A universal method to evaluate groundwater pollution risk with the site scale has not been established. The groundwater pollution risk is commonly evaluated by overlaying the results of groundwater vulnerability and pollution sources, but this method is inapplicable for contaminated sites. Considering the current supervision demands of groundwater pollution in China, this paper suggests a risk evaluation method for contaminated sites. The present and forecasted groundwater pollutant concentrations and the variation trend of groundwater quality were used to get the risk classes of groundwater pollution. Chromium is one of the most common groundwater contaminants found at industrially contaminated sites. A chromium residue disposal site in Baotou, China was selected as the study area. The groundwater flow model and solute transport model were constructed by MODFLOW and MT3DMS. Based on the simulation results of the numerical model, the groundwater pollution risk of the contaminated site was divided into four classes, and the regulation measures of groundwater pollution were suggested. This study could provide guidance for the environmental supervision and management of contaminated sites.

INTRODUCTION

Groundwater pollution is a major environmental problem in the process of urbanization. Remediation of polluted aquifers is prohibitively costly and often impractical, so the groundwater vulnerability assessment is important for a rational management of groundwater resources (Thirumalaivasan et al. 2003; Almasri 2008). The pollution sources and intrinsic vulnerability are the main factors in the groundwater pollution risk assessment, while a universal approach to estimate groundwater vulnerability and pollution risk has not been developed yet (Dimitriou et al. 2008; Wang et al. 2012).

DRASTIC is one of the most widely used models for groundwater vulnerability and pollution risk evaluation. Al-Adamat et al. (2003) assessed the potential risk of groundwater to pollution using DRASTIC and a land use map in the Azraq groundwater basin. Nobre et al. (2007) used GIS, modeling and a fuzzy logic tool to get the groundwater vulnerability and risk map, the groundwater vulnerability was assessed by DRASTIC, and the potential contaminant source index was evaluated by a fuzzy hierarchy methodology. Dimitriou et al. (2008) assessed the groundwater risk of a wetland catchment using hazard originating from land uses and vulnerability as a physical property.

Evaluation of groundwater pollution risk at contaminated sites has great significance for the utilization and supervision of groundwater resources. The hydrogeologic settings in the DRASTIC model have been chosen to represent areas larger than 100 acres in size, thereby limiting the DRASTIC system to use as a screening tool and not as a site assessment methodology (Aller et al. 1985). Therefore, a universal method to evaluate groundwater pollution risk with the site scale has not been constructed. DRASTIC cannot be used for groundwater vulnerability and pollution risk evaluation with the site scale.

Numerical models are often used for the management of groundwater resources. In this study, the process-based method (numerical simulation method) was applied to evaluate the groundwater pollution risk of contaminated site. MODFLOW and MT3D are probably the most frequently used models for simulating groundwater flow and solute transport (Shaki & Adeloye 2007; Abu-El-Sha'r & Hatamleh 2007). MODFLOW mainly focuses on simulating the variations of groundwater exploitation under external influences (Saeed & Ashraf 2005; Gedeon et al. 2007; Shaki & Adeloye 2007). In contrast, MT3D aims at simulating contaminant migration in groundwater, such as uranium migration in groundwater (Brown et al. 1998), nitrate migration in groundwater (Schilling et al. 2006), arsenic migration in groundwater (Davis et al. 2006; Moldovan et al. 2008), and solute exchange in surface water and groundwater (Lautz & Siegel 2006).

Chromium is one of the most common groundwater contaminants at industrially contaminated sites (Hellerich et al. 2008). Two stable oxidation states of chromium, Cr(VI) and Cr(III), are present in the environment (Bartlett 1991). In groundwater, the predominant form of chromium is the oxidized form, hexavalent chromium (Cr(VI)), present as chromate (CrO42−) and dichromate (Cr2O72−) (Piau & Aspray 2011). Cr(VI) is toxic and mobile, while Cr(III) is less toxic and less mobile because it can be easily adsorbed onto the surface of iron oxides and oxyhydroxides (Biedermann & Landolph 1990; Boni & Sbaffoni 2009; Guan et al. 2011). In the past decades, intensive interest has been drawn to the Cr contamination in soil and groundwater owing to various anthropomorphic activities (Ryan et al. 2002; Shanker et al. 2005).

Recently, more and more attentions have been paid to the Cr(VI) migration in groundwater caused by the depletion of groundwater resources. Shashidhar et al. (2007) proposed a one-dimensional mathematical model based on the laboratory column experiments to study the effectiveness of bio-barriers for the Cr(VI) contaminated confined aquifers. Hellerich et al. (2008) employed a one-dimensional transport model to simulate Cr(VI) migration through the sub-wetland aquifer at a chromium contaminated site in Connecticut. Dhakate et al. (2008) constructed a two-dimensional model via Visual MODFLOW to assess the chromite mining impact on groundwater in Sukinda chromite mining area, Orissa, India. Wanner et al. (2012a, 2012b) conducted two-dimensional model simulations to evaluate the Cr(VI) reduction efficiency of a permeable reactive barrier. Few studies have been directed to the Cr(VI) migration in groundwater at the Cr(VI) contaminated sites in China.

This study focuses on the evaluation of groundwater pollution risk at contaminated site. A chromium residue disposal site in Baotou, Inner Mongolia, China was selected as the study area. The objectives of this study are: (1) to suggest the risk evaluation method of groundwater pollution for contaminated sites; (2) to construct the groundwater flow model and solute transport model at the Cr(VI) contaminated site in Baotou; and (3) to assess the groundwater pollution risk by the constructed numerical model.

METHODS

The study area

Description of the chromium residue disposal area

The study area is located in Baotou, Inner Mongolia, which is about 900 km2 (Figure 1). In Figure 1, the top figure shows the relative location of the study area, and the bottom figure shows the cross-section from north to south (A-B as shown in top figure). The study area is enclosed by the Piedmont fault in the north, the Yellow River in the south, Xiheyan region in the west, and Dengkou region in the east. The study area consists of piedmont and alluvial plain. The elevation of the piedmont plain is between 1,020 and 1,100 meters with an average slope about 8‰, while the alluvial plain has an elevation ranging from 1,000 to 1,020 meters, with an average slope of about 1.5‰.
Figure 1

Location of the study area and hydrogeologic profile.

Figure 1

Location of the study area and hydrogeologic profile.

The Cr(VI) contaminated site is located in Hanqingba region, the central part of the study area. Approximately 50 thousand tons of chromium residues were buried in 1976. The depth of the pit was about 10 meters with a clay layer at the bottom. No other anti-seepage measures were taken at that time. On top of the chromium residue, a clay layer with a thickness of 30 cm was deposited as the impermeable layer. This contaminated site had been partially destroyed during the past 30 years, and the groundwater had been polluted by Cr(VI).

Hydrogeological conditions

The study area has two main aquifers, the confined aquifer and the unconfined aquifer. The confined aquifer is located between the piedmont fault and Lan'a fault, and composed primarily of middle, lower Pleistocene gravel and sandstone. The thickness of confined aquifer ranges from 10 to 50 meters, and gradually decreases from north to south. The unconfined aquifer consists of upper Pleistocene to Holocene sand and gravel, as well as pebbles in the north and central area, and sand in the south area. The thickness of unconfined aquifer is 20 to 30 meters in the northern and central area, while 5 to 10 meters in the southern area. The unconfined groundwater flows from northeast to southwest. Between the confined and unconfined aquifer exists the broadly distributed, thick and impermeable Pleistocene muddy clay. The present work focuses on the unconfined aquifer, because the confined aquifer has not been contaminated.

Cr(VI) contamination in groundwater

In 1999, the groundwater color in more than 20 wells near the COPR site was found to be yellow or green. Since then, the local government started to pay great attention to the groundwater Cr(VI) pollution, and groundwater quality has been regularly monitored by the Environmental Protection Bureau and Environmental Monitoring Station of Baotou. The groundwater monitoring results showed that the area with more than 0.05 mg/L Cr(VI) in groundwater was 0.166 km2 in 1999. In 2002, the area with more than 1.0 mg/L Cr(VI) in groundwater increased to 0.135 km2, in which the Cr(VI) concentrations within a 0.123 km2 area was between 0.05 and 0.1 mg/L. Further information on the groundwater Cr(VI) contamination was gained from the groundwater quality monitoring data in March 2009. Specifically, the Cr(VI) concentrations were 0.05–0.1 mg/L, 0.1–0.5 mg/L and 0.5–1 mg/L within the areas of 0.023 km2, 0.063 km2 and 0.041 km2, respectively. The Cr(VI) concentrations were 1–5 and >5 mg/L within the areas of 0.089 and 0.141 km2. It can be concluded that the Cr(VI) contamination plume had spread from north to south, which could potentially threat the downstream groundwater.

Numerical simulation method

Hydrogeological conceptual model

Based on the hydrogeological conditions of the study area, the aquifer can be divided into two types: unconfined and confined. Numerical simulation of the unconfined groundwater contamination is the major objective of this study. The southern boundary of the study area is Yellow River. The river and groundwater have closely hydraulic connection in the numerical simulation process, and the river can be identified as the specified head boundary. The western boundary is perpendicular to the contour lines of groundwater level, which is thus defined as no-flow boundary. The northern and eastern boundaries are considered as general head boundary. The top elevation of the unconfined aquifer is the upper boundary, and the bottom elevation of the aquifer is defined as the lower boundary. The groundwater is recharged by rainfall infiltration, surface water leakage and irrigation seepage through the upper boundary, and discharged by evaporation and artificial exploitation through this boundary. The groundwater flow can be considered as a two-dimensional flow, which follows Darcy's law at normal temperatures and pressures. Also, the spatial and temporal variations in groundwater level were small in the study area, so the groundwater flow can be considered as a steady flow. In addition, the groundwater system in the study area can be considered as heterogeneous and isotropic.

Mathematical model of groundwater flow system

Mathematical model
The unconfined water flow model includes the control equation, the initial conditions and boundary conditions: 
formula
1
where represents the study area; denotes the Dirichlet boundary; denotes the Neumann boundary; is the initial water level, L; is the water level along the Dirichlet boundary, L; is the flow rate for the Neumann boundary, ; x and y are the space coordinates, L; t is time, T; is the direction normal to the boundary; h is groundwater level, L; represents the sources and sinks, ; k is the hydraulic conductivity, ; B is the bottom elevation of unconfined aquifer, L; is the specific yield.
Groundwater flow model

The groundwater flow model was constructed using MODFLOW (Harbaugh & McDonald 1996). The study area was divided into fine rectangular grids, and the layer has a total of 98,793 effective units. The grid size is 20 m × 20 m in the adjacent area of COPR disposal site, and 100 m × 100 m in other areas.

The top elevation of the layer was imported from the terrain data, while the bottom elevation of the layer was imported from borehole data, and the initial head was imported from the measured data. The aquifer layer was divided into 18 partitions according to the hydrogeological conditions (Figure 2). Each partition has the same hydrogeological parameters, and the initial parameters were obtained from the exploration data. The unconfined aquifer is mainly composed of gravel, pebbles, coarse sand and fine sand. The hydraulic conductivity was 9–50 m/d, and the specific yield was 0.20–0.26 (Table 1). The recharge terms included rainfall infiltration, surface water leakage, irrigation seepage, lateral inflows and the lateral seepage recharges from Yellow River. The discharge terms included the evaporation and artificial exploitation. The unconfined groundwater flows from northeast to southwest.
Table 1

Hydraulic conductivity and specific yield in the hydrogeological partitions

Partitions 
Hydraulic conductivity (m/d) 50 50 30 20 20 10 25 30 
Specific yield 0.26 0.26 0.20 0.20 0.20 0.20 0.20 0.20 0.20 
Partitions 10 11 12 13 14 15 16 17 18 
Hydraulic conductivity (m/d) 25 40 40 40 30 30 12 12 0.5 
Specific yield 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.07 
Partitions 
Hydraulic conductivity (m/d) 50 50 30 20 20 10 25 30 
Specific yield 0.26 0.26 0.20 0.20 0.20 0.20 0.20 0.20 0.20 
Partitions 10 11 12 13 14 15 16 17 18 
Hydraulic conductivity (m/d) 25 40 40 40 30 30 12 12 0.5 
Specific yield 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.07 

In Table 1, the hydrogeological parameters were obtained from the numerical model, and the initial parameters were from the exploration data.

Figure 2

Hydrogeological conditions partition of the aquifer layer.

Figure 2

Hydrogeological conditions partition of the aquifer layer.

Solute transport modeling

Numerical model
Based on the migration characteristics of Cr(VI) in groundwater and the hydrogeological conditions, the solute transport model, including the control equation, initial conditions and boundary conditions, was established: 
formula
2
where C is the concentration of Cr(VI), ; t is time, T; represents the distance along the Cartesian coordinate axis, L; is the hydrodynamic dispersion coefficient tensor, L2T−1; is the seepage or linear pore water velocity, ; is the volumetric flow rate per unit volume of aquifer representing fluid sources and sinks, ; is the concentration of the source or sink flux for Cr(VI), ; is the porosity of the subsurface medium, dimensionless; is the chemical reaction term, ; is the initial concentration, ; is a known function representing the dispersive flux normal to the boundary, ; denotes the entire model domain; denotes the specified-concentration boundary.
Solute transport model

Based on the historical investigation results, the initial time for the solute transport model was set as the COPR burial time (1976). The solute transport model was constructed using MT3DMS (Zheng & Wang 1999).

We calculated Cr(VI) concentrations in groundwater at different times using MT3DMS, and calibrated the solute transport model using the Cr(VI) concentrations from 1999, 2002 and 2009, respectively. The initial data of dispersion coefficient were obtained from the soil column leaching test conducted in the laboratory. The adsorption equilibrium constant was measured through the laboratory adsorption experiment. The soil samples were collected in the study area, followed by the subsequent preparation of solutions of different Cr(VI) concentrations. Based on the experiment results, the Freundlich and Langmuir equations were employed to generate the adsorption equilibrium constant. The initial dispersion coefficient and adsorption were adjusted according to the groundwater quality during the model calibration.

Groundwater pollution risk assessment

The present and forecasted pollutant concentrations and the variation trend of groundwater quality were considered in the risk evaluation of groundwater pollution. The pollution risk was assign values from 1 to 5, based on the classes from very low to very high. The present and forecasted groundwater quality was evaluated by the Chinese groundwater quality standard (GB/T 14848-93). Table 2 shows the assigned values of groundwater quality index.

Table 2

The assigned values of groundwater quality index

FCPC Class I or II Class III Class IV Class V 
Class I, II or III 3 (0 < ≤ 0.25) 
4 (0.25 < ≤ 0.33) 
5 ( > 0.33) 
Class IV 2 ( ≤ 0) 4 (0 < ≤ 0.25) 
3 ( > 0) 5 ( > 0.33) 
Class V 4 ( ≤ 0) 
5 ( > 0) 
FCPC Class I or II Class III Class IV Class V 
Class I, II or III 3 (0 < ≤ 0.25) 
4 (0.25 < ≤ 0.33) 
5 ( > 0.33) 
Class IV 2 ( ≤ 0) 4 (0 < ≤ 0.25) 
3 ( > 0) 5 ( > 0.33) 
Class V 4 ( ≤ 0) 
5 ( > 0) 

In Table 2, FC represents the forecasted pollutant concentrations in groundwater, and PC is the present pollutant concentrations.

The change rate of groundwater quality in the given period is computed by the formula: 
formula
where represents the change rate index of groundwater quality; is the pollutant concentration at a specified time in the future; is the pollutant concentration at present time, and is the given time period.

The groundwater quality is divided into five categories in the Chinese groundwater quality standard (GB/T 14848-93). Classes I and II mainly reflect the natural background content of groundwater chemical composition, class III reflect the human health benchmarks, class IV was classified by agricultural and industrial water demands, and class V reflects the undrinkable water. The classes of groundwater pollution risk are determined by the groundwater quality index. Groundwater pollution risk is divided into five classes from ‘very low’ to ‘very high’ based on the groundwater quality index from 1 to 5.

Results and discussion

Groundwater flow simulation

Under the given conditions, the spatial and temporal distribution of groundwater level can be simulated by the groundwater flow model using MODFLOW. In this simulation, the hydrogeological parameters, boundary conditions, recharge and discharge values were adjusted through model calibration, so that the model can meet the actual hydrogeological conditions. In this study, the groundwater flow model was calibrated using the water level. For the groundwater flow model, the main calibration parameters were the hydraulic conductivity and the specific yield, and the simulated values were suggested in Table 1. Deviations between the simulated and observed values in the groundwater observation wells were suggested in Figure 3.
Figure 3

Deviations between the simulated and observed values in the groundwater observation wells.

Figure 3

Deviations between the simulated and observed values in the groundwater observation wells.

As shown in Figure 3, there were some slight deviations between the simulated groundwater levels and the measured levels, among which 81% of the deviations were less than 0.5 m, and 19% were between 0.5 m and 1 m. Overall, the calibrated groundwater flow model could reflect the real hydraulic system of groundwater, and the simulation results were close to those of real groundwater flow system.

Solute transport simulation

The spatial and temporal distributions of Cr(VI) in groundwater were simulated by MT3DMS under the given conditions. The model was calibrated to meet the real hydrogeological conditions by comparing the calculated Cr(VI) concentrations with those measured over a certain period. 15 Cr(VI) concentrations in 1999, 32 Cr(VI) concentrations in 2002, and 16 Cr(VI) concentrations in 2009 were selected to calibrate the model. For the solute transport model, the primary calibration parameters were the dispersion and adsorption coefficient. Figure 4 shows the calibration plots of measured Cr(VI) concentrations versus simulated data, with the corresponding parameter values shown in Table 3. In Figure 4, A and a show the simulated and observed values of Cr(VI) concentrations in 1999, B and b show the simulated and observed values of Cr(VI) concentrations in 2002, and C and c show the simulated and observed values of Cr(VI) concentrations in 2009. The number and position of observation wells in 1999 are consistent with those in 2002, but are not the same as those in 2009.
Table 3

Observed and simulated Cr(VI) concentrations in groundwater observation wells in 1999, 2002 and 2009

1999
 
2002
 
2009
 
Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error 
S1 2.580 2.389 −0.191 S1 1.973 1.908 −0.065 S18 0.018 0.034 0.016 S34 0.394 0.412 0.018 
S3 1.794 1.494 −0.300 S2 N.D. 0.000 S19 0.018 −0.018 S35 0.020 0.038 0.018 
S4 0.005 0.000 −0.005 S4 N.D. 0.000 S20 0.012 −0.012 S36 0.047 0.069 0.022 
S5 0.120 0.080 −0.040 S5 0.125 0.017 −0.108 S21 N.D. 0.000 S37 4.238 4.044 −0.194 
S6 21.02 21.193 0.173 S6 13.821 13.806 −0.015 S22 N.D. 0.000 S38 0.380 0.448 0.068 
S7 0.094 0.166 0.072 S7 N.D. 0.038 0.038 S23 0.009 −0.009 S39 0.274 0.225 −0.049 
S8 0.005 0.008 0.003 S8 N.D. 0.003 0.003 S24 N.D. 0.000 S40 0.220 −0.220 
S9 0.011 0.47 0.459 S9 N.D. 0.416 0.416 S25 N.D. 0.019 0.019 S41 4.171 4.036 −0.135 
S10 0.014 0.025 0.011 S10 0.022 0.013 −0.009 S26 N.D. 0.000 S42 0.020 −0.020 
S12 0.020 0.000 −0.020 S11 0.352 0.390 0.038 S27 N.D. 0.000 S43 0.033 0.025 −0.008 
S13 2.580 2.74 0.160 S12 N.D. 0.000 S28 0.027 −0.027 S44 0.027 −0.027 
S14 0.008 −0.008 S13 2.167 2.137 −0.030 S29 0.009 −0.009 S45 0.027 −0.027 
S23 0.012 −0.012 S14 2.704 2.897 0.193 S30 0.029 −0.029 S46 0.033 0.001 −0.032 
S25 0.010 −0.010 S15 0.011 −0.011 S31 0.009 −0.009 S47 0.020 −0.020 
S28 N.D. 0.000 S16 9.272 9.134 −0.138 S32 N.D. 0.000 S48 0.020 −0.020 
    S17 N.D. 0.000 S33 0.309 0.212 −0.097 S49 0.320 0.048 −0.272 
1999
 
2002
 
2009
 
Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error Number of wells Observed value Simulated value Error 
S1 2.580 2.389 −0.191 S1 1.973 1.908 −0.065 S18 0.018 0.034 0.016 S34 0.394 0.412 0.018 
S3 1.794 1.494 −0.300 S2 N.D. 0.000 S19 0.018 −0.018 S35 0.020 0.038 0.018 
S4 0.005 0.000 −0.005 S4 N.D. 0.000 S20 0.012 −0.012 S36 0.047 0.069 0.022 
S5 0.120 0.080 −0.040 S5 0.125 0.017 −0.108 S21 N.D. 0.000 S37 4.238 4.044 −0.194 
S6 21.02 21.193 0.173 S6 13.821 13.806 −0.015 S22 N.D. 0.000 S38 0.380 0.448 0.068 
S7 0.094 0.166 0.072 S7 N.D. 0.038 0.038 S23 0.009 −0.009 S39 0.274 0.225 −0.049 
S8 0.005 0.008 0.003 S8 N.D. 0.003 0.003 S24 N.D. 0.000 S40 0.220 −0.220 
S9 0.011 0.47 0.459 S9 N.D. 0.416 0.416 S25 N.D. 0.019 0.019 S41 4.171 4.036 −0.135 
S10 0.014 0.025 0.011 S10 0.022 0.013 −0.009 S26 N.D. 0.000 S42 0.020 −0.020 
S12 0.020 0.000 −0.020 S11 0.352 0.390 0.038 S27 N.D. 0.000 S43 0.033 0.025 −0.008 
S13 2.580 2.74 0.160 S12 N.D. 0.000 S28 0.027 −0.027 S44 0.027 −0.027 
S14 0.008 −0.008 S13 2.167 2.137 −0.030 S29 0.009 −0.009 S45 0.027 −0.027 
S23 0.012 −0.012 S14 2.704 2.897 0.193 S30 0.029 −0.029 S46 0.033 0.001 −0.032 
S25 0.010 −0.010 S15 0.011 −0.011 S31 0.009 −0.009 S47 0.020 −0.020 
S28 N.D. 0.000 S16 9.272 9.134 −0.138 S32 N.D. 0.000 S48 0.020 −0.020 
    S17 N.D. 0.000 S33 0.309 0.212 −0.097 S49 0.320 0.048 −0.272 

In Table 3, the unit of Cr(VI) concentrations is mg/L (milligrams per liter). N.D. = not detected. In calculating the error, N.D. is considered as zero.

Figure 4

Simulated and observed values of Cr(VI) concentrations in groundwater observation wells in 1999, 2002 and 2009, respectively.

Figure 4

Simulated and observed values of Cr(VI) concentrations in groundwater observation wells in 1999, 2002 and 2009, respectively.

Apparently, the differences between the measured and simulated data were within the error range. The error range was suggested as 0.01 mg/L (Cr(VI) concentration less than 0.1 mg/L), 0.1 mg/L (Cr(VI) concentration between 0.1 and 1 mg/L) and 0.2 mg/L (Cr(VI) concentration greater than 1 mg/L). The simulated migration rates of Cr(VI) in 1976–1999 and 1999–2009 periods were 47.9 m/yr, 30.4 m/yr, respectively. Overall, the accuracies of the calibration results of the groundwater flow model and the solute transport model are acceptable. The established solute transport model can reflect the real hydrogeological conditions and the Cr(VI) concentrations in groundwater, and therefore the calibrated model is expected to be able to forecast the changes of Cr(VI) concentrations in groundwater.

Groundwater pollution risk assessment

Remediation of polluted soil and groundwater needs a huge amount of funds, and the different periods of groundwater pollution risk and management countermeasures for contaminated sites should be considered in China. The groundwater pollution risk is commonly evaluated by overlaying the results of groundwater vulnerability and pollution sources, and the time factor is rarely considered. This method is applicable for the regions with large area, and unapplicable for the contaminated sites. The very high class of groundwater pollution risk could be usually obtained by overlaying the results of groundwater vulnerability and pollution sources, while these results cannot meet the current demand of groundwater pollution regulations in China. Therefore, the risk evaluation method considering the time factor was suggested in this research, and the regulation measures of groundwater pollution were put forward.

Cr(VI) concentrations in groundwater under different conditions could be simulated and computed by the numerical model, and the computed concentrations could be used to get the risk classes. The spatial and temporal distributions of Cr(VI) concentrations in groundwater under different scenarios, including chromium residue removal in 2010, chromium residue removal in 2030, and non-removal of chromium residue, were investigated in this study. The simulation period spans from 2010 to 2060, which was divided into 100 stress periods, and each period included 180 days. Figure 5 shows the predicted Cr(VI) concentrations under these three scenarios. In scenario 1, the chromium residue had been removed in 2010, and the Cr(VI) concentrations in 2015, 2020, 2030, 2040, 2050 and 2060 are shown in figure A, B, C, D, E and F, respectively. In scenario 2, the chromium residue will be removed in 2030, and the Cr(VI) concentrations in 2035, 2040, 2045, 2050, 2055 and 2060 are shown in figure A, B, C, D, E and F, respectively. In scenario 3, the chromium residue will not be removed, and Cr(VI) concentrations in 2015, 2020, 2030, 2040, 2050 and 2060 are shown in figure A, B, C, D, E and F, respectively.
Figure 5

Spatial and temporal distribution of Cr(VI) concentrations in groundwater under three scenarios.

Figure 5

Spatial and temporal distribution of Cr(VI) concentrations in groundwater under three scenarios.

Based on the simulation results, we conclude that the Cr(VI) contaminated groundwater will migrate along with the groundwater flow. The migration rate of Cr(VI) will be about 23 m/yr from 2010 to 2060, and the Cr(VI) pollution zone varies under the three scenarios. These results show that the Cr(VI) pollution area under the COPR removal scenarios is smaller than that under COPR non-removal scenario. The sooner the removal of COPR, the smaller the Cr(VI) pollution area will be. Based on the computed results by the numerical model, the groundwater pollution risk could be divided into four classes (very low, low, high and very high) as follows (Figure 6).
Figure 6

Groundwater pollution risk classes at the Cr(VI) contaminated site in 2015.

Figure 6

Groundwater pollution risk classes at the Cr(VI) contaminated site in 2015.

The regions with very high pollution risk

The Cr(VI) concentrations in groundwater were greater than 0.1 mg/L in 2009, and the Cr(VI) concentrations will be greater than 1 mg/L in 2015. The pollution risk of groundwater in this area is extremely large, and the costs of pollution remediation and management is very high. The chromium residue should be removed, and the contaminated groundwater and soil should be remediated in this area.

The regions with high pollution risk

The Cr(VI) concentrations in groundwater were less than 0.5 mg/L in 2009, and the Cr(VI) concentrations will be 0.1–1 mg/L in 2015. The pollution risk of groundwater in this area is large, and the groundwater pollution control costs is high. The groundwater quality should be closely monitored, and the effective measures should be taken to control the groundwater pollution.

The regions with low pollution risk

The Cr(VI) concentrations in groundwater were less than 0.05 mg/L in 2009, and the Cr(VI) concentrations will be 0.05–0.1 mg/L in 2015. The pollution risk of groundwater in this area is low, and the variation of groundwater quality should be monitored.

The regions with very low pollution risk

The Cr(VI) concentrations in groundwater will be less than 0.05 mg/L before 2015. The groundwater pollution risk is very low, and the groundwater quality near the regions with low pollution risk should be monitored.

Therefore, we strongly recommend the local government to remove the COPR as soon as possible, and the polluted groundwater should not be directly used for drinking, agricultural irrigation and industrial application. Monitoring and management measures should be taken to prevent groundwater pollution.

CONCLUSIONS

The objective of this study is to suggest the risk evaluation method of groundwater pollution at contaminated sites. The main conclusions of this study include the following.

The risk evaluation method considering the time factor for groundwater pollution at contaminated sites was suggested. The present and forecasted pollutant concentrations and the variation trend of groundwater quality were considered in the suggested risk evaluation method.

The numerical model of groundwater flow and solute transport at the Cr(VI) contaminated site was constructed. Based on the simulation results, Cr(VI) in contaminated groundwater will migrate along with the groundwater flow regardless of chromium residue removal or not. The migration rate of Cr(VI) is about 23 m/yr from 2010 to 2060.

The groundwater pollution risk of the contaminated site could be divided into four classes (very low, low, high and very high). Upon the groundwater flow, Cr(VI) will migrate fast and Cr(VI) will be retained in groundwater for a long time, which may result in severe pollution in groundwater. The groundwater in the study area is facing high risk of Cr(VI) contamination. Therefore, we strongly recommend the local government to remove the chromium residue as soon as possible. Moreover, effective measures should be taken to prevent and control groundwater contamination, so that the agriculture, industry and domestic water use in the study area cannot be threatened by the groundwater pollution.

ACKNOWLEDGEMENT

This work was financially supported by the National Natural Science Foundation of China (No. 41471266).

REFERENCES

REFERENCES
Abu-El-Sha'r
W. Y.
Hatamleh
R. I.
2007
Using Modflow and MT3D groundwater flow and transport models as a management tool for the Azraq groundwater system
.
Jordan Journal of Civil Engineering
1
,
153
172
.
Aller
L.
Bennett
T.
Lehr
J. H.
Petty
R. J.
1985
DRASTIC: A standardized system for evaluating groundwater pollution potential using hydrogeologic settings
.
EPA 600/2–85/018, Washington, DC
.
Bartlett
R. J.
1991
Chromium cycling in soils and water: links, gaps, and methods
.
Environmental Health Perspectives
92
,
17
24
.
Biedermann
K. A.
Landolph
J. R.
1990
Role of valence state and solubility of chromium compounds on induction of cytotoxicity, mutagenesis, and anchorage independence in diploid human fibroblasts
.
Cancer Research
50
,
7835
7842
.
Brown
P. L.
Guerin
M.
Hankin
S. I.
Lowson
R. T.
1998
Uranium and other contaminant migration in groundwater at a tropical Australian uranium mine
.
Journal of Contaminant Hydrology
35
,
295
303
.
Davis
A.
Bellehumeur
T.
Hunter
P.
Hanna
B.
Fennemore
G. G.
Moomaw
C.
Schoen
S.
2006
The nexus between groundwater modeling, pit lake chemogenesis and ecological risk from arsenic in the Getchell Main Pit, Nevada, U.S.A
.
Chemical Geology
228
,
175
196
.
Dimitriou
E.
Karaouzas
I.
Sarantakos
K.
Zacharias
I.
Bogdanos
K.
Diapoulis
A.
2008
Groundwater risk assessment at a heavily industrialised catchment and the associated impacts on a peri-urban wetland
.
Journal of Environmental Management
88
,
526
538
.
Gedeon
M.
Wemaere
I.
Marivoet
J.
2007
Regional groundwater model of north-east Belgium
.
Journal of Hydrology
335
,
133
139
.
Guan
X. H.
Dong
H. R.
Ma
J.
Lo
I. M. C.
2011
Simultaneous removal of chromium and arsenate from contaminated groundwater by ferrous sulfate: Batch uptake behavior
.
Journal of Environmental Sciences
23
(
3
),
372
380
.
Harbaugh
A. W.
McDonald
M. G.
1996
User's documentation for MODFLOW-96, an update to the U.S. Geological Survey modular finite-difference groundwater flow model
.
Open-File Report 96–485
,
US Geological Survey
,
Reston VA
.
Hellerich
L. A.
Nikolaidis
N. P.
Dobbs
G. M.
2008
Evaluation of the potential for the natural attenuation of hexavalent chromium within a sub-wetland ground water
.
Journal of Environmental Management
88
,
1513
1524
.
Lautz
L. K.
Siegel
D. I.
2006
Modeling surface and ground water mixing in the hyporheic zone using MODFLOW and MT3D
.
Advances in Water Resources
29
,
1618
1633
.
Nobre
R. C. M.
Filho
O. C. R.
Mansur
W. J.
Nobre
M. M. M.
Cosenza
C. A. N.
2007
Groundwater vulnerability and risk mapping using GIS, modeling and a fuzzy logic tool
.
Journal of Contaminant Hydrology
94
(
3–4
),
277
292
.
Ryan
M. P.
Williams
D. E.
Chater
R. J.
Hutton
B. M.
McPhail
D. S.
2002
Why stainless steel corrodes
.
Nature
415
,
770
774
.
Shanker
A. K.
Cervantes
C.
Loza-Tavera
H.
Avudainayagam
S.
2005
Chromium toxicity in plants
.
Environment International
31
,
739
753
.
Thirumalaivasan
D.
Karmegam
M.
Venugopal
K.
2003
AHP-DRASTIC: software for specific aquifer vulnerability assessment using DRASTIC model and GIS
.
Environmental Modelling & Software
18
(
7
),
645
656
.
Wang
J. J.
He
J. T.
Chen
H. H.
2012
Assessment of groundwater contamination risk using hazard quantification, a modified DRASTIC model and groundwater value
.
Beijing Plain
,
China
.
Zheng
C.
Wang
P. P.
1999
MT3DMS: A modular three-dimensional multispecies transport model for simulation of advection, dispersion, and chemical reactions of contaminants in groundwater systems: documentation and user's guide. SERDP-99–1, US Army Engineer Research and Development Center, Vicksburg, MS
.