Spatiotemporal distributions and ecological risk assessment of pharmaceuticals and personal care products in groundwater in North China

The contamination of surface water by pharmaceuticals and personal care products (PPCPs) has attracted widespread attention, but data regarding their impacts on groundwater (GW) are sparse. In river–GW interaction areas, rivers are likely an important source of PPCPs in aquifers, especially rivers impacted by sewage treatment plant effluent. Understanding the characterization, transport, and risk is valuable for the effective protection of vital aquatic ecosystem services, environmental health, and drinking water supplies. To attain this objective, statistics with spatial analysis and ecological risk were used to assess the effects of artificial recharge (AR) engineering on 16 PPCPs in aquifers in North China. The results indicated that 15 PPCPs were detected in unconfined and confined aquifers, with a few PPCPs being detected up to 1,000 ng/L. The most frequently detected PPCPs were sulfisoxazole, sulfachloropyridazine, sulfamerazine, sulfamethazine, sulfamethoxazole, and ibuprofen. In addition, the spatial and seasonal variations in most PPCPs were significant. Furthermore, the maximum concentrations were compared to the predicted no-effect concentrations to evaluate the ecological risk, and four PPCPs were found to be of medium or high ecological risk. This study highlights that AR engineering has a significant ecological effect on GW. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). doi: 10.2166/nh.2020.001 ://iwaponline.com/hr/article-pdf/51/5/911/775286/nh0510911.pdf Jin Wu Jingchao Liu Zenghui Pan Boxin Wang Dasheng Zhang (corresponding author) Hebei Institute of Water Science, Shijiazhuang 050051, China E-mail: skyzhangdasheng@126.com Jin Wu Jingchao Liu College of Architecture and Civil Engineering, Beijing University of Technology, Beijing 100124, China


INTRODUCTION
However, such infiltration processes are not expected to impact unconfined aquifers and deteriorate GW quality.
A few works have investigated the occurrence of PPCPs in confined aquifers, examining the selected PPCPs (diethyltoluamide, crotamiton, ethenzamide, propyphenazone, carbamazepine, and caffeine) in confined aquifers (up to 500 m in depth) in Tokyo due to leakage from decrepit sewer networks (Kuroda et al. ). Generally, the GW in unconfined aquifers is used as a drinking water source in North China and is considered a key factor for the steady development of the national economy (Bo et al. ; Li et al. ). Therefore, in developing management strategies to control GW pollution by PPCPs, the contamination and ecological risk of PPCPs in unconfined and confined aquifers both need to be assessed.
Considering the issues raised above, the objectives of this study were to investigate the occurrence of PPCPs in different types of aquifers, including unconfined and confined aquifers, and to characterize the ecological risk for the development of resistance to PPCPs in the GW as well as to optimize the management of GW recharging.

Study area
The studied river-GW interaction (RGI) area is located in North China (Figure 1). The area is characterized by hot, humid summers and generally cold, windy, dry winters.
Its annual temperature is ∼11.5 C, and the average precipitation is ∼600 mm/a. AR engineering is using reclaimed water which is produced by a membrane bioreactor with wastewater treatment plant effluent. Then the reclaimed water, about 38,000,000 m 3 /a, is introduced to the river.  depth of buried bedrock increases from north to south, the thickness of the quaternary strata is between 50 and 300 m, and gradually increases from northeast to southwest.
The lithology changes from coarse particles to fine particles, and the layers change from a single layer to multiple layers.
The quaternary strata mainly include a sand pebble layer, a sand gravel layer, and a silty clay layer.
Two subzones of the study area were identified, including the northern part (N zone) and southern part (S zone), based on the hydrogeological conditions. The N zone of the study area was dominated by gravel and sand with good permeability, while the S zone consisted of silty clay with poor permeability. The lithology of the aquifer changed from sandy gravel to fine sand from north to south. The shallow aquifer in the N zone is an unconfined aquifer (UA-N). ). In short, GW samples were pumped into 2 L glass bottles using a stainless-steel submersible pump. All samples were kept in precleaned containers at a cool temperature and then immediately transported to the laboratory for treatment. In the laboratory, the water samples were commonly concentrated by preconditioned solid-phase extraction. The target antibiotics were subsequently analyzed using high/ ultra/ultrahigh-performance liquid chromatography-tandem mass spectrometry. Appropriate quality assurance and quality control procedures were followed, usually including solvent blank, procedure blank, and independent check standard.
Method detection limits (limits of detection, LODs) and quantification limits (limits of quantification, LOQs) were generally determined as the minimum detectable amount of an analyte with a signal-to-noise ratio. Recoveries obtained by spiking the analytes into GW ranged from 65 to 128%. The LOQs were 0.2-6 ng/L for pharmaceuticals in GW.  Table S1.

Leaching potential assessment
Leaching potential assessment models were adopted to assess the leaching potential of the selected PPCPs in the vadose zone. The model provides a quantitative value for representing the leaching potential. The model is described by Equation (1) by considering both the mobility and the persistence of chemicals.
where GUS is the groundwater ubiquity score, K OC is the organic carbon partition coefficient, and t 1/2 is the degradation half-life in the soil (days). The adjusted criteria were as follows: low leaching potential (GUS 1.8), moderate leaching potential (1.8 < GUS 2.8), and high leaching potential (2.8 GUS).

Ecological risk assessment
The ecological risk assessment has been used to quantify the ecological effect exposed by environmental pollutants in previous studies. In this study, risk quotients (RQs) were applied to assess the ecological risk of PPCPs in GW; a high RQ suggests a high ecological risk and vice versa. The RQs for each PPCP in the GW sample were calculated with the following equation: where MEC is the measured concentration, and PNEC is the predicted no-effect concentration. In this study, the chronic or acute toxicity data of the target antibiotics were collected from previous studies (Supplementary material, Table S2). The PNEC was obtained from the lowest no observed effect concentration (NOEC) with bold marks in Table S2. According to the recommendations of the European technical guidance document (TGD), NOECs were priority toxicity data (EC ). Three levels of ecological risks were classified: between 0.01 and 0.1 is low risk; between 0.1 and 1 is medium risk; and larger than 1 is high risk.

Statistical analysis
Statistical tests were performed to identify major factors that likely affect the occurrence and contribution of PPCPs in

Overview of PPCPs in GW
The frequency of detection and the maximum concentrations of the PPCPs in aquifers are summarized in   (Table S3).

Temporal variations in PPCPs
The type of PPCP detection frequency can be categorized into three classes. The CAF is different from other PPCPs with a higher detection frequency in the dry season (  Figure 2 shows the mean concentrations  As shown in Table 2, most of the targeted PPCPs appeared to have high leaching potential. Among the high leaching potential group, EFX, SMM, and CAF were most likely to leach because of their high t 1/2 and low K OC . TMP and IBU had moderate potential for leaching.
TCS would be the compound most resistant to leaching. It is noted that due to relatively small chemical inputs, the tar-

Spatial variations in PPCPs in different aquifers
Because information on confined aquifers in the northern part was not available, spatial variations in the PPCPs in different aquifers were analyzed in the southern part. As shown in Table 1, PPCPs in the dry season were more frequently detected in unconfined aquifers (50% for average frequency) than in the first confined aquifers (42% for average frequency in confined aquifers) and in the second confined aquifers (46% for average frequency).
With regard to PPCPs in the wet season, the average frequency of detection in the second confined aquifers (48%) was lower than that in the unconfined aquifers (50%). For PPCPs in the first confined aquifers, the average frequency of detection (58%) was higher than that in the unconfined aquifers (50%). Figure     Ecological risk of PPCPs The highest concentration and lowest PNEC were simultaneously used as a worst-case scenario assuming ecological risk. Due to the lack of predictive toxicity data on the chronic effects of SMM N4AcSMX and DIF, their RQs were ignored.
As shown in Table 3, eight PPCPs were detected in aquifers at low-risk levels. The ecosystem risks of SFS, CAF, IBU, and TCS were found to be at least middle risk by using a RQ. Similar to the concentration of PPCPs, the potential ecological risks also exhibited spatial and seasonal variations to some extent.
Overall, the potential ecological risks posed by PPCPs to GW could be a serious issue. The results of this study indicated that PPCPs in GW in the RGI area also need to be monitored for regulation and control by legislation due to their wide distribution and significant adverse ecological effects.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this paper is available online at https://dx.doi.org/10.2166/nh.2020.001.