- Email: [email protected]
A simple methodology for the evaluation of groundwater pollution risks
Science of the Total Environment 378 (2007) 67 – 70 www.elsevier.com/locate/scitotenv
A simple methodology for the evaluation of groundwater pollution risks J. Fernández-Gálvez 1 , E. Barahona 1 , A. Iriarte 1 , M.D. Mingorance ⁎ Department of Earth Science and Environmental Chemistry, Estación Experimental del Zaidín (CSIC), Profesor Albareda 1, E-18008 Granada, Spain Available online 6 February 2007
Abstract Groundwater represents a very significant source of fresh water for irrigation and drinking purposes and therefore preserving the availability and quality of this resource is extremely important. A first assessment of the amount of pollutants that can be exported to groundwater via soil drainage can be made by a) measuring the amount of contaminants present in the soil solution at the bottom of the soil after a prolonged simulated rainfall event, and b) estimating the amount of drainage water passing the soil bottom during a period of time long enough to include sufficient instances of both, wet and dry episodes inherent to the local climate. Drainage water was estimated by means of a simple infiltration model (“bucket model”) which computes on a daily basis the inputs and outputs of soil water through rainfall and evapotranspiration generated by a stochastic model of the local climate along a period of 50–100 years. The methodology was applied in the Guadiamar valley after the toxic spill of a pyrite mine in Aznalcóllar (Spain). The results show that Zn is the dominant contaminant at the site with a 1.2 g m− 2year− 1 contribution to groundwater. The presence of a gravel rich horizon below 50 cm depth reveals an increase in drainage and the threat to groundwater. © 2007 Elsevier B.V. All rights reserved. Keywords: Groundwater pollution; Infiltration; Simulation
1. Introduction When there is a change in soil use and management, and even more, when there is a soil pollution episode, it is important to account for groundwater pollution risk. To evaluate this risk, it is necessary to account for a number of factors such as a) the contaminant mobility, which could vary in time and it is a function of the interaction with the active soil components, primarily clay, organic matter, carbonates and aluminium and iron oxy-hydro⁎ Corresponding author. Tel.: +34 958 18 16 00; fax: +34 958 12 96 00. E-mail address: [email protected] (M.D. Mingorance). 1 Tel.: +34 958 18 16 00; fax: +34 958 12 96 00. 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.01.015
xides. The fixation of contaminants in the soil is not an instantaneous process, it can have a complex temporal dynamic depending on physic-chemical processes and in most cases it is difficult to predict. A simple and direct way to assess contaminants mobility is by determining their abundance in the soil solution. Although a progressive fixation could take place with time, the abundance could provide an idea of what would happen in the worst case, which, unquestionably, is very valuable information for evaluating the risk. b) On the other hand, it is also necessary to account for the intrinsic variability in rainfall and evapotranspiration, being both variables directly linked with drainage. It is not possible to compute drainage based only on average climatic values, without
68
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
taking in account extreme events such as intensive rains or droughts. c) Finally, other factors determining the amount of water reaching the water table are the soil water retention and transmission characteristics. These characteristics depend basically on particle size distribution and the degree of compaction of the fine earth. At pedon level, it is possible to obtain accurate measurements of the physical characteristics of the fine earth; but this is not the case for coarse fragments, which usually are only roughly estimated by visual observations. The paradox is that the accuracy for calculating soil water retention capacity is limited by the accuracy on calculating coarse fragments. Besides this, it is also important to note the wide variety of soils that can occur in a relatively small stretch landscape. For convenience due to the large number of uncertainties, it would be ideal to use relatively simple and easy to use tools for measuring a wide variety of conditions. A methodology for evaluating groundwater pollution risk with easy experiments is presented in this work.
(∼ 200 mm of H2O) to simulate contaminant leaching by rainfall. Rainfall was applied at low intensity to avoid surface runoff. The saturation of the soil column was completed in about 20 h. Extraction of the soil solution started when the wetting front reaches the Rhizom filters. 3) Samples from the same horizon were mixed together (3 samples per horizon), acidified with HNO3 and stored in refrigerated chamber before transport to the laboratory. As, Cu and Zn were determined in soil solution by ICP-MS. 2.2. Soil analysis
Experiments described in this work took place in the Guadiamar valley at the site called “El Vicario” after the toxic spill of a pyrite mine in Aznalcóllar (Spain) but the methodology is equally applicable to other sites.
Soil samples were taken during the excavation of the soil pit. Fine earth samples were taken for particle size analysis (Robinson's pipette, Soil Conservation Service, 1972), organic matter determinations (Walkley and Black, 1934) and water retention at 33 and 1500 kPa using the Richards' (1954) membrane. Undisturbed samples were also taken for bulk density determinations using the cylinder method. In one of the horizons where coarse fragment was very abundant bulk density was measured by excavation of a known soil volume. The amount of soil extracted from this excavation was also used for gravimetric determination of coarse fragments.
2.1. Contaminants mobility determination
2.3. Stochastic simulation for infiltration
At the Aznalcollar toxic spill, contamination primary affected the upper soil layers and contaminants were released into the soil mainly through pyrite mud oxidation (Aguilar et al., 2003). Elements released were mainly As, Cu, and Zn. The procedure to simulate the elements mobility due to rainfall infiltration throughout the soil profile is describe as follows:
Simulations were carried out on a program written in PYTHON language (van Rossum, 1990) with interface to the statistical language R (R Development Core Team, 2004). Python and R are free software available for all platforms (the script used for simulations is available from authors upon request). The movement of water through the soil is computed based on a 1 cm thick layers bucket model where the following assumptions are made:
2. Material and methods
1) A 1.2 m depth pit was dig taking both fine earth and undisturbed samples for bulk density determinations. The front site of the pit was cover with SARAN (Trademark of The Dow Chemical Company, http:// www.dow.com/saran/resins/index.htm) resin dissolved in ethyl-methyl-ketone. This mix when dry produces a waterproof flexible plastic coating. Rhizom filters were inserted in each of the principal soil horizons for soil solution extraction. Each filter is connected to a syringe that produces vacuum where the soil solution is stored. 2) The squared drop-former chamber (50 cm long side with drippers spaced 1.5 cm) of a rainfall simulator was put in place over the soil surface and continuous rainfall was applied until soil column saturation
1) All rainfall from water infiltrates into the soil; there is no surface runoff. 2) Water that drains from each layer is only the one above field capacity. Drainage stops when field capacity is reached. 3) For each layer, water losses through evaporation stop when the permanent wilting point is reached. 4) Continuous descending flow of water will occur either while water is present at the soil surface or if drainage water is present at any within of the soil profile. 5) Evapotranspiration is driven by the potential evapotranspiration and its magnitude is either the potential
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
69
Table 1 Soil analytical characteristics Horizon
Depth cm
Sand %
Silt %
Clay %
Organic matter %
Bulk density g cm− 3
Gravels %
W1/3 %
W15 %
Ap AB B C
0–20 20–30 30–50 N50
55.9 68.4 81.3 95.8
1.5 0.4 0.7 0.5
42.6 31.2 18.0 3.7
1.2 0.6 0.2 0.1
1.28 1.29 1.25 1.30
6.34 4.80 0.42 35.50
23.9 20.5 13.8 4.5
13.0 9.9 6.5 4.1
value, or the maximum amount of water that could be evaporated from the amount stored in the soil profile (soil water above permanent wilting point). Water balance is computed at daily time steps. During dry days, when there is no rainfall, only water losses due to evapotranspiration are computed. Concerning the climatic stochastic simulation, inputs data are obtained by climatic records form nearby meteorological stations. Inputs used are: average monthly rainfall, number of rainy days per month, maximum daily rainfall within each month and average monthly evapotranspiration. The number of rainy days within a month is described by a Poisson distribution and the monthly rainfall by a Weibull distribution. The magnitude parameter in the Weibull distribution is obtained from monthly rainfall and the shape parameter is chosen to reproduce the maximum rainfall within each month. Satisfactory results are obtained in most cases with a shape parameter between 0.8 and 1.2. Simulation is carried out for a large number of years (50 to 100 years) in order to reproduce climate variability at the site. Principal program outputs are the amount of water infiltrated at each depth along the simulated period and
Fig. 1. Contaminants present in the drainage water.
drainage at the bottom of the profile. Other complementary outputs are: monthly rainfall and evapotranspiration stochastically generated and runoff. Runoff could appear if there is a prolonged saturation of the soil profile. 3. Results and discussion Table 1 shows soil analytical characteristics at the experimental site. Only characteristics related with water retention in the soil profile are presented. It can be observed that the available water capacity (W1/3–W15) notably decreases for horizon C (sandy layer with high gravel content). The only data needed for simulations are bulk density, gravel content and water retention at 1/3 and 15 bars (W1/13, W15). In Fig. 1 the concentration of contaminants extracted with the Rhizom filters from the soil profile is presented. Zn is the dominant contaminant with a maximum concentration at 12 cm depth over 15 mg/L decreasing to around 6 mg/L at the C horizon. At this last horizon, concentrations of As and Cu are negligible. Data from
Fig. 2. Accumulated mean yearly drainage at each depth for both soil profiles including the sandy and gravelly C horizon and sandy loam profile.
70
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
this last horizon are the most important groundwater pollution risk to evaluate. Fig. 2 represents the yearly average water infiltrated at each depth (mm) obtained from 50 year simulations. Water infiltrated in the soil suddenly decreases from 600 to nearly 200 mm/year at 50 cm depth, exactly where the sandy and gravelly C horizon starts. At 150 cm depth the average water infiltrated keeps close to 200 mm. Under this circumstances, if Zn mobility is expected to be fixed, exportation of Zn to groundwater would be 1.2 g·m− 2year− 1 (6 mg/L Zn ⁎ 200 L m− 2 year− 1). It is important to note the effect of the gravel layer in the soil profile. Due to its low water retention capacity, evapotranspiration from this layer is also low, increasing drainage and therefore the risk of groundwater pollution. Fig. 2 represents simulation results considering that the sandy loam B horizon extends down to 150 cm depth. For this last case water drainage decreases about 100 mm and the amount of Zn exported from the soil is approximately half. 4. Conclusions An easy to use methodology for evaluation of groundwater pollution risk is developed that could be
used as a first evaluation for risk under a wide range of scenarios. The findings pointed out that the presence of a gravel rich horizon below 50 cm depth makes known an increase in drainage and consequently in the risk of groundwater contamination. References Aguilar J, Dorronsoro C, Bellver R, Fernández E, Fernández J, García I, et al. Contaminación de los suelos tras el vertido tóxico de Aznalcóllar. Departamento de Edafología y Química Agrícola. Universidad de Granada; 2003. 184 pp. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria; 2004. http://www. R-project.org. Richards LA. Diagnosis and improvement of saline and alkali soils. U.S. Salinity Laboratory, USDA, vol. 60. Handbook; 1954. 160 pp. Soil Conservation Service. Soil Survey Laboratory Methods and Procedures for Collecting Samples. Washington: USDA; 1972. 6 pp. van Rossum G. Python's integrated development environment; 1990. http://www.python.org. Walkley A, Black IA. An examination of the Degtjareff method for determining soil organic matter and a proposed modification of the chromic acid titration method. Soil Sci 1934;37:29–38.
A simple methodology for the evaluation of groundwater pollution risks J. Fernández-Gálvez 1 , E. Barahona 1 , A. Iriarte 1 , M.D. Mingorance ⁎ Department of Earth Science and Environmental Chemistry, Estación Experimental del Zaidín (CSIC), Profesor Albareda 1, E-18008 Granada, Spain Available online 6 February 2007
Abstract Groundwater represents a very significant source of fresh water for irrigation and drinking purposes and therefore preserving the availability and quality of this resource is extremely important. A first assessment of the amount of pollutants that can be exported to groundwater via soil drainage can be made by a) measuring the amount of contaminants present in the soil solution at the bottom of the soil after a prolonged simulated rainfall event, and b) estimating the amount of drainage water passing the soil bottom during a period of time long enough to include sufficient instances of both, wet and dry episodes inherent to the local climate. Drainage water was estimated by means of a simple infiltration model (“bucket model”) which computes on a daily basis the inputs and outputs of soil water through rainfall and evapotranspiration generated by a stochastic model of the local climate along a period of 50–100 years. The methodology was applied in the Guadiamar valley after the toxic spill of a pyrite mine in Aznalcóllar (Spain). The results show that Zn is the dominant contaminant at the site with a 1.2 g m− 2year− 1 contribution to groundwater. The presence of a gravel rich horizon below 50 cm depth reveals an increase in drainage and the threat to groundwater. © 2007 Elsevier B.V. All rights reserved. Keywords: Groundwater pollution; Infiltration; Simulation
1. Introduction When there is a change in soil use and management, and even more, when there is a soil pollution episode, it is important to account for groundwater pollution risk. To evaluate this risk, it is necessary to account for a number of factors such as a) the contaminant mobility, which could vary in time and it is a function of the interaction with the active soil components, primarily clay, organic matter, carbonates and aluminium and iron oxy-hydro⁎ Corresponding author. Tel.: +34 958 18 16 00; fax: +34 958 12 96 00. E-mail address: [email protected] (M.D. Mingorance). 1 Tel.: +34 958 18 16 00; fax: +34 958 12 96 00. 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.01.015
xides. The fixation of contaminants in the soil is not an instantaneous process, it can have a complex temporal dynamic depending on physic-chemical processes and in most cases it is difficult to predict. A simple and direct way to assess contaminants mobility is by determining their abundance in the soil solution. Although a progressive fixation could take place with time, the abundance could provide an idea of what would happen in the worst case, which, unquestionably, is very valuable information for evaluating the risk. b) On the other hand, it is also necessary to account for the intrinsic variability in rainfall and evapotranspiration, being both variables directly linked with drainage. It is not possible to compute drainage based only on average climatic values, without
68
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
taking in account extreme events such as intensive rains or droughts. c) Finally, other factors determining the amount of water reaching the water table are the soil water retention and transmission characteristics. These characteristics depend basically on particle size distribution and the degree of compaction of the fine earth. At pedon level, it is possible to obtain accurate measurements of the physical characteristics of the fine earth; but this is not the case for coarse fragments, which usually are only roughly estimated by visual observations. The paradox is that the accuracy for calculating soil water retention capacity is limited by the accuracy on calculating coarse fragments. Besides this, it is also important to note the wide variety of soils that can occur in a relatively small stretch landscape. For convenience due to the large number of uncertainties, it would be ideal to use relatively simple and easy to use tools for measuring a wide variety of conditions. A methodology for evaluating groundwater pollution risk with easy experiments is presented in this work.
(∼ 200 mm of H2O) to simulate contaminant leaching by rainfall. Rainfall was applied at low intensity to avoid surface runoff. The saturation of the soil column was completed in about 20 h. Extraction of the soil solution started when the wetting front reaches the Rhizom filters. 3) Samples from the same horizon were mixed together (3 samples per horizon), acidified with HNO3 and stored in refrigerated chamber before transport to the laboratory. As, Cu and Zn were determined in soil solution by ICP-MS. 2.2. Soil analysis
Experiments described in this work took place in the Guadiamar valley at the site called “El Vicario” after the toxic spill of a pyrite mine in Aznalcóllar (Spain) but the methodology is equally applicable to other sites.
Soil samples were taken during the excavation of the soil pit. Fine earth samples were taken for particle size analysis (Robinson's pipette, Soil Conservation Service, 1972), organic matter determinations (Walkley and Black, 1934) and water retention at 33 and 1500 kPa using the Richards' (1954) membrane. Undisturbed samples were also taken for bulk density determinations using the cylinder method. In one of the horizons where coarse fragment was very abundant bulk density was measured by excavation of a known soil volume. The amount of soil extracted from this excavation was also used for gravimetric determination of coarse fragments.
2.1. Contaminants mobility determination
2.3. Stochastic simulation for infiltration
At the Aznalcollar toxic spill, contamination primary affected the upper soil layers and contaminants were released into the soil mainly through pyrite mud oxidation (Aguilar et al., 2003). Elements released were mainly As, Cu, and Zn. The procedure to simulate the elements mobility due to rainfall infiltration throughout the soil profile is describe as follows:
Simulations were carried out on a program written in PYTHON language (van Rossum, 1990) with interface to the statistical language R (R Development Core Team, 2004). Python and R are free software available for all platforms (the script used for simulations is available from authors upon request). The movement of water through the soil is computed based on a 1 cm thick layers bucket model where the following assumptions are made:
2. Material and methods
1) A 1.2 m depth pit was dig taking both fine earth and undisturbed samples for bulk density determinations. The front site of the pit was cover with SARAN (Trademark of The Dow Chemical Company, http:// www.dow.com/saran/resins/index.htm) resin dissolved in ethyl-methyl-ketone. This mix when dry produces a waterproof flexible plastic coating. Rhizom filters were inserted in each of the principal soil horizons for soil solution extraction. Each filter is connected to a syringe that produces vacuum where the soil solution is stored. 2) The squared drop-former chamber (50 cm long side with drippers spaced 1.5 cm) of a rainfall simulator was put in place over the soil surface and continuous rainfall was applied until soil column saturation
1) All rainfall from water infiltrates into the soil; there is no surface runoff. 2) Water that drains from each layer is only the one above field capacity. Drainage stops when field capacity is reached. 3) For each layer, water losses through evaporation stop when the permanent wilting point is reached. 4) Continuous descending flow of water will occur either while water is present at the soil surface or if drainage water is present at any within of the soil profile. 5) Evapotranspiration is driven by the potential evapotranspiration and its magnitude is either the potential
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
69
Table 1 Soil analytical characteristics Horizon
Depth cm
Sand %
Silt %
Clay %
Organic matter %
Bulk density g cm− 3
Gravels %
W1/3 %
W15 %
Ap AB B C
0–20 20–30 30–50 N50
55.9 68.4 81.3 95.8
1.5 0.4 0.7 0.5
42.6 31.2 18.0 3.7
1.2 0.6 0.2 0.1
1.28 1.29 1.25 1.30
6.34 4.80 0.42 35.50
23.9 20.5 13.8 4.5
13.0 9.9 6.5 4.1
value, or the maximum amount of water that could be evaporated from the amount stored in the soil profile (soil water above permanent wilting point). Water balance is computed at daily time steps. During dry days, when there is no rainfall, only water losses due to evapotranspiration are computed. Concerning the climatic stochastic simulation, inputs data are obtained by climatic records form nearby meteorological stations. Inputs used are: average monthly rainfall, number of rainy days per month, maximum daily rainfall within each month and average monthly evapotranspiration. The number of rainy days within a month is described by a Poisson distribution and the monthly rainfall by a Weibull distribution. The magnitude parameter in the Weibull distribution is obtained from monthly rainfall and the shape parameter is chosen to reproduce the maximum rainfall within each month. Satisfactory results are obtained in most cases with a shape parameter between 0.8 and 1.2. Simulation is carried out for a large number of years (50 to 100 years) in order to reproduce climate variability at the site. Principal program outputs are the amount of water infiltrated at each depth along the simulated period and
Fig. 1. Contaminants present in the drainage water.
drainage at the bottom of the profile. Other complementary outputs are: monthly rainfall and evapotranspiration stochastically generated and runoff. Runoff could appear if there is a prolonged saturation of the soil profile. 3. Results and discussion Table 1 shows soil analytical characteristics at the experimental site. Only characteristics related with water retention in the soil profile are presented. It can be observed that the available water capacity (W1/3–W15) notably decreases for horizon C (sandy layer with high gravel content). The only data needed for simulations are bulk density, gravel content and water retention at 1/3 and 15 bars (W1/13, W15). In Fig. 1 the concentration of contaminants extracted with the Rhizom filters from the soil profile is presented. Zn is the dominant contaminant with a maximum concentration at 12 cm depth over 15 mg/L decreasing to around 6 mg/L at the C horizon. At this last horizon, concentrations of As and Cu are negligible. Data from
Fig. 2. Accumulated mean yearly drainage at each depth for both soil profiles including the sandy and gravelly C horizon and sandy loam profile.
70
J. Fernández-Gálvez et al. / Science of the Total Environment 378 (2007) 67–70
this last horizon are the most important groundwater pollution risk to evaluate. Fig. 2 represents the yearly average water infiltrated at each depth (mm) obtained from 50 year simulations. Water infiltrated in the soil suddenly decreases from 600 to nearly 200 mm/year at 50 cm depth, exactly where the sandy and gravelly C horizon starts. At 150 cm depth the average water infiltrated keeps close to 200 mm. Under this circumstances, if Zn mobility is expected to be fixed, exportation of Zn to groundwater would be 1.2 g·m− 2year− 1 (6 mg/L Zn ⁎ 200 L m− 2 year− 1). It is important to note the effect of the gravel layer in the soil profile. Due to its low water retention capacity, evapotranspiration from this layer is also low, increasing drainage and therefore the risk of groundwater pollution. Fig. 2 represents simulation results considering that the sandy loam B horizon extends down to 150 cm depth. For this last case water drainage decreases about 100 mm and the amount of Zn exported from the soil is approximately half. 4. Conclusions An easy to use methodology for evaluation of groundwater pollution risk is developed that could be
used as a first evaluation for risk under a wide range of scenarios. The findings pointed out that the presence of a gravel rich horizon below 50 cm depth makes known an increase in drainage and consequently in the risk of groundwater contamination. References Aguilar J, Dorronsoro C, Bellver R, Fernández E, Fernández J, García I, et al. Contaminación de los suelos tras el vertido tóxico de Aznalcóllar. Departamento de Edafología y Química Agrícola. Universidad de Granada; 2003. 184 pp. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria; 2004. http://www. R-project.org. Richards LA. Diagnosis and improvement of saline and alkali soils. U.S. Salinity Laboratory, USDA, vol. 60. Handbook; 1954. 160 pp. Soil Conservation Service. Soil Survey Laboratory Methods and Procedures for Collecting Samples. Washington: USDA; 1972. 6 pp. van Rossum G. Python's integrated development environment; 1990. http://www.python.org. Walkley A, Black IA. An examination of the Degtjareff method for determining soil organic matter and a proposed modification of the chromic acid titration method. Soil Sci 1934;37:29–38.