Polycyclic aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB) in urban soils of Greater London, UK

Surface soils from a 19 km 2 area in east London, UK were analysed for polycyclic aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB) ( n = 76). P 16 PAH ranged from 4 to 67 mg/kg (mean, 18 mg/ kg) and P 50 PAH ranged from 6 to 88 mg/kg (mean, 25 mg/kg). P 7 PCB ranged from 1 to 750 l g/kg (mean, 22 l g/kg) and P tri-hepta PCB ranged 9 to 2600 l g/kg (mean, 120 l g/kg). Compared to other international cities concentrations were similar for PAH but higher for PCB. Normal background concentra- tions (NBC) were calculated and compared to risk-based human health generic assessment criteria (GAC). Benzo[ a ]pyrene NBC for urban (6.9 mg/kg), semi-urban (4.4 mg/kg) and urban + semi urban (6 mg/kg) domains exceed residential (1 mg/kg) and allotment (2.2 mg/kg) LQM/CIEH GAC (at 6% SOM) and the Indeno[ 1,2,3-cd ]pyrene NBC for urban (6.8 mg/kg) and urban + semi-urban (5.2 mg/kg) domains exceed the residential (4.2 mg/kg) LQM/CIEH GAC (at 6% SOM). Capsule Abstract: Normal background concentrations of polycyclic aromatic hydrocarbons and poly- chlorinated biphenyls are elevated in east London soils and in some cases exceed regulatory assessment criteria. (cid:2) 2014 The Authors. Published by Elsevier Ltd. ThisisanopenaccessarticleundertheCCBYlicense(http:// creativecommons.org/licenses/by/3.0/).


Introduction
London is Europe's largest city by population and one of the top five richest by GDP in the world. The capital has a long industrial history which has contributed to sources of organic pollutants in soil such as polycyclic aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB). These compounds are strongly hydrophobic and lipophilic as well as having a propensity to accumulate in plant and animal tissues. PAH are present in soil as a result of natural and anthropogenic processes which produce both point source and diffuse emissions. PCB are not found naturally in the environment and are mainly present in soils due to point source releases (Creaser et al., 2007a). However, studies have also reported that PCB are prone to long-range atmospheric transport and deposition, controlled by temperature induced volatilisation (Jones and De Voogt, 1998;United Nations -Economic Commission for Europe, 1994). PAH and PCB tend to remain in soil for extended periods of time due to their comparatively long half-life, which can range from years to decades (Jones and De Voogt, 1998;Environment Agency, 2008). These characteristics are important since both PAH and PCB have been implicated in adverse effects on human and environmental health (Environment Agency, 2009a;Nathanail et al., 2009;ATSDR, 2011).
Previous studies published by the Environment Agency for England and Wales reported elevated concentrations of PAH and PCB in UK urban soil (Creaser et al., 2007a(Creaser et al., , 2007b. PAH sources were thought to be mixed and predominantly include those from traffic and domestic combustion, while PCBs were likely to have arisen from atmospheric deposition as a consequence of original point source releases. Similar findings have been reported for cities such as: Glasgow, Scotland (Morillo et al., 2007); Birmingham, England (Smith et al., 1995); Boston, US (Bradley et al., 1994); and Xiamen in China (Xing et al., 2005). In addition, many geo-environmental site investigations take place each year for regulatory purposes that give rise to a large volume of soil contamination information that is not published in peer reviewed journals. Notwithstanding, there is paucity of comprehensive background data on soil PAH and PCB concentrations for London. Our study bridges this knowledge gap by systematically sampling an area of 19 km 2 in the Greater London Authority (GLA) (n = 76) (Fig. 1). Moreover, this study recognises that the source of PAH and PCB is especially important were concentrations are elevated, which is why where http Analyses of each sample comprised 33 PAH (parent and alkylated), 7 PCB congeners, 5 PCB homologous series and total organic carbon (TOC). TOC was selected as a soil property that is known to affect soil-compound sorption interactions (Reid et al., 2000). PAH and PCB concentrations are evaluated statistically, including the use of a recently proposed method for deriving normal background concentrations (NBC) of contaminants in soil (Ander et al., 2013). Concentration data are compared to those recorded in similar urban areas to identify whether there are any obvious differences for London. A discussion of the potential sources of soil PAH is presented to characterise the likely contribution of point and diffuse emissions. The PCB congener profile is also discussed along with findings from other studies. NBCs are calculated for 16 parent PAH, P 16 PAH, P 50 PAH, 7 PCB congeners, P 7 PCB, 5 homologous PCB series and P tri-hepta PCB. This study also compares NBCs with selected human health generic assessment criteria (GAC) to put them in a of risk-based land management context.

Soil sampling
Surface soil samples (n = 76) were collected between 4th and 18th April 2009 from a 19 km 2 area covering Abbey Wood, Thamesmead, Erith, Belvedere and Jenningtree Point in the GLA of London, UK (Fig. 1). Four sites were selected every kilometre square and the position of each location recorded using a handheld Garmin GPS device (accuracy ±5 m). At each site soil samples were recovered from four corners and the centre of a 20 m by 20 m square grid at a depth of 5-20 cm using a hand-held Dutch auger. The samples were combined, sealed in a Rilsan bag, stored at $4°C for 1-3 h and then frozen at À18°C. Sampling locations covered the following eleven land-use types: commercial and residential, domestic garden, golf course, industrial, park, playground, road verge, rough grazing, school, urban open space and water treatment works.

Laboratory analysis
All soils were freeze-dried, disaggregated and passed through a brass sieve with an aperture of 2 mm before being agate ball milled (Vane et al., 2007b). This method was chosen as a statistically robust method for high precision and accuracy PAH quantification (Beriro et al., 2014).
Concentrations of PCB were measured using GC/MS. The MS was operated in selected ion mode (m/z 71, 220, 256, 258, 292, 326, 328, 360, 362, 394, 396) with a scan time of 0.5 s. Congener profile distributions for tri to hepta-chlorinated biphenyls were determined using 5 separate individual relative response factors for each congener group. A standard consisting of 2 Â 5 individual PCBs based on first and last eluting congener was used to define the 'retention time window' for each group. The sum of these 5 groups is P tri-hepta PCB and P 7 PCB comprised congeners 28, 52, 101, 118, 153, 138, 180.

Total organic carbon
Total organic carbon (TOC) content was determined using an Elementar VarioMax C, N analyser after acidification with HCl (50% v/v) to remove carbonate. The limits of quantification for a typical 300 mg sample were 0.18%. Details of this method have been described previously (Vane et al., 2007a).

Data evaluation
The data were grouped into urban and semi-urban domains. The urban domain (n = 42) was defined by land-use, i.e. developed land and the semi-urban domain (n = 34), i.e. undeveloped land were assigned using information presented in 1953-present Ordnance Survey maps. Differences in PAH and PCB concentrations between the two domains were tested using Levene's and Bartlett's test for homogeneity of variance followed by a Student's t-test to decide which domains should be used to derive NBC and compare with GAC.
An NBC is the upper confidence limit of the 95th percentile, providing a conservative estimate of the background concentration over a given area accounting for geogenic and diffuse sources of pollution (Cave et al., 2012;Ander et al., 2013). NBCs have recently been introduced into the contaminated land regime in England and Wales under Part 2a of the Environmental Protection Act 1990 (EPA) (as amended) by secondary statutory guidance in each country (DEFRA, 2012). The regime supports the derivation of normal levels of contaminants as a way of identifying concentrations that represent low rather than unacceptably high levels of risk. NBCs should not be used in support of the National Planning Policy Framework (NPPF), which states that the developer or site owner needs to ensure the land is safe for its intended use. (Paragraph 121: CLG, 2012). The data underpinning the NBCs could contribute to one of many lines of evidence that may be used to assist with risk-based land management under these and other regimes (Environment Agency, 2004).
GAC are scientific risk-based conservative estimates of chemical concentrations in soil, below which the level of risk posed under specific exposure scenarios to people may be negligible or minimal (Environment Agency, 2009d). A GAC is the concentration of contaminant that would give rise to an estimated Average Daily Exposure that is equal to the Health Criteria Value for that contaminant (Environment Agency, 2009c). Far from representing thresholds of unacceptable risk GAC indicate concentrations below which the risk is negligible or minimal and whose exceedance should trigger a detailed quantitative risk assessment. The level of unacceptable risk varies among different legal regimes. For Part 2a of the EPA it is the significant possibility of significant harm (SPOSH) such as death, life threatening diseases, other diseases likely to have serious impacts on health, serious injury, birth defects and impairment of reproductive function (DEFRA, 2012). The NPPF states that unacceptable risks should be prevented by ensuring land is appropriate for its location, it is safe and after remediation not capable of being defined as Contaminated Land under Part 2a of the EPA (CLG, 2012). LQM/CIEH GAC have been published for 16 PAH for three land-use types (residential with plant uptake, allotment and commercial) and three soil organic matter (SOM) contents (1%, 2.5% and 6%) (Nathanail et al., 2009). There is an Environment Agency derived GAC, called a soil guideline value (SGV), for dioxin like PCB (Environment Agency, 2009b). For this study the calculated NBCs are compared with published LQM/CIEH GAC for 16 PAH and the SGV for PCB.

Data summary
The spatial distribution of P 16 PAH across the study area is summarised in Fig. 2

Comparison of data with other urban areas
Other studies of PAH in urban soils report similar concentrations to those presented in this research. A UK study reported P 22 PAH concentrations in the range of <1 to 55 mg/kg (n = 366) (Creaser et al., 2007c). Similarly, Glasgow has soils with P 15 PAH concentrations in the range of 1-52 mg/kg (n = 15) (Morillo et al., 2007;Vane et al., 2011). In contrast, lower amounts of PAH have been observed in urban soils from Wales and the Midlands/Birmingham (Butler et al., 1984;Smith et al., 2001;Wild and Jones, 1995). Comparable amounts of PAH have also been recorded outside the UK: e.g. New Orleans, USA (<1 to 41 mg/kg; mean 6 mg/ kg), Boston, US (highest conc. 360 mg/kg; mean 19 mg/kg), Moscow, Russia (<1 to 1 mg/kg), Torino, Italy (1-24 mg/kg; mean 2 mg/kg), Switzerland (highest conc. 8 mg/kg) (Bradley et al., 1994;Desaules et al., 2008;Morillo et al., 2007;Wang et al., 2004;Wilcke et al., 2005). The PAH data in this study are far lower than those encountered for soils which have suffered specific point sources releases of pollution. For example, a former tar works site in Newcastle upon-Tyne, UK, yielded surface soil PAH concentrations ranging from 66 to 1400 mg/kg (Lorenzi et al., 2010). This suggests that the study data presented herein for an area of London represents diffuse pollution sources.
3.1.3. Source apportionment of PAH PAH distribution data for all samples showed minor contributions of low molecular weight PAH (2-3 ring) such as naphthalene, methylnaphthalenes and acenapthylene (Fig. 3). This may be due to their increased susceptibility to weathering by oxidation and increased losses due their higher aqueous solubility (Marynowski et al., 2011). In contrast, major contributions from higher molecular weight PAH (4-6 ring) including fluoranthene, pyrene and benzo[a]pyrene were observed (Fig. 3), although it is acknowledged that high molecular weight compounds may also experience losses due to weathering. The sole exception to this finding is phenanthrene, the most thermodynamically stable of the three-ringed parent PAH, which was also present at relatively high amounts in all London soils. Overall, the soils were dominated by 4-6 ring PAH which generally signifies PAH from mainly fossil fuel combustion (McCready et al., 2000;Vane et al., 2007a it should be borne in mind that weathering processes such as leaching (solubilisation), evaporation/volatilisation and biodegradation can modify PAH distribution (Wilcke, 2000). For example, it has been shown that parent PAH and low molecular weight PAH are more susceptible to microbial degradation than alkylated PAH and high molecular weight compounds (Douglas et al., 1996). On balance, the distribution of parent PAH suggests an environmentally modified pyrolytic source. Alkylated PAH compounds e.g. 1-methylnaphthalene (1-MN) and 2-methylnaphthalene (2-MN) are important constituents of diesel fuels and are also found in wood smoke, burnt peatlands and unburnt coal in minor amounts (Conde et al., 2005;Yang and Lu, 2005;Achten and Hofmann, 2009;Vane et al., 2013). The low concentrations of 1-MN and 2-MN relative to other PAH suggests that non-combusted petroleum sources (e.g. fugitive diesel spills) are an unlikely source. The ratio of parent to alkylated PAH has been widely utilised as a simple yet diagnostic tool in environmental source apportionment studies (Ou et al., 2004;White et al., 2005;Stout and Emsbo-Mattingly, 2008). These have shown that petrogenic sources produce greater amounts of alkylated PAH relative to parent PAH, whereas pyrogenic sources yield parent PAH in greater abundance than their alkylated counterparts. Accordingly, crude oil and coal show parent/alkyl PAH values of approximately 0.2 whereas those from coal tar and creosote are about 3. The ratios of the C1-benzo[a]antracene + C1-chrysene/ benzo[a]antracene + chrysene showed a mean of 0.11 which is commensurate with pyrogenic sources.
A complementary approach to characterising PAH source (petrogenic vs. pyrolitic) in soil and sediments is to use PAH isomeric ratios: phenanthrene/anthracene (Phen/Anth) compared with fluroanthene/pyrene (Fanth/Pyr) (Budzinski et al., 1997), benzo[a]anthracene/(benzo[a]anthracene + chrysene) (BaA/(BaA + Chrys)) compared with fluoranthene/(fluoranthene + pyrene) (Fl/(Fl + Pyr)) (Yunker et al., 2002) (Figs. 4 and 5). Isomer ratios, are to a certain extent, able to distinguish PAH from petrogenic as compared to incomplete combustion/pyrolitic sources because phenanthrene and fluoranthene are more thermodynamically stable than anthracene and pyrene isomeric counterparts (Vane et al., 2013). Consequently, PAH formed at low temperature (e.g. slow thermal maturation of organic matter in petroleum) produce Phen/Anth ratios of >10 and Fanth/Pyr of <1 whereas PAH formed at higher temperatures (e.g. partial combustion of fossil fuels) provide Phen/Anth ratios of <10 and Fanth/Pyr of > 1. Inspection of the Phen/Anth to Fanth/Pyr isomeric plot shows that all of the soils plot within the pyrolytic quadrant, supporting the assertion that the PAH were mainly derived from organic matter that has been subject to high temperature processes (Fig. 4). The second plot using BaA/BaA + Chrys and Fanth/Fanth + Pyr also suggests that PAH are from pyrogenic sources such as petroleum, grass, wood and coal combustion (Fig. 5). Atmospheric PAH ratios measured in London between 1991 and 2008 show BaA/BaA + Chrys to be much lower and Fanth/Fanth + Pyr to be consistent with our findings (Katsoyiannis et al., 2011). These results suggest that soil PAH is a mixed source comprising atmospheric pyrogenic PAH as well as additional direct pyrogenic inputs from other sources.
Traffic exhaust can be a major source of pyrolytic PAH producing a relatively high concentration of benzo[g,h,i]perylene (BghiP) compared to benzo[a]pyrene. Accordingly, it has been reported that traffic sources can be inferred when the ratio of BghiP to BaP exceeds 3.14 (Creaser et al., 2007b). It has recently been shown that estuarine sediments deposited in close proximity to major carriageways can yielded BghiP/BaP values of >9, demonstrating that traffic sourced PAH are conserved in the sedimentary record (Creaser et al., 2007c;Vane et al., 2011). For the current study none of the soils gave values greater than 3.14, and the range of values found (0.6-1.1, with a mean of 0.8) were rather invariant -perhaps supporting a more recent study that questions the validity of these ratios (Katsoyiannis et al., 2011). Overall, the ratios suggest mixed pyrolytic sources that cannot be solely ascribed to traffic sources. Other potential sources include combustion in power stations, combustion by industry or possibly domestic and or natural fires (Wilcke, 2000(Wilcke, , 2007. An explanation for the dominance of high molecular weight combustion derived PAH could be that they originated from the now decommissioned Belvedere oil burning power station which operated from 1962 to mid-1980s near Jenningtree Point (Fig. 2). However, inspection of total PAH concentrations for the study area revealed no obvious evidence of a PAH pollution plume gradient as would be expected from a single dominant point source (Fig. 2).
As has been shown, isomeric ratio plots can be useful in providing possible explanations of the processes that result in the formation of PAH present in the study samples. However, the method alone is not definitive and has even been reported to be a source of uncertainty (Katsoyiannis et al., 2011(Katsoyiannis et al., , 2007. Alternative exploratory data analysis methods may be used as additional lines-ofevidence, e.g. Principal Components Analysis (PCA) (Hu et al., 2013;Carlon et al., 2001;Dahle et al., 2003;Stout et al., 2001;Wang and Fingas, 2003;Thavamani et al., 2012) and regression analysis (Schauer et al., 1996(Schauer et al., , 2007. Moreover, analytical methods such as compound specific isotope analysis are reported to be reliable determinants of the source of PAH (McRae et al., 2000;Snape et al., 2001;Górka et al., 2014).
The relationship between TOC and P 16 PAH was examined using Pearson's correlation coefficient. Results show a weak  positive relationship (0.02) whereas in sediments correlations are typically stronger (Vane et al., 2009). Weak correlations between TOC and PAH are common in soils (Bucheli et al., 2004;Heywood et al., 2006;Ribes et al., 2003), though on occasions some statistically significant relationships have been encountered (Cousins et al., 1997). Weak correlations have been ascribed to the operation of a non-equilibrium situation (Bucheli et al., 2004). When PAH from combustion processes are released to the atmosphere they do so in association with soot particles. Once in the air the more volatile, low molecular weight PAH i.e. the 2-ringed PAHs (naphthalene, 2-methylnaphthalene) and the 3-ringed PAHs (acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene), partition from the soot particles more rapidly into the vapour phase than do the less volatile high molecular weight PAH i.  (Bucheli et al., 2004). Upon deposition to the soil, the low molecular weight PAHs, being in the vapour phase, are free to partition into, and equilibrate more readily with, the soil TOC than the more strongly particle bound high molecular weight PAH. Accordingly, for the London soils, where high molecular weight PAH predominate, a lack of correlation of TOC with total PAH concentration may be anticipated. This finding is indicative of a diffuse source of pollution, further substantiating the likelihood that the concentrations recorded are typical of background concentrations rather than a dominant single source.  (USEPA, 1987). NBCs for the urban domain are higher than the semi-urban domain. NBCs for the urban + semi-domain (all data) account for these differences by representing an approximate average of the two domains. The urban + semi-urban NBCs are proposed as suitably representative of the whole study area given the mostly urban setting of this part of London.
NBCs for 16 parent PAH were compared with published residential, allotment and commercial land-use LQM/CIEH GACs (6% SOM) (Nathanail et al., 2009) (Table 1). The 6% SOM GACs were chosen on the basis of the TOC data. TOC was recorded between 3% and 21% with a mean of 7%. Using a SOM conversion factor of 1.74 (Pribyl, 2010), calculated conservative values ranging from 5% to 37% with a mean of 12%. The benzo[a]pyrene NBC for all domains exceeds the LQM/CIEH GAC for residential and allotment land-uses (Table 1). The Indeno[1,2,3-cd]pyrene NBC for the urban domain exceeds the residential LQM/CIEH GAC (Table 1). Fig. 6 shows the relationship between the benzo[a]pyrene NBC and LQM/CIEH GAC for each domain. These results suggest that if data were to be reviewed in the context of the regulatory requirements of the NPPF for proposed residential development, further investigation would probably be needed to show that sites are safe and suitable for use. Similarly, if sites within the study area were to be examined by the local authority in the context of Part 2a of the EPA then further investigation may also be necessary.

Polychlorinated biphenyl concentrations
3.2.1. Data summary PCB analyses comprised 7 congeners (PCB 28, PCB 52, PCB 101, PBB 118, PCB 138, PCB 153, PCB 180), 5 homologous series (tri, tetra, penta, hexa, hepta) and P tri-hepta PCB. Quality control procedure resulted in removing 2 samples from the dataset, leaving 74 samples for evaluation. The individual congeners selected are the ICES-7 defined during the International Conference of the Environment of the Sea based on their prominence and persistence (Environment Agency, 2003). The spatial distribution of P tri-hepta PCB across the study area is summarised in a diagram (Fig. 7).
P tri-hepta PCB concentrations ranged from 9 to 2642 lg/kg with a mean of 123 lg/kg.  Fig. 8 and Fig. 9). Descriptive statistics for all PCB data are presented in the Supporting Information (SI: Table 2).

Comparison of data with other urban areas
Comparing recorded PCB concentrations with other studies is generally complicated by differences between the selected congeners and associated statistics. However, ICES-7 PCB data for urban and suburban soils in the UK, Russia and France show that median values are higher than another UK study (1.77 lg/kg) (2007a) ( Table 2). Mean values for P 7 ICES-7 PCB for this study (21.4 lg/kg) are comparable with suburban data for France (21.5 lg/kg) (Motelay-Massei et al., 2004) but slightly higher than data recorded in Moscow (9.9 lg/kg) (Wilcke et al., 2006) (Table 2).
A review of 12 studies were not comparable because of differences between congeners analysed (Table 3). As expected, the concentrations recorded in this study are considerably lower than point source releases associated with the manufacture and industrial uses of PCB. For example, a former telecommunications manufacturing site in south west England showed mean concentrations of 120,000 lg/kg for P 7 PCB (Norris et al., 1999) and a PCB manufacturing facility in Sauget, US showed concentrations ranging from 150 to 27,000 lg/kg total PCB (congener types not identified) distributed within a 1500 m radius of the site (Stratton, 1976).

PCB congener distribution
A summary of both individual congeners and homologous series concentrations shows that London soil samples are enriched in penta-, hexa-, and hepta-chlorobiphenyls (CB) (87% in mass) (Figs. 8 and 9). This finding agrees with previous reports of 85% (which included mono-, di-, and octa-CB) in urban soils from Cedar Rapids, USA (Martinez et al., 2012). Similar profiles have also been identified in Glasgow, Scotland in urban soils where a higher proportion of penta-CB was also observed (Cachada et al., 2009). However, significant contribution of tri-CB reported in the Tianjin region of China was attributed to the recent localised production of this congener group (Li et al., 2010). Losses of tri-and tetra-CB congeners (via volatilisation) and the relative dominance of the heavier chlorinated biphenyls over time (25 years) have also been reported . Temporal change is the likely explanation of the congener pattern in our study where tri-CB constitutes a minor component (Fig. 9). This is probably because a large proportion of these congeners are lost via volatilisation whilst in-situ (Harrad et al., 1994;Johnson et al., 2006). Similarly, the depletion of lighter PCB may be associated with urban soils in comparison with rural sites, as observed in France (Motelay-Massei et al., 2004). Using a   . 8). The land-use ascribed to each sample is either industrial or commercial which together with the concentrations suggests point source releases. Prior to calculating PCB NBCs, these 4 samples were filtered out of the dataset to reduce bias. It is suggested that point sources are commonly due to leakages from sealed and open systems, disposal and accidental spills and leaks (Breivik et al., 2002). Such sources are likely to be the main contributor to atmospheric PCB and therefore an important contributor to diffuse pollution sources (Creaser et al., 2007a). Concentrations of P tri-penta PCB are similar between the urban domain and semi-urban domain (Fig. 7). Data were applied to Levenne's and Bartlett's test for homogeneity of variance (p P 0.05). The null hypothesis was accepted for Levene's test, suggesting the data are of equal variance. The data were then applied to the Student's t-test. Results suggest there is no difference between the mean concentrations of P tri-penta PCB for the two domains (p P 0.01). Only PCB concentrations for the urban + semi-urban domain (n = 71) were used to calculate NBCs for individual congeners and homologous series (Table 4). A comparison of the calculated NBCs with the SGV for dioxins, furans and dioxin-like PCBs in soil is limited to PCB 118 only. This is because the SGV is designed to be compared with the sum of dioxins, furans and dioxin-like PCBs in soil. For this study, PCB 118 was the only dioxin-like PCB analysed (PCB 118). The calculated NBC of 4.4 lg/kg for the filtered urban + semi-urban domain does not exceed the residential and allotment SGV (8 lg/kg) or the commercial SGV (240 lg/kg) (Environment Agency, 2009b). The PCB data alone would probably not immediately call for regulatory measures. However, there are many persistent compounds covered by the SGV for dioxins, furans and dioxin-like PCBs that were not quantified as part of this study.

Conclusions
This study reports the concentrations of PAH and PCB in soil samples taken from a 19 km 2 area in the east of London, UK. Key quantitative findings include: (1) P 16 PAH ranged from 4 to 67 mg/kg with a mean of 18 mg/kg; (2) P 50 PAH ranged from 6 to 88 mg/kg with a mean of 25 mg/kg; (3) P 7 PCB ranged from 1 to 750 lg/kg with a mean of 22 lg/kg; (4) P tri-hepta PCB concentrations ranged 9-2600 lg/kg with a mean of 120 lg/kg. PAH concentrations were comparable with other cities, whereas PCB concentrations were higher. Isomeric PAH ratios suggest multiple pyrogenic sources, inferred to be representative of diffuse pollution. PCB congener profiles were generally dominated by Ppenta-CB congeners. Selected NBCs were compared to human health GACs. The benzo[a]pyrene NBC for the urban (6.9 mg/kg), semi-urban (4.4 mg/kg) and urban + semi urban (6 mg/kg) domains exceeded the residential (1 mg/kg) and allotment (2.2 mg/kg) LQM/CIEH GAC (at 6% SOM) and that the Indeno[1,2,3-cd]pyrene NBC for urban (6.8 mg/kg) and urban + semi-urban (5.2 mg/kg) domains exceeded the residential (4.2 mg/kg) LQM/CIEH GAC (at 6% SOM). The NBC for benzo[a]pyrene and Indeno[1,2,3-cd]pyrene suggest that further investigation would be required for future risk-based land management under Part 2a of the EPA for residential or allotment land-use within the study area. Although not directly relevant, NBC exceedances of the GAC suggest these substances should be considered where a change in land-use is being considered. The PCB 118 NBC (a dioxin-like PCB) did not exceed the SGV, however, further monitoring of dioxins, furans and other dioxin-like PCB within the study area would be prudent to inform ongoing risk-based land management decision-making.  Wilcke et al. (2006) Moscow Urban and peri-urban 11 17 13.9 9.9 3.1-42 a All values in parentheses are standard deviations of the mean P PCB within each study; Urban 1 -urban public spaces, away from known contamination.