Assessing the sorption and leaching behaviour of three sulfonamides in pasture soils through batch and column studies

Science of the Total Environment 493 (2014) 535–543

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Assessing the sorption and leaching behaviour of three sulfonamides in pasture soils through batch and column studies Prakash Srinivasan, Ajit K. Sarmah ⁎ Department of Civil & Environmental Engineering, Faculty of Engineering, The University of Auckland, Private Bag 92019, Auckland, New Zealand

H I G H L I G H T S • • • •

Sulfonamides exhibited a mobility pattern similar to that of conservative tracer. Considerable retardation was observed for the Hamilton soil. No resident concentration of sulfamethazine was detected at any soil depth. Model estimated and the experimental retardation factors deferred.

a r t i c l e

i n f o

Article history: Received 30 April 2014 Received in revised form 29 May 2014 Accepted 10 June 2014 Available online 27 June 2014 Editor: D. Barcelo Keywords: Sulfamethoxazole Sulfachloropyridazine Sulfamethazine Breakthrough curves Partition coefficients Retardation factor

a b s t r a c t We investigated the sorption potential and transport behaviour of three sulfonamides, namely, sulfamethoxazole (SMO), sulfachloropyridazine (SCP) and sulfamethazine (SM), and a conservative bromide tracer (Br−) in two undisturbed soil columns collected from the dairy farming regions in the North Island of New Zealand. Based on the low log Koc values obtained from the sorption study, all three sulfonamides are likely to have high mobility, making them a potential threat to surface and ground water. Soil column studies also showed that the mobility of the sulfonamides varied among soils and antibiotic type. Sulfonamides exhibited a mobility pattern similar to that of conservative Br− tracer. Considerable retardation was observed for the Hamilton soil, and the delayed peak arrival time (or maxima) was due to the role of sorption-related retention processes under saturated flow conditions. Residual antibiotic concentrations for SMO and SCP were detected in all soil sections including at 18 cm depth, while no resident concentration of SM was detected at any depth in the entire length of the core for both soils. The deterministic, physical equilibrium model (CXTFIT) described the peak arrival time as well as the maximum concentration of the antibiotic breakthrough curves reasonably, but showed some underestimation at the advanced stages of the leaching process. There was a significant difference in the model estimated retardation factors obtained from column study and the experimental retardation factors obtained from the conventional batch sorption experiments. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Sulfonamides are one of the most commonly used veterinary antibiotics for therapeutic and sub-therapeutic purposes (Sukul and Spiteller, 2006). According to the New Zealand Food Safety Authority, the sulfonamide group contributes ~10% of total antibiotic usage (Srinivasan et al., 2014), and is a common class of antibiotics widely used in livestock industries in New Zealand (NZ). Land-application of animal waste effluent is a permitted activity in NZ as long as the farmers follow the prescribed conditions set out by the respective regional councils (Sarmah et al., 2006). This, along with the direct excretal inputs from the grazing ⁎ Corresponding author at: Department of Civil & Environmental Engineering, Faculty of Engineering, Private Bag 92019, Auckland 1142, New Zealand. Tel.: + 64 9 9239067; fax: +64 9 3737462. E-mail address: [email protected] (A.K. Sarmah).

http://dx.doi.org/10.1016/j.scitotenv.2014.06.034 0048-9697/© 2014 Elsevier B.V. All rights reserved.

animals, constitutes a direct source of exposure for antibiotics in the pasture environment and makes the assessment of its retardation and mobility in soil environment necessary to establish their risk to humans and environment. Recent fate studies conducted on sulfonamides in pasture soils collected from dairy farming region of NZ suggest that sulfonamides have a medium to high potential to leach through the soil profile and be transported to the groundwater, drainage water, and by surface runoff to surface waters (Srinivasan et al., 2014). Widespread occurrence of sulfonamides in aquatic and terrestrial environments has been reported in a number studies (Batt et al., 2006; Garcia-Galan et al., 2011; Kim et al., 2011). For example sulfamethoxazole antibiotic, which belongs to this group, was detected at a concentration of up to 0.48 μg L−1 in surface water samples in Germany (Hirsch et al., 1999). SMO was also detected in groundwater at a concentration of 0.22 μg L−1 (Lindsey et al., 2001) and in monitoring well samples (Barber et al., 2009), and streams (Kolpin et al., 2002) in the USA. These

536

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

studies show that there is a possibility that these chemicals may be encountered in drinking water supply. The transport of veterinary antibiotics in soils depends on several factors such as the chemical properties of the antibiotics, temperature, moisture content, ionic strength (Chen et al., 2011), soil physicochemical properties (Strauss et al., 2011), soil charge density, and contact time (Kurwadkar et al., 2011). Chemical properties such as water solubility, dissociation constants, sorption–desorption processes, stability, binding to the soils, and the partitioning coefficients at various pH values can all affect the mobility of antibiotics in the soil environment (Kurwadkar et al., 2011; Tolls, 2001; Unold et al., 2010). Other factors that can influence the mobility of veterinary antibiotics are the timing of manure application, prevailing weather conditions, preferential flow via desiccation cracks and worm channels as recently demonstrated in a UK field study (Kay et al., 2004; 2005a). In addition to sorption, the fate and transport of sulfonamides in subsurface environments can also be governed by other abiotic reduction due to the presence of reduced sulfur compounds (e.g., bisulfide and polysulfides) in soil/sediment pore waters (Zeng et al., 2011). Results obtained from overseas studies cannot be extrapolated to reflect the behaviour of the antibiotics in New Zealand conditions. This is primarily due to differences in pedo-climatic conditions, rainfall pattern, rainfall intensity and other temporal and spatial variability of soils including carbon content, mineralogical characteristics (such as clay and silica content) or soil pH. The predominance of volcanicallyderived soils with a high percentage of allophonic clay mineral having high surface area, and high organic carbon content is known to strongly influence the fate of other organic contaminants such as pesticides (Sarmah et al., 2005) and veterinary antibiotics (Srinivasan et al., 2014). However, to date there is no published information available on the transport behaviour of sulfonamides in NZ soils. The main objective of this study was to investigate the transport behaviour of three sulfonamides, namely SMO, SCP and SM in two different undisturbed soil cores under saturated conditions. In an earlier study it was observed that the mobility of mixed sulfonamides was not impacted by competitive sorption (Kurwadkar et al., 2011), which meant that the mobility of all the three sulfonamides could be evaluated simultaneously relative to a conservative Br− tracer. Sulfonamide sorption affinity onto these two soils was also investigated through a series of batch sorption experiments by extracting the residues from solid phase using a solvent extracted scheme. In order to test the applicability of the column test to other polar contaminants, results from the batch and column studies were compared against each other. 2. Materials and methods 2.1. Chemicals SMO, SCP, and SM (Table 1) of N99% purity, and calcium chloride dihydrate (CaCl2·2H2O N 99% purity) were obtained from Sigma Aldrich, Australia. Acetonitrile (Mallinckrodt ChromAR, ≥ 99.8% purity) and

dichloromethane (Mallinckrodt UltimAR, ≥99.9% purity) were obtained from Thermo Fischer Scientific Ltd. New Zealand. HPLC grade deionized water was obtained from an Arium® 61316 high performance reverse osmosis system (Sartorius Stedim Biotech GmbH, Germany). Nitrogen gas oxygen free (gas code 152) was purchased from BOC (New Zealand). 2.2. Soils Two top soils (0–5 cm) from the North Island of New Zealand, Hamilton clay loam from the Waikato dairy farming regions, and Matawhero silt loam from the Gisborne region were freshly collected, air-dried to room temperature and sieved (b2 mm). Some important physico-chemical properties of the soils used in this study are summarized in Table 2. Detailed physico-chemical properties of these soils can be found in an earlier study (Srinivasan et al., 2014). 2.3. Batch sorption studies A stock solution of SMO, SCP and SM antibiotic each at a concentration of 1000 mg L−1 was prepared in methanol. By adding an appropriate amount of the stock solution separately to 5 mM CaCl2 solution, six different initial aqueous solution concentrations were prepared in duplicate. Batch studies were performed using a similar experimental protocol earlier adopted by Sarmah et al. (2008). Briefly, duplicate samples of air-dried soils (2 g) were weighed into glass centrifuge tubes (35 mL) with Teflon-lined screw caps. Aliquots (30 mL) of six concentrations of each sulfonamides (1.5, 3, 5, 7.5, 10 & 15 mg L−1) were added to the respective tubes, wrapped in aluminum foil, placed in the dark and shaken in an end-over-end shaker for 24 h. The solution pH for the batch sorption studies ranged from 5.22 to 5.41 for Matawhero and from 5.60 to 6.03 for Hamilton soil. Preliminary investigation carried out on the sorption kinetics showed that an apparent equilibrium time of 24 h was sufficient for the batch sorption studies performed (Srinivasan et al., 2014). 2.4. Preparation of the soil lysimeters The soil lysimeters used in this study were made of polyvinylchloride (PVC) with a length of 20 cm and an internal diameter of 7.5 cm. Undisturbed soil cores of Matawhero and Hamilton soil were obtained in situ by hand carving from the ground surface. The internal cylindrical surface of the lysimeter casings was coated with petroleum jelly and the casings were slowly pushed into the soil under pressure. When the desired length of core was reached (18 cm) the lysimeters were carefully cut off from the bottom with a sharp metal blade. Immediately after excavation, both lysimeters were covered with polythene sheet and PVC end caps were attached to prevent water loss. The lysimeters were transported to the laboratory and the bottom of the cores was cleaned and secured with a fly screen (ø 1 mm) to allow leachate percolation, and sealed with a PVC cap containing a number of holes that acted as sampling ports for leachate collection. On the outer

Table 1 Selected chemical properties of the antibiotics used in the study. Properties

SMO

SCP

SM

Molecular formula Molecular weight (g mol−1) Log Kow pKa1/pKa2 at 25 °C Structure

C10H11N3O3S 253.28 0.89a 1.83a/5.57a

C10H9N4 ClO2S 284.72 0.31b 1.87/5.5b

C12H14N4O2S 278.3 0. 89a 2.28a/7.42a

a b

Figueroa-Diva et al. (2010). Srinivasan et al. (2014).

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

537

Table 2 Selected properties of soils used in the sorption study. Soils

pH 1:2 H2O

OC (%)

CEC (cmolc kg−1)

Sand %

Silt %

Clay %

SSA (m2g−1)

Matawhero Hamilton

6.1 5.8

2.1 4.0

15.4 21.5

11 19

62 51

27 30

60 36

edge of the cores, petroleum jelly was applied to seal the gaps between the PVC column and soil to prevent surface/edge flow during the experiment. The lysimeters were then slowly saturated over a 24 h period by placing them in two large trays containing water; this reduced the air trapped and also maintained the soil structure (Fan et al., 2011). Soil column physical parameters such as porosity, bulk density, water content and the pore volumes for each soil lysimeter are summarized in Table 3.

2.5. Transport experiment

different time intervals, and samples of the leachate fraction were taken for solvent extraction. At the end of the experiment, the PVC cap at the bottom of the lysimeters was carefully opened and the soil core was carefully pushed out intact. The soil core was cut into 6 parts (3 cm each). The soil from each section was thoroughly mixed and triplicate samples were taken to calculate its water content. Another set of triplicate soil samples (5 g) were placed into 35 mL Kimax centrifuge tubes in order to determine the residual antibiotic concentrations. 2.6. Extraction and analysis

The soil lysimeters were set up at a constant temperature room (22 °C ± 2) and transport studies were performed with three sulfonamide antibiotics SMO, SCP, and SM using a mediator solution of 5 mM CaCl2 solution. The top open end of the lysimeter was fitted with a tight cylindrical Teflon® compartment equipped with an irrigation head comprised of a reservoir and its lid consisting of 41 hypodermic needles to irrigate the lysimeter uniformly. The irrigation head was connected to a peristaltic pump by means of appropriate tubing, and the pump supplied a constant head of 5 mM of CaCl2 solution from a reservoir. The tubes and other accessories of the lysimeter were covered with aluminium foil to avoid photodegradation. A collecting funnel was attached to the bottom of the soil column for the collection of leachate into a beaker from which leachate fractions were sampled at regular intervals (15–20 min) for analysis. One litre of antibiotic solution containing all the three sulfonamides at a concentration of 0.5 mg L−1 each was prepared by adding an appropriate amount of methanolic stock solutions to CaCl2. Also 1 L of Br− conservative tracer (KBr 5.35 g L− 1) was prepared separately in deionized water. Once the lysimeters were completely saturated, flow was established from the top using 5 mM CaCl2 solution, and the flow rate in the controller was set to at a flow rate of 7.5 mL min−1. Each lysimeter was irrigated with at least 4 pore volumes (PV) of CaCl2 solution and a constant head was maintained in the compartment, which gave a constant downward flow. Once a steady-state pore velocity was achieved (12.5 mL min−1), a pulse input of Br− tracer was applied (1 L) and eluted with the 5 mM CaCl2. The leachate was collected at regular time intervals, and each fraction was analysed for the bromide concentration using an ion-selective electrode (Metrohm 6.0502.100, Herisau, Switzerland). The Br− breakthrough curve (BTC) was used to identify the physical transport characteristics of each soil column. Once the Br− BTC was obtained, i.e., when the bromide concentration becomes zero or consistently low, several pore volumes of the 5 mM CaCl2 solution were flushed through the soil column. Following this once a steady-state pore velocity was achieved a pulse input consisting of 1 L of the mixed antibiotic solution made in 5 mM CaCl2 solution was applied. This was followed by the application of the mediator solution (without antibiotic) until complete breakthrough was achieved. The leachate was collected in the beaker at

For the sorption experiment following equilibration, it is assumed that the concentration of the sulfonamides in the supernatant is in equilibrium to those sorbed to soil. The supernatant (2 mL) was filtered through a 0.45 μm microporous membrane. The filtered supernatant was immediately analysed by HPLC using UV detection at a detection wavelength of 275 nm. The remaining supernatant was carefully decanted and the weight of the remaining residual liquid phase was measured gravimetrically. It was assumed that antibiotic concentration in the residual soil was the same as that in the bulk supernatant solution. The residual antibiotic in the soil was extracted with 5 mL of DCM following shaking for 12 h. Following shaking, 1 mL of aliquot of each solvent extract was evaporated to dryness under N2, reconstituted in methanol (0.5 mL) and analysed using HPLC. For the transport studies, an aliquot of 10 mL of the leachate was extracted with 5 mL dichloromethane (DCM), and 1 mL of the extracted DCM was evaporated in a HPLC vial to dryness under a gentle stream of N2, reconstituted in methanol (0.5 mL) and analysed using HPLC-UV. The residual antibiotics in the soil cores were extracted with dichloromethane (10 mL) with overnight shaking in a rotary drum shaker. An aliquot DCM extract (1.5 mL) was taken into a HPLC vial and evaporated to dryness under a gentle stream of N2. The contents were reconstituted with methanol (0.5 mL), vortexed, and analysed by means of HPLC-UV at a detection wavelength of 275 nm. 2.7. HPLC analysis A gradient elution scheme was employed to measure the three sulfonamides simultaneously; the mobile phase at the start of the run consisted of 5% ACN, 90% TFA, and 5% THF; and this composition was maintained for 4 min. The composition changed linearly between 4 and 10 min to give a composition of 37.5% ACN, 57.5% TFA and 5% THF. Between 10 and 12 min the composition changed linearly back to its initial composition of 5% ACN, 90% TFA, and 5% THF and remained unchanged until 13 min. The flow rate was set to 1 mL min−1, and the injection volume of 20 μL. Separation of SM, SCP, and SMO was achieved on a Luna 5 μ RP-C18 column at 6.0, 8.8 and 9.5 min respectively with a total run time of 13 min. The limits of detection (LOD) were calculated

Table 3 Soil physical parameters and bromide breakthrough results. MSE is the mean squared error. R2 is a measure of the relative magnitude of the total sum of squares associated with the fitted equation. Soil

Porosity

Bulk density (g cm−3)

Moisture content (v/v)

Pore volume (cm3)

Pore water velocity (cm h−1)

Estimated pore water velocity (cm h−1)

R2

λ (cm)

MSE

Matawhero Hamilton

0.45 0.51

1.22 1.16

0.520 0.442

1432 1644

172.3 110.7

148.6 147.5

0.975 0.981

2.416 1.892

0.53 E−3 0.42 E−3

538

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

on a signal to noise ratio of 3 to 1. LOD for the three sulfonamides was 0.02 μg mL−1. Standard linearity was assessed in the concentration range of 0.1 to 1.0 μg mL−1 and 1.0 to 20 μg mL−1 and quantification was achieved by using external standards prepared from serial dilution of the stock solution. A full description of the analytical method along with the quantitative recovery obtained can be found in Srinivasan et al. (2012). No transformation product was detected in the leachate and residual soil samples.

The initial condition for this problem assumed that the soil column was bromide-free and is given as: C ðz; t ¼ 0Þ ¼ 0 for 0 b z b L

ð2Þ

The upper boundary condition of the soil column is a third-type boundary and is given as: νC ðz; t ¼ 0Þ−D

∂C ðz ¼ 0; t Þ ¼ νC 0 ðt Þ for 0 b t b t 0 ∂z

ð3Þ

νC ðz; t ¼ 0Þ−D

∂C ðz ¼ 0; t Þ ¼ 0 for t Nt 0 ∂z

ð4Þ

2.8. Sorption modelling Sorption isotherms were modelled using the Freundlich model: Cs = Kf CwN, where Cs (mg kg−1) and Cw (mg L−1) are the equilibrium sorbed and aqueous phase concentrations respectively, Kf (mg1 − N LN kg−1) is the Freundlich sorption coefficient and N (dimensionless) is the measure of sorption non-linearity (N = 1 represents a linear isotherm). Kf was quantified by the empirical Freundlich equation in the log transformation form: Log Cs = log Kf + N log Cw. In order to compare sorption across soils, a concentration-dependent effective distribution coefficient N−1 ) was calculated with an environmental concentration (Keff d = Kf Cw of 5 ng L− 1 (Table 4). Like-wise, concentration-dependent organic carbon-normalized partition coefficient values (Koc, mg L− 1) in each soil were calculated for all three sulfonamides at 5 ng L− 1 using the −1 /foc, where foc is the fraction of soil non-linear equation Koc = Kf CN w organic carbon. 2.9. Transport modelling The breakthrough curves of the bromide ion (Br−), a non-reactive conservative tracer, and the antibiotics were analysed using the physical equilibrium convective dispersive equation (CDE) assuming onedimensional, steady-state flow in a homogeneous soil (van Genuchten and Alves, 1982). The governing equation can be written as

R f ∂C ∂2 C ∂C ¼λ 2− ν ∂t ∂z ∂z

ð1Þ

where t is time (h), z is distance (cm), λ is dispersivity (cm), ν is the average pore water velocity (cm h− 1), C is the bromide concentration (mg L− 1), and Rf is the dimensionless retardation factor that reflects solute sorption. The coefficient Rf is set to 1 for the bromide tracer because it is assumed that bromide ions have minimal sorption onto soil particles (Fan et al., 2011). Eq. (1) contains one unknown parameter λ, where, λ = D/ν and D is the dispersion coefficient (cm2 h− 1 ). Dispersivity (λ) was inversely estimated using CXTFIT 2.0 (Toride et al., 1995). The values used to calculate this parameter were the input solution concentration (Co); the volumetric water content (θ), which was obtained from the average of the six sections; soil bulk density (ρ) of 1.16–1.22 g cm−3 was taken from Pang et al. (2008); and the soil column length (L).

and, the lower boundary condition of the soil column is given as: ∂C ðz ¼ L; t Þ ¼ 0 for t Nt 0 ∂z

ð5Þ

The variable t0 is the pulse duration (h), and C0 (t) is the dimensional concentration from time zero to t0. The CXTFIT program within the software STANMOD version 2.0 (Simunek et al., 1999) was used to obtain the analytical solution for the CDE by means of inverse fitting of the obtained bromide BTCs. The parameters initially estimated were ν and λ. These parameters were later used as input values for the antibiotic BTCs, so that Rf and D were estimated. It was assumed that the parent sulfonamide was the only compound present in the leachate, and the transformation products or metabolites were not included in the transport model. 3. Results and discussions 3.1. Batch sorption isotherms Plots for sorption isotherms of SMO, SCP, and SM are shown in Fig. 1. The model derived sorption parameters for each soil and compound are summarized in Table 4 along with their N and R2 values. The Freundlich model could describe the isotherms well as indicated by the high coefficient of determinant (R2) values. The overall recovery from the soil and aqueous phase was found to be acceptable with values of 94 (±5.1) and 92 (±1.7) % for SMO; 93.4 (±1.1) and 71.2 (±0.76) % for SCP; 82 (± 2.2) and 91 (± 4.9) % for SM respectively in Matawhero and Hamilton soils. The sorption affinity for Hamilton soil (OC 4.0%) was slightly higher than Matawhero soil (2.1%). Among the sulfonamides, SCP had highest sorption affinity. For the three sulfonamides, the Kf values ranged from 0.34 to 1.10 for Matawhero soil and from 2.0 to 2.52 for Hamilton soil. Among the sorbates examined, Keff d for the three sulfonamides ranged from 0.37 to 2.63 L kg− 1 across the soils investigated. The N values for the sulfonamides varied between 0.88 and 1.21 for the 2 soils displaying both linear and non-linear behaviour. Concentration-dependent effective distribution coefficient (Keff d ) values reported in our study were similar to the Kd (0.03–0.47) values reported in a recent study by Zhang et al. (2014).

Table 4 The model-estimated parameters for the deterministic physical equilibrium model fitted to the antibiotic breakthrough data. Soils

SA

Rf ± SE estimated

R2

MSE

Kf

N

R2

Keff d

Log Koc

Rf calculated

Matawhero

SMO SCP SM SMO SCP SM

8.20 (0.21) 7.36 (0.27) 2.69 (0.08) 12.44 (0.12) 3.98 (0.12) 7.74 (0.20)

0.91 0.81 0.85 0.96 0.85 0.84

0.26 E−3 0.58 E−4 0.14 E−3 0.78 E−4 0.74 E−3 0.13 E−3

0.34 0.97 1.10 2.00 2.52 2.19

1.21 1.08 0.99 0.88 1.11 0.92

0.99 0.99 0.97 0.99 0.99 0.99

1.19 1.00 1.10 1.90 2.63 2.12

1.14 1.64 1.72 1.74 1.77 1.76

1.06 3.17 3.70 7.53 8.85 8.16

Hamilton

Rf is the retardation factor; MSE is the mean squared error; R2 is a measure of the relative magnitude of the total sum of squares associated with the fitted equation; Keff d is the concentration dependent effective sorption distribution coefficient.

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

16

12

Cs (mg kg-1)

Matawhero

SMO SCP SM

14

539

Hamilton

30 25

10

20

8

15

6 10 4 5

2

0

0 0

2

4

6

8

10

12

14

Cw (mg L-1)

16

0

2

4

6

8

10

12

14

16

Cw (mg L-1)

Fig. 1. Multiple-concentration batch sorption isotherms for SMO, SCP and SM in Matawhero and Hamilton soils. Symbols represent measured data, while solid lines represent Freundlich model fits.

Nonlinear sorption isotherms have been reported for tylosin and sulfonamides under neutral pH conditions (Lertpaitoonpan et al., 2009) in contrast to some studies which reported linear isotherms (Boxall et al., 2002). In the present study, the soil properties such as texture, soil organic carbon content, and CEC and soil pH were different across the soils investigated. As a result the sorption affinity of the antibiotics varied from soil to soil. Recent studies have reported the influence of soil organic carbon on the sorption of sulfonamide antibiotic (FigueroaDiva et al., 2010). Kahle and Stamm (2007) reported sorption to inorganic sorbents (clay) to be an order of magnitude lower than sorption onto organic matter at neutral pH. It's a known fact that at natural soil pH sulfonamides tend to be mostly in neutral and anionic form (Srinivasan et al., 2014). While the neutral species are readily adsorbed on to the organic matter, the ionic form is readily taken up by negatively charged clay particles. Therefore these findings lend further support on the assumption that sulfonamide antibiotic sorption to soils is primarily due to hydrophobic partitioning mediated by soil organic carbon, and partially due to anion exchange to negatively charged soil particles. Based on the pKa (Table 1), and soil pH data (Table 2), the expected net charges for each antibiotic under the pH values of the soils would be: SCP and SMO as anions, while SM would always be present in neutral form. Concentration-dependent average log Koc values for the three sulfonamides based on the environmental concentration of 5 ng L−1 ranged from 1.14 to 1.77 in both soils. Based on the low log Koc values obtained in this study, all three sulfonamides are expected to exhibit high mobility, making them a potential threat to surface and ground water. The nonlinear Freundlich partition coefficient (Kf) values obtained in this study for these antibiotics were used to calculate the experimental retardation factors (Rf) and is discussed later. 3.2. Transport studies 3.2.1. Bromide-tracer breakthrough curve The breakthrough curves of the conservative tracer (Br−) in the two soils are shown in Fig. 2. Bromide initiates its breakthrough at approximately 0.85 and 1.05 pore volumes for Matawhero and Hamilton soil, which was expected for a conservative tracer. The maximum amount recovered for the bromide tracer was calculated to be 93% and 94% for Matawhero and Hamilton soil respectively. These results were similar to the findings of Kurwadkar et al. (2011), who showed that the breakthrough of Br− occurred between 1 and 2 pore volumes in 3 US soils with varying physical and chemical properties. The authors also reported bromide recoveries to range between 95 and 100%. In this present

study no preferential flow was observed in either of the soil columns, as the maximum breakthrough for the bromide tracer was similar on both occasions, and no significant tailing was observed in the bromide BTCs for either soil, which is an indicative of a well-saturated soil column. 3.2.2. Antibiotic breakthrough curve Fig. 2 shows the breakthrough curves for the three antibiotics in Matawhero and Hamilton soil respectively. The breakthrough for the antibiotics in Matawhero soil occurred between 1.06 and 1.60 pore volumes, and between 2.36 and 2.65 pore volumes for Hamilton soil. The maximum amounts recovered in the leachate for SMO, SCP, and SM for Matawhero soils were 65.4%, 78.1%, and 45.4% respectively. For Hamilton soil, however, maximum recoveries for the three antibiotics were much lower and ranged from 17 to 58%. The final leachate concentrations for SMO, SCP, and SM from the Matawhero soil core were 3.37, 2.67 and 2.50 μg L−1 and for Hamilton soil core it was 3.57, 8.58, and 6.90 μg L−1 respectively. Recently, Kurwadkar et al. (2011) reported the recoveries in the column effluent for sulfamethazine and sulfathiazole to be 50 to 90% and 34 to 75%, respectively. In another study 69–99.7% of SM, along with its primary metabolite, was recovered in column effluents (Fan et al., 2011). Elsewhere, the maximum relative concentrations of sulfadiazine differed as well as the eluted mass fractions, ranging from as low as 8 to 83% after 500 h of leaching (Wehrhan et al., 2007). In a much earlier study in the UK, from a spiked soil column experiment, Kay et al. (2005b) reported SCP concentration of 703.2 μg L− 1 in rainfallsimulated leachate. Past studies have shown that the transport of sulfadiazine was found to be dependent on the input mass and duration; the mass recovery of sulfadiazine in soil column effluent increased as the solute input duration increased (Wehrhan et al., 2007). In this present study the antibiotic BTC's peaks shifted only marginally (by 0.2 to 0.85 pore volumes) for Matawhero soil, whereas for Hamilton soil antibiotic BTC peaks showed significant shifts (1.6 to 1.9) in terms of pore volumes when compared to the conservative Br− tracer. These results clearly demonstrate that sulfonamides could potentially behave in a similar manner to a conservative tracer and their transport is dependent on soil types. The slightly late arrival of the antibiotics maxima in Matawhero soil indicates that retardation still played some part in antibiotic transport through the soil column, and this could be attributed to the antibiotic sorption to the soil particles. This effect was much more pronounced in Hamilton soil (Fig. 2) where late arrival of the antibiotics maxima was about 0.1–0.2 days when compared to Br− tracer. These shifts in

540

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

0.5

0.12

Matawhero

Matawhero Br-

0.4

SCP SMO SM

0.10

CXTFIT 0.08

Ct/Co

0.3 0.06 0.2 0.04 0.1

0.02

0.0

0.00 0.20

0.5

Hamilton

Ct/Co

Hamilton 0.4

0.16

0.3

0.12

0.2

0.08

0.1

0.04

0.0

0.00 0.0

0.2

0.4

0.6

0.8

1.0

Days

1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Days

Fig. 2. Breakthrough curves for Br− tracer (left column) and three sulfonamides (right column) for Matawhero and Hamilton soils. Symbols represent measured data, while the solid lines indicate the CXTFIT model fits.

the BTC peaks for the sulfonamides varied among the soils investigated and this correlated well to the results obtained batch sorption studies which showed Hamilton soil having more sorptive tendency towards the sulfonamides than Matawhero soil. Recent studies have shown that peak antibiotic concentration in the BTCs can occur simultaneously with bromide breakthrough (Kurwadkar et al., 2011). However, in the present study, for Hamilton soil, there was considerable retardation in the antibiotic transport through the soil column. In general, the BTCs for the sulfonamides were broader when compared with the BTCs of Br− tracer, which is indicative of the influence of the retardation process. Hamilton soil has a greater percentage of clay compared with Matawhero soil, which is loamy by nature. The higher clay content in Hamilton soil could be responsible for increased retardation. Hamilton soil has almost twice the organic carbon content of Matawhero soil (Table 2), thereby providing a favourable environment for neutral forms of the antibiotic to partition into the organic carbon domain. Of the three antibiotics studied, SM breakthrough was the fastest in both soils, indicating the highly mobile nature of this compound. SMO and SCP breakthrough was more or less identical in both soils with similar pore volumes. The BTCs observed for the sulfonamides in the present study were similar to the BTC observed by Kurwadkar et al. (2011) at natural soil pH for sulfamethazine and sulfathiazole antibiotic. In our study the natural pH values of soils were 5.1 and 6.1 respectively for Matawhero and Hamilton soil. Therefore it is likely that all three sulfonamides would exist as 50% neutral and 50% anionic forms making them highly mobile and with little to no retention on the soil column. A visual observation of the BTCs (Fig. 2) also lends further support to the argument that the sorption and mobility of the sulfonamide group

of antibiotics is pH dependent (Kurwadkar et al., 2011; Lertpaitoonpan et al., 2009). 3.2.3. Antibiotic retardation The results of the soil extraction from the sectioned lysimeters are shown in Fig. 3. For both Matawhero and Hamilton soil, SMO and SCP were detected in each portion of the sectioned core up to the first 18 cm of the column, and were well above the method detection limit (13 μg kg−1). No residual concentration for SM was detected at any depth in the entire core for both the soils, suggesting loss of the compound due to biodegradation in the soil. A recent laboratory soil column study showed that 33–70% of SM degraded within 6 h during a transport study involving soil columns (Fan et al., 2011). It is also possible that the compound might have a greater potential to penetrate even deeper into the soil profile. Elsewhere, no residual concentrations for sulfadiazine were detected in a rainfall simulated column study, suggesting that sulfadiazine might have leached to a greater depth (Kim et al., 2010). A plausible explanation for not detecting SM in samples collected from the column at any depth could be a limitation of our analytical method, and/or is a result of inefficient extraction method to extract SM from the soil samples using DCM (Srinivasan et al., 2012). Burkhardt and Stamm (2007) reported the depth distribution of three sulfonamides applied to a loamy grassland soil; with levels of antibiotic concentrations at 30–50 cm depth being equal to the concentration found in the top 5 cm regardless of whether application of antibiotics was in manure or in aqueous solution. In the current study for both soils, the antibiotic concentration profile decreased with depth, indicating that possible preferential flow through cracks, or worm holes did not occur. SCP was found in greater proportion than SMO at any

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

541

MDL

Matawhero 0-3

Depth (cm)

3-6

SMO SCP SM

6-9

9-12

12-15

15-18

0.0

0.1

0.2

0.3

-1

C s (mg kg ) MDL

Hamilton 0-3

Depth (cm)

3-6

SMO SCP SM

6-9

9-12

12-15

15-18

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-1

Cs (mg kg ) Fig. 3. Residual antibiotic concentration in the sectioned soil cores respectively. Average values of n = 3 samples are displayed with one standard deviation. The dotted line indicates the method detection limit (MDL) of 0.013 mg kg−1.

given depth for either of the soils, suggesting that SCP is more readily retained than SMO; these results again correlate well with the batch sorption results. The variations in the antibiotic resident concentrations when compared to other reported studies may be due to varying soil properties such as pH, chemical properties and experimental protocol such as column length and pulse input duration (Kurwadkar et al., 2011; Wehrhan et al., 2007). 3.2.4. Transport modelling The physical equilibrium model available within the STANDMOD (Toride et al., 1995) was used inversely to fit the Br− breakthrough data, and to obtain transport parameters ν, and D, respectively. Unlike some recent studies (Fan et al., 2011; Jeong et al., 2012), no tailings and asymmetrical breakthroughs were observed for the bromide BTCs, indicating the absence of physical non-equilibrium conditions during the transport of the antibiotics through the soil columns. The nonsorbing bromide breakthrough curves were all quite symmetrical (Fig. 2) as expected for a non-reactive solute, and were described well (R2 values of 0.97 and 0.98 for the Matawhero and Hamilton soils

respectively) by the model setting Rf = 1. Fan et al. (2011) used a two-site non-equilibrium transport model to describe the breakthrough curves of sulfamethazine, while Wehrhan et al. (2007) showed that a three-site sorption model fitted the breakthrough curves for sulfadiazine the best. Unlike these studies an attempt to use a two-site nonequilibrium transport model did not yield favourable results for either datasets (Matawhero and Hamilton) in our work. The pore water velocity used in this study, was much higher (110–190 cm h−1) compared with the velocities (0.44–0.54 cm h− 1) used by Wehrhan et al. (2007). High pore water velocity is known to limit the contact time of aqueous-dissolved organic contaminant with the sorbed phase (Lee et al., 2002), and thus it can be inferred that velocity is the reason why neither the two-site nor the three-site non-equilibrium transport model would fit the data well. However, using the equilibrium CDE with the boundary conditions as described earlier, and with the assumption of one dimensional steady-state flow in homogeneous soil yielded a good description of the datasets as indicated by the goodness of fit (R2) values (Table 4). The CXTFIT model described well the peak arrival time as well as the

542

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

maximum concentration for the three sulfonamides. However, it exhibited some underestimation at advanced stages of sulfonamide (SCP and SMO) leaching, especially in the Matawhero soil. Overall, the measured BTCs, peak arrival time, and the maximum concentration of the three sulfonamides in the final effluent for both soils, with contrasting properties, were adequately described by the CXTFIT model. BTCs for the sulfonamides were not entirely symmetric, with some broadening of peaks when compared with the conservative tracer. These broadened peaks for antibiotics are unlikely to be due to the hydrodynamic characteristics of the soil columns because the Br− tracer exhibited relatively ideal BTCs. The behaviour of the BTCs suggests that sulfonamide transport in both soils may have been retarded (Fig. 2) to some extent. It was also evident that the BTCs of the antibiotic for the Hamilton soil had much delayed peaks and longer tailing (SCP) than for the Matawhero soil. In addition, there were relatively higher peaks of relative concentration (C/Co) observed in the Hamilton soil, compared with the Matawhero soil, while the maximum concentration for the bromide and SM occurred at the same time in the case of Matawhero soil. CXTFIT model parameters were obtained by using ν, and λ values from the bromide BTC, while only Rf was optimized. The estimated dispersion coefficient along with the estimated pore water velocities for Matawhero and Hamilton data were in good agreement to the experimentally determined values, and small differences can be attributed to the heterogeneous nature of the soil with varying pore sizes affecting the pore water velocity or possibly due to the presence of cracks, or fissures in the soils allowing macropore flow. The retardation factor (Rf) values obtained with the CXTFIT model for the three sulfonamides for the Matawhero and Hamilton soils are summarized in Table 4. The estimated Rf values of CXTFIT were compared with the experimental Rf values obtained from the non-linear partition coefficients (Kf) for these antibiotics obtained from the sorption experiment. The experimental retardation factor is often written as ρ R f ¼ 1 þ Kd θ

ð6Þ

where R (the dimensionless retardation factor) accounts for fully reversible reactions of a solute species in the soil system (Jeong et al., 2012), ρ is the soil bulk density (kg m−3), θ is volumetric water content (v/v), and Kd, the linear sorption coefficient (L kg−1). However, for nonlinear sorption coefficients obtained from Freundlich isotherms the equation can be written as Rf ¼ 1 þ

ρ θ

  1 N1 −1 Kf C N

ð7Þ

where Kf is the Freundlich sorption coefficient, C is the concentration of the contaminant (mg L−1) and N is dimensionless and represents the degree of linearity (Sharma and Reddy, 2004). The results obtained from the column studies correlate to the findings from batch sorption studies performed on the same set of soils and antibiotics. In general all the three sulfonamides were weakly sorbed to the soils, with the sorption capacity of Matawhero soil being lower than that of Hamilton clay soil; therefore it is not surprising to expect some retardation in Hamilton soil. SM, with the lowest Kf values for both soils, had the quickest arrival time (breakthrough), and SCP, with the highest Kf in both soils, had the slowest breakthrough. There was a marked difference in the model estimated retardation factors and the experimental retardation factor. These variations could be due to the different experimental protocols adopted. Often the sorption process for organic chemicals in batch experiments tends to occur faster than in a leaching experiment conducted in the laboratory using packed or intact soil columns, and this may be explained by the optimized contact between the continuously agitated (e.g., by shaking) soil and solution in such experiments (Vereecken et al., 2011). Higher contact time indicates lower mobility due to increased interaction with soil material. Recently Maszkowska et al. (2013) reported the results of the batch and

Table 5 Summary of results from lysimeter studies with veterinary antibiotics. Sulfonamides

Sample

Concentrations/% recovery

Reference

Sulfadiazine

Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate Leachate

4.1% 0.78 μg L−1 95–97% 0.08 to 56.7 μg L−1 18 to 83% 34–75% 50–90% 69–99.7% 41–97% 56–92% 100% 0.08 to 56.7 μg L−1 38%–86% 63.6–98%

Kreuzig and Holtge (2005) Blackwell et al. (2007) Unold et al. (2009) Aust et al. (2010) Wehrhan et al. (2007) Kurwadkar et al. (2011) Kurwadkar et al. (2011) Fan et al. (2011) Strauss et al. (2011) Strauss et al. (2011) Chen et al. (2011) Aust et al. (2010) Strauss et al. (2011) Zhang et al. (2014)

17–65.4% 35–78.1% 45.4–58%

This study

Sulfathiazole Sulfamethazine

Sulfamethoxazole Sulfadimethoxine Sulfapyridine Sulfameter Sulfadimethoxine Sulfamethoxazole Sulfachloropyradazine Sulfamethazine

column tests to be comparable to a large extent, with slightly higher concentrations being obtained in the column test experiments of finegrained soils with a high organic matter content. However the authors did not report the values for Rf, and in the absence of compoundspecific (sulfonamides) literature data on retardation factors, a comparison with the Rf values obtained in the present study is not possible. However, the Rf values obtained in this study were of a similar order of magnitude to the Rf values reported for another sulfonamide, sulfathiazole with Rf values ranging from 2.84 to 16.0 in three US agricultural soils having pH values of 5.2 to 7.8 (Kurwadkar et al., 2005). Table 5 summarizes results from various laboratory lysimeter studies conducted in the recent past as well as the findings from our study with high leaching potential for the sulfonamides investigated. For example, Kurwadkar et al. (2011) reported the mobility of anionic form of SM and SCP to be similar to that of conservative bromide tracer. The authors reported considerable retention at low pH and high charge density, and this was attributed to the increased fraction of cationic and neutral forms of the sulfonamides. Another recent study indicated that the transport of sulfonamides in the soil column to be mainly influenced by sorption capacity (Zhang et al., 2014). 4. Conclusions Sulfonamide transport through soils is of concern after direct excretal inputs from grazing animals and/or subsequent effluent application on the field. There is a potential for these antibiotic residues to migrate through the soil profile and cause eventual contamination of surface water and groundwater. From this study it can be concluded that sulfonamides have weak sorption affinity and relatively high mobility in soils. The results of the soil column studies showed that all three sulfonamides were mobile and exhibited nearly the similar conservative behaviour as that of the Br− tracer. The BTCs of the antibiotic varied depending on soil type, and the type of chemical. For the three sulfonamides studied, the order of breakthrough was generally observed to be SMO N SM N SCP for both Matawhero and Hamilton soils. The CXTFIT model described the peak arrival time as well as the maximum concentration of the antibiotic breakthrough curves but showed some underestimation at advanced stages of sulfonamide leaching, especially in the Hamilton soil. There was a significant difference in the model estimated retardation factors and the experimental retardation factor. The high mobility of sulfonamide antibiotics can be explained with the compounds dissociation constant and the corresponding pH values of the media. For example, at soil pH (pH N N pKa1), they are likely to exist mostly in their anionic form and thus can attain higher solubility and

P. Srinivasan, A.K. Sarmah / Science of the Total Environment 493 (2014) 535–543

can become highly mobile in the environment. In addition, under realistic field scenarios, processes such as preferential flow or colloid or dissolvedorganic carbon-facilitated transport, could also play an important role in the rapid transport of sulfonamides. Acknowledgements The authors wish to thank the three anonymous reviewers for their useful comments on the manuscript. Oisik Das of University of Auckland is thanked for the initial proof reading of this manuscript. References Aust MO, Thiele-Bruhn S, Seeger J, Godlinski F, Meissner R, Leinweber P. Sulfonamides leach from sandy loam soils under common agricultural practice. Water Air Soil Pollut 2010;211:143–56. Barber LB, Keefe SH, Leblanc DR, Bradley PM, Chapelle FH, Meyer MT, et al. Fate of sulfamethoxazole, 4-nonylphenol, and 17 beta-estradiol in groundwater contaminated by wastewater treatment plant effluent. Environ Sci Technol 2009;43:4843–50. Batt AL, Snow DD, Aga DS. Occurrence of sulfonamide antimicrobials in private water wells in Washington County, Idaho, USA. Chemosphere 2006;64:1963–71. Blackwell PA, Kay P, Boxall ABA. The dissipation and transport of veterinary antibiotics in a sandy loam soil. Chemosphere 2007;67:292–9. Boxall ABA, Blackwell P, Cavallo R, Kay P, Tolls J. The sorption and transport of a sulphonamide antibiotic in soil systems. Toxicol Lett 2002;131:19–28. Burkhardt M, Stamm C. Depth distribution of sulfonamide antibiotics in pore water of an undisturbed loamy grassland soil. J Environ Qual 2007;36:588–96. Chen H, Gao B, Li H, Ma LQ. Effects of pH and ionic strength on sulfamethoxazole and ciprofloxacin transport in saturated porous media. J Contam Hydrol 2011;126:29–36. Fan ZS, Casey FXM, Hakk H, Larsen GL, Khan E. Sorption, fate, and mobility of sulfonamides in soils. Water Air Soil Pollut 2011;218:49–61. Figueroa-Diva RA, Vasudevan D, Mackay AA. Trends in soil sorption coefficients within common antimicrobial families. Chemosphere 2010;79:786–93. Garcia-Galan MJ, Diaz-Cruz MS, Barcelo D. Occurrence of sulfonamide residues along the Ebro river basin removal in wastewater treatment plants and environmental impact assessment. Environ Int 2011;37:462–73. Hirsch R, Ternes T, Haberer K, Kratz KL. Occurrence of antibiotics in the aquatic environment. Sci Total Environ 1999;225:109–18. Jeong CY, Wang JJ, Dodla SK, Eberhardt TL, Groom L. Effect of biochar amendment on tylosin adsorption–desorption and transport in two different soils. J Environ Qual 2012;41:1185–92. Kahle M, Stamm C. Sorption of the veterinary antimicrobial sulfathiazole to organic materials of different origin. Environ Sci Technol 2007;41:132–8. Kay P, Blackwell PA, Boxall ABA. Fate of veterinary antibiotics in a macroporous tile drained clay soil. Environ Toxicol Chem 2004;23:1136–44. Kay P, Blackwell PA, Boxall ABA. A lysimeter experiment to investigate the leaching of veterinary antibiotics through a clay soil and comparison with field data. Environ Pollut 2005a;134:333–41. Kay P, Blackwell PA, Boxall ABA. Transport of veterinary antibiotics in overland flow following the application of slurry to arable land. Chemosphere 2005b;59:951–9. Kim SC, Davis JG, Truman CC, Ascough JC, Carlson K. Simulated rainfall study for transport of veterinary antibiotics—mass balance analysis. J Hazard Mater 2010;175:836–43. Kim KR, Owens G, Kwon SI, So KH, Lee DB, Ok YS. Occurrence and environmental fate of veterinary antibiotics in the terrestrial environment. Water Air Soil Pollut 2011;214: 163–74. Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, Barber LB, et al. Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999– 2000: a national reconnaissance. Environ Sci Technol 2002;36:1202–11. Kreuzig R, Holtge S. Investigations on the fate of sulfadiazine in manured soil: laboratory experiments and test plot studies. Environ Toxicol Chem 2005;24:771–6. Kurwadkar ST, Adams CD, Meyer MT, Kolpin DW. Modeling transport of sulfathiazole in three loamy soils. Proc Water Environ Fed 2005:6957–63. [2005].

543

Kurwadkar ST, Adams CD, Meyer MT, Kolpin DW. Comparative mobility of sulfonamides and bromide tracer in three soils. J Environ Manage 2011;92:1874–81. Lee J, Hundal LS, Horton R, Thompson ML. Sorption and transport behavior of naphthalene in an aggregated soil. J Environ Qual 2002;31:1716–21. Lertpaitoonpan W, Ong SK, Moorman TB. Effect of organic carbon and pH on soil sorption of sulfamethazine. Chemosphere 2009;76:558–64. Lindsey ME, Meyer M, Thurman EM. Analysis of trace levels of sulfonamide and tetracycline antimicrobials in groundwater and surface water using solid-phase extraction and liquid chromatography/mass spectrometry. Anal Chem 2001;73:4640–6. Maszkowska J, Kołodziejska M, Białk-Bielińska A, Mrozik W, Kumirska J, Stepnowski P, et al. Column and batch tests of sulfonamide leaching from different types of soil. J Hazard Mater 2013;260:468–74. Pang L, McLeod M, Aislabie J, Simunek J, Close M, Hector R. Modeling transport of microbes in ten undisturbed soils under effluent irrigation. Vadose Zone J 2008;7: 97–111. Sarmah AK, Close ME, Pang LP, Lee R, Green SR. Field study of pesticide leaching in a Himatangi sand (Manawatu) and a Kiripaka bouldery clay loam (Northland). 2. Simulation using LEACHM, HYDRUS-1D, GLEAMS, and SPASMO models. Aust J Soil Res 2005;43:471–89. Sarmah AK, Meyer MT, Boxall ABA. A global perspective on the use, sales, exposure pathways, occurrence, fate and effects of veterinary antibiotics (VAs) in the environment. Chemosphere 2006;65:725–59. Sarmah AK, Northcott GL, Scherr FF. Retention of estrogenic steroid hormones by selected New Zealand soils. Environ Int 2008;34:749–55. Sharma HD, Reddy KR. Geoenvironmental engineering, site remediaition, waste containment, and emerging waste management technologies. Hoboken, NJ: John Wiley; 2004. Simunek J, van Genuchten MT, Sejna M, Toride N, Leij FJ. The STANMOD computer software for evaluating solute transport in porous media using analytical solutions of convection–dispersion equation. Riverside, California: U. S. Salinity Laboratory, USDA; 1999. Srinivasan P, Sarmah AK, Manley-Harris M, Wilkins AL. Development of an HPLC method to analyze four veterinary antibiotics in soils and aqueous media and validation through fate studies. J Environ Sci Health A 2012;47:2120–32. Srinivasan P, Sarmah AK, Manley-Harris M. Sorption of selected veterinary antibiotics onto dairy farming soils of contrasting nature. Sci Total Environ 2014;472:695–703. Strauss C, Harter T, Radke M. Effects of pH and manure on transport of sulfonamide antibiotics in soil. J Environ Qual 2011;40:1652–60. Sukul P, Spiteller M. Sulfonamides in the environment as veterinary drugs. Rev Environ Contam Toxicol 2006;187:67–101. Tolls J. Sorption of veterinary pharmaceuticals in soils: a review. Environ Sci Technol 2001;35:3397–406. Toride N, Leij FJ, van Genuchten MT. The CXTFIT code for estimating transport parameters from laboratory or field tracer experiments. Version 2.0. Riverside, California: United States Salinity Laboratory, United States Department of Agriculture; 1995. Unold M, Kasteel R, Groeneweg J, Vereecken H. Transport and transformation of sulfadiazine in soil columns packed with a silty loam and a loamy sand. J Contam Hydrol 2009;103:38–47. Unold M, Kasteel R, Groeneweg J, Vereecken H. Transport of sulfadiazine in undisturbed soil columns: effects of flow rate, input concentration and pulse duration. J Environ Qual 2010;39:2147–59. van Genuchten MT, Alves WJ. Analytical solutions of the one-dimensional convectivedispersive solute transport equation. Technical Bulletin. Agricultural Research Service. U.S.D.A.;no. 1661. Washington: U.S.D.A; 1982. Vereecken H, Vanderborght J, Kasteel R, Spiteller M, Schäffer A, Close M. Do lab-derived distribution coefficient values of pesticides match distribution coefficient values determined from column and field-scale experiments? A critical analysis of relevant literature. J Environ Qual 2011;40:879–98. Wehrhan A, Kasteel R, Simunek J, Groeneweg J, Vereecken H. Transport of sulfadiazine in soil columns—experiments and modelling approaches. J Contam Hydrol 2007;89: 107–35. Zeng T, Ziegelgruber KL, Chin Y-P, Arnold WA. Pesticide processing potential in prairie pothole porewaters. Environ Sci Technol 2011;45:6814–22. Zhang Y-L, Lin S-S, Dai C-M, Shi L, Zhou X-F. Sorption–desorption and transport of trimethoprim and sulfonamide antibiotics in agricultural soil: effect of soil type, dissolved organic matter, and pH. Environ Sci Pollut Res 2014:1–9.