Quantification of benzene in groundwater sources and risk analysis in a popular South Indian Pilgrimage City – A GIS based approach

2.1 Details of the study area

Madurai, a world famous classic South Indian Pilgrimage City of more than 2500 years old is located at 9°58′N latitude and 78°10′E longitude and occupies an area of about 140 km² .This metropolitan city, situated on the banks of the River Vaigai, is the 24th largest town in India (population wise) and third largest in the state of Tamil Nadu after Chennai and Coimbatore. As per the 2001 census, the estimated population of the city of Madurai is 1,194,665.

2.2 Sampling and analysis

One hundred and twenty groundwater samples from bore well sources were collected in pre-cleaned bottles in the months of Jan 11, Feb 11 and Mar 11 from corporate and residential locations of Madurai District. The six corporate locations were Arasaradi (A), TPK road (TPK), Periyar Bus Stand (PBS), Theppakulum road (TR), North Veli Street (NVS) and KK Nagar (KKN) and the ten residential locations were Iyyar Banglow, Anna Nagar, Tirunagar, North Car Street, Kannadasan Street, SS Colony, Pudhur, Kochadai, Krishnapuram and Simmakkal. The collected samples were stored in the refrigerator (8–10 °C) before the estimation of BTEX. At the time of estimation, the chilled water samples were equilibrated to room temperature followed by vacuum filtration using a 0.4 μm filter paper. All the reagents were of HPLC grade purchased from Sigma–Aldrich Chemie GmbH (Steinheim, Germany). The sampling spots of commercial and residential locations are shown in Fig. 1.

Close

Figure 1

Details of groundwater sampling locations in Madurai District, South India.

Export to PPT

Benzene was estimated using High Pressure Liquid Chromatography (HPLC), Shimadzu Model SPD-M20A (Liquid Chromatography module 20AT) with class VP software. Standard solutions of benzene, toluene, ethyl benzene and xylene of 2000 mg L⁻¹ were prepared and diluted to 500 mg L⁻¹ (primary dilution standard) in acetonitrile (as per the certified standard mixture 1:4) and stored in amber colored bottle at 4 °C. Fresh aqueous working standard solutions were prepared daily by diluting the primary dilution standard solution in water of HPLC grade. After the zero correction was done, 20 μl of the standard was injected using a micro syringe (pre cleaned with isopropyl alcohol) with initially set run time of 20 min. From the knowledge of retention time (RT), the samples in triplicates were injected and the corresponding chromatograms were saved in the data path. A representative standard chromatogram for benzene, toluene, xylene and benzene–toluene–ethyl benzene mixture is given (Figs. S1 and S2). The BTEX concentration was estimated using the formula as follows.

The total alkalinity of all the groundwater samples was estimated by the volumetric method using hydrochloric acid of 0.02 N and methyl orange as indicator.

2.3 Risk analysis

Risk assessment process is defined by a model which may use different mathematical and empirical models in arriving at the estimate of risk. A conceptual exposure model is derived by tracing the chemical from its source to the receptor. The life time risk for adults and children can be calculated using the formula given below. Chronic Daily Intake (CDI) was calculated for 70 years life time (mg kg⁻¹ day⁻¹).

2.4 Interpretation through statistical and GIS tools

In this study, hierarchical cluster analysis was used to group bore wells into separate clusters based on a similar benzene concentration. Although there are a number of hierarchical clustering techniques, all of which are not regularly applied the most widely used measure of ordering is Ward’s criterion, which uses an analysis of variance approach that minimizes the sum of squares within the clusters and maximizes the variance between separate clusters (Ward, 1963). The results are represented via a dendrogram, or tree plot and are separated into different clusters. The sample locations are grouped on the vertical axis, and the linkage distances, representing the relative differences between clusters, are shown on the horizontal axis.

Geographic Information Systems (GIS) provide platforms for managing data, computing spatial relationships such as distance, connectivity and directional relationships between spatial units, and visualizing both the raw data and spatial analytic results within a cartographic context.

The GIS based analysis of the spatiotemporal behavior of benzene level in groundwater in the study area was done using the spatial analyst module of Arc GIS 9.3 powerful spatial analysis is feasible once the database is established. The interpolation technique used in the analysis is the inverse distance weighted (IDW) method. IDW is an algorithm for spatially interpolating or estimating values between measurements. Each value estimated in an IDW interpolation is a weighted average of the surrounding sample points. Weights are computed by taking the inverse of the distance from an observation location to the location of the point being estimated (Burrough and Mcdonnell, 1998). The inverse distance can be raised to a power (e.g. linear, squared, and cubed) to model different (e.g. line, area, volume) geometries (Guan et al., 1999). In a comparison of several different deterministic interpolation procedures, a squared distance term yielded results most consistent with original input data using IDW. This method is suitable for datasets where the maximum and minimum values in the interpolated surface commonly occur at sample points (ESRI, 2002). Toposheets were used to prepare the base map and to understand the general nature of the study area. GPS is used to map the location of each sampling well and finally the results were taken to GIS for further analysis.

2.5 Student’s test for difference of means

A student’s (t) test was carried out for testing the significant difference between the means of factors for Jan–Feb 2011, Feb–Mar 2011 and Jan–Mar 2011 periods against the left sided alternative hypothesis. The test statistic, which follows t distribution with (n − 1) degrees of freedom, is given by

3.1 Benzene contamination level at corporate locations

The benzene concentration levels estimated in the groundwater samples at different corporate locations in the months of Jan 2011, Feb 2011 and Mar 2011 are shown in Table. 1. The average values of groundwater samples at corporate locations were in the range of 0.016–0.041 mg L⁻¹, 0.060–0.099 mg L⁻¹, 0.026–0.138 mg L⁻¹, 0.044–0.100 mg L⁻¹, 0.030–0.099 mg L⁻¹ and 0.028–0.057 mg L⁻¹ for Arasaradi (A01–A05), Thirupparankundram road (TPK01–TPK05), Periyar Busstand (PBS01–PBS05), Theppakulam road (TR01–TR05), North Veli Street (NVS01–NVS05) and KK Nagar (KKN01–KKN05), respectively. A representative chromatogram for a groundwater sample is shown in Fig. S3. The recommended permissible level (PL) of benzene in groundwater by World Health Organization is 0.01 mg L⁻¹. The average values recorded in the months of Jan 2011, Feb 2011 and Mar 2011 were 0.100 mg L⁻¹ (10 times of PL), 0.138 mg L⁻¹ (14 times of PL) and 0.060 mg L⁻¹ (6 times of PL), respectively. The groundwater samples PBS03 and A01 were estimated to be within the safe limit of World Health Organization (WHO) in Jan 2011 and Feb 2011, respectively. Fig. 2 depicts the recorded benzene concentrations (average values) at various sampling spots of Madurai District. From the estimated benzene levels, it was observed that 7% of samples in Jan 2011 decrease to 3% in Feb 2011 (Fig. 2) with respect to the samples recommended under the safe limit of WHO. But, in March 2011, no sample was recorded under the safe limit of WHO. Based on the permissible limit for drinking, prescribed by Environmental Protection Agency (EPA) and Canadian Council of Ministers of Environment (Canadian Council of Ministers of the Environment, 1999a,b), all the groundwater samples at corporate locations registered above the permissible limit of 0.005 mg L⁻¹. The fluctuating percentage levels in the three months with respect to CCME (fresh, marine and estuary aquatic lives) are shown in Fig. 5.

Close

Figure 2

Average benzene levels in the groundwater samples of Madurai District, Tamil Nadu.

Export to PPT

Close

Figure 5

The distribution (average) of total alkalinity in the groundwater samples of Madurai District, South India.

Export to PPT

3.2 Benzene contamination level at residential locations

The groundwater samples from bore well sources were also identified with benzene contamination at 10 residential locations in Madurai District. Fig. 3 illustrates the average value of estimated benzene levels in bore wells at residential locations in the months of Jan 11, Feb 11 and Mar 11. In accordance with WHO guidelines, 20% of the residential samples were found to be fit for drinking. But, taking into account, the guidelines of EPA and CCME, the benzene concentration (0.005 mg L⁻¹) of groundwater samples in all the residential locations exceeded the recommended limit. The recommended quality criteria by CCME (for fresh aquatic, marine and estuary lives) reveal that 90% of the residential groundwater samples were found to be under the permissible limit (Fig. 4). The percentage of groundwater samples within the prescribed limit of WHO and CCME was greater in the residential locations than that in the corporate locations. In both the locations, no sample was registered within the prescribed drinking limit (USEPA) of 0.005 mg L⁻¹. The fluctuations in the benzene concentration levels with respect to months may be attributed to the natural attenuation (Lee and Lee, 2003). Research studies prompted that on- and off-site contamination of groundwater due to benzene can cause cancer mortality and birth defects (Logue and Fox, 1986). In the study area, a total of 21 petrol bunks were located within a stretch of 30 km between 17 and 28 years old in Madurai District. On counting the age of petrol bunks, leakage of hydrocarbons in the subsurface region is possible where benzene is recalcitrant under anaerobic condition but degrades moderately in the presence of oxygen (Johnson et al., 2003). Thus the subsurface leakage of benzene causing deterioration of groundwater quality may also be accounted in addition to the other anthropogenic factors (Guo et al., 2012). Evidently, to support the subsurface leakage, it was reported that the leaching of benzene (ATSDR, 2007; IPCS, 1993) was possible from the soil based on its aqueous solubility and octanol–water partition coefficient (K_ow). Also the sorption behavior of benzene onto Suspended Organic Matter (SOM) varies with other soil properties and environmental conditions (Environment Agency, 2003).

Close

Figure 3

Average benzene concentration levels (Jan–Feb–Mar 2011) in some residential locations of Madurai District.

Export to PPT

Close

Figure 4

Percentage of groundwater samples within the recommended level of CCME for marine and estuary aquatic life.

Export to PPT

The possible mechanism for the deterioration of groundwater by benzene with reference to Zhang et al. (2010) is as follows. In the first stage, the movement of benzene occurs by gravitational forces (Freeze and Cherry, 1979). At this seepage stage, lateral migration of benzene takes place by capillary forces which causes a zone called ‘oil-wetting zone’ around the core of infiltration body. In the oil-wetting zone, the capillary forces are dominant with lessening outward saturation of oil. In the second stage, the seepage of benzene ceases when it reaches the water table and ultimately floats on the groundwater. The third stage involves the contemporary mass with separate phase migration, vaporization and dissolution of some of the benzene into groundwater. The dissolution further leads to the transport of benzene along with groundwater flow. The major reported sources of benzene in water are atmospheric deposition, spills of petrol and other petroleum products, and chemical plant effluents (WHO, 1996). The role of petroleum industries in groundwater contamination by benzene was reported at Love Canal (Paigen et al., 1987), Superfund site, Pennsylvania (Budnick et al., 1984) and New Jersey (Najem et al., 1994).

Fifteen traffic signals in this 30 km stretch may also be another valid reason which attributes to the benzene contamination in groundwater samples. The accumulation of benzene due to cigarette smoke, vehicular emissions, combustion of solid biomass (Sinha et al., 2006), petrol evaporation from automobiles and petrol stations may be the associated risk causing factors in the study area and, as benzene is soluble in water (Bennett et al., 1993; Eganhouse et al., 1993), its transport into soil may also be due to rainfall in the study area. Evidently, influence of rainfall (as a vehicle) could facilitate the movement of benzene into the subsurface of the soil (Wadge and Salisbury, 1997). Various possible sources leading to benzene concentration in air are represented in Table 2 such that the fate of groundwater sources is registered with respect to deterioration by benzene concentration (Ritchie et al., 2003). The consistency in the association of benzene due to high traffic density (Wheeler et al., 2008; Bruno et al., 2006; Kwon et al., 2006) and tobacco smoke (Maestri et al., 2004) was already reported. The presence of hydrocarbons was reflected by the alkaline nature (Fig. 5) of groundwater samples in the study area. The correlation value between benzene and total alkalinity was 0.287. The present observation was supported by earlier reports (Cozzarelli et al., 2001; Lee et al., 2001; Borden et al., 1995; Basberg et al., 1998).

To examine the month-wise effect of benzene in groundwater, t-tests were carried out for the combined data of Jan–Feb 2011, Feb–Mar 2011 and Jan–Mar 2011. The results of the t-test for the combined data are 1.439, 2.322 and 0.745 for Jan–Feb 2011, Feb–Mar 2011 and Jan–Mar 2011, respectively. As the computed value of t-test was greater than the critical value of 1.699 (df = 29), the difference of means was significant at 5% level for Feb–Mar 2011. The difference of means for Jan–Feb 2011 and Jan–Mar 2011 was insignificant with respect to the values obtained from t-test. Hence, the results reflected a significant variation in the benzene level between February and March 2011.

The dendrograms resulting from the hierarchical cluster analysis shown in Fig. 6(A–C) display the presence of four clusters in March 2011 and three clusters in Jan and Feb 2011. The bore well samples are grouped according to the benzene concentrations. Table 3 represents the minimum, maximum, average and SD of benzene concentration for each cluster.

Close

Figure 6

Dendograms using Hierarchical cluster analysis for the distribution of benzene in groundwater in Jan 2011, Feb 2011 and Mar 2011.

Export to PPT

The average values of each cluster revealed the heterogeneous distribution of benzene between the average values of 0.028 and 0.193 mg L⁻¹in the bore wells in Jan 2011. Three clusters for the month of Feb 2011 were obtained with two clusters showing the average benzene level almost equal but the other in four-folds. The prevalence of unsymmetric distribution of benzene level was studied in a close range. Four different contamination ranges were corroborated in the month of Mar 2011 which were approved from the clusters. The minimum and maximum differences in the benzene level were 0.017 mg L⁻¹ and 0.081 mg L⁻¹, respectively. The division of clusters in the month of Mar 2011 was greater than Jan 2011 and Feb 2011 which reflects that benzene distribution was more heterogeneous in Mar 2011.

3.3 Cancer risk analysis among adults and children

The estimated cancer risk among adults and children due to benzene contamination in groundwater in the corporate locations was calculated and shown in Fig. 7. In every 10⁴ adults, an average of 16, 23 and 13 persons were found to be under cancer risk in the months of Jan 11, Feb 11 and Mar 11, respectively. The risk analysis among children was in seven folds as compared to the risk of adults. The children who were at risk were in the average range of 90–160 per 10,000 in the study area between Jan 11 and Mar 11. The risk analysis revealed that a high magnitude of adults and children were at cancer risk in the corporate locations namely Thirupparankundram, Periyar Bus stand and KK Nagar in the months of Jan 11, Feb 11 and Mar 11, respectively. The cancer risk analyzed in the residential samples corroborated the possibility of a greater risk among adults than children (Fig. 8A and B). The cancer risk in every 10⁴ adults was evaluated in the following order: North Car Street > Tirunagar > Krishnapuram > Iyyar banglow > Kochadai > SS Colony > Anna Nagar > Kannadasan Street.

Close

Figure 7

Estimated cancer risk analysis among adults and children around commercial locations of Madurai District, South India.

Export to PPT

Close

Figure 8

Number of adults at cancer risk per 10⁴ (A) and children at cancer risk per 10⁷ (B) due to the benzene contaminant level in the residential locations of Madurai District, South India.

Export to PPT

3.4 Geographic information system (GIS) analysis

The thematic maps for the benzene contaminated bore wells in six corporate locations in the months of Jan 11, Feb 11 and Mar 11 was integrated using the addition function available in the Arc GIS software. We created a final benzene contaminated groundwater map by overlaying these thematic maps which are produced as a result of inverse distance weighted (IDW) interpolations. The spatial integration for benzene contaminated groundwater zone mapping was carried out using Arc GIS Spatial Analyst extension. The areas were delineated based on the benzene contamination level of bore well sources in accordance with nine benzene concentration levels as shown in Fig. 9(A–G).

Close

Figure 9

GIS based thematic maps depicting the benzene distribution in groundwater at Arasaradi (A), TPK Road (B), Periyar Bus Stand (C), Theppakulum (D), North Veli Street (E) and KK Nagar (F) in Jan–Feb 2011 and Mar 2011 (G).

Export to PPT

The map shows mostly the distribution of benzene level in the range 0.030–0.164 mg L⁻¹ in the TPK corporate locations during Jan 11. But in the month of Feb 11, a heterogeneous pattern of distribution in benzene was depicted from the contours with a high level of concentration at the center point (at Bharat Petroleum) of TPK area.

The Jan 11 contour pattern at Theppakulum shows the increased level of benzene from Thiagarajar College to Kamarajar Road between 0.030 and 0.332 mg L⁻¹. But the pattern of Feb 11 illustrates that the benzene level in this area was observed with variations in the contour pattern with a dominance level of 0.231–0.332 mg L⁻¹. At Northveli Street, the contours were observed with appreciable differences in the months of January and February 2011. In particular, the estimated benzene levels at NV Street, Bharat Petroleum and Bell hotel were with a striking change. Two sampling points, Bharat Petroleum (near Karimedu) and Jail Road at Arasaradi were spotted with a great reciprocation in Jan 11 and Feb 11 in the benzene levels. In March 2011, the concentration levels recorded at Theppakulum road, KK Nagar and TPK road were 0.231–0.332 mg L⁻¹. Periyar Bus Stand and Arasaradi (Jail road) registered between 0.030 and 0.198 mg L⁻¹ and North Veli Street was observed with the benzene concentration range of 0.164–0.265 mg L⁻¹. TPK road–Theppakulum road–KK Nagar locations and Arasaradi–North Veli Street locations were observed with a homogeneous pattern of distribution in the benzene level in March 2011.