Assessment of heavy metal contamination of road dusts from industrial areas of Hyderabad, India

Ramavati Mathur+, V. Balaram, M. Satyanarayanan, S.S. Sawant
CSIR-National Geophysical Research Institute, Hyderabad 500007. India.
+E-mail: mathurramavati@yahoo.com

Abstract
Road dust in industrial areas carries high levels of toxic heavy metals. Exposure to such polluted dust significantly affects the health of people residing in these areas, which is of major concern. The present study was taken up with an aim to highlight the magnitude and potential sources of accumulation of heavy metals in 32 road dust samples collected from six industrial areas of Hyderabad. Acid digested sample solutions were analyzed by ICP-MS for Cu, Zn, Cr, Co, Pb, Ni, V, Zr, Ce, Y, Hf. Validation of analytical data was done by analysis of CRMs SO-1, SO-2 and JG-2. The road dusts exhibit significantly high mean metal levels which are much above their crustal abundances, e.g. Zr (526.9µg/g), Zn (314.3µg/g) and Pb (166.8µg/g). The relative ordering of mean metal contents is Zr>Zn>Pb>Cr>Ce>Cu>V>Ni>Y>Co>Hf. Elevated pollution indices (Igeo, EF, Cif and Cdeg) reveal that the road dusts are pollution impacted showing varying degree of heavy metal contamination. Strong positive correlations exhibited by metal pairs Cu-Zn, Cr-Ni, Ce-V, Y-Ce, Hf-Zr imply their origin from common anthropogenic sources. Principal component analysis grouped the metals according to the sources which contributed to their accumulation. The present study confirms to an intensive anthropogenic impact on the accumulation of heavy metals in the studied road dusts attributable mainly to strong influences of vehicular and industrial activity and partly to domestic and natural processes. The results obtained imply the need for further investigations to assess their ecological implications and human health risks.

Key Words  Road dusts, heavy metal pollution, anthropogenic activity, contamination assessment, statistical analysis.

Introduction
Rapid urbanization, unregulated industrialization and fast expanding transport systems have resulted in contamination of the urban environment by several pollutants. Road dust that accumulates along the impervious roadways of the urban environment acts as a sink for these pollutants and gets enriched in many harmful components, especially heavy metals that originate from natural geochemical processes and a vast array of anthropogenic sources. The latter are mainly associated with vehicular traffic, industrial emissions, domestic and commercial activity (Brown et al. 1990; Pacyna et al. 2007). Heavy metals are known for their toxicity and have been coined as “Chemical time bombs” (Stigliani et al. 1991; Alloway and Ayres 1997). Even some of the biologically essential heavy metals can cause toxic effects when their concentrations exceed certain thresholds. Ingestion, inhalation or dermal contact of heavy metal laden polluted dust from roadways has been implicated as detrimental to human health. Exposure to heavy metals leads to kidney and liver damage, respiratory, cardiovascular and skin disorders, developmental retardation, failure of the central nervous system, different types of cancers (Fergusson 1990; WHO 2007). Accumulation of these heavy metals in road dusts is one of the major processes through which they enter the soils, from where they get into the living tissues of plants and animals and subsequently enter the food chain of humans affecting their health and well being (Adriano 2001).
Heavy metal pollution of urban dusts is one of the fastest growing types of environmental pollution and it has raised serious concerns in the last decades. Extensive research carried out in different cities of the developed countries has provided important information on the content, source, accumulation and distribution patterns, mobility and interaction of heavy metals in different segments of the environment (Charlesworth et al. 2003; Wei et al. 2010; Shi et al. 2013). However, such information from developing countries like India is scarce. Previous studies in India have focused mainly on the heavy metal contamination of urban soils, aerosols, sediments and water (Govil et al. 2008; Krishna et al. 2009) and very few pertain to toxic metal contamination of road dusts, especially data on the content and distribution of heavy metals in road dusts from industrial areas is very limited (Banerjee 2003; Singh 2011; Reddy et al. 2012; Balaram et al. 2013). Unregulated industrial activity adversely affects the quality of air, soil and ground water resulting in severe environmental pollution. Environmental as well as occupational exposure of people residing in these industrial areas results in a considerable increase of heavy metals in their tissues and blood (Brown et al. 1990; Ferguson 1990). Extensive development of industrial establishments in and around the city of Hyderabad in the past decades has become a subject of concern because the industrial activity has led to contamination of the environment with several pollutants which are posing a serious threat not only to people residing within these industrial areas but also to neighboring human settlements due to atmospheric transport of the pollutants. These areas are also polluted due to vehicular emissions because most of the vehicles plying on these roads, especially the transport vehicles, do not fulfill established fitness criteria that are necessary for a clean environment.
In consideration of the harmful health impact on the workers and residents of the industrial areas there is an increasing need to have an in-depth understanding of its environmental quality. In view of this the present work, comprising of the study of road dust samples collected from six severely polluted industrial areas in and around the city of Hyderabad, was taken up. The main objectives of this work are: (i) To assess the concentration levels and distribution of eleven heavy metals in the road dust samples (ii) To identify the different sources of heavy metals to the dusts (iii) To evaluate the contribution of natural and anthropogenic sources and understand the association of the heavy metals using statistical methods (iv) To assess the magnitude of heavy metal contamination of the road dusts using various indices like Index of Geoaccumulation (Igeo), Enrichment Factor (EF), Contamination Factor (Cif) and Degree of Contamination (Cdeg) (v) To provide comprehensive information about the heavy metal pollution of the study area which can be used by the authorities to exactly locate contaminated sites for remediation, establish reliable protection approaches and plan strategies to achieve better urban environmental quality.
Materials and Methods
Study Area
Hyderabad is at present the joint capital of southern Indian states of Telangana and Andhra Pradesh. Located at 17◦ 23’N and 78◦ 28’E, it is situated on the Deccan Plateau, at an average altitude of 542 meters. It is one of the largest metropolitan areas of India, spreading over 625 Sq.Kms. It has a population of about seven million (2003 census) which is expected to be double by 2021. The climate of the region is semi arid with three distinct seasons i.e. summer, rainy and winter. The temperature varies from 20◦ to 39◦ C and wind speed ranges from 0.2 to 9.5 KMPH. Average annual rainfall is ~700mm which is mainly due to the southwest monsoon. Hyderabad lies on a predominantly sloping terrain of grey and pink granites of Achaean age. They are not well exposed but appear as small hills at places (Balakrishna and Raghava Rao 1961; Gnaneshwar and Sitaramayya 1998). The soil cover is mainly derived from weathered granites. In recent decades the city has rapidly expanded and now has densely packed buildings on either side of narrow roads. With an increase of its population there was also a tremendous increase in motorized transport resulting in frequent road congestions and a sharp increase of vehicular emissions. Many industrial areas, comprising of both multinational as well as small scale industries, have come up in and around the city. These industrial areas comprise of units manufacturing chemicals, pharmaceuticals, pesticides, paints, petrochemicals, batteries, textiles, paper, plastic and rubber products, different types of metallic goods, alloys, automobile spares and machine tools. Besides this, the industrial areas also house scrap dealers, incineration plants, electroplating, tanning, welding, galvanizing and dye molding workshops and automobile servicing centers. These factories have operated without strict environmental regulations for many years. With the increased industrial activity many residential areas and commercial complexes have established within the industrial zones as a result of which many industrial areas that were located at the outskirts have now become integral part of the city. This has led to considerably increased industrial and municipal waste discharge and other pollutant emissions in the area. As a consequence of rapid urbanization and increased economic and uncontrolled industrial development Hyderabad is experiencing a declining trend of its environmental quality at an alarming rate. Air quality studies have been carried out by different regulatory bodies and the central and state pollution control boards. In the past years research has also been carried out to monitor the levels, distribution and behavior of different pollutants especially heavy metals in soils, sediments and water bodies of the industrial areas of Hyderabad (Govil et al. 2008; Krishna et al. 2009). However environmental research had not been focused on road dusts of these industrial areas.

Sample collection and Analysis
A total of 32 road dust samples were collected from six industrial areas in and around the city of Hyderabad namely: Balanagar, Sanathnagar, Jeedimetla, Patancheru, Nacharam and Katedan (Fig.1), details of which are given in Table 1. At each sampling site 500g to 1kg (fresh weight) of composite sample was collected by combining five sub-samples taken along the road edges on either side. The sub-samples were thoroughly mixed to obtain the composite sample. All samples were taken wearing clean latex gloves and using a plastic scoop and dustpan. Each sample was placed in a clean labeled, self locking polythene bag taking care to collect all the fine dust with a brush. The polythene bag was taken to the laboratory for analysis. Samples were air-dried for five days and sieved at 2mm to remove any extraneous materials like stones and other debris. They were again sieved through a sieve shaker (Fritsch, Germany) and the very fine fraction (< 75 µm) was taken for analysis. This fraction was homogenized in an agate mortar and oven dried at 70º C for 12h before analysis. The finer fraction of the dust samples was analysed and studied because earlier studies which focused on heavy metals in different size fractions showed that many metals generally tend to be associated with and are concentrated in the finer particle sizes and pose a higher health risk than larger particles (Lin et al. 2005). Sample solutions were prepared by closed vessel acid digestion using Savillex pressure decomposition vessels (M/s Savillex Corporation, Minnetonka, MN, USA) according to the procedure given in Roy et al. (2007). Electronic grade HF, analytical reagent grade chemicals, distilled acids and high purity water (18 MΩ), obtained from a Milli-Q system, were used for solution preparation. Rh was used as an internal standard. International certified reference materials (CRMs) SO-1, SO-2 and JG-2 were also prepared in an identical manner and used for quality control validation (Table 2).

An ICP-MS Perkin-Elmer, Sciex (Model Elan DRC II, Toronto, Canada) was used for sample analysis to obtain data of eleven heavy metals viz: Cu, Zn, Cr, Co, Pb, Ni, V, Zr, Ce, Y, Hf. Scandium concentrations also obtained for the samples (Table 1) were used to get normalized enrichment factor. The sample introduction system consisted of a standard Meinhard nebulizer with a cyclonic spray chamber attached to an auto sampler (CETAC ASX-500). All quantitative measurements were performed using instrument software. The instrumental parameters and operating conditions of the ICP-MS were set using a procedure described by Balaram and Rao (2003), such that a uniform sensitivity was obtained for all the masses across the entire mass range. Synthetic standard solutions prepared from a 1000µg/ml stock solution (Perkin-Elmer) were used for checking daily instrument performance. The precision of the measurement was <5% RSD. The detection limits were 0.004-0.016 ng/g. Several well known isobaric interferences were programmed and corrections automatically applied through instrument software. All data acquisitions were made in peak hopping mode covering all the analyte masses as well as internal standards. Online interference corrected and blank subtracted raw counts were used for calculating the concentrations of different elements. External calibration was performed using international reference materials SO-1, SO-2 and JG-2. The certified values for the reference materials were obtained from Govindaraju (1994) (Table 2).

Data Analysis
Basic statistical data
Descriptive statistical data analysis was performed using SPSS version 7.5 for Windows. The basic statistical attributes (Min., Max., Mean, Median, Range, SD, Kurtosis, Skewness, Coefficient of Variation (CV)) (Table 3) were computed to have an elementary understanding of the heavy metal data and to know the activeness of the metals in the examined environment.
Contamination Assessment Methods
The assessment of contamination levels of all studied heavy metals of the road dusts was carried out using the Index of geoaccumulation, Enrichment factor, Contamination factor and Contamination degree. These indices evaluate the intensity of heavy metal pollution by comparing observed heavy metal concentrations to background concentrations and are used to assess the extent of contamination of the examined environment (Loska et al. 2004). In this study the upper continental crust (UCC) values for metals given by Taylor and McLennan (1985) (Table 1) were adopted as background values (reference values).
The Index of geoaccumalation (Igeo) introduced by Muller (1969) was calculated using the equation:                                     Igeo = log2    Cn / 1.5 Bn
where Cn represents the measured concentration of the examined element (n) in the sample (road dust) and Bn is the background value (UCC) of the element (n). The constant 1.5 is the background matrix correction factor due to lithogenic variability (Loska et al. 2004). The descriptive classification for Igeo (Muller 1981) includes seven grades starting from class 0 (uncontaminated) to class 6 (extremely contaminated) as follows: Igeo≤0 practically uncontaminated; 0<Igeo<1 uncontaminated to moderately contaminated; 1<Igeo<2 moderately contaminated; 2<Igeo<3 moderately to heavily contaminated; 3<Igeo<4 heavily contaminated; 4<Igeo<5 heavily to extremely contaminated; 5<Igeo extremely contaminated.
Enrichment factor (EF) is based on normalization of a tested element against a reference element which is often a conservative one and is not subject to contamination by anthropogenic sources. The most common reference elements are Al, Fe, Ti, Sc and Mn (Reimann and de Caritat 2005). In this study scandium (Sc) was chosen as the reference element. The EF value was calculated using the modified formula given by Loska et al. (2004) based on the equation suggested by Buat-Menard and Chesselet (1979).
                    EF = [Cn (sample) / Cref (sample)] / [Bn (background) / Bref (background)]
where Cn (sample) is the content of the examined element in the examined environment (road dust), Cref  (sample) is the content of the reference element in the examined environment, Bn (background) is the content of the examined element in the reference environment (UCC) and Bref  (background) is the content of the reference element in the reference environment. Five contamination categories are recognized on the basis of the EF (Sutherland 2000) as follows: EF<2 Deficiency to minimal enrichment; EF = 2-5 Moderate enrichment; EF = 5-20 Significant enrichment; EF = 20-40 Very high enrichment; EF>40 Extremely high enrichment.
Contamination factor (Cif) and Degree of contamination (Cdeg) and their respective categories were suggested by Hakanson (1980). The Cif  is the single element index; the sum of contamination factors for all elements examined represents the contamination degree of the environment. Thus Cdeg indicates the overall contamination of the studied road dusts. The contamination factor (Cif) was calculated using the equation:
Cif = Ci0-1 / Cin
where Ci0-1 is the mean content of metals from at least five sampling sites and Cin is the pre industrial concentration of individual metal. In the present study, modification of the factor (Loska et al. 2004) was used wherein the reference values were of UCC. Cif   is defined according to four categories as follows: Cif <1 Low contamination; 1≤Cif <3 Moderate contamination; 3≤Cif <6 Considerable contamination; 6≤Cif Very high contamination. The contamination degree (Cdeg) of the environment has four classes which are: Cdeg<8 Low degree of contamination; 8≤Cdeg<16 Moderate degree of contamination; 16≤Cdeg<32 Considerable degree of contamination; 32≤Cdeg Very high degree of contamination.
Statistical Analysis
Pearson’s correlation coefficient analysis and Principal component analysis (PCA) were carried out using the software SPSS version 7.5 for Windows. Pearson’s correlation analysis was applied to evaluate the associations of the studied heavy metals in the road dusts and establish the strength of their interrelationship for identifying the sources and pathways of these metals. Principal component analysis is widely used to reduce complexity of the data and to extract a small number of latent factors (principal components) for analyzing relationships among the observed variables (Banerjee 2003; Shi et al. 2013). In the present study PCA was carried out to ascertain the possible contributing factors on the elemental concentrations of the studied road dusts and to assign natural vs. anthropogenic contributions. Varimax rotation with Kaiser normalization (Kaiser 1960) was applied to maximize the variances of the factor loadings across variables for each factor for better interpretation of results.

Results and Discussion
Elemental concentrations
The heavy metal data of the studied road dusts as well as crustal (UCC) abundances given by Taylor and McLennan (1985), which have been used as reference values, are presented in Table 1. Data obtained for the international certified reference materials SO-1, SO-2 and JG-2 analyzed along with the samples are in good agreement with certified values (Table 2). Replicate measurements of the standard reference materials helped to monitor accuracy of data. Table 3 summarizes the basic statistical attributes of these road dusts. Elemental data shows a large variation (Table 1). Zn and Zr record the highest levels with values ranging from 82.4 to 990.2 µg/g and 156.3 to 1085.8 µg/g respectively. About 38% of the samples have Zn contents >300 µg/g, 53% of them exhibit Zr concentrations >500 µg/g but only six samples have Cu levels above 200 µg/g (Table 1). The mean levels of Ce (123.7µg/g), Zr (526.9µg/g), Y (29.8 µg/g) and Hf (8.7 µg/g) obtained for the present study are higher in comparison to the mean levels of these metals reported for road dusts from different traffic areas of Hyderabad (Mathur et al. 2011) which indicate that industrial activity mainly contributed to the enrichment of these metals. The ranges of the contents of the heavy metals are also quite wide. The maximum values of Cu and Cr are 9.5 and 7 times respectively higher than their minimum values while maximum Zn and Pb levels are 12 times greater than their minimum levels (Table 3). The relative ordering of mean contents of metals is Zr > Zn >Pb > Cr > Ce > Cu > V > Ni >Y > Co >Hf. The mean concentrations of Pb (166.8 µg/g), Zr (526.9 µg/g) and Ni (50.8 µg/g) are 8.3, 2.8 and 2.5 times respectively higher while the mean contents of Cu, Zn and Cr are about 4 times higher than their corresponding crustal (UCC) abundances (Tables 1 and 3). The mean concentrations of Co, V, Ce, Y and Hf are 1.4 to 1.9 times higher than the reference values.
All the industrial sites display a wide range for most elements with samples from each area exhibiting both high as well as low values (Table 1). However, road dusts from Balanagar and Sanathnagar industrial areas are identified as having higher levels of Zn, Cu, Cr and Pb while those from Patancheru have higher Ni, Co and Cr contents. Samples from Jeedimetla exhibit moderate levels of most elements in comparison to other areas. Highest mean concentrations of Pb and Hf are displayed by Katedan road dusts while those from Nacharam and Katedan have higher mean concentrations of Ce, Y, Zr and V (Table 1). Govil et al. (2008) had also reported high concentrations of Pb, Cr, Ni and Zn in soils from the Katedan industrial area. Thus, it can be inferred that the heavy metal concentrations differ depending on the predominant activities of the sampling sites and local pollution sources. Comparison of the metal concentrations of the present work with some of the other studies reveal that the mean concentrations of Cr and Ni (Table 3) are quite comparable to those reported by Ahmed et al. (2007) and Shi et al. (2013) for road dusts of Dhaka city, Bangladesh and Xianyang city, China respectively. The mean Co and V concentrations obtained by Singh (2011) for road dusts of Dhanbad and Bokaro regions, India and mean Cu, Ni and Co values of Wei et al. (2010) (Urumqi, China) are quite comparable but mean Cu, Zn, Cr, Ni and Pb values given by them and Cu, Zn, Zr and Pb by Ahmed et al. (2007) are much lower than those obtained for the present study. However, the average values of Zn, Cr and Pb reported by Yongming et al. (2006) for urban dusts of Xi’an, China and mean levels of Cu and Zn reported by Shi et al. (2013) are higher than those of the present work. In their analysis of soils from Manali industrial area, India, Krishna and Govil (2008) also reported elevated concentrations of Cr and Cu but their Pb and Zn levels were lower and Co, V and Ni contents were in comparison to those of this work. Maximum concentrations of Cu, Cr, Pb and Ni reported for street dusts of Delhi, India (Banerjee 2003) are several times higher than those of the present study (Table 3).
The mean and median values of Co, Ni, Zr and Y are same (Table 3). Regardless of their skewed distribution, the mean and median values of V and Hf are also close. The coefficient of variation for Cu, Zn and Pb are more than 0.6 in comparison to CV’s of other metals (Table 3). Skewness coefficient of Zn, Cr, Pb and Ce exceed one, which measure the asymmetry and indicate that these metals are positively skewed towards lower concentrations as revealed by the fact that the median concentrations of these metals are lower than their mean concentrations. Zn contents show larger variability with a range of 907.8 (82.4 to 990.2 µg/g) and SD of 235.6 of which five samples have high concentrations >650 µg/g (Tables 1 and 3). The highest values of skewness and Kurtosis are found for lead (Table 3) suggesting its distribution is not normal and it has a broad range of concentrations characterizing a strong anthropogenic influence. Kurtosis calculation of data shows negative value for Co implying a more even distribution of this metal while positive values of most heavy metals (Table 3) indicate a peaked distribution with sporadic high emissions of these elements from certain sources.

Assessment of pollution levels
Determination of the Index of geoaccumulation, the enrichment factor, and contamination factor helped for the quantification of heavy metal accumulation in the studied road dusts from Hyderabad. The contamination degree of the dusts was useful in assessing the overall pollution level of metals in the study area.
Index of geoaccumulation
The minimum, maximum and mean Igeo values obtained are presented in Table 4 and the box plot of Igeo for metals are displayed in Fig. 2. Based on Igeo results the road dusts can be categorized as practically uncontaminated to moderately contaminated with respect to the studied heavy metals except Pb. Y exhibits the lowest mean Igeo value (-0.24) followed by V ( mean Igeo -0.20) (Table 4) as most of the samples have Igeo <0 for these metals (Fig. 2). None of the samples have Igeo values >1 for V and Y. With respect to Co, V, Y and Hf these dusts are practically uncontaminated with mean Igeo <0 while Ni, Zr and Ce exhibit mean Igeo values which categorize them to be uncontaminated to moderately contaminated. Igeo data obtained reveals that most of the contaminant metals are considerably enriched when compared to their crustal abundances, though none of the samples exhibit Igeo values that indicate extreme contamination. Igeo values of dust samples for Cu, Zn, Cr and Pb (Fig. 2) indicate pollution due to these heavy metals is wide spread, especially samples from Balanagar and Sanathnagar industrial areas exhibit higher values for these metals (Table 1). Road dusts from Katedan industrial area are heavily contaminated with respect to Pb. Igeo data reveals that high levels of toxic metals in the study area are derived predominantly due to anthropogenic activities with differential accumulation of these metals.
Enrichment factor
The enrichment factor calculated for the studied heavy metals are shown in Table 5 and the distribution of each metal’s EF is displayed in Fig.3.  Pb shows the highest values of EF ranging from 2.63 to 30.86 with a mean of 12.44 (Table 5). The EFs for Cu, Zn and Cr range from 1.40 to 16.60; 1.17 to 19.67 and 1.86 to 15.62 with mean values of 7.21, 6.91 and 6.18 respectively and along with Pb are ranked as a group showing significant enrichment based on the classification of Sutherland (2000). The elevated EFs observed for Cu, Zn, Cr and Pb are mainly due to significant enrichment of these metals in dust samples from Balanagar, Sanathnagar and Katedan. The lowest mean EF of 2.08 is obtained for Y, with values ranging from 0.97 to 4.32, followed by V which exhibits mean EF of 2.12 (Table 5). This is similar to the results observed for Igeo (Table 4). The minimum EFs displayed by all the metals except Pb are < 2 indicating deficiency to minimal enrichment (Sutherland 2000). None of the analyzed heavy metals have mean EF showing very high enrichment (Table 5). Based on the mean EFs the road dusts can be categorized as moderate to significantly enriched with respect to all the metals studied. Though the mean EFs of Ni and Zr are higher showing values of 4.05 and 4.35 respectively when compared to the mean EFs of Co, V, Ce, Y and Hf (Table 5), they all are grouped into the same class showing moderate enrichment, with samples from Nacharam being more enriched in Ce, Y, V and Hf. Samples from Jeedimetla and Patancheru display moderate to significant enrichment of Cu, Zn, Cr, Pb, Ni and Zr. The lowest mean EF found for Y implies contribution also from natural sources. According to Sutherland (2000) EF value <2 indicates deficiency to minimal enrichment implying contribution of the metal from natural processes. On the other hand Cu, Zn, Cr, and Pb with maximum EF >15 (Table 5) suggest that anthropogenic sources contributed a substantial fraction of these elements to the road dusts. The maximum EF for all the studied metals ranges from 4.01 to as high as 30.86 (Table 5) due to the difference in the input of each metal to the dusts reflecting varying degree of anthropogenic pollution. Maximum EF of 30.86 for Pb indicates serious contamination due to site specific pollution sources.
Contamination factor and Degree of contamination
Table 6 displays the contamination factor evaluated for the studied heavy metals in this work. Highest mean Cif of 8.34 displayed by Pb implies very high contamination level of this metal according to the classification categories given by Hakanson (1980). Pb also displays highest levels for Igeo and EF. On the basis of mean Cif the road dusts are classified as considerably contaminated by Cu, Zn and Cr but moderately contaminated with respect to all other studied metals. However, maximum Cif  for Cu (12.89), Zn (13.95) and Cr (10.23) (Table 6) indicate very high contamination levels attributable to road dusts mainly from Balanagar and Sanathnagar industrial areas, as discussed earlier. Maximum Cif values of Co, Ni, Zr, Ce and Hf (Table 6) indicate considerable contamination (Hakanson 1980). On the whole Cif trends confirmed the results achieved applying Igeo and EF indices (Tables 4, 5 and 6). The Cdeg for the mean metal levels of the road dusts is 34.47 (Table 6) denoting very high degree of contamination (Hakanson 1980). Data presented in Table 6 indicates that Pb contributed most to the degree of contamination (24.19%) followed by Cu (13.46%), Zn (12.85%) and Cr (11.37%) respectively. Individual contributions of Co, V, Y and Hf to Cdeg were less than 5%.
Based on the evaluated pollution indices (Igeo, EF and Cif) comparative trends of metal pollution levels in decreasing order are:
                        Igeo:  Pb > Cu > Cr > Zn > Zr > Ni > Ce > Co >Hf > V > Y.
                       EF:   Pb > Cu > Zn > Cr > Zr > Ni > Ce > Co > Hf > V > Y.
                       Cif :  Pb > Cu > Zn > Cr > Zr > Ni > Ce > Co > Hf > V > Y.
The evaluated pollution indices reveal that road dusts from the industrial areas of Hyderabad are pollution impacted showing varying degree of heavy metal contamination.

Correlation coefficient analysis
Pearson’s correlation analysis was applied to the total concentration of heavy metals to establish their interrelationship. The correlation matrix (Table 7) shows that there is a strong positive correlation (r > 0.7) between the metal pairs of Cu-Zn, Cr-Ni, Ce-V, Y-Ce and Hf-Zr, with maximum value of 0.82 displayed by Cu and Zn. Significant positive correlations (between 0.7 and 0.5) are also exhibited between Cr- Cu, Zn, Co; Ni- Cu, Co; V- Co, Y; Hf- Pb, Y at the 0.01 level. Zr and Y also present a positive correlation. The positive correlation coefficients for these metals imply that they originated from common anthropogenic sources. Pb on the other hand displays a less pronounced positive correlation with Cu, Zn, Zr and Y. Association of Pb with Cu and Zn as well as Hf, Zr and Y indicates its contribution to the dusts from multiple sources since Cu and Zn unlike Pb display very low to negative correlation with Hf, Zr and Y (Table 7). The correlation of Cu, Zn and Cr in the dusts demonstrates that these metals are deposited from anthropogenic sources since there is no known geogenic source in the study area which can contribute to this type of association. The correlation matrix shows that there is very low to negative correlation between the other metal pairs especially of Cu, Zn, Cr and Ni individually with V, Zr, Ce, Y and Hf which signify that the sources for accumulation of these metals are much different and they may have accumulated through independent or multiple pathways. Since the studied heavy metals are commonly used as components of automobile parts and for different industrial processes, their accumulation in the road dusts is attributed mainly to industrial activity and vehicular emissions with minor contributions also from domestic activity and natural sources.

Principal Component Analysis
In the present study PCA helped to identify significantly important variables and the possible sources of the heavy metals to the road dusts. According to the Kaiser criterion, the number of significant principal components with eigen values higher than 1, were selected. The factor loadings, eigen values of the factors and % variances are listed in Table 8. Factor loadings > 0.5 (bold characters in Table 8) are regarded as significant in the interpretation of data. Three factors are obtained accounting for 80.49% of the total variance (Table 8). All the elements are well represented by the three factors. 32.64% of total variance is controlled by factor 1 showing higher loadings for Cu, Cr, Zn, Ni and Co. Association of these metals in the group imply that they mainly originated from similar pollution sources. Correlation analysis (Table 7) also displays significant positive correlations between them which are an indicator of the same source. It is also observed that the loading of Co (0.61) is not as high as the loading of the other metals of this group implying that Co may have an additional source which contributed to its concentration. This factor suggests a source relating to both vehicular and industrial emissions (anthropogenic factor) since the elements grouped under this factor are associated with vehicle engine exhaust, wearing and corrosion of automobile parts besides their emission from metal working and fabrication units, chemical and other industries (DeMiguel et al. 1997; Hjortenkrans et al. 2007; Thorpe and Harrison 2008).
The variance explained through the second factor is 26.94% of the total variance and is mainly characterized by high positive loadings of Hf (0.93), Pb (0.77), Y (0.61) and Zr (0.82) (Table 8). In this group Y has a loading lower than that of other metals of the group while Pb exhibits a higher loading (0.77) for this factor in comparison to its less significant loading (0.39) for factor 1. The metals exhibiting high loadings for factor 2 are employed in different industrial processes (Kogel et al. 2006) and hence can be defined as anthropogenic components related mainly to industrial emissions. Zr, Y and Hf are also components of the automobile catalytic converters (Helmers 1996; Jarvis et al. 2001) and may have been partially contributed by autocatalyst derived traffic emissions. Though the metals showing significant correlation with factor 1 are also identified as traffic and industry related metals but their lower to negative correlation with factor 2 (Table 8) imply sources which are different from those which contributed Hf, Zr and Y to the road dusts in the study area. Results of correlation analysis also substantiate this inference as Cu, Zn, Cr and Ni exhibit very low to negative correlation with these metals (Table 7). Correlation of Pb both with factors 1 and 2 and of Y with high loadings for factors 2 and 3 (0.61 and 0.67 respectively) indicate their association with heavy metals of these factors and release from similar sources. Factor 3 is dominated by Ce, Co, V and Y explaining 20.91% of total variance. The loading of Co (0.57) is lower than the other metals of this group. Rotated component matrix shows that Co displays a combined relationship with factor 1 and factor 3 indicating that it has a combined origin. It exhibits a strong correlation especially with Cr, Ni and V but a very low to negative correlation with Ce and Y (Table 7). V on the other hand displays a strong positive correlation not only with Co but also with Ce and Y indicating its contribution from multiple sources. This factor can be assigned to a source of mixed origins with contributions not only from industrial sources but also from natural sources, vehicular and domestic activity. Ce, Y, Co and V all have different uses in vehicular systems and industrial processes especially in the metal processing industries and as fuel additives (Helmers 1996; Jarvis et al. 2001). In general results of PCA agreed well with that of the correlation coefficient analysis. Thus, PCA results indicate that different kinds of industrial activity, exhaust and non exhaust traffic emissions clubbed with contributions from domestic activity and less significantly from geogenic (natural) sources might have contributed to the accumulation of heavy metals in the studied road dusts.

Source identification
The industrial areas of Hyderabad are characterized by the presence of different types of industries, incineration plants and servicing units which are intermingled with residential as well as commercial establishments. These areas are thus frequented by vehicles, including transport vehicles, which do no adhere to strict fitness norms. The industrial operations are also carried out without following any criteria to maintain a clean environment. Due to lack of sophisticated management of wastes and effluents that are released from the industries, mechanic workshops, servicing centers and domestic activity, heaps of solid waste are strewn all over the area and heavy metal laden effluents are discharged directly into streams and near by drains. Owning to these characteristics of the sampling sites many pollutant sources with overlapping source contributions are recognized and it is difficult to quantify the relative contribution of each source. Results obtained in the present study indicate that anthropogenic influence on the concentration of heavy metals in the road dusts is more pronounced. Correlation coefficient analysis and principal component analysis results (Tables 7 and 8) are quite consistent and give major information about the sources of these metals which can be assigned mainly to industrial and traffic emissions with contributions also from domestic and natural sources. Cu, Zn, Cr, Ni and Co grouped together under factor 1 of PCA (Table 8) are strongly correlated (Table 7) and  exhibit a wide range of concentrations and high mean levels in comparison to their corresponding crustal abundances (Tables 1 and 3) which imply to their origin from similar anthropogenic sources. They are used in fabrication of automobile metallic parts like body panels, steel studs, tyres, brake pads, brake lining systems and are also components of lubricating oil (engine oil) and fuel especially diesel fuel (Hjortenkrans et al. 2007; Pacyna et al. 2007; Thorpe and Harrison 2008). Zn and Co are additives of road marking paint. Zn compounds are used as anti oxidants and as dispersant improvers for lubricating oils while Ni is used as a fuel additive (DeMiguel et al. 1997; Wik and Dave 2009). Thus these metals are emitted to the roadway environment due to both exhaust and non exhaust vehicular emissions resulting from fuel combustion, lubricating oil spills, mechanical abrasion and corrosion of metallic automobile parts. Cu, Zn, Cr, Ni and Co also represent the galvanizing, chemical, electroplating, oil refining and metallurgy industry (Adriano 2001; Romic and Romic 2003). Zn, Cu and Cr are part of different alloys especially steel and brass alloys while Cr, Co, Ni are used in tanning, textiles and manufacture of plastic products, paints, glass and ceramics (Rodriguez et al. 2004). Ni is also used in manufacture of batteries and fertilizers (Sutherland 2000). Hence, metallurgical and chemical processes, manufacture of goods, fuel combustion, panel beating, rusting of metal scrap and waste incineration release these metals which accumulate in the roadway environment. In keeping with the observations of Banerjee (2003), Ahmed et al. (2007), Romic and Romic (2003) it is considered that these heavy metals have originated from industrial and traffic emissions.
Pb is associated with vehicular, domestic and industrial pollution (Parekh et al. 2002; Charlesworth et al. 2003; Wei et al. 2010) which is reflected by its positive correlation with Zn, Cu, Zr, Y and Hf (Table 7) and loadings for factors 1 and 2 of PCA (Table 8). It is a component of brake lining material and wheel weights (Hjortenkrans et al. 2007). Lead was being used as an octane enhancer in gasoline which has been phased out (DeMiguel et al. 1997). However, Pb containing petrol and diesel are still being used in vehicles at unchecked points in the study area (Mathur et al. 2011). A significant amount of Pb is employed in the manufacture of paints and dyes, pesticides, fertilizers, batteries, glass and as an impurity of Zn is present in galvanized metals (Parekh et al. 2002; Yongming et al. 2006). Its high concentration and wide range, in the studied dusts could be a result of emissions from these multiple sources. Part of the Pb could have also accumulated from domestic activity like burning of fuel, use of Pb based products and Pb painted construction material. Zr, Hf and Y are well known to be contributed by industrial as well as traffic sources (Helmers 1996; Charlesworth et al. 2003; Kogel et al. 2006). They are employed as catalysts in the chemical and metallurgical industry, in electronics and to improve the properties of metal alloys (Schaller 1991; Kogel et al. 2006) which explains for their loading and association with factor 2 of PCA (Table 8). Hf is used for making cutting tools and Zr for abrasives and leather tanning. Hf is always associated with Zr indicating to their chemical association which explains for the observed high positive correlation (0.79) of these two metals (Table 7). Y is used in automobile oxygen sensors, in certain spark plugs and along with Ce is employed in ceramic and glass industry which could be the reason for its variation and loading in both factors 2 and 3 of PCA (Table 8). It is employed as a Pb replacement in anti corrosion coatings and hence it exhibits a moderate correlation with Pb (Table 7) (Goering et al. 1991). Zr, Hf and Y are important components of catalytic converters in automobiles and along with Pt, Pd and Rh are used for the catalyst’s wash coat (Helmers 1996; Zereini et al. 2001). Zr oxides are added to the washcoat as stabilizers to improve the thermal stability and poison resistance. Zr oxides and Zr silicates are used in brake lining materials along with Pb (Schaller 1991). But autocatalyst derived traffic emissions do not appear to have significantly contributed these metals in view of the fact that the type of vehicles, especially the transport vehicles, trucks and auto rickshaws plying in these industrial areas are more than 15 years old and most of them may not be fitted with catalytic converters. Unlike in the developed countries, catalytic converters fitted to automobiles have been introduced very late in India. This is also substantiated by the fact that Ce which is also an important component of automobile catalytic converters (Helmers 1996) shows a very weak loading (0.11) for factor 2 and also very low correlation with Zr and Hf which are also components of autocatalysts. The higher mean levels of these metals obtained for the present work in comparison to those obtained for road dusts from traffic areas of Hyderabad (Mathur et al. 2011) also support this inference. The accumulation of Zr, Hf, Y and Pb which strongly correlate with factor 2 of PCA (Table 8) can thus be traced back mainly to industrial processes and less significantly to automobile emissions.
Vanadium exhibits a strong correlation with Co, Ce and Y (Table 7) and is also grouped along with these metals under factor 3 of PCA. Ce and V are fuel additives and along with Co are released due to petroleum and vegetation burning and domestic combustion processes. This inference is in consistence with that of Rodriguez et al. (2004). V, Co, Ce and Y are markers of industrial emissions as they are used in the manufacture of glass, ceramics and electronics and in dye making, chemical and metallurgical processes. They are also employed in different vehicle parts like V and Co are parts of steel studs of tyres (Edwards et al. 1995; Helmers 1996; Zereini et al. 2001) which implies their release partly due to automobile emissions and wear and tear of automobile parts. V is also released due to abrasion of asphalt. Co, Y and Zr may have also originated from the parent rocks and are thus considered to have the influence of natural sources (Ferguson 1990; Govil et al. 2008; Wei et al. 2010). Thus associations of the studied heavy metals with different products and processes and inter element correlations (Table 7) and respective factor loadings (Table 8) imply to their origin from multiple sources which could be mainly linked to vehicular and industrial activity and in part to domestic and geogenic/lithogenic sources. Igeo, EF and Cif   results (Tables 4, 5 and 6) also indicate considerable enrichment of these metals in the road dusts and contamination of the study area. Other studies also reported the combined pollution of soils and road dusts by multi heavy metals (Govil et al. 2008; Krishna and Govil 2008; Wei et al. 2010; Shi et al. 2013).
Conclusion
Excessive levels of heavy metals in the environment severely disturb the natural geochemical cycling of the eco system. People in residential areas surrounded by industries are to a large extent exposed to heavy metal laden polluted road dust which is detrimental to their health. In view of this the present study was taken up to assess the extent of heavy metal contamination of road dusts from industrial areas of Hyderabad and to delineate their sources. The results obtained reveal that the studied road dusts have elevated concentrations of Cu, Zn, Cr, Co, Pb, Ni, V, Zr, Ce, Y and Hf. These heavy metals are significantly enriched in comparison to their crustal abundances which demonstrate an intensive impact of various anthropogenic sources on their accumulation. Wide variations of the metal concentrations reflected by their maximum and minimum values also substantiate this finding. Similar trends of mean metal levels obtained applying the three pollution indices (Igeo, EF and Cif ) indicate the relative magnitude of contamination of each toxic metal. On the basis of Igeo the road dusts are categorized as practically uncontaminated to moderately contaminated with respect to all metals except Pb. Mean EFs of the studied metals classify the dusts to be moderately to significantly enriched while their mean Cif imply moderate to considerable contamination. Highest mean levels displayed by Pb for the three pollution indices suggest very high level of contamination of this metal in the study area. The results of correlation analysis and PCA also provide the same information on heavy metal accumulation confirming a strong anthropogenic influence from multiple sources, attributable mainly to industrial and vehicular activity besides they being partly contributed by domestic and natural sources. Release of untreated industrial effluents produced by processing plants and manufacturing of goods, combustion of fuels and emissions from smoke stacks, random dumping of solid waste as well as vehicular exhaust emissions and corrosion/wear and tear of vehicle components appear to have contributed significantly to the accumulation of heavy metals in the road dusts.  As there are no designated areas for different types of activities especially due to intermingling of residential settlements and industrial units, emissions from different sources may have carried different metals. The results of this study imply the need for periodic investigation of the study area to understand the spatial distribution and pathways of heavy metals and the relative significance of different emission sources to assess their ecological implications and human health risks. The results can serve as a useful guide for the management to plan effective strategies for pollution control like implementing improved treatment techniques for solid waste and industrial effluents, assigning specific areas/drains for waste disposal, bioremediation, phytoremediation and instituting strict laws for vehicle fitness and operation. It is suggested that government should establish monitoring stations and launch awareness programs for the public about the health risks associated with the heavy metal pollutants.

Acknowledgements
The authors are thankful to the Director, CSIR-National Geophysical Research Institute, Hyderabad, for his support and permission to publish this paper. The present study was carried out under the institute project MLP-6201-28 (CM).

References
Adriano, D. C. (2001). Trace elements in terrestrial environments. Biogeochemistry, bioavailability and risks of metals. (2nd ed., p. 867). Berlin: Springer.
Ahmed, F., Bibi, M. H., & Ishiga, H. (2007). Environmental assessment of Dhaka city (Bangladesh) based on trace metal contents in road dusts. Environmental geology, 51, 975-985.  
Alloway, B. J., & Ayres, D. C. (1997). Chemical principles of environmental  pollution. (2nd ed., p.395). London: Blackie academic and professional.
Balakrishna, S., & Raghava Rao, M. (1961). Pink and gray granites of Hyderabad. Current Science, 30, 264-274.
Balaram, V., & Rao, T. G. (2003). Rapid determination of REEs and other trace elements in geological samples by microwave acid digestion and ICP-MS. Atomic Spectroscopy, 24, 206-212.
Balaram, V., Kamala, C. T., Mathur, R., Rao, S., & Satyanarayanan, M. (2013). PGE in road dust/ Top soil from major Indian cities: Implications on human health and environment. In: Mu. Ramkumar (ed.), On a sustainable future of the earth’s natural resources (pp. 85-107). Springer Earth System Sciences, Berlin : Springer Verlag.
Banerjee, A. D. K. (2003). Heavy metal levels and solid phase speciation in   street dusts of Delhi, India. Environmental Pollution, 123, 95-105.
Brown, H. S., Kaspertan, R. E., & Raymonds, S. (1990). Trace Pollutants. In: Turner, B.L., Clarke, W.C., Yates, R. W., Richards, J.F., Mathews, J.T., & Meyer, W.B. (eds.), The Earth as transformed by human action. (pp. 437-455). Cambridge: Cambridge University Press.
Buat-Menard, P., & Chesselet, R. (1979). Variable influence of the atmospheric flux on the trace metal chemistry of oceanic suspended matter. Earth and Planetary Science Letters, 42, 398-411.
Charlesworth, S., Everett, M., McCarthy, R., Ordonez, A., & De Miguel, E. (2003). A comparative study of heavy metal concentration and distribution in deposited street dusts in a large and a small urban area: Birmingham and Coventry, West Midlands, UK. Environment International, 29, 563-573.
De Miguel, E., Llamas, J. F., Chacon, E., Berg, T., Larssen, S., Royset, O., & Vadset, M. (1997). Origin and patterns of distribution of trace elements in street dust: Unleaded petrol and urban lead.  Atmospheric Environment, 31, 2733-2740.
Edwards, R., Lepp, N. W., & Jones, K. C.  (1995). Other less abundant elements of potential significance. In: Alloway, B. J. (ed.), Heavy Metals in Soils. (2nded. pp.307-352). London, U.K.: Blackie Academic/Chapman and Hall.
Fergusson, J. E. (1990). The heavy elements: chemistry, environmental impacts and health effects (p.614). Oxford, U.K.: Pergamon press.
Gnaneshwar, P., & Sitaramayya, S. (1998). Petrochemistry and origin of archean granitic rocks of Hyderabad city. Indian Journal of Geology, 70, 249-264.
Goering, P. L., Fisher, B. R., & Fowler, B. A. (1991). The Lanthanides. In: Merian, E. (ed.), Metals and their compounds in the environment. (pp.959). V.C.H., Weinheim.
Govil, P. K., Sorlie, J. E., Murthy, N. N., Sujatha, D., Reddy, G. L. N., Lund, R. K., Krishna, A. K., & Ram Mohan, K. (2008). Soil contamination of heavy metals in the Katedan Industrial Development Area, Hyderabad, India. Environmental Monitoring and Assessment, 140, 313-323.
Govindaraju, K. (1994). Compilation of working values and sample description for 383 geostandards. Geostandards Newsletter, 18, (special issue), 1-158.
Hakanson, L. (1980). An ecological risk index for aquatic pollution control: A sedimentological approach. Water Research, 14, 975-1001.
Helmers, E. (1996). Elements accompanying platinum emitted from automobile catalysts.     Chemosphere, 33, 405-419.
Hjortenkrans, D. S. T., Bergback, B. G., & Haggerud, A. V. (2007). Metal emissions from brake linings and tires: case studies of Stockholme, Sweden, 1995 / 1998 and 2005. Environmental Science and Technology, 41, 5224-5230.
Jarvis, K. E., Parry, S. J., & Piper, J. M. (2001). Temporal and spatial studies of autocatalyst derived platinum, rhodium and palladium and selected vehicle derived trace elements in the environment. Environmental Science and Technology, 35, 1031-1036.
Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational and psychological measurement, 20,141-151.
Kogel, J. E., Trivedi, N. C., Barker, J. M., & Krukowski, S. T. (2006). Industrial minerals and rocks: Commodities, Markets and Uses. (7th ed.  p.1548). Colorado, USA: Society for Mining, Metallurgy and Exploration, Inc.
Krishna, A. K., & Govil, P. K. (2008). Assessment of heavy metal contamination in soils around Manali industrial area, Chennai, Southern India. Environmental Geology, 54, 1465-1472.
Krishna, A. K., Satyanarayanan, M., & Govil, P.K. (2009). Assessment of heavy metal pollution in water using multivariate statistical techniques in an industrial area: a case study from Patancheru, Medak District, Andhra Pradesh, India. Journal of Hazardous Materials, 167, 366-373.
Lin, C.C., Chen, S.J., Huang, K.L., Hwang, W.I., Chang-Chien, G.P., & Lin, W.Y. (2005). Characteristics of metals in nano / ultrafine / fine/ coarse particles collected beside a heavily trafficked road. Environmental Science and Technology, 39, 8113-8122.
Loska, K., Wiechula, D., & Korus, I. (2004). Metal contamination of farming soils affected by industry. Environment International, 30, 159-165.
Mathur, R., Balaram. V., Satyanarayanan, M., Sawant, S. S., & Ramesh, S. L. (2011). Anthropogenic platinum, palladium and rhodium concentrations in road dusts from Hyderabad city, India. Environmental Earth Sciences, 62, 1085-1098.
Muller, G. (1969). Index of geoaccumulation in sediments of the Rhine River. Geojournal, 2, 108-118.
Muller, G. (1981). Die Schwermetallbelastung der sedimentation des Neckars und Seiner Nebenflusse : Eine Bestandsaufnahme. Chemiker-Zeitung, 6, 157-164.
Pacyna, E. G., Pacyna, J. M., Fudala, J., Strzelecka-Jastrzab, E., Hlawiczka, S., Panasiuk, D., Nitter, S., Pregger, T., Pfeiffer, H., & Friedrich, R. (2007). Current and future emissions of selected heavy metals to the atmosphere from anthropognic sources in Europe. Atmospheric Environment, 41, 8557-8566.
Parekh, P. P., Khwaja, H. A., Khan, A. R., Naqvi, R. R., Malik, A., Khan, K. M., & Hussain, G. (2002). Lead content of  petrol and diesel and its assessment in an urban environment . Environmental Monitoring and Assessment, 74, 255-262.
Reddy, M. V., Babu, K. S., Balaram, V., & Satyanarayanan, M. (2012). Assessment of the effects of municipal sewage, immersed idols and boating on the heavy metal and other elemental pollution of surface water of the eutrophic Hussainsagar lake (Hyderabad, India). Environmental Monitoring and Assessment, 184, 1991-2000.
Reimann, C., & de Caritat, P. (2005). Distinguishing between natural and anthropogenic sources for elements in the environment: Regional geochemical surveys versus enrichment factors. Science of the Total Environment, 33, 91-107.
Rodriguez, S., Xavier, Q., Alastuey, A., Viana, M., Alarcon, M., Mantilla, E., & Ruiz, C. R. (2004). Comparative PM10 - PM2.5 source contribution study at rural ,urban and industrial sites. PM episodes in eastern Spain. Science of the Total Environment, 328, 95-113.
Romic, M., & Romic, D. (2003). Heavy metal distribution in agricultural top soils in urban area,    England.  Environmental Geology, 43, 795-805.
Roy, P., Balaram, V., Kumar, A., Satyanarayanan, M., & Rao, T. G. (2007). New REE and trace element data on two kimberlitic reference materials by ICP-Ms. Geostandards and Geoanalytical Research, 31, 3, 261-273.
Schaller, K. H. (1991). Zirconium. In: Merian, E. (ed.), Metals and their compounds in the environment. (pp.1343-1348). VCH, Weinheim. 
Shi, X., Chen, L., & Wang, J. (2013). Multivariate analysis of heavy metal pollution in street dusts of Xianyang city , NW China. Environmental Earth Sciences, 69, 1973-1979.
Singh, A.K. (2011). Elemental chemistry and geochemical partitioning of heavy metals in road dust from Dhanbad and Bokaro regions, India. Environmental Earth Sciences, 62, 1447-1459.
Stigliani, W. M., Doelman, P., Salomons, W., Schulin, R., Schmidt, G.R.B., & van der Zee, S.E.A.T.M. (1991). Chemical time bombs: predicting the unpredictable. Environment, 33, 26-30.
Sutherland, R.A. (2000). Bed sediment associated trace metals in an urban stream, Oahu, Hawaii. Environmental Geology, 39, 611-627.
Taylor, S.R., & McLennan, S.M. (1985). The continental crust: Its composition and evolution. (p.312). Oxford:  Blackwell Scientific Publications.
Thorpe, A., & Harrison, R. M. (2008). Sources and properties of non-exhaust particulate matter     from road traffic: a review. Science of the Total Environment, 400, 270-282.
Wei, B., Jiang, F., Li, X., & Mu, S. (2010). Contamination levels assessment of potential toxic metals in road dust deposited in different types of urban environment. Environmental Earth Sciences, 61, 1187-1196.
WHO. (2007). Health risks of heavy metals from long- range transboundary air pollution. Copenhagen, Denmark: World Health Organization, Regional Office for Europe.
Wik, A., & Dave, G. (2009). Occurrence and effects of tire wear particles in the environment –A critical review and an initial risk assessment. Environmental Pollution, 157, 1-11.
Yongming, H., Peixuan, D., Junji, C., & Posmentier, E. S. (2006). Multivariate analysis of heavy metal contamination in urban dusts of Xi’an, Central China.  Science of the Total Environment, 355, 176-186.
Zereini, F., Wiseman, C., Beyer, J. M., Artelt, S., & Urban, H. (2001). Platinum, lead and cerium concentrations of street particulate matter (Frankfurt, Germany). Journal of Soils and Sediments, 1, 188-195.