Multimedia fate modeling and risk assessment of antibiotics in a water-scarce megacity

Journal of Hazardous Materials 348 (2018) 75–83

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Multimedia fate modeling and risk assessment of antibiotics in a waterscarce megacity ⁎

T

⁎

Haiyang Chen , Lijun Jing, Yanguo Teng , Jinsheng Wang Engineering Research Center of Groundwater Pollution Control and Remediation, Ministry of Education, College of Water Sciences, Beijing Normal University, No 19, Xinjiekouwai Street, Beijing, 100875, China

A R T I C L E I N F O

A B S T R A C T

Keywords: Fugacity model Multimedia fate Environmental risk Monte Carlo simulation Antibiotics

As a result of the widespread use of antibiotics, a large amount of excretion from human and animals containing antibiotic residues was discharged into the environment with wastewaters and manures, leading to potential adverse effects on ecosystem health. To understand the environmental fate of antibiotics, a dynamic level IV fugacity model was established here by introducing the novel process of nondiffusive wastewater irrigation from water to soil, and applied to a large-scale water-scarce region, the megacity Beijing. Furthermore, a Monte-Carlo based risk assessment approach was employed to evaluate the potential risks posed by antibiotics in water, sediment and soil, combined with the soil-water equilibrium partitioning method. Model validation, sensitivity and uncertainty analysis suggests that the fugacity model can successfully simulate the reported concentration data within an average difference of 0.2 logarithmic units. Results showed that more than one hundred tonnes of antibiotics were estimated to be discharged into the environment of Beijing in 2013, and, resulted in high antibiotics levels and posed high potential risks on the aquatic environment. On the other hand, although wastewater irrigation increased the antibiotics concentrations in soil and even dominated the total transfer fluxes, the overall risk levels of antibiotics in the soil were acceptable.

1. Introduction Since the discovery of penicillin in 1928, antibiotics have been extensively used in human and veterinary medicine to prevent or treat bacterial infections and to promote growth in animal husbandry and aquaculture [1,2]. However, following administration, antibiotics are only partially metabolized and, therefore, a large amount of antibiotic residues is excreted with urine and faeces as unchanged and active species [3]. Furthermore, these antibiotic excretion from human and animals cannot always be efficiently removed by wastewater treatment plants (WWTPs) and livestock farms due to the limitations of treatment technologies and infrastructure [4]. Therefore, a considerable proportion of antibiotics used by human and animals are discharged into the receiving environment [5]. Some previous reports indicated that nearly 70 antibiotics had been detected in various compartments, and even in drinking water worldwide [6–10]. Because of their persistence mode of toxic action in environmental microorganisms, antibiotics can pose a potential risks on ecosystem health [11]. Especially, some of antibiotics are environmentally pseudopersistent due to their continuous release from the increasing consumption of human and animals [12]. On the other hand, in comparison

⁎

to other chemicals, antibiotics are a group of special xenobiotics that can pose a potential systematic effect on human physiology by gut microbiota, which has been linked to many diseases related to immune and metabolism [13]. More importantly, a growing body of studies suggested that the selection of antibiotic-resistant bacteria in the environment had caused an increasing risk to human health [14,15]. Thus, with the concern for possible threats to the ecosystem and human health, antibiotics have been listed as one group of emerging environment contaminants [16]. Beijing, located within the water-scarce area of the North China Plain, is one of the most populous cities in the world and owns a huge population of more than 20 million. With the rapid development of urbanization and industrialization in the recent decades, the megacity Beijing faces increasingly severe water scarcity and serious water pollution problems [17–19]. It was estimated that the per capita water availability of Beijing in 2007 dropped to 230 m3/a, which was only 2.7% of the 8500 m3/a world average [20]. To relieve the increasing pressure on water scarcity due to extensive overuse and pollution of the available water resources, the wastewaters from various sources are often used for agricultural irrigation in this region [20], which may further cause potential environmental effects [21].

Corresponding authors. E-mail addresses: [email protected] (H. Chen), [email protected] (Y. Teng).

https://doi.org/10.1016/j.jhazmat.2018.01.033 Received 9 August 2017; Received in revised form 12 January 2018; Accepted 15 January 2018 Available online 31 January 2018 0304-3894/ © 2018 Elsevier B.V. All rights reserved.

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

where the subscript i or j is 1, 2, 3 and 4 for air, water, soil and sediment, respectively; V indicates compartment volume (m3); Z indicates fugacity capacity (mol/m3 Pa); f indicates fugacity (Pa); E indicates source emission rate (mol/h); Q indicates advective input (mol/h); Dji is the transfer coefficient from compartment j to compartment i (mol/ Pa h); DA indicates transfer coefficient of advective output (mol/Pa·h); DR indicates transfer coefficient of chemical reactive (mol/Pa·h). The definitions of the fugacity capacity Z and the transfer coefficient D are shown in Table S2 and Table S3, respectively. Among them, the transfer coefficient for the process of nondiffusive wastewater irrigation is defined with the symbol of D32 using the formula:

In recent, a number of studies have shown that various antibiotics were increasingly detected in the surface water and soil of Beijing due to the intermittent discharge from sewage wastes and wastewater reuse [7,10,17–19]. To prevent the ecosystem risks caused by antibiotics and protect human health for the megacity Beijing, it is essential to understand the multimedia concentrations of antibiotics and their effects in the environment [22]. However, the available information concerning this issue is very limited to date. In this study, a dynamic level IV fugacity model was established to simulate the distribution of 22 typical antibiotics (Text S1) in four compartments (air, water, soil, and sediment) and the transfer fluxes between adjacent compartments. Fugacity model has been shown to be proven useful to investigate the transport, distribution and fate of chemicals in the environment [5,23–27]. Here, the novel process of nondiffusive wastewater irrigation from water to soil was introduced into the model, and applied to the water-scarce megacity, Beijing. Sensitivity analysis and uncertainty analysis were conducted to determine the parameters’ influence and output variation. Furthermore, a Monte-Carlo based risk assessment approach was employed to evaluate the probable risks posed by antibiotics in the environment of Beijing with the predicted concentrations derived from the fugacity model. As very few studies reported the risks of antibiotics in the terrestrial compartment, a soil-water equilibrium partitioning method was adopted to estimate the toxicity of contaminants in soil and sediment [28]. We hope the results of the current study can provide a basis for the pollution prevention and control of antibiotics in the megacity Beijing, and provide a systematic methodology to understand the multimedia fate of antibiotics and their environmental effects in other wastewater irrigation region worldwide.

2

where A2 is the area of water (m ); USI is the wastewater irrigation rate (m/h); Zw is the fugacity capacity in water (mol/m3 Pa). Eq. (1) can also be written in a matrix form:

df / dt = Mf + N

⎛ f1 ⎞ ⎛ (E1 + Q1)/ V1 Z1 ⎞ ⎜f ⎟ (E + Q2)/ V2 Z2 ⎟ f = ⎜ 2 ⎟, N = ⎜ 2 , M f3 ⎜ ⎟ E3/ V3 Z3 ⎜ ⎟ ⎜ ⎟ 0 ⎝ ⎠ ⎝ f4 ⎠ 0 D21/ V1 Z1 D31/ V1 Z1 ⎛ − (D12 + D13 ⎞ ⎜ + DA + DR ) ⎟ 1 1 ⎜ ⎟ ⎜ / V1 Z1 ⎟ ⎜ − (D21 + D23 D12 / V2 Z2 D32 / V2 Z2 D42 / V2 Z2 ⎟ ⎜ ⎟ + D24 + DA2 ⎟ =⎜ ⎜ ⎟ + DR2)/ V2 Z2 ⎜ ⎟ − (D31 + D32 0 D13/ V3 Z3 D23/ V3 Z3 ⎜ ⎟ + )/ DR V Z 3 3 3 ⎜ ⎟ ⎜ − (D42 + DA 4 ⎟ 0 0 D24 / V4 Z4 ⎜ ⎟ + DR 4 )/ V4 Z4 ⎠ ⎝

2.1. Study area Beijing, the capital of China, locates between latitudes 39°26′N and 41°03′N and longitudes 115°25′E and 117°30′E, with an area of 16,411 km2 (Fig. S1). Characterized by hot, humid summers and generally cold, windy, dry winters, the area has a typical monsoon-influenced climate. Its annual temperature is ∼11.5 °C, and the average precipitation is ∼ 600 mm/a. In this region, the types of soil mainly include mountain meadow soil, mountain brown earth, cinnamon soil, moisture soil, bog soil, paddy soil, and aeolian sandy soil. While the major land-use types are agricultural land, garden land, forest land, grassland, residential and industrial land, transportation land, and water area. Beijing has a total of more than 200 rivers and streams all over the city, generally categorized into 5 main river systems: Yongding River, Chaobai River, Beiyun River, Juma River and Gou River.

The established model was programmed using MATLAB R2009b (7.9.0). The differential equations of Eq. (3) were solved with the Runge-Kutta method at monthly time steps to obtain fugacity values which were then used to calculate the antibiotic concentrations Ci = fi × Zi in the four compartments. The initial fugacity values were derived from a steady-state fugacity model (level III). 2.3. Model parameters Input parameters describing the physical environment (i.e., area/ depth of the water, soil and sediment), and physicochemical properties of the chemicals (i.e., the organic carbon-water partition coefficients, KOC; degradation half-lives) are required to conduct the model simulation. Herein, the environmental characteristics of the study area were collected from the Beijing Statistical Yearbook and from literatures [5,30–32], while default values were mainly taken from Mackay and Paterson [33] and Mackay [23] in absence of reliable literature data (Table S4). The physicochemical properties of antibiotics were mainly adopted from USEPA EPI suit v4.1 and from literatures (Table S1, S5). Besides the parameters mentioned above, the emission loads of antibiotics are crucial for the fugacity model because they are among the major driving forces for the simulation [34]. Here, the consumption-based estimates [35] were adopted to calculate the emission loads of antibiotics, by considering five classes of potential sources: urban populations, rural populations, swine, poultry and other animals (i.e. cattle, sheep, and fish). Among them, it was assumed that the human excretion from rural populations and all animal excretion were directly discharged into the environment because of the limitations of wastewater treatment facilities in rural areas and most livestock farms in China [5]. Therefore, the emission loads of antibiotics can be estimated

2.2. Multimedia fugacity model A level IV multimedia fugacity model was established and applied to quantitatively simulate the fate of the target antibiotics in the study area based on the approach of Mackay [23]. It is a dynamic non-steady state model for intermittent exposure in four bulk compartments including air, water, sediment and soil, which is a typical scenario in the environmental system [27]. Besides the general processes, such as antibiotic emissions to water and soil from human and animal sources, advection air/water flow in/out of area, exchange between inter-compartment, and degradation in the four bulk compartments, the environmental process of nondiffusive wastewater irrigation from water to soil was introduced into the model (Fig. S2). The non-steady-state mass balance equations for a specific time (t) can be established in terms of the transfer fluxes as below:

∑ (Dji f j ) − (DA (i) + DR (i) + ∑ Dij) × fi

(3)

where, f is the fugacity vector, M is the fate matrix, N is the emission rate vector:

2. Materials and methods

Vi Zi dfi / dt = Ei + Qi +

(2)

D23 = A2 USI ZW

(1) 76

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

Table 1 Emission loads of the target antibiotics in Beijing in 2013. Humana (tonneb)

Antibiotic

Animala (tonneb)

Name

Abbrev

Urban (WWTPs)

Urban (untreated)

Rural

Subtotal

Swine (urine)

Swine (feces)

Poultry

Other animals

Subtotal

Sulfachlorpyridazine Sulfadiazine Sulfamonomethoxine Sulfamethoxazole Sulfamethazine Sulfathiazole Trimethoprim Chlortetracycline Doxycycline Oxytetracycline Tetracycline Ciprofloxacin Difloxacin Enrofloxacin Fleroxacin Lomefloxacin Norfloxacin Ofloxacin Clarithromycin Erythromycin-H2Od Roxithromycin Tylosin

SCPc SDZ SMM SMX SMZ STZ TMP CTC DC OTC TC CFX DIFc EFXc FL LFX NFX OFX CTM ETM RTM TYLc

– 0.373 0.003 0.002 0.244 0 1.527 0.028 0.112 0.323 1.974 0.817 – – 0.485 0.554 1.872 4.993 0.114 2.351 0.751 –

– 0.312 0.002 0.001 0.113 0.001 0.759 0.056 0.235 0.321 1.837 0.614 – – 0.227 0.406 1.563 2.445 0.056 1.092 0.308 –

– 0.267 0.002 0.001 0.097 0.001 0.65 0.048 0.201 0.275 1.574 0.526 – – 0.194 0.348 1.339 2.095 0.048 0.936 0.264 –

– 0.952 0.007 0.004 0.454 0.002 2.936 0.132 0.548 0.919 5.385 1.957 – – 0.906 1.308 4.774 9.533 0.218 4.379 1.323 –

0.409 0.968 0.227 0.691 0.335 0.125 0.179 0.267 4.523 1.539 0.22 4.033 0.314 2.287 0.059 0.637 2.981 2.39 0.049 0.702 0.095 0.044

0.016 0.037 0.01 0.006 0.012 0.005 0.007 0.127 2.148 0.733 0.105 3.693 0.898 2.091 0.054 0.584 2.733 2.193 0.07 1.013 0.137 4.203

0.13 0.258 0.387 0.079 0.074 0.016 0.063 0.099 1.675 0.539 0.087 2.28 0.37 2.474 0.046 0.478 2.067 1.79 0.194 1.536 0.183 2.855

0.058 0.123 0.091 0.067 0.04 0.013 0.025 0.048 0.812 0.271 0.041 1.005 0.193 1.001 0.018 0.181 0.813 0.677 0.046 0.421 0.032 0.855

0.613 1.386 0.715 0.843 0.461 0.159 0.274 0.541 9.158 3.082 0.453 11.011 1.775 7.853 0.177 1.88 8.594 7.05 0.359 3.672 0.447 7.957

a b c d

Data was derived from Zhang et al. [5]. When preforming the fugacity model, it is necessary to convert tonne to mol. The antibiotic is not for human. Erythromycin-H2O is a major degradation product of erythromycin.

values, assuming that the parameters followed the log-normal distributions which were typically used to describe the environmental variables [26,31]. Generally, an assigned confidence factor of 2 or 3 is used to estimate the 2.5%-tile and 97.5%-tile to define the shape of the log-normal distribution, which will generate a distribution that have 95% of the values occurring within the estimation range [34]. In this study, the dispersion factor value of 3 was selected for KOC and reaction half-lives (Table S5), and the value of 2 for other parameters. For a lognormal distribution, the assigned confidence factor values of 2 and 3 are roughly equivalent to coefficients of variation of about 37% and 65% respectively, which are commonly considered to be reasonable high-biased estimates of the uncertainty [34].

as follows:

Eh = Ch × Xh × (1 − Ru × Rt × R c )

(4)

Ea (k ) = Ca (k ) × Xa (k )

(5)

E2 = Eh + Ea (1) × Pu + Ea (1) × (1 − Pu ) × Pfw

(6)

E3 = Ea (1) × (1 − Pu ) × (1 − Pfw ) + Ea (2) + Ea (3)

(7)

where Eh and Ea represent the antibiotic emissions from human and animals respectively (tonne/a); the subscript k is 1, 2, and 3 for swine, poultry, and other animals, respectively; Ch and Ca represent the consumption amount of antibiotics for human and animals, respectively (tonne/a); Xh and Xa represent the metabolism fraction of the unchanged and glucuronide conjugates of target antibiotics for human and animals, respectively; Ru, Rt and Rc respectively represent the proportion of urban populations, the treatment ratio of urban wastewater, and the removal efficiency of target chemicals in WWTPs; E2 and E3 respectively represent the emission loads into the water compartment and soil compartment (tonne/a); Pu and Pfw represent the excretion ratio of swine urine and the flush ratio of swine feces, respectively.

2.5. Risk assessment method To evaluate the potential ecological effects of antibiotics, the risk quotients (RQs) were employed using the formula[36]:

RQ =

Sensitivity coefficient was used to analyse the influence of parameter variation on the output of the fugacity model, by defining as the ratio of the relative variation of the predicted results to that of the input parameter:

ΔYi / Yi ΔXi / Xi

(9)

where PEC and PNEC are the abbreviations of predicted environmental concentration and predicted no effect concentration for antibiotics, respectively; Z indicates fugacity capacity (mol/m3 Pa); f indicates fugacity (Pa); M indicates molar mass (g/mol). The values of Z and f are estimated by the fugacity model, while a PNEC value in water compartment (PNECwat) for a given chemical is generally derived from toxicity data divided by an assessment factor [7]. As very few studies reported the risks of antibiotics in the soil and sediment compartments, the values of PNEC in sediment (PNECsed) and soil (PNECsoil) were estimated from PNECwat values based on the soilwater equilibrium partitioning method [28]:

2.4. Sensitivity and uncertainty analysis

SCi =

Z×f×M PEC = PNEC PNEC

(8)

where SCi represents the sensitivity coefficient of input parameter i; Xi and Yi represent the input parameter i and corresponding predicted result, respectively. After identifying the sensitive parameters, Monte Carlo simulation was employed to assess the overall uncertainty of the model output. The values for input parameters were randomly selected to replace discrete

PNECsed = PNECwat × Ksed

(10)

PNECsoil = PNECwat × Ksoil

(11)

where Ksed and Ksoil are the partition coefficient for sediment-water and 77

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

antibiotics in water associated with the study area were collected from previous studies and showed in Table 2 with the values of mean or median through the statistical analysis for the determination of their normal or lognormal distribution [31].

soil-water, respectively. Additionally, to deal with the uncertainty of risk analysis caused by the model parameters, Monte Carlo Simulation was applied once again to describe the variability of the input parameters and its propagation throughout the calculation. Consequently, the final risk levels were obtained in terms of probability distributions of the expected values rather than single point estimates [37].

3.2.1. Model performance evaluation Comparisons between the predicted annual mean concentrations and measured concentrations in water were considered herein, because few measured values of antibiotics in soil and sediment were available. In general, the fugacity model successfully simulated the reported concentration data for most antibiotics. The ratios of predicted annual mean concentrations to measured ones (P/M ratios) were ranged from 0.4 (SCP) to 3.0 (SDZ), except for DC, FL and LFX (Fig. S3). Previous studies showed that many factors would influence the predicted results of the fugacity model in environment, including emission rates, physicochemical parameters of chemicals in various compartments [26,27]. In this study, it was assumed that the animal emissions were discharged directly into the environment because of the limitations of treatment facilities in livestock farms in China, which might overestimate, to some extent, the emission loads for some antibiotics (i.e. SDZ, TMP, TC, and DIF). In Beijing, some simple treatment facilities (i.e. lagoons, sedimentation tanks) were available in a small portion of livestock farms, although they were not effective enough to remove the antibiotics [41]. On the other hand, as for the underestimation for SCP, SMX and EFX, this might be explained by the fact that the predicted concentrations represented the average of the whole study area, while the measured data were taken from limited sampling sites. Additionally, due to lack of local data, the measured concentrations for some antibiotics were collected around the study area within Hai River basin, which would lead to a large deviation in measured concentrations. Overall, comparisons between predicted and measured concentrations of most antibiotics in water showed reasonable agreement with an average difference of 0.2 logarithmic units. The largest variations for DC, FL and LFX might be related to the uncertainty of the input parameters (i.e. KOC), most of which were from the literature not from specific field investigations for the purpose of modeling.

3. Results and discussions 3.1. Overall characteristic of emission loads of antibiotics Based on the survey data of antibiotic usages in China [5], the consumption amount of antibiotics in Beijing was estimated, and then used to calculate the environmental loads of antibiotics. Results showed that the annual total emission for the 22 target antibiotics was estimated to be 104.2 tonnes (Table 1). It can be seen that the emission amounts varied much from each other, ranging from 0.16 tonnes (STZ) to 16.58 tonnes (OFX). The three fluoroquinolones (OFX, NFX, and CFX) showed the highest annual loads (> 10 tonnes), which might be related to their large consumption amounts due to their wide antimicrobial spectrum property and remarkable antibacterial ability [32]. Doxycycline was the other chemical with an annual load of about 10 tonnes. Among all potential sources considered, swine emission occupied the largest contribution (42%) to the total loads of antibiotics, followed by human (34%), poultry (17%) and other animals (7%). Relatively, the animal emission had major impacts on the local environment with more than 65% of the total emission loads. To some extent, the large discharge quantities for animals were most likely to be related with the high consumption of veterinary antibiotics in China. In the past several decades, antibiotics were widely and frequently used in animal feeds for growth promotion and disease treatment or prevention to satisfy the increase of human demand for animal products [30,32]. Comparing the per capita consumptions for the specific antibiotics, China showed higher than many countries in the world (i.e. Australia, France, Germany, Spain, and Sweden) [5]. Only with regard to the 22 antibiotics considered in this study, it was estimated that the consumption amounts for animals reached more than 240 tons in Beijing, about 60% of the animal usages in UK [38]. For human emission, the effluent of WWTPs, untreated urban wastewater, and rural wastewater contributed about 16%, 10%, and 9% of the total emission loads, respectively. The large discharge quantities for urban populations might be related with the large population amount in Beijing (> 20 million) and the limitation of the removal efficiency for antibiotics in WWTPs [39,40]. Previous study demonstrated that the removal efficiencies for fluoroquinolones, sulfonamides, and macrolides in eight WWTPs in Beijing were only 48–72%, 39–64%, and −34–69%, respectively [40]. In addition, besides the emission from WWTPs, the discharges of untreated urban wastewater should not be ignored, although most domestic and hospital wastewater in Beijing had been collected and treated by municipal sewage system [30,31]. As a sample, the total wastewater volume of Beijing (1.55 billion tonnes in 2013) was still greater by 18% than the water treatment capacity (1.31 billion tonnes in 2013) [30], which might bring a large proportion of antibiotics in untreated wastewater directly entering into the environment.

3.2.2. Predicted concentrations and seasonal trends As shown in Table 2, it can be seen that the antibiotics in air had extremely low concentrations due to their low volatilities. In water compartment, ETM showed the highest predicted concentrations (421.5 ng/L), followed by DC (341.0 ng/L). The other chemicals with relatively high concentrations were SDZ, TMP, NFX, and OFX in water (> 100 ng/L). This was consistent with the monitoring results in the Beiyun River which covers the most populated area of Beijing with a population of 14 million [19,31,42]. As a whole, the concentration levels for antibiotics in the water of Beijing were comparable with those in other rivers of China (i.e. Hai River, Huangpu River, Pearl River) [8], but were higher than those in Australia [43], Japan [44], USA [6], and many European countries (i.e. Italy, France) [45,46]. In the soil and sediment compartments, the highest concentrations were dominated by fluoroquinolones. This might be related with the fact that most fluoroquinolones had a stronger sorption affinity for particles [47]. In sediments, OFX showed the highest predicted level with mean concentration of 61.5 ng/g, followed by EFX (50.0 ng/g), CFX (41.1 ng/g), and NFX (40.5 ng/g), while the concentrations of other chemicals were less than 40 ng/g. Overall, the concentrations of antibiotics in sediments of the study area were generally higher than those in USA [48,49], which might be explained that the usage proportion for fluoroquinolones in China was generally higher than that in USA. Meanwhile, with the dynamic model, the temporal trends of antibiotics concentrations were simulated based on a non-steady-state assumption. It can be seen from Fig. 1 that differences in predicted concentrations for the target antibiotics existed among the 12 months. Changes in the temperature with season were significantly reflected in

3.2. Predicted concentrations of antibiotics and model validation Based on the emission data estimated above, the predicted concentrations of antibiotics in various compartments were simulated by the level IV fugacity model. The temporal trends of predicted concentrations were showed in Fig. 1 at monthly steps to characterize the seasonal variant of antibiotics levels, grouped into 4 categories. Additionally, to evaluate the model performance, the measured data of 78

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

Fig. 1. Temporal trends of the predicted concentrations of target antibiotics in 2013.

Table 2 Predicted annual mean concentrations of the target antibiotics in various compartments in 2013, and measured concentrations (median/mean) in water within or around Beijing. Antibiotic

SCP SDZ SMM SMX SMZ STZ TMP CTC DC OTC TC CFX DIF EFX FL LFX NFX OFX CTM ETM RTM TYL

Air (ng/L)

*

Water (ng/L)

Soil (ng/g)

Sediment (ng/g)

Measured value (ng/L)

Mean

S.D.**

Mean

S.D.

Mean

S.D.

Mean

S.D.

1.26E-12 2.82E-13 1.46E-14 1.17E-13 2.57E-10 5.26E-15 2.24E-13 1.02E-26 2.26E-26 1.35E-28 4.39E-25 1.09E-20 7.87E-21 3.59E-20 2.29E-18 8.34E-19 2.34E-19 6.31E-23 4.45E-30 2.38E-29 3.12E-32 2.60E-40

3.84E-12 9.75E-13 5.04E-14 3.83E-13 3.73E-10 1.82E-14 7.66E-13 1.87E-26 7.63E-26 4.64E-28 9.73E-25 3.78E-20 2.73E-20 1.24E-19 7.93E-18 2.89E-18 8.09E-19 2.19E-22 1.54E-29 8.22E-29 8.55E-32 8.58E-40

42.57 166.01 26.88 55.30 69.42 10.81 117.70 5.68 340.95 98.26 65.46 87.44 7.79 36.08 95.72 91.59 250.99 139.89 41.15 421.52 32.05 39.94

5.99 25.32 7.90 12.04 11.53 1.71 5.40 0.03 31.98 24.48 15.17 1.38 0.23 0.53 4.07 4.82 9.13 19.91 7.55 8.80 0.61 2.07

0.004 0.014 0.031 0.000 0.010 0.003 0.167 0.014 0.702 0.061 0.532 0.705 0.594 1.721 0.070 0.735 0.460 1.360 0.060 0.082 0.425 0.176

0.002 0.010 0.021 0.000 0.007 0.002 0.131 0.011 0.529 0.050 0.481 0.542 0.454 1.317 0.042 0.549 0.356 1.117 0.028 0.062 0.325 0.135

0.063 0.110 0.103 0.006 0.043 0.062 2.204 0.788 15.806 0.856 23.155 41.097 9.037 49.669 3.658 26.689 40.459 61.496 0.643 30.148 24.075 6.346

0.009 0.017 0.030 0.001 0.007 0.010 0.101 0.004 1.482 0.213 5.365 0.650 0.271 0.725 0.156 1.405 1.472 8.754 0.118 0.629 0.456 0.330

121.0 (n = 31) ***, 56.8 (n = 68) b 10.8 (n = 9) a 90.8 (n = 211) a 22.80 (n = 68) b 8.0 c 54.2 (n = 77) e 7.77 (n = 3) f 46.9 c 39.5 (n = 188) a 26.0 (n = 188) a 64.2 (n = 154) a 3.4 d 60.50 (n = 80) a 17.6 d 10.7 d 112.8 (n = 228) a 192.2 (n = 246) a 36.70 g 300.2 (n = 68) b 30.0 (n = 186) a 19.4 d

a

The units of predicted concentrations are generally converted from mol/m3 to ng/L (air and water) or ng/g (soil and sediment). S.D. indicates standard deviation. *** n in brackets means sample number. Data from a: Chen et al [7]; b: Ma et al [19]; c: Bu et al [8]; d: Gao et al [47]; e: Wang et al [31]; f: Xue et al [57]; g: Zhang et al [32].

*

**

efficiencies of antibiotics are also the factors influencing the seasonal variant of concentration levels [7]. Generally, the removal rates of conventional treatment process for antibiotics by WWTPs may decrease dramatically with decreasing temperature [51].

temporal changes of antibiotic concentrations. For each of the environmental compartments, most antibiotics in low temperature season had relatively higher levels than in high temperature season (May to September), which coincided with previous investigations [19,50]. It would be attributed to lower runoff, decreased flow rate of river, slower photolysis, thermal degradation and biodegradation in low temperature season [7,29]. In Beijing, the average water temperature during winter season is about 5 °C, much lower than 25 °C in summer season. While the precipitation in summer season accounts for over 80% of the annual amount. Additionally, the consumption pattern and removal

3.3. Transfer fluxes and mass distribution The transfer fluxes of antibiotics between various compartments were estimated from the fugacity model results (Table S6), and the mass contents of antibiotics in water, soil and sediment were calculated using 79

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

Fig. 2. Distribution ratios of the transfer fluxes of antibiotics for (a) source emission, (b) degradation and (c) inter-media transport; and (d) the mass contents among the water, soil and sediment compartments. A2W: air to water; A2S: air to water; W2A: water to air; W2S: water to soil; W2sed: water to sediment; S2A: soil to air; S2W: soil to water; Sed2W: sediment to water.

to transporting or partitioning from water to the sediment, with more than 50% of the total mass. Besides sediment, the soil was another important destination compartment for these three categories of antibiotics, accounting for more than 40% of the total mass. It might be attributed to the wastewater irrigation and animal waste application [17]. The results also agreed with the previous study which found that antibiotics predominated in the sediment and soil matrix for all the 58 basins in China [5]. In water compartment, only some sulfonamides (SMX, SDZ, SCP, SMZ, and STZ) were found containing more than 10% of the total mass. Additionally, it should be noted that the ratio of the total degradation mass of antibiotics to the emission loads was about 81.5%, suggesting that antibiotics did not have a strong potential to accumulate in the environment. Nevertheless, antibiotics still can exist in the environment “persistently” due to the continuous discharge from human and animals because of their widespread use [7].

the predicted concentrations and volumes of the bulk compartments. Fig. 2 shows the distribution ratios of transfer fluxes and mass contents for the target antibiotics. It can be seen that, among various pathways, source emission made the most contribution for the input fluxes of antibiotics, with ∼63% received by water and 37% by soil compartment. By contrast, degradation was the dominant eliminating pathway from the study area, and degradation in soil was the major loss process, accounting for ∼74% of the total degradation amount (Fig. 2b). The results implied that the rate of disappearance for antibiotics in the environment mainly determined by their degradation rate in the soil. The dominant transfer process of antibiotics in environment of the study area was the wastewater irrigation from water to soil, with approximately 49% of the total transfer fluxes between adjacent compartments (Fig. 2c). According to the second national survey of the wastewater irrigation, the Beijing region was one of the representative regions of wastewater irrigation on farmland [5]. Previous study also showed that over 90% of the wastewater discharge from Beijing was used for irrigation in agricultural areas within approximately 100 km downstream of Beijing [20], and formed a mega-scale wastewater irrigation system from the up and downstream areas [17]. Additionally, for tetracyclines, fluoroquinolones and macrolides, the sedimentation from water to sediment contributed more than 36% of total transfer fluxes, suggesting these three categories of antibiotics were prone to enter into the sediment due to their hydrophobicity in surface water. While for sulfonamides, besides wastewater irrigation, erosion from soil to water was the other dominant inter-media transfer process with 37% of total transfer fluxes. As shown in Fig. 2d, tetracyclines, fluoroquinolones, and macrolides were mainly distributed into the sediments because of their being prone

3.4. Sensitivity and uncertainty analysis A total of 13 input parameters for the fugacity model were identified with the absolute sensitivity coefficients (ASC) exceeding 0.1. The significant parameters, with the ASC value higher than 0.5, were source emission, KOC, reaction half-lives in soil and sediment, area and depth of soil, depth of sediment, volume fractions of water in sediment, and water velocity (Fig. S4). Among them, source emission was the most important parameter influencing the model outputs (ASC > 0.9), followed by KOC, and reaction half-lives in soil and sediment, which all made a positive contribution to the antibiotics concentrations. Generally, the higher KOC can increase the portion of chemicals in sediments, and the longer reaction half-lives will cause heavier Fig. 3. Uncertainty analysis for the fugacity model with the logarithmic transformed differences (LogSQR) between the first and the third quartiles of the predicted concentrations of antibiotics.

80

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

Table 3 Risk quotients of antibiotics in the water, soil and sediment in Beijing based on annual mean concentrations of antibiotics predicted by the fugacity model in 2013. Antibiotic

Water Mean

SCP SDZ SMX SMZ STZ TMP CTC DC OTC TC CFX EFX LFX NFX OFX CTM ETM RTM TYL

1.2E-02 8.0E-02 2.8E-02 4.2E-01 7.8E-03 1.1E-04 7.5E-03 1.1E-01 5.4E-03 3.1E-01 8.0E-01 4.5E + 00 7.8E-03 7.3E-01 9.6E-02 4.1E + 00 4.1E + 00 7.6E + 00 9.3E-01

Soil S.D.

1.1E-02 3.4E-02 2.7E-02 3.1E-01 5.3E-03 1.9E-04 2.6E-03 1.0E-01 1.4E-03 1.2E-01 2.1E-01 1.2E + 00 6.5E-03 2.8E-01 5.9E-02 1.8E + 00 9.4E-01 1.7E + 00 7.9E-01

Percentiles

Mean

75th

90th

1.4E-02 9.6E-02 3.3E-02 5.2E-01 9.8E-03 1.2E-04 8.9E-03 1.4E-01 6.2E-03 3.8E-01 9.2E-01 5.2E + 00 9.7E-03 8.7E-01 1.2E-01 5.0E + 00 4.7E + 00 8.6E + 00 1.2E + 00

2.3E-02 1.2E-01 5.4E-02 7.7E-01 1.4E-02 2.5E-04 1.1E-02 2.3E-01 7.2E-03 4.7E-01 1.1E + 00 6.1E + 00 1.5E-02 1.1E + 00 1.7E-01 6.4E + 00 5.4E + 00 9.8E + 00 1.8E + 00

7.9E-04 1.6E-03 7.7E-03 2.1E-03 3.1E-04 4.8E-06 1.1E-04 1.6E-04 3.3E-04 3.8E-03 8.9E-03 1.7E-02 6.3E-04 1.5E-02 6.9E-03 4.1E-01 5.1E-02 4.7E-02 3.2E-01

Sediment S.D.

Percentiles

1.1E-03 1.2E-03 6.4E-03 2.9E-03 3.8E-04 1.3E-05 3.9E-05 1.3E-04 7.6E-05 1.4E-03 2.4E-03 3.7E-03 3.0E-04 3.7E-03 5.8E-03 1.6E-01 1.0E-02 1.0E-02 3.2E-01

Mean

75th

90th

9.4E-04 1.9E-03 9.6E-03 2.4E-03 3.5E-04 4.8E-06 1.3E-04 2.0E-04 3.8E-04 4.5E-03 1.0E-02 1.9E-02 7.7E-04 1.7E-02 8.2E-03 4.9E-01 5.7E-02 5.3E-02 3.9E-01

1.7E-03 2.9E-03 1.5E-02 4.6E-03 6.0E-04 9.7E-06 1.6E-04 3.0E-04 4.3E-04 5.6E-03 1.2E-02 2.2E-02 1.0E-03 2.0E-02 1.2E-02 6.2E-01 6.4E-02 6.0E-02 6.5E-01

1.6E-02 1.1E-02 2.4E-02 2.7E-02 1.1E-03 9.7E-05 1.4E-03 7.7E-03 7.0E-03 5.4E-02 4.0E-01 8.8E-01 8.6E-03 3.9E-01 3.8E-01 1.4E + 01 4.1E + 00 2.1E + 00 3.8E + 00

S.D.

1.5E-02 4.8E-03 2.3E-02 2.2E-02 7.3E-04 1.6E-04 4.7E-04 7.1E-03 1.8E-03 2.0E-02 1.0E-01 2.4E-01 7.1E-03 1.5E-01 2.3E-01 6.0E + 00 9.4E-01 4.6E-01 3.2E + 00

Percentiles 75th

90th

2.0E-02 1.3E-02 2.9E-02 3.4E-02 1.4E-03 1.1E-04 1.6E-03 9.8E-03 8.0E-03 6.4E-02 4.6E-01 1.0E + 00 1.1E-02 4.7E-01 4.7E-01 1.7E + 01 4.7E + 00 2.4E + 00 4.7E + 00

3.2E-02 1.7E-02 4.7E-02 5.1E-02 1.9E-03 2.2E-04 2.0E-03 1.6E-02 9.3E-03 8.0E-02 5.3E-01 1.2E + 00 1.7E-02 5.9E-01 6.6E-01 2.1E + 01 5.4E + 00 2.7E + 00 7.3E + 00

S.D. indicates standard deviation.

Comparatively speaking, the estimated concentrations in sediment and soil had the larger uncertainties than those in air and water compartments, which were possibly associated with the influential parameters mainly related to the soil and sediment compartments (i.e. reaction half-lives of antibiotics in soil and sediment). When compared with other categories of antibiotics, sulfonamides showed larger uncertainties in water. This could be explained by their relatively great variation of the input parameters (KOC and reaction half-lives) in water compartment. Overall, although the variation of parameters might bring moderate uncertainty within one order of magnitude, the model could successfully generate statistical concentration distribution of antibiotics. However, it should be noted that some of the input parameters used in the model were the average estimates from the study area. Therefore, a better simulation would require a more thorough and possibly site-specific treatment. In order to minimize the model uncertainties and then obtain a consistency with the real environment, detailed field investigations and experimental programs should be devised and developed especially focusing on those most influential parameters (i.e. source emission, KOC, water velocity).

Fig. 4. Cumulative probabilities of the risk quotients for LFX and CTM in the soil.

contamination due to slower decomposition. On the contrary, the area and depth of soil, depth of sediment, and water velocity were identified as the negatively influencing parameters determined the total volume of bulk compartments and water change. The volume dilution and higher outflow rate of water will certainly reduce the antibiotic levels. Relatively, the other parameters (area and depth of water, molecular diffusivity in sediment) posed less influence on the model outputs, with the ASC value between 0.1 and 0.3. Furthermore, uncertainty analysis was performed using Monte Carlo simulation to reflect the influence of the parameter variability and to address the antibiotic concentration variation. The dispersion factor value of 3 was selected for KOC and reaction half-lives (Table S5), the value of 2 for other significant parameters identified above (i.e. emission rate, water velocity, depths of water, soil and sediment, and so on). The simulation was carried out repeatedly 10,000 times to obtain the probability distributions of the antibiotic concentrations. The logarithmic transformed differences between the third and the first quartiles of the predicted concentrations were employed to quantify the uncertainties [5]. As shown in Fig. 3, the dispersions of the concentrations for most of antibiotics in water, soil and sediment compartments were moderately large, and generally covered 0.1–1.1 orders of magnitude.

3.5. Environmental risk of antibiotics According to the most sensitive values of effect concentrations, the PNEC values were calculated with the acute or chronic toxicity data of antibiotics obtained from the available literatures (Table S7), and the risk quotients (RQs) were then estimated utilizing the Monte-Carlo based risk assessment approach. Due to lack of the toxicity data for SMM, FL and DIF, their RQs were ignored. The Monte Carlo simulation was done by programming an evaluation algorithm with Matlab R2009b software, considering 10,000 iterations. Preliminary evaluations suggested that > 3000 simulations were sufficient to ensure consistent results with smooth probability distributions. To better elucidate the risk levels posed by antibiotics, the calculated RQs were classified into three risk levels: 0.01–0.1, low risk; 0.1–1, medium risk; and > 1, high risk [52]. It can be seen from Table 3 that several fluoroquinolones and macrolides, including CFX, EFX, LFX, NFX, OFX, CTM, ETM and TYL, were likely to pose high risk to aquatic organisms in the water and sediment compartments, due to their high concentrations and/or toxicity (Fig. S5a, S5b). The average RQs of these chemicals accounted for 81

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

Appendix A. Supplementary data

more than 80% of the total hazard index (THI) which was calculated by summing the individual hazard index for all antibiotics. Therefore, it is worthwhile to pay attention to the elimination of these antibiotics in aquatic environment and the control of their potential hazard, because some of them (i.e. fluoroquinolones) could even be taken up and accumulated by food chain and then posed potential risk to human health [53]. Additionally, TC, OTC, SMX, and RTM also would cause at least medium risks in the water and sediment. In contrast to the twelve antibiotics mentioned above, the RQs for other antibiotics were mostly less than 0.1 in the aquatic environment, suggesting that they were unlikely to cause significantly adverse toxic effects on the selected organisms. Comparatively speaking, the potential ecological risks posed by the target antibiotics considered in this study in aquatic environment decreased in the order of fluoroquinolones > macrolides > tetracyclines > sulfonamides. On the other hand, even though, sorption to soil appears to be a dynamically continuous process, the predicted concentrations in soil for most antibiotics were unlikely to cause significant risk, even when considering the worst scenario case analysis (Fig. S5c). As a whole, the risk level of antibiotics in soil was acceptable. However, comparing the predicted concentrations of specific antibiotics in soil with the toxicity data, the presence of LFX and CTM would represent medium risk to soil microorganisms (Fig. 4). Furthermore, it should be noticed that the environmental organisms are generally exposed to mixed substances including different antibiotics and other coexisting pollutants (i.e. metals, pesticide), which can be more significant than individual effects due to synergistical actions [54–57]. For example, Miguel et al. [55] observed strong synergistic interaction between ERY and TET both in the cyanobacterium and the green alga. Christensen et al. [56] also found synergistic effects of the mixture, OTC and ETM, in the green alga. Therefore, special attention still should be paid to these antibiotics in soil to target the lowest threats to the ecosystem safety and human health.

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jhazmat.2018.01.033. References [1] M. Hvistendahl, China takes aim at rampant antibiotic resistance, Science 336 (2012) 795. [2] C. Nathan, O. Cars, Antibiotic resistance–problems, progress, and prospects, N. Engl. J. Med. 371 (2014) 1761–1763. [3] E.J. Rosi-Marshall, J.J. Kelly, Antibiotic stewardship should consider environmental fate of antibiotics, Environ. Sci. Technol. 49 (2015) 5257–5258. [4] H.X. Wang, N. Wang, B. Wang, Q. Zhao, H. Fang, C.W. Fu, C.X. Tang, F. Jiang, Y. Zhou, Y. Chen, Q.W. Jiang, Antibiotics in drinking water in Shanghai and their contribution to antibiotic exposure of school children, Environ. Sci. Technol. 50 (2016) 2692–2699. [5] Q.Q. Zhang, G.G. Ying, C.G. Pan, Y.S. Liu, J.L. Zhao, Comprehensive evaluation of antibiotics emission and fate in the River basins of China: source analysis, multimedia modeling, and linkage to bacterial resistance, Environ. Sci. Technol. 49 (2015) 6772–6782. [6] D.W. Kolpin, E.T. Furlong, M.T. Meyer, E.M. Thurman, S.D. Zaugg, L.B. Barber, H.T. Buxton, Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999–2000: a national reconnaissance, Environ. Sci. Technol. 36 (2002) 1202–1211. [7] H.Y. Chen, L.J. Jing, Y.G. Teng, J.S. Wang, Characterization of antibiotics in a largescale river system of China: occurrence pattern, spatiotemporal distribution and environmental risks, Sci. Total Environ. 618 (2018) 409–418. [8] Q. Bu, B. Wang, J. Huang, S. Deng, G. Yu, Pharmaceuticals and personal care products in the aquatic environment in China: a review, J. Hazard. Mater. 262 (2013) 189–211. [9] Y.H. Jiang, M.X. Li, C.S. Guo, D. An, J. Xu, Y. Zhang, B.D. Xi, Distribution and ecological risk of antibiotics in a typical effluent–receiving river (Wangyang River) in north China, Chemosphere 112 (2014) 267–274. [10] C.Q. Chen, J. Li, P.P. Chen, R. Ding, P.F. Zhang, X.Q. Li, Occurrence of antibiotics and antibiotic resistances in soils from wastewater irrigation areas in Beijing and Tianjin, China, Environ. Pollut. 193 (2014) 94–101. [11] P.A. Segura, M. Francois, C. Gagnon, S. Sauvé, Review of the occurrence of antiinfectives in contaminated wastewaters and natural and drinking waters, Environ. Health Perspect. 117 (2009) 675–684. [12] M.S. Kostich, A.L. Batt, J.M. Lazorchak, Concentrations of prioritized pharmaceuticals in effluents from 50 large wastewater treatment plants in the US and implications for risk estimation, Environ. Pollut. 184 (2014) 354–359. [13] Y. Belkaid, T.W. Hand, Role of the Microbiota in immunity and inflammation, Cell 157 (2014) 121–141. [14] N.J. Ashbolt, A. Amézquita, T. Backhaus, S.P. Borriello, K.K. Brandt, P. Collignon, A. Coors, R. Finley, W.H. Gaze, T. Heberer, J.R. Lawrence, D.G.J. Larsson, S.A. McEwen, J.J. Ryan, J. Schönfeld, P. Silley, J.R. Snape, C. van den Eede, E. Topp, Human health risk assessment (HHRA) for environmental development and transfer of antibiotic resistance, Environ. Health Perspect. 121 (2013) 993–1001. [15] K.K. Brandt, A. Amézquita, T. Backhaus, A. Boxall, A. Coors, T. Heberer, J.R. Lawrence, J. Lazorchakk, J. Schönfeld, J.R. Snape, Y.G. Zhu, E. Topp, Ecotoxicological assessment of antibiotics: a call for improved consideration of microorganisms, Environ. Int. 85 (2015) 189–205. [16] J.L. Liu, M.H. Wong, Pharmaceuticals and personal care products (PPCPs): a review on environmental contamination in China, Environ. Int. 59 (2013) 208–224. [17] F. Heeb, H. Singer, B. Pernet-Coudrier, W. Qi, H. Liu, P. Longre´ıe, B. M¨uller, M. Berg, Organic micropollutants in rivers downstream of the megacity Beijing: sources and mass fluxes in a large-scale wastewater irrigation system, Environ. Sci. Technol. 46 (2012) 8680–8688. [18] G.H. Dai, B. Wang, J. Huang, R. Dong, S.B. Deng, G. Yu, Occurrence and source apportionment of pharmaceuticals and personal care products in the Beiyun River of Beijing, China, Chemosphere 119 (2015) 1033–1039. [19] R.X. Ma, B. Wang, L.N. Yin, Y.Z. Zhang, S.B. Deng, J. Huang, Y.J. Wang, G. Yu, Characterization of pharmaceutically active compounds in Beijing, China: occurrence pattern, spatiotemporal distribution and its environmental implication, J. Hazard. Mater. 323 (2017) 147–155. [20] B. Pernet-Coudrier, W.X. Qi, H.J. Liu, B. Müller, M. Berg, Sources and pathways of nutrients in the semi-arid region of Beijing−Tianjin, China Environ. Sci. Technol. 46 (2012) 5294–5301. [21] I. Muñoz, M.J. Gómez-Ramos, A. Agüera, A.R. Fernández-Alba, J.F. García-Reyes, A. Molina-Díaz, Chemical evaluation of contaminants in wastewater effluents and the environmental risk of reusing effluents in agriculture, Trends Anal. Chem. 28 (2009) 676–694. [22] C.E. Chen, H. Zhang, G.G. Ying, L.J. Zhou, K.C. Jones, Passive sampling: a costeffective method for understanding antibiotic fate, behaviour and impact, Environ. Int. 85 (2015) 284–291. [23] D. Mackay, Multimedia Environmental Models: The Fugacity Approach, 2nd ed., CRC, 2001. [24] H. Cao, S. Tao, F. Xu, R.M. Coveney, J. Cao, B. Li, W. Liu, X. Wang, J. Hu, W. Shen, Multimedia fate model for hexachlorocyclohexane in Tianjin, China Environ. Sci. Technol. 38 (2004) 2126–2132. [25] S.J. Khan, J.E. Ongerth, Modelling of pharmaceutical residues in Australian sewage

4. Conclusion The level IV fugacity model and Monte-Carlo based risk assessment approach were, in turn, established to understand the multimedia fate of antibiotics in Beijing and to evaluate their potential environmental risks, respectively. Based on the consumption survey and statistics, more than one hundred tonnes of antibiotics were estimated to be annually discharged into the environment from swine (42%), human (34%), poultry (17%) and other livestock (7%), and, to some extent, led to considerable concentration levels for some antibiotics in the aquatic environment. Relatively, the predicted concentrations of most antibiotics in low temperature season were higher than those in high temperature season. Sensitivity analysis indicated that source emission, KOC, degradation rate in soil, and water velocity had the strongest influence on the model outputs. Degradation, especially in soil, was found to be the dominant eliminating pathway from the study area. Although the process of nondiffusive wastewater irrigation from water to soil dominated the largest proportion of the total transfer fluxes between adjacent compartments, the risk level of antibiotics in soil was acceptable, which might be related to that more than 70% of the total degradation amounts were disappeared in the soil. However, results of Monte-Carlo based risk assessment showed that fluoroquinolones and macrolides were likely to pose high potential risk on the aquatic environment. Effort should be dedicated to further investigate the antibiotic contamination in the aquatic environment of Beijing. Acknowledgments This study was financially supported by Beijing Natural Science Foundation (8172030) and Major Science and Technology Program for Water Pollution Control and Treatment of China (2017ZX07302). 82

Journal of Hazardous Materials 348 (2018) 75–83

H. Chen et al.

across China, Environ. Pollut. 204 (2015) 223–232. [43] A.J. Watkinson, E.J. Murby, D.W. Kolpin, S.D. Costanzo, The occurrence ofantibiotics in an urban watershed: from wastewater to drinking water, Sci. Total Environ. 407 (2009) 2711–2723. [44] A. Murata, H. Takada, K. Mutoh, H. Hosoda, A. Harada, N. Nakada, Nationwide monitoring of selected antibiotics: distribution and sources of sulfonamides, trimethoprim, and macrolides in Japanese rivers, Sci. Total Environ. 409 (2011) 5305–5312. [45] V. Emmanuelle, C.O. Cécile, Screening of pharmaceuticals and hormones at the regional scale, in surface and groundwaters intended to human consumption, Environ. Pollut. 159 (2011) 2929–2934. [46] I.T. Carvalho, L. Santos, Antibiotics in the aquatic environments: A review of the European scenario, Environ. Int. 94 (2016) 736–757. [47] L.H. Gao, Y.L. Shi, W.H. Li, H.Y. Niu, J.M. Liu, Y.Q. Cai, Occurrence, distribution and bioaccumulation of antibiotics in the Haihe River in China, J. Environ. Monit. 14 (2012) 1248–1255. [48] R.T. Pei, S.C. Kim, H.C. Kenneth, P. Amy, Effect of River landscape on the sediment concentrations of antibiotics and corresponding antibiotic resistance genes (ARG), Water Res. 40 (2006) 2427–2435. [49] S.C. Kim, K. Carlson, Temporal and spatial trends in the occurrence of human and veterinary antibiotics in aqueous and River sediment matrices, Environ. Sci. Technol. 41 (2007) 50–57. [50] Y. Luo, L. Xu, M. Rysz, Y. Wang, H. Zhang, P.J. Alvarez, Occurrence and transport of tetracycline, sulfonamide, quinolone, and macrolide antibiotics in the Haihe River basin, China Environ. Sci. Technol. 45 (2011) 1827–1833. [51] Q. Sui, J. Huang, S.B. Deng, W.W. Chen, G. Yu, Seasonal variation in the occurrence and removal of pharmaceuticals and personal care products in different biological wastewater treatment processes, Environ. Sci. Technol. 45 (2011) 3341–3348. [52] W. Xu, W. Yan, X. Li, Y. Zou, X. Chen, W. Huang, L. Miao, R. Zhang, G. Zhang, S. Zou, Antibiotics in riverine runoff of the pearl River delta and pearl River Estuary, China: concentrations, mass loading and ecological risks, Environ. Pollut. 182 (2013) 402–407. [53] X.Y. Li, Y.F. Xie, C.L. Li, H.N. Zhao, H. Zhao, N. Wang, Investigation of residual fluoroquinolones in a soil–vegetable system in an intensive vegetable cultivation area in northern China, Sci. Total. Environ. 468–469 (2014) 258–264. [54] K. Chen, J.L. Zhou, Occurrence and behavior of antibiotics in water and sediments from the Huangpu River, Shanghai, China, Chemosphere 95 (2014) 604–612. [55] G.P. Miguel, G. Soledad, R.P. Ismael, L. Francisco, R. Roberto, B. Karina, M. Eduardo, F.P. Francisca, Toxicity of five antibiotics and their mixtures towards photosynthetic aquatic organisms: implications for environmental risk assessment, Water Res. 47 (2013) 2050–2064. [56] A.M. Christensen, F. Ingerslev, A. Baun, Ecotoxicity of mixtures of antibiotics used in aquacultures, Environ. Toxicol. Chem. 25 (2006) 2208–2215. [57] D.L. Wang, X.D. Wu, Z.F. Lin, Y.Y. Ding, A comparative study on the binary and ternary mixture toxicity of antibiotics towards three bacteria based on QSAR investigation, Environ. Res. 162 (2018) 127–134.

by quantities of use and fugacity calculations, Chemosphere 54 (2004) 355–367. [26] J.T. Ao, J.W. Chen, F.L. Tian, X.Y. Cai, Application of a level IV fugacity model to simulate the long-term fate of hexachlorocyclohexane isomers in the lower reach of yellow River basin, China, Chemosphere 74 (2009) 370–376. [27] J. Chi, Q. Yang, Effects of Potamogeton crispus L. on the fate of phthalic acid esters in an aquatic microcosm, Water Res 46 (2012) 2570–2578. [28] X.L. Wu, L. Xiang, Q.Y. Yan, Y.N. Jiang, Y.W. Li, X.P. Huang, H. Li, Q.Y. Cai, C.H. Mo, Distribution and risk assessment of quinolone antibiotics in the soils from organic vegetable farms of a subtropical city, South China Sci. Total Environ. 487 (2014) 399–406. [29] H.Y. Chen, L.J. Jing, Y.G. Teng, K. Kummerer, Antibiotics in the aquatic environment-A review-Part I, Chemosphere 5 (2009) 417–434. [30] Q.F. Xia, Beijing Statistical Yearbook, China Statistics Press, Beijing, China, 2013. [31] B. Wang, G.H. Dai, S.B. Deng, H. Huang, Y.J. Wang, G. Yu, Linking the environmental loads to the fate of PPCPs in Beijing: considering both the treated and untreated wastewater sources, Environ. Pollut. 202 (2015) 153–159. [32] Z.X. Zhang, B. Wang, H.L. Yuan, R.X. Ma, J. Huang, S.B. Deng, Y.J. Wang, G. Yu, Discharge inventory of pharmaceuticals and personal care products in Beijing, China, Emerg. Contam. 2 (2016) 148–156. [33] D. Mackay, S. Paterson, Evaluating the multimedia fate of organic chemicals: a level III fugacity model, Environ. Sci. Technol. 25 (1991) 427–436. [34] J. Kim, D.E. Powell, L. Hughes, D. Mackay, Uncertainty analysis using a fugacitybased multimedia mass-balance model: application of the updated EQC model to decamethylcyclopentasiloxane (D5), Chemosphere 93 (2013) 819–829. [35] C. Marx, V. Muhlbauer, S. Schubert, R. Oertel, M. Ahnert, P. Krebs, V. Kuehn, Representative input load of antibiotics to WWTPs: predictive accuracy and determination of a required sampling quantity, Water Res. 76 (2015) 19–32. [36] ECA (European Chemical Agency), Guideline on the Environmental Risk Assessment of Medicinal Products for Human Use, Doc Ref. EMEA/CHMP/SWP/ 4447/00, (2006). [37] M. Mari, M. Nadal, M. Schuhmacher, J.L. Domingo, Exposure to heavy metals and PCDD/Fs by the population living in the vicinity of a hazardous waste landfill in Catalonia, Spain: health risk assessment, Environ. Int. 35 (2009) 1034–1039. [38] S. Eckford, K. Grace, C. Harris, H. Reeves, C. Teale, C. Tallentire, UK Veterinary Antibiotic Resistance and Sales Surveillance, (2013) https://www.gov.uk/ government/publications/veterinary-antimicrobial-resistance-and-salessurveillance-2013. [39] W. Li, Y. Shi, L. Gao, J.M. Liu, Y.Q. Cai, Occurrence and removal of antibiotics in a municipal wastewater reclamation plant in Beijing, China, Chemosphere 92 (2013) 435–444. [40] L.H. Gao, Y.L. Shi, W.H. Li, H.Y. Niu, J.M. Liu, Y.Q. Cai, Occurrence of antibiotics in eight sewage treatment plants in Beijing, China, Chemosphere 86 (2012) 665–671. [41] L.J. Zhou, G.G. Ying, S. Liu, R.Q. Zhang, H.J. Lai, Z.F. Chen, C.G. Pan, Excretion masses and environmental occurrence of antibiotics in typical swine and dairy cattle farms in China, Sci. Total Environ. 444 (2013) 183–195. [42] Z. Wang, X.H. Zhang, Y. Huang, H. Wang, Comprehensive evaluation of pharmaceuticals and personal care products (PPCPs) in typical highly urbanized regions

83