, Smith A. Treatment of highway runoff using constructed wetlands}an interim manual. Environment Agency, Thames Region, Reading, 1998. National Rivers Authority. Policy and practice for the protection of groundwater. Bristol: NRA, 1992. Shutes RBE, Revitt DM, Mungur AS, Scholes LNL. Design of wetland systems for the treatment of urban runoff.. Water Sci Technol 1997;35Ž5.:19]25.