FATE OF HYDROCARBON POLLUTANTS IN SOURCE AND NON-SOURCE CONTROL
SUSTAINABLE DRAINAGE SYSTEMS
Georgios Roinas*, Cath Mant, John B. Williams
University of Portsmouth, School of Civil Engineering and Surveying, Portland Building, Portland
Street, Portsmouth, Hants, PO1 3AH, UK.
* Corresponding author. Tel.: +44 (0)23 9284 2428 Fax: +44 (0)23 9284 2525 georgios.roinas@port.ac.uk
Abstract
Sustainable Drainage (SuDs) is an established method for managing runoff from developments and source control is part of accepted design philosophy. However there are limited studies into the contribution source control makes to pollutant removal, especially for roads. This study examines organic pollutants, Total Petroleum Hydrocarbons (TPH) and Polycyclic Aromatic Hydrocarbons
(PAHs), in paired source and non-source control full-scale SuDs systems. Sites were selected to cover local roads, trunk roads and housing developments, with a range of SuDs including porous asphalt, swales, detention basins and ponds. Soil and water samples were taken bi-monthly over 12 months to assess pollutant loads. Results show first flush patterns in storm events for solids, but not for TPH. The patterns of removal for specific PAHs were also different, reflecting varying physicochemical properties. The potential of trunk roads for pollution was illustrated by peak runoff
TPH >17,000 μg/l. Overall there was no significant difference between pollutant loads from source and non source control systems, but the dynamic nature of runoff means that longer term data is required. The outcomes of this project will increase understanding of organic pollutants behaviour in SuDs. This will provide design guidance about the most appropriate systems for treating these pollutants.
Keywords
SuDs, Vegetated detention ponds, Hydrocarbons, PAHs, TPH.
Introduction
Industrialization, based on increasing fossil fuel use, has resulted in widespread environmental contamination with petroleum and petroleum by-products (Bauman, 1991). Many of these compounds have mutagenic and carcinogenic effects so there is concern over their environmental fate (Luch, 2005). Their transport and ultimate fate depends on the form of release and a range of factors such as solubility, volatilisation, and effects of biodegradation and weathering. These in turn are affected by the temperature, solar radiation and the availability of nutrients, oxygen and water.
Many of the pollutants in runoff arise from non-point source emissions. Roads are a major contributor to this load and contain a range of hydrocarbons (Ellis, 1991). Total Petroleum
Hydrocarbons (TPHs:C
10
-C
40
) and Polycyclic Aromatic Hydrocarbons (PAHs) are two hydrocarbon groups commonly detected and can therefore enter water courses from runoff (Viklander, 1999).
PAHs have been shown to have a higher sorption potential on to the finer fractions of solids in road runoff (<45 μm) and therefore the relative proportions of size fractions in runoff is an important factor affecting overall PAH levels (Aryal Rupak et al ., 2005). The organic pollutants in urban

runoff are often not included in monitoring and further investigation is needed to assess the sources and types (Makepeace et al., 1995).
Conventional urban and road drainage has traditionally been most concerned with removing surface water quickly. This exacerbated flooding problems and did little to manage the pollutant load in runoff. Thus, this problem led to the requirement to install Sustainable Drainage Systems (SuDs) at many locations, especially in new developments. SuDs attempt to recreate the pre-development hydrology by using a range of techniques, including infiltration, swales, wetlands and vegetated detention ponds providing management of water quantity and quality, while also providing an amenity value. Managing runoff close to the water, or source control (SC) is an accepted part of
SuDs philosophy.
The complex ecology of vegetated SuDs systems exposes pollutants to a range of conditions and processes. This is especially useful for treating the wide variety of petroleum hydrocarbons as treatment can occur through processes such as adsorption, volatilization, photolysis and biodegradation in the same part of the SuDs system.
Some studies have examined TPHs and PAHs in SuDs. However, there has been limited comparison between source and non-source control systems in terms of treatment efficiency. This study examines the fate of organic pollutants, TPHs and PAHs in a range of paired source and nonsource control full-scale SuDs to assess the fate of these compounds and the influence of source control techniques.
Methods
Sites were selected to cover housing developments and trunk roads served by a range of SuDs options, such as porous asphalt, swales, detention basins and vegetated ponds. Where possible source control (SC) and non-source control (NSC) sites have been selected in the same or similar catchments to allow comparison. Soil and water samples were collected bi-monthly over a 12 month period. The study was undertake across a range of sites in the South of England and logistics meant that a “spot” sample design was undertaken and storm monitoring was restricted to the location closest to the University laboratories (Case study 1).
Case Study 1 – Commuter Access Road
This location was used to develop the laboratory methodology for the project and is part of the
Newlands development of over 2,000 new houses at Waterlooville (Hampshire, UK Latitude =
50.881315, Longitude = -1.037575) and SuDs were designed by Mayer Brown Ltd.
The SC SuDs system is comprised of a vegetated pond system planted with Phragmites australis which is fed by
100 m swale receiving runoff from an access and commuter road (Figure 1A).
The swale receives the runoff from the urban road via two pipes and supplies it to the pond system.
The storage capacity of the vegetated pond is 304 m
3
and the permanent water depth is 1 m, rising up to 1.60 m. More details about this system were presented by Seekkubadu et al . (2009). This site was fitted with flow gauging devices but these failed during the project and the manufacturer was unable fix them, so storm hydrographs were not available.
Case Study 2 – Trunk roads
A non-source control system was selected on the M27, a major motorway at Townhill ( Hampshire,
UK Latitude = 50.937128, Longitude = -1.347590) . This is a rectangular pond with a section planted with Phragmites australis and receives runoff directly from a concrete 6 lane highway
(NSC) (Figure 1C). It is compared with detention pond C at the A34 Newbury Bypass (Wiltshire,
UK Latitude = 51.364868, Longitude = -1.359183) which receives road runoff from permeable asphalt through an oil interceptor, which is located before the inlet of the pond (SC) (Figure 1B).
This vegetated pond was planted with Glyceria maxima . Details of the A34 systems and their initial

performance can be found in Pontier et al. (2003). Both vegetated ponds with different shapes allow the comparison of source control on major roads with high traffic loadings.
Case Study 3 – Residential Development
Cambourne (Cambridgeshire, UK Latitude = 52.213356, Longitude = -0.065998) is a major housing development which was used as a demonstration site for SuDs in a previous project. The catchments compared in this study are located in Friar Way (NSC) and Lambs Drove (SC). The
NSC catchment is a conventional kerb/gully and manhole system serving a residential area (layout not shown). The SC catchment studied consisted of a sequential treatment train of two detention basins (numbered 1 and 2) connected by swales (details can be found in Royal Haskoning, 2011)
(Figure 1D). The water from both systems is discharged into a large vegetated detention pond planted with a variety of reeds of Phragmites australis, Glyceria maxima and Typha latifolia
Figure 1. Schematic Plans of the Case Study SuDs Systems, showing the main features and sampling locations. Flow paths are shown by dotted lines. A = Case Study 1 Commuter Access
Road: B + C= Case Study 2 Trunk roads: D= Case Study 3 Residential Development Source
Control system.
Sampling
Water samples were collected using a hand pump to avoid air bubbles and transferred to amber glass bottles to avoid degradation. Soil samples were collected using metal tubes (60 mm) and wrapped in foil for transport to the labs. The cores were divided into three layers (0-20 mm, 20-40 mm and 40-60 mm) for analysis. The regular monitoring has been supplemented by targeted monitoring of storm events and combined with general measurements of water quality (biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended solids (TSS), volatile suspended solids (VSS), conductivity, pH, TPHs and PAHs). The original 16 EPA PAH priority pollutants were monitored along with naphthalene,2-methyl and naphthalene,1-methyl which area commonly found from incomplete fuel combustion (Lerda, 2010). This covers a range of compounds with varying molecular weight and physico-chemical properties. Total PAHs are reported as the sum of these compounds.
Extraction Methods
Water Extraction
For TPH a suitable volume was passed through Solid phase extraction on C18 Discs (3M) using hexane as the solvent (EPA method 1664 revision A, 2011). The extract was then passed through a
1 g anhydrous Na
2
SO
4
cartridge (Bond Elut) to remove residual water. 50 µl of nonane was added and samples concentrated down to 1 ml at 40˚C in a stream of N
2
prior injection on the Gas
Chromatograph-Mass Spectrometer (GC-MS). The extraction of the PAHs followed the same procedure but using dichloromethane (DCM) as a solvent, based on EPA Method 550.1 using application note 54 from SUPERLCO (1998) for C18 discs.
Soil Extraction
For the extraction of TPHs and PAHs associated with solids an Accelerated Solvent Extractor (ASE
200 Dionex) was used, the samples were mixed with equal parts of drying agent HYDROMATRIX and packed between washed sand and cellulose filters in the metal cells and placed in the ASE. For
TPHs a 50:50 solvent mixture of hexane:acetone at a pressure of 1500 psi was applied using N
2
gas at an oven a temperature of 200˚C (Dionex, 2011a). The solvent extract was passed through a Bond
Elut anhydrous Na
2
SO
4
1GM cartridges to remove any remaining water. The extract was filtered
(0.45 µm Chromacol filters) and 50 µl of nonane was added. Heat and N
2
were used to blow down the samples to 1 ml prior to injection in to the GC-MS. PAH extraction was similar but a 50:50 mixture of acetone:DCM was used at a pressure of 1500 psi and an oven temperature of 100˚C

(Dionex, 2011b). The extract was dried using anhydrous Na
2
SO
4
Bond Elut tubes. After concentration down to 1 ml there was a further clean up stage using Bond Elut silica gel cartridges
500 MG eluting the PAHs with 60:40 hexane:DCM (EPA Method 3630C), then 50 µl of nonane was added and samples blown down to 1 ml again before GC-MS analysis.
Recovery Rates
Diesel fuel was used to determine TPH recovery rates. Clean soil was spiked with diesel and taken through the whole method, giving recovery rates ranging from 116-122%. For water samples recovery tests were made using distilled water spiked with diesel, giving 60% recovery. For PAH recovery tests 4 deuterated compounds (dAcenapthene, dFluorene, dPyrene and dBenzo(a)anthracene) were spiked to clean soil and distilled water. For soils the recovery rates were between 91-94% and for water samples recovery rates ranged from 73-76%.
GC-MS
For the examination and identification of the traces of TPHs and PAHs a Varian 43GC and a Varian
21-MS Ion trap with an auto-sampler CP-8400 with a VF5 column was utilized. Ion trap conditions were: Trap 220˚C; Manifold; 80˚C; Transferline; 300˚C. The injection was a split/splitless injection of 1
 l.
Column conditions for TPH: Injector temperature: 280˚C; Initial temperature 50˚C for 1.5 minutes
Increase @ 15˚C/min to 300˚C hold for 10 minutes.
Column conditions for PAH: injector temperature 250˚C; initial temperature 60˚C hold for 1 minute; increase to 150˚C at rates of 30˚C/minute; then increase to 186˚C at rate of 6˚C/minute then increase to 280˚C at rate of 4˚C/minute; and hold for 20 minutes.
Results and Discussion
Case Study 1 – Commuter Access Road
Extensive monitoring was undertaken under normal and storm flow conditions at this site. Figure
2B shows that there was a progressive reduction of TSS road runoff over the first three hours of storm events, clearly indicating the phenomenon of the first flush .
Initial values of over 100 mg/l pose a significant pollution risk if directly discharged. VSS had a similar pattern of reducing over time, but tended to make up a larger proportion of the total SS as storms progressed. This suggests that proportion of organic matter, possibly the smaller size fraction with lower specific gravity, increases as storms progress and possibly flow velocity decreases. However, TPH concentrations
(Figure 2A) did not show such a clear pattern. The filtered and unfiltered lines show that the majority of TPHs were attached to particulate matter, but this shows variability over time with highest concentrations seen at 120 minutes after the start of the storms. The difference in the patterns of TSS and TPHs with time may indicate that the TPHs are attached to particular types of solid matter and that the bulk TSS parameter alone does not represent the process. Aryal Rupak et al. (2005) noted that the composition of TSS changed over storm duration as different solid size fractions had different transport characteristics. Future monitoring will examine the size fractions and types of solids during storm events to investigate this observed pattern. The concentrations of
TPHs can also be influenced by naturally occurring compounds that are extractable in hexane, so care needs to be taken in assigning all this to road derived pollution.

Figure 2. Mean concentrations of (A) TPH and (B) SS in road runoff with time after the start of a storm events (n=3) at Case Study Site 1.
Systematic monitoring of the swale showed that it significantly reduced the amount of TSS that reached the pond, by the end of the 100 m swale the TSS had reduced by 99% showing highly efficient retention of runoff solids. Figure 3 shows how the behaviour of 4 PAH’s in the swale varied depending on their physico-chemical properties. The heavier PAHs (phenantranene, fluoranthene and pyrene) show significant reductions in the swale. These compounds tend to be hydrophobic and probably are attached to the particles filtered out during passage through the swale. However, naphthalene tends to stay at similar levels during passage. Napthalene has a lower molecular weight and tends to be more water soluble, by several orders of magnitude, than the other
PAHs shown (e.g. approximately 30 mg/l for naphthalene compared to a range of 0.0003 – 3 mg/l for the others, according to Manoli and Samara (1999)). Therefore naphthalene is likely to be found mainly in the water phase and be less affected by sedimentation in the swale.

Figure 3. Concentrations of four PAHs in road runoff along the Swale at Case Study Site 1.
The solid bars show the inter-quartile ranges, the cross bars the median, the whisker lines the
95% confidence limits and astrices the outliers.
The accumulations of PAHs and TPHs in the soils and sediments in the system have also been monitored. Figure 3 shows the accumulations of TPHs in various parts of the systems by season.
Concentrations tend to be highest in preliminary parts of the system with a general trend for a progressive decline nthrough the swale and pond. Although the highest inlet concentrations were seen in winter and autumn, they decrease more rapidly in the system than in the summer.
Figure 4. Total TPH concentrations in the soil and sediments of the vegetated SuDs at Case
Study 1 by Season.
The TPHs cover a wide range of hydrocarbons with varying properties. Many are hydrophobic and are likely to partition onto settlable solids. Ongoing monitoring will assess the accumulations over time and how these relate to water phase concentrations. PAHs will also be examined in soils as these may be more approriate as indicator/tracer compounds for automobile emissions, especially in vegetated systems with componds that can interfear with more general hydrocarbon grouping measurements.

Case Study 2 – Trunk roads
Figure 5 compares the influent and effluent of two vegetated detention ponds on the trunk road network. Inflow and outflow hydrographs were not monitored so a mass balance is not possible.
The runoff here was also highly variable. It was not possible to sample runoff from the A34 SC systems directly and the influent values refer to runoff that has passed through the porous asphalt and sediment traps. This may account for the significantly lower BOD and COD than that entering the M27 NSC systems and demonstrate the effectiveness of the SC. However, the SC had higher
TSS and conductivity, the TSS could be affected by the microbial growth or other source in the open sediment traps and the conductivity winter icing events. Significant removals of BOD, COD,
TSS and conductivity were seen in both systems indicating the general effectiveness of vegetated detention systems. Moreover, dead vegetation and organic debris was seen in both systems which could influence these general water quality parameters.
Figure 5. Water quality in the influent and effluent of the vegetated detention ponds at the
Case Study 2 Trunk Road Sites (M27 and A34) (n>5). NS = Non Source Control; SC=Source
Control.
The TPHs in the influent and effluent of these ponds were generally similar in the order of 100-300
µg/l. However spikes were observed at both locations during one of the winter sampling locations.
The runoff on the M27 had a peak value of over 17,000 µg/l and the A34 over 5,950 µg/l. The ponds reduced these values by 67 and 58% respectively, but this does indicate very high levels of contamination being released. Further monitoring will investigate the relationship and correlation between hydrocarbon concentrations, the particulate fractions of TSS and factorsa ffecting the spikes.
Residential Developmet
Figure 6 shows that there was a little difference in water quality between SC and NSC catchments at the outlets. However, significantly higher levels of COD and SS were seen in the initial parts of the vegetated SuDs. This may be due to organic debris from the plants and soil, but generally the
PAHs (apart from basin 2) did not follow this pattern. There was progressive reduction of COD and significant reduction of TSS along the source control system. Nevertheless, no clear pattern of the
TPHs and PAHs exists in this case study. Further work and extended data analysis will investigate the patterns and treatment in more detail.

Figure 6. Water quality in the Source Control (SC) and Non Source Control (NS) catchments at the Residential Development Case Study 3
General Discussion
This survey has shown that road runoff poses a much higher risk than residential runoff to watercourses due to high levels of pollutants. However, there is a high degree of variation in pollutant levels due to seasonal factors and the first flush effect.
All the systems studied showed potential for removal of pollutants. Vegetated ponds and basins, such as the ones at the A34 Newbury Bypass and Cambourne, appeared to affected by increases of organic matter and TSS, due to dead plants ( Glyceria maxima) and organic debris. This means removals of specific water quality variables is not always progressive. The limitations of vegetated systems in achieving conventional discharge consents has been recognised in previous studies, particularly when plant residues are counted as SS (Williams et al ., 1996).
A range of treatment processes appear to be occurring, but for hydrophobic pollutants attachment to solids and subsequent sedimentations appear to predominate. The use of “bin” range parameters based on integration across GC retention times, such as TPH, appear to be of limited value due to interference and inclusion of natural organic compounds, PAHs are suggested as a better way of assessing fate and performance as they are more specific to automobile contamination. As PAHs have different physico-chemical properties individually they can help indicate the predominant treatment mechanisms (e.g. photolysis, volatilization, adsorption and sedimentation).
This study has 6-months more monitoring to undertake, followed by a multivariate data analysis phase. It is hoped that the final report will extend the understanding of the mechanisms affecting the fate of hydrocarbons in vegetated systems and lead to improvements of the design codes.
Conclusions

The purpose of this paper is to provide a better understanding of the differences between SC and
NSC systems, in terms of hydrocarbon treatment efficiency:

The behavior of the hydrocarbon constituents is related to their physical and chemical properties, in terms of susceptibility and solubility.

Variation in the runoff quality makes the comparison difficult. Seasonal variation and the dynamic nature of storm events imply that a large data set is required to make comparison.

Generally, SC systems appear to offer benefits and treatment trains show improvements in quality.

Some SC measures may give short term increases in certain pollutants.

Wetland systems have shown good removals of pollutants, but their dynamic nature means that this is not always progressive.
References
Aryal R. K., Furumai H., Nakajima F. and Boller M. 2005 Dynamic behaviour of fractional suspended solids and particle-bound polycyclic aromatic hydrocarbons in highway runoff. Water Research ,
39( 20), 5126-5134.
Bauman B. 1991 Research needs: motor fuel contaminated soils. Hydrocarbon Contaminated Soils. E.
J. Calevrese and P. T. Kostecki. Lewis Publishers: Chelsea, 41-56.
DIONEX. 2011a Accelerated Solvent Extraction (ASE) of Hydrocarbon Contaminants (BTEX, Diesel and TPH) in Soils. Application Note 324.
DIONEX. 2011b Extraction of PAHs from Environmental Samples by Accelerated Solvent Extraction
(ASE). Application Note 313.
Ellis J. B. 1991 Urban runoff quality in the UK: problems, prospects and procedures. Applied
Geography , 11 (3), 187 - 200.
EPA Method. 1664, Revision A: N-Hexane Extractable Material (HEM; Oil and Grease) and Silica
Gel Treated N-Hexane Extractable Material (SGT-HEM; Non-polar Material) by Extraction and
Gravimetry. United States Environmental Protection Agency. http://water.epa.gov/scitech/swguidance/methods/oil/1664.cfm
(accessed 27 January 2011)
EPA. Method 3630C Silica Gel Cleanup, Dec1996. Revision 3
EPA. Method 550.1 Determination of polycyclic aromatic hydrocarbons in drinking water by liquid solid extraction and HPLC with coupled ultraviolet and fluorescence detection 1990. Environmental monitoring systems laboratory office of research and development. U.S Environmental Protection
Agency. Cincinnati, Ohio.
Lerda, D. 2010. Polycyclic Aromatic Hydrocarbons Factsheet. 3 rd
Edition. Institute for Reference
Materials and Measurements. European Union. http://irmm.jrc.ec.europa.eu/EURLs/EURL_PAHs/about_pahs/Documents/JRC%2060146_Factsheet%
20PAH_3rd%20edition.pdf
(Accessed 4 th
September 2013)
Luch A. 2005 The carcinogenic effects of polycyclic aromatic hydrocarbons. Massachusetts Institute of Technology. Imperial College press . USA .
Makepeace D. K., Smith, D. W. and Stanley S. J. 1995 Urban stormwater quality: Summary of contaminant data. Critical Reviews in Environmental Science and Technology , 25 (2), 93-139.
Manoli E., Samara C. 1999 Polycyclic aromatic hydrocarbons in the natural waters:source, occurrence and analysis. Trends in Analytical Chemistry, 6 , 417-428
Pontier H., Williams J. B. and May E. 2003 Behaviour of metals associated with sediments in a wetland based system for road runoff control. Water Science and Technology , 48 (5), 291-298
Royal Haskoning 2011 Lamb Drove SuDS Showcase Project Cambourne . Second Interim Monitoring
Report. Cambridgeshire County Council.
Seekkubadu S., Mant C., Williams J., Stewart P., May E., Aldridge S. and Brown 2009 An Assessment of the Performance of Sustainable Drainage Systems Integrated into the Design of a Major
Development Area. SUDSnet National Conference , Coventry UK. Proceedings available at:

http://www.sudsnet.abertay.ac.uk/Nov%202009/powerpoint%20pdfs/session%201/Seekkubadu_As sessment%20Of%20Performance%20of%20SUDS.pdf
SUPERLCO 1998 Extract Polynuclear Aromatic Hydrocarbons from Water, Using Solid Phase
Extraction Disks. Application note 54 . Sigma-Aldrich. http://www.sigmaaldrich.com/etc/medialib/docs/Supelco/Application_Notes/4600.Par.0001.File.tm
p/4600.pdf (accessed September 13 2011)
Viklander M. 1999 Substances in Urban Snow. A comparison of the contamination of snow in different parts of the city of Lulea, Sweden. Water Air & Soil Pollution , 114 (3), 377-394.
Williams J.B., Bahgat M. May E. Butler J.E. and Ford, M.G. 1996. Mineralisation and pathogen removal in GBH wetlands for wastewater treatment. Water Science and Technology. 32 (3), 49-58