Journal of Hazardous Materials 448 (2023) 130876 Contents lists available at ScienceDirect Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat Research Paper Chemicals of emerging concern in coastal aquifers: Assessment along the land-ocean interface Daniel Gutiérrez-Martín a, Rubén Gil-Solsona a, *, Maarten W. Saaltink b, c, Valentí Rodellas d, Rebeca López-Serna e, f, Albert Folch b, c, Jesús Carrera a, b, Pablo Gago-Ferrero a, * a Institute of Environmental Assessment and Water Research (IDAEA), CSIC, Jordi Girona 18, 08034 Barcelona, Spain b Associated Unit: Hydrogeology Group (UPC-CSIC), Spain Department of Civil and Environment, Spain c Department of Civil and Environmental Engineering (DECA), Universitat Politècnica de Catalunya (UPC), Jordi Girona 1-3, 08034 Barcelona, Spain d Institut de Ciència i Tecnologia Ambiental, Universitat Autònoma de Barcelona, Bellaterra, Spain e Department of Analytical Chemistry, Faculty of Sciences, University of Valladolid, 47011 Valladolid, Spain f Institute of Sustainable Processes, Dr. Mergelina s/n, Valladolid 47011, Spain H I G H L I G H T S G R A P H I C A L A B S T R A C T • Submarine groundwater discharge was evaluated as a source of sea contamination. • 92 chemicals of emerging concern (CECs) were detected in coastal aquifer. • Unique spatial characterization of the CECs present in the land-ocean interface. • 6 PFAS and 2 pharmaceuticals revealed to be tracers of contamination in groundwater. • A PFAS, non-previously detected in environmental waters, was tentatively identified. A R T I C L E I N F O A B S T R A C T Editor: Lee Blaney Submarine Groundwater Discharge (SGD) is recognized as a relevant source of pollutants to the sea, but little is known about its relevance as a source of chemicals of emerging concern (CECs). Here, both the presence and Keywords: distribution of a wide range of CECs have been evaluated in the most comprehensive manner to date, in a well- Coastal aquifer characterized Mediterranean coastal aquifer near Barcelona (Spain). Samples from coastal groundwater and Submarine groundwater discharge (SGD) seawater allowed for the unique spatial characterization of the pollutants present in the land-ocean interface, an Emerging pollutants outstanding research gap that required attention. The main goals were (1) to determine CECs in the aquifer, so as Non-target analysis High-resolution mass spectrometry (HRMS) to evaluate the SGD as a relevant source of marine pollution, and (2) to identify new tracers to improve our Perfluoroalkyl and polyfluoroalkyl substances understanding of SGD dynamics. To this end, 92 CECs were located in the aquifer by using wide-scope analytical (PFAS) target methodologies (>2000 chemicals). Among them, the perfluoroalkyl and polyfluoroalkyl substances (PFAS), along with the pharmaceuticals carbamazepine and topiramate, were revealed to be good markers for tracing anthropogenic contamination in ground- and seawater, in concrete situations (e.g., highly contaminated sites). Additionally, non-target analysis expanded the number of potential tracers, making it a promising tool for identifying both the source and the fate of pollutants. * Corresponding authors. E-mail addresses: ruben.gil@idaea.csic.es (R. Gil-Solsona), pablo.gago@idaea.csic.es (P. Gago-Ferrero). https://doi.org/10.1016/j.jhazmat.2023.130876 Received 23 September 2022; Received in revised form 11 January 2023; Accepted 25 January 2023 Available online 26 January 2023 0304-3894/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 1. Introduction high reactivity of the seawater-freshwater interface suggests that many CECs will be transformed. In summary, the behavior of CECs in coastal Human beings have always made use of the freshwater provided by and marine environments, their ultimate compartment, has been poorly aquifers. Unfortunately, anthropogenic activities have also negatively studied in comparison to other water bodies such as sewage water or impacted groundwater (GW), which represents about 99% of the surface freshwater [10,24,6]. worldwide liquid freshwater (FW) [35]. Pollution is depleting the We find it self-evident that traditional target strategies need to be quality of this essential resource. In fact, more than 300 chemicals complemented with a broad range of organic compounds to properly (pharmaceuticals, pesticides or disinfection by-products among others) understand the presence, distribution, transformation, and mobility of have been detected in groundwater and/or drinking water to date [5]. pollutants in costal systems. In this context, recent advances in high- The presence of chemicals of emerging concern (CECs) in these systems resolution mass spectrometry (HRMS) facilitate a more holistic charac­ may have an impact on aquatic life and human health, even at low terization of chemicals by expanding the available domain [21]. This concentration levels [63]. Thus, both from an environmental and human non-target analytical strategy is applied after a sample pre-treatment health perspective, it is paramount to characterize the presence and fate focuses on quantitatively extracting as many chemicals as possible. of CECs in aquifers. Unfortunately, physicochemical and biological Based on the exact mass and the information provided by the high res­ transformation of CECs in groundwater is poorly understood. This is olution (HR) data, the substances of interest are statistically prioritized especially true in coastal aquifers, where mixing of the usually oxic or and can be tentatively identified. The retrospective character of this suboxic fresh GW with the usually anoxic saline GW leads to high analytical tool should be highlighted. Hence, with this forward-looking reactivity conditions [49,57], which may contribute to the trans­ approach, analysis is not limited to the data provided by the current formation of CECs in coastal systems. research environment, as it will allow the evaluation of additional Two complementary processes occur along the coast: (1) Seawater substances of interest at a future time and without the need to run intrusion (SWI) and (2) Submarine groundwater discharge (SGD). SWI further data acquisitions [43]. refers to the landwards displacement of the subterranean freshwater/ The purpose of this study is to take advantage of recent advances in saltwater interface, which salinizes aquifer freshwater. GW exploitation HRMS analysis, combining wide-scope target (>2000 chemicals) and lowers water levels and enhances SWI, thus threatening coastal aquifers non-target strategies, to broaden the understanding of coastal aquifer worldwide [18]. SGD refers to the aquifer discharge of fresh GW, systems regarding (1) SGD relevance as a source of organic CECs to the circulated seawater, or their mixture [44]. SGD is important for coastal sea, by identifying chemicals present in coastal groundwaters and their biogeochemical cycles [52]. SGD is also becoming a source of global reactivity in the mixing zone, and (2) the identification of chemicals concern [68]. which could potentially act as SGD tracers. To this end, groundwater (n SWI is complex because of the density dependence of water flow and = 17) and seawater (n = 2) samples were taken from a coastal site, its sensitivity to heterogeneity, which, together with its relevance and located on the Spanish Mediterranean Coast, near Barcelona. The site ubiquity, has motivated numerous hydrologists to strive to understand has been extensively studied in an effort to characterize SWI and SGD at it. SWI has been profusely studied from a theoretical and numerical a typical Mediterranean coastal plain. As a result, it is characterized in point of view [33,38,48,8], as well as by means of laboratory tank terms of geology, hydraulics, biogeochemistry, geophysics, and hydro­ simulations [29,65]. In-situ analyses of SWI are possible by means of chemistry [13,14,20,28,27,40,46]. This site is therefore apt to test and hydraulic, chemical, electrical and thermal measurements. Therefore, evaluate the analytical methodologies we are proposing here. SWI is a multidisciplinary topic which includes other disciplines such as geology, [1,66], geophysics [16,19,25], hydrodynamics [11,30,64] or 2. Material and methods hydrochemistry [45,59]. Because the impact of SGD on the dynamics and composition of the 2.1. Chemicals and materials oceans was previously considered negligible, the study of SGD is a relatively new phenomenon. This misconception, along with the diffi­ Methanol (MeOH) (HPLC-grade), water (HPLC-grade), formic acid culties associated with implementation, delayed SGD assessments until (>99% purity) and hydrochloric acid (HCl) (37%) were purchased from the mid-1990 s [44,60]. It is now widely acknowledged that SGD plays Merck (Darmstadt, Germany). Distilled water was obtained by a Milli-Q an essential role in the dynamics of marine ecosystems, as it represents a purification system (Aurium, PRO-VFT, Sartorius, Göttingen, Germany). relevant source of nutrients, trace metals and contaminants [52,61]. Ammonium acetate was acquired from Merck. Analytical standards, This source may be essential for many coastal habitats but may also including the internal standards (IS), were purchased from Sigma- negatively impact coastal ecosystems [4]. An excessive nutrient load Aldrich (Steinheim, Germany) and LGC Standards (Barcelona, Spain). may lead to eutrophication. SGD may also account for a wide variety of Further information regarding analytical standards and ISs was sum­ pollutants affecting the sea, many of them potentially persistent [52]. marized in the supplementary data (Table A.1, SI-1). A mix of ISs (1 Therefore, it is not surprising that marine scientists have been paying ppm) was prepared by mixing appropriate aliquots of each standard more attention in recent times to SGD, investigating suitable tracers for stock solution in MeOH. Water samples were filtered through What­ identifying and quantifying chemical fluxes driven by SGD [44]. man™ Glass Microfiber Filters (GF/F) 0.7 µm, 47 mm. SPE cartridges Radioactive isotopes of Ra or Rn, silica, or the salinity gradient, have Bond Elut-PPL (500 mg, 6 mL) from Agilent were also used. been used for this purpose [17,23,53], but their applicability is limited. A more comprehensive picture can be gained by coupling traditional 2.2. Site description and sampling hydrological methods (e.g., hydraulic calculations, seepage meters) with alternative and complementary tracers [60]. We share the view of 2.2.1. Site description and characterization Adyasari et al. [2], that organic chemicals, including CECs, represent a The experimental site is located along the Maresme coastal plain, threat but also an opportunity to gain understanding about SGD pro­ specifically in the alluvial aquifer of the Argentona ephemeral stream cesses [2]. (Barcelona, Spain; 41◦ 31′ 11.2′′ N, 2◦ 25′ 25.3′′ E) (Fig. 1a). The climate is Several sources for the presence of CECs in the sea have been iden­ typically Mediterranean, marked by warm, dry summers and wet, mild tified, including atmospheric deposition, surface water inputs or winters. Precipitation ranges from around 600 mm/y at the plain to groundwater discharges [10,32]. But only a few studies have evaluated some 700 mm/y in the coastal range behind the plain. Rainfall is un­ the role of SGD as a source of CECs in coastal seawater [41,42,58]. equally distributed. It concentrates in spring and, especially, autumn, Furthermore, they focused on a limited number of chemicals, barely when monsoon-like rainfall events cause floods, the only occasion when assessing their transformation under coastal aquifer conditions. The the stream actually flows to the sea. The stream runs between 2 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 agricultural fields on the SW side and the city of Mataró (actually its traditional definition in fresh (salinity less than 1 PSU, Practical Salinity suburban industrial zone) to the NE. The wastewater treatment plant Units), brackish (salinity between 1 and 10 PSU) and saline (salinity (WWTP) of the city is located adjacent to the stream, some 300 m up­ greater than 10 PSU). Note however that the salinities we have stream from the site. Its effluent discharges some 2000 offshore through measured within the aquifer are well below Mediterranean Sea salinities an outfall that runs close to the site. Some effluent may leak off the (averaging 38 PSU at our site). Also, the diffuse and unstable nature of outfall into the aquifer. Moreover, the sewer system of Mataró is unitary FW SGD makes it very difficult to sample actual SGD. Instead, what we (it collects both wastewater and surface runoff) and bypasses the WWTP sample is GW next to the sea, close to its discharge point, as well as two when the flow rate exceeds the WWTP capacity during storms, causing superficial seawater samples in the zone influenced by SGD. some untreated wastewater, diluted by surface runoff, to infiltrate the aquifer. 2.2.2. Sampling Characterization of the sited involved drilling 17 piezometers on the Sampling was carried out during three consecutive days in December right bank of the stream. These piezometers allowed for monitoring 2021. Autumn is usually the wettest period of the year, so that heads and collecting groundwater samples along the freshwater- groundwater levels are usually highest during December. Even if 2021 seawater mixing zone (often referred to as subterranean estuary for was a relatively dry year, heavy storms which occurred in October the marine science community). The piezometers used in this study were contributed to the explanation of the distribution of salinities in the located along a section perpendicular to the Mediterranean Sea, at dis­ aquifer. Sampling was performed according to standard protocols. In tances ranging from 18 to 102 m to the shoreline and depths from 7 to short, firstly a submersible pump was placed at the screened intervals of 22 m below the ground surface. Except for two stand-alone piezometers the piezometers to extract groundwater, and a multi-parameter probe (PP10 and PP18), piezometers were grouped in nests (N1, N2, N3 and with a flow cell was connected to the pump to measure in-situ temper­ NMar) consisting of 3–4 piezometers of different depths (Fig. 1b) to ature, dissolved oxygen (DO), pH, redox potential (ORP) and salinity. facilitate monitoring and testing along the vertical course, so as to best Second, three times the volume of the screened interval of the piezom­ characterize vertical mixing processes. Most piezometers were equipped eter for each sampling point was purged. Third, pumping continued at a with pressure sensors, as well as distributed temperature sensors and constant rate, until some 4.5 L had been collected, filtered through electrodes attached to the outside of the PVC tubing (see [20] for de­ 20 µm in-situ and stored at 4 ◦ C in amber glass bottles (pre-cleaned with tails). The tubing (casing) was slotted over a short 2 m interval, adjacent acetone and Milli-Q water, and rinsed with the sampling point to sandy sediments, to minimize mixing within the piezometer and groundwater twice). Additionally, two surface seawater samples were during sampling, leaving at least 2 m of blind pipe at the bottom of the taken offshore at 60 and 130 m from the shoreline in the direction casing. aligned with the piezometers. Field blanks (n = 3) analyses were per­ The main findings relating to the hydrogeological context are sum­ formed with 4 L of Milli-Q water every sampling day. Additional data marized in Fig. 1. The aquifer consists of sandy fluvial layers separated regarding the sampling sites, including shoreline distance, depth, and in- by thin silty layers. Fluvial deposits overlie weathered granite of low situ measured parameters, is provided in Table A.2 (SI-2). permeability but are often underlaid by a conductive layer at the point of contact between the weathered and unweathered granite [39], below 2.3. Sample preparation and instrumental analysis the monitored section. Freshwater (FW) discharges sub-horizontally into the sea, along the sand layers and vertically across the silt layers. The water samples (4L) were filtered through glass fiber filters (pore Seawater (SW) flows within the aquifer under the influence of three size 0.7 µm, Whatman GF/F, pre-heated (450 ◦ C, 4 h)) and the pH was mechanisms: (1) according to the traditional SWI paradigm, the dense adjusted to 2 with HCl. Then, two isotopically labeled chemicals (ben­ SW flows onshore at depth due to its relatively high density, and returns zotriazole-d4 and methyl-paraben-d3) were added as surrogate to seawards with a sharp turn as it mixes with FW; (2) SW enters near the ensure good SPE reproducibility. PPL columns were activated by leaving shore due to storm surges, and; (3) SW fingers down into FW zones, and them to soak in 3 mL of methanol overnight. Methanol was then drained vice versa, due to tidal fluctuations and the instability caused by the and 3 mL of Milli-Q water (pH=2, HCl) was passed through for column dense SW overlying the light FW. conditioning. Filtered samples were loaded onto the PPL columns and The location and extent of the FW flow paths below the silt layers, eluted (< 3 mL⋅min− 1). Salts were subsequently removed from the PPL which hold SW above, has been ascertained by Electrical Resistivity columns by passing 6 mL of Milli-Q water (pH=2, HCl). Next, the car­ Tomography [46], induction logging [40], and tidal response analyses tridges were dried under vacuum for 20 min and further eluted with [27]. This is shown in Fig. 1, where water is divided according to the 3 mL of methanol. The extract volume was finally adjusted to 2 mL with Fig. 1. Site description. A) Google Earth image of the site, including location of sampling piezometers (red circles near the shore) and land use. B) Schematic vertical cross-section, including slotted intervals, lithology, grain size, flow patterns, and water salinity (fresh, brackish, and saline) zones. Adapted from [40]. 3 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 HPLC-grade MeOH and stored at − 20 ◦ C pending analysis. Before high-energy MS2 spectra. In all cases the tentative identifications were analysis, samples were diluted (1:1, v/v) with HPLC-grade water and the manually checked, and, when available, compared with the spectra IS mix was added (Table A.1, SI-1) to correct for matrix effects. observed in databases including MassBank, MassBank of North America, Instrumental analysis was performed using an Acquity UHPLC sys­ Metlin, HMDB and mzCloud. MetFrag was also used for in silico frag­ tem (Waters, Milford, USA) coupled with a Q-Exactive Orbitrap mass ment annotation in some identifications [50,51]. The identification analyser (Thermo Fisher Scientific, Dreieich, Germany). The chro­ confidence was reported using the hierarchical levels previously matographic separation was carried out by a Waters Cortecs C18 described by Schymanski et al., [54]. In all cases, procedural blanks (2.1 ×100 mm, 2.7 µm) column headed by a Waters Cortecs C18 were carefully assessed to avoid false positives. (2.1 ×5 mm, 2.7 µm) precolumn. An electrospray ionization (ESI) interface operated at 3000 and 2800 V in positive and negative mode, 3. Results and discussion respectively. Instrumental parameters for ESI were: 350 ◦ C capillary temperature, 40 sheath gas flow, 10 auxiliary gas flow, 100 of maximum 3.1. Occurrence and distribution of the target chemicals in the aquifer spray current, 350 ◦ C probe heater temperature and 60 SLens RF level. In positive ionization mode, HPLC-grade water (aqueous phase) and A total of 92 chemicals out of the 748 initially targeted were detected methanol (organic phase), both with 0.1% formic acid, were employed in the aquifer (Table 1) and their abundance is represented as a clustered as a mobile phase, and underwent the following elution gradient: 5–75% heatmap in Fig. A.2, SI-4. A wide variety of industrial chemicals were B (0–7 min), 75%− 100% B (7–10 min), 100% B (10–17 min), 100–5% determined in the groundwater under study, including plastic additives B (17–18 min). In negative ionization mode, HPLC-grade water (e.g., bisphenol AF, phthalates), perfluoroalkyl substances (e.g., PFOA, (aqueous phase) and methanol (organic phase), both with 5 mM PFOS), surfactants (e.g., oleamide, lauryl sulfate), flame retardants (e.g., ammonium acetate, followed the elution gradient described here: Tris(1-chloro-2-propyl) phosphate), and benzothiazoles and benzo­ 5–50% B (0–3 min), 50%− 90% B (3–6 min), 90–100% B (6–13 min), triazole derivatives, among others. Also, several pharmaceuticals (e.g., 100% B (13–17 min) and 100–5% B (17–18 min). In both modes, a carbamazepine, dextrorphan), personal care products (e.g., cetrimo­ constant flow of 0.3 mL⋅min− 1, an injection volume of 5 µL and, a nium, ensulizole), food-related compounds (e.g., caffeine), biocides (e. constant column temperature of 40 ◦ C was applied. The instrument g., atrazine, clothianidin) and drugs of abuse (e.g., cocaine and its executed two acquisition modes: data independent (DIA) and data metabolite benzoylecgonine) were detected. These compounds suggest a dependent acquisition (DDA). DIA consisted of a full scan with low broad range of sources, ranging from infiltrated surface runoff or waste collision energy (4 eV) and a full scan with a high collision energy water, return flow of irrigated lands, or legacy industrial pollution [62]. (25 eV), in a mass-to-change ratio (m/z) range from 67 to 1000 and Among these anthropogenic chemicals, a general decrease in abundance resolving power of 60,000. In contrast, DDA mode acquired full scans at of those compounds, was observed with the increase of salinity, due to low collision energy (4 eV), and MS2 spectra for the 5 most intense ions the higher proportion of seawater. in each scan were, then, registered (35 eV applied as collision energy In seawater samples, a total of 44 chemicals were detected. Although and 30,000 as resolving power), excluding those ions previously in most cases they showed the lowest abundances among all the eval­ selected in the last 30 s uated sampling sites (Fig. A.2, SI-4), it indicated that a high number of chemicals were present in the coastal waters, which was in line with the 2.4. Target and non-target analysis and data processing data reported by the few other studies carried out in the field [3,7]. However, the observed minor variations may be due to the difference in Data Independent Acquisition (DIA) files were converted to “.mzml” the biogeochemical processes taking place in the mixing zone. files with MSConvert (Table A.3, SI-3). Target and non-target strategies Regarding SGD processeses, target substances with the potential to were applied to assess the presence of organic pollutants and possible provide useful information were priorized (Fig. 2). Out of the detected organic tracers in the aquifer, as described below. The target and non- target compounds, a total of 32 chemicals were present in more than target performance workflow is illustrated in Fig. A.1, SI-3. 75% of the analyzed aquifer samples. Only these 32 chemicals were prioritized as potential tracers, as their presence in a high number of 2.4.1. Target analysis samples indicates persistence under aquifer conditions. Thus, their peak Target analysis evaluation was conducted for 748 chemicals area in the chromatograms, directly related with their abundance, were (Table A.1, SI-1) in a similar manner as described elsewhere [22]. Raw compared with the salinity values (Table A.2, SI-2) to find those that files were processed with MZMine2 software [47]. Parameters and linearly correlate. In this manner, two linear regressions (salinity vs workflow for MZMine2 are summarized in Table A.4, SI-3. corrected peak area and Ln(salinity) vs corrected peak area) were per­ Only the chemicals meeting the following criteria were reported as formed. The term “corrected peak area” refers to the peak area of each positive findings: i) Peak area in the samples > summatory of the field compound corrected by the peak area of the IS. Eight anthropogenic blank peak area for each sampling day plus three times the standard chemicals followed a reverse trend with salinity (R2 >0.7) (Table 2). deviation; ii) Peak area for the parent ion (and its main fragment ion, These chemicals were of special interest as their abundance decreased when available) > 5E5 in the most intense sample (to avoid the occur­ mainly as a consequence of dilution, indicating a minimum degradation. rence of false positives due to interferences or noise). These compounds included 6 perfluoroalkyl substances (PFAS, also known as ‘forever chemicals’), that followed similar and consistent 2.4.2. Non-target-based prioritization strategy and identification of trends among them, and two pharmaceuticals: topiramate and carba­ unknowns mazepine (CBZ). All of them are highly recalcitrant chemicals and In the non-target workflow, generic (.mzml) files were imported to therefore little affected by physicochemical degradation processes MZMine2 and a mass list for each scan was generated. Parameters and within the aquifer. Consequently, its distribution in the different zones workflow for MZMine2 are summarized in Table A.5, SI-3. The final of the aquifer might help in monitoring the aquifer system and fresh­ table was exported to a “.csv” file and blank subtraction was performed water SGD, if they indeed behave as conservative tracers for monitoring in the same manner described in the target approach. Finally, a corre­ SGD. lation between the salinity levels and each feature at all the sampling The spatial distribution of the whole amount of PFAS throughout the points was calculated. The features with a R2 higher than 0.7 were aquifer, as well as the specific distribution of PFOS, are represented in prioritized. The feature list was exported to SIRIUS4.20 [15]. Both Fig. 3a and Fig. 3b, respectively. PFAS commonly enter the environment molecular formula and tentative compound annotations were obtained, through production or contaminated waste streams. PFAS are highly based on the information on accurate mass, isotopic pattern, and the mobile and can be easily transported through the aquifer. Their presence 4 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 Table 1 Chemicals detected in the aquifer. Compound Molecular formula Type of product/industrial category RT (min) DR%b IMc LogPd ESI (+) / ESI (-) 1,3-Dicyclohexylurea C13H24N2O Tire rubber – derived contaminants 7.49 / n.a. 94 + 3 1,3-Diphenylguanidine C13H13N3 Accelerator in rubber industry 2.97 / n.a. 100 + 2.4 1 h,1 h,2 h,2 h-perfluorooctanesulfonic acid C8H5F13O3S PFAS n.a. / 5.41 12 - 3.9 1 H-Benzotriazole C6H5N3 Anticorrosion agent 3.76 / 3.22 100 +/ - 1 2,4,6-trimethylbenzenesulfonic acid C9H12O3S Industrial chemical n.a.a / 3.34 82 - 1.8 2-Benzothiazolesulfonic acid C7H5NO3S2 Benzothiazole family n.a. / 2.57 94 - 1.5 2-Hydroxybenzothiazole C7H5NOS Benzothiazole family 5.46 / 4.34 100 +/ - 1.8 3,5-di-tert-butyl-4-hydroxybenzaldehyde C15H22O2 Benzaldehyde family 8.64 / n.a. 41 + 4.4 4-Acetamidoantipyrine C13H15N3O2 Pharmaceutical (metabolite) 3.34 / n.a. 88 + 0 4-Dodecylbenzenesulfonic acid C18H30O3S Surfactant n.a. / 6.73 47 - 7 4-formylaminoantipyrine C12H13N3O2 Pharmaceutical (metabolite) 3.27 / n.a. 41 + 0 4-Hydroxybenzoic acid C7H6O3 Preservative n.a. / 1.85 18 - 1.6 4-Methyl-1 H-benzotriazole C7H7N3 Benzotriazol family 4.98 / 4.02 100 +/ - 1.4 4-Nitrophenol C6H5NO3 Manufacture drugs / biocides n.a. / 3.64 29 - 1.9 4-Octylphenol C14H22O Surfactant n.a. / 6.94 41 - 5.3 5-Chlorobenzotriazole C6H4ClN3 Benzotriazol family n.a. / 4.25 100 - 1.7 6-(methoxyphenyl)pyrimidine-2,4-diamine C11H12N4O Member of pyrimidines 2.23 / n.a. 12 + 1.1 Acridine C13H9N Dyes/coating 4.06 / n.a. 29 + 3.4 Acridone C13H9NO Acridine derivarive 6.07 / n.a. 12 + 3 Antipyrine C11H12N2O Pharmaceutical 4.30 / n.a. 100 + 0.4 Atrazine C8H14ClN5 Biocide 6.81 / n.a. 35 + 2.6 Atrazine-desisopropyl C5H8ClN5 Biocide 3.80 / n.a. 82 + 1.1 Benzoic acid C7H6O2 Preservative n.a. / 1.56 53 - 1.9 Benzothiazole C7H5NS Rubber production 6.04 / n.a. 29 + 2 Benzoylecgonine C16H19NO4 Drugs of abuse (metabolite) 3.71 / n.a. 71 + -0.3 Bisphenol AF C15H10F6O2 Plastic additive n.a. / 5.80 24 - 4.5 Bisphenol S C12H10O4S Plastic additive 4.73 / 3.65 18 +/ - 1.9 Caffeine C8H10N4O2 Food related chemical 3.59 / n.a. 12 + -0.1 Caprolactam C6H11NO Synthesis of fibers 3.23 / n.a. 12 + -0.1 Carbamazepine C15H12N2O Pharmaceutical 6.33 / n.a. 88 + 2.5 Cetrimonium C19H42N+ Surfactant, PCPs 8.54 / n.a. 12 + 8 Clothianidin C6N5H8SO2Cl Biocide n.a. / 3.37 53 - 1.3 Cocaine C17H21NO4 Drugs of abuse 3.47 / n.a. 71 + 2.3 Cotinine N-oxide C10H12N2O2 Tobacco related chemical 0.96 / n.a. 35 + -1.3 Decaethyleneglycol C20H42O11 Petroleum derivate 4.32 / n.a. 12 + n.a.e Desethylterbuthylazine C7H12ClN5 Biocide (metabolite) 6.27 / n.a. 82 + 2 Desvenlafaxine C16H25NO2 Pharmaceutical 3.38 / n.a. 94 + 2.6 Dextrorphan C17H23NO Pharmaceutical 3.37 / n.a. 29 + 3.1 Dibutyl phosphate C8H19PO4 Antifoaming agent, catalyst 8.85 / n.a. 82 + 1.5 Diethyltoluamide C12H17NO Biocide 6.95 / n.a. 53 + 2 Diisobutyl phthalate C16H22O4 Plastic additive 8.97 / n.a. 24 + 4.1 Dimethyl phthalate C10H10O4 Plastic additive 5.75 / n.a. 35 + 1.6 Dimethyl sebacate C12H22O4 Plastic additive 6.33 / n.a. 24 + 3 Di-n-Butyl phtalate C16H22O4 Plastic additive 9.07 / n.a. 24 + 4.7 Diuron C9H10Cl2N2O Biocide 6.92 / 5.30 82 +/ - 2.7 Dodecaethylene glycol C24H50O13 Petroleum derivate 4.72 / n.a. 12 + n.a. Ensulizole C13H10N2O3S PCPs (sunscreen agent) 3.30 / n.a. 59 + 2 Icaridin C12H23NO3 Insect repellant 7.17 / n.a. 41 + 2 Imidacloprid C9H10ClN5O2 Biocide 4.09 / n.a. 53 + 1.2 Lamotrigine C9H7Cl2N5 Pharmaceutical 3.29 / n.a. 71 + 1.4 Laurolactam C12H23NO Synthesis of fibers 7.40 / n.a. 24 + 2.8 Lauryl diethanolamide C16H33NO3 Personal care product 8.77 / n.a. 35 + 3.5 Lauryl sulfate C12H26O4S Surfactant n.a. / 6.28 12 - 4.7 Lidocaine C14H22N2O Pharmaceutical 2.40 / n.a. 41 + 2.3 Metalaxyl C15H21NO4 Biocide 6.87 / n.a. 59 + 1.6 Metolachlor ethanesulfonic acid C15H23NO5S Biocide 7.70 / n.a. 53 + 1.4 N-Butyldiethanolamine C8H19NO2 Neutralizing additive 0.81 / n.a. 29 + 0.1 Nicotine C10H14N2 Tobacco related chemicals 0.78 / n.a. 71 + 1.2 Octapropylene glycol (PPG n8) C24H50O9 Moisturizer 7.79 / n.a. 18 + 1 Octyl hydrogen phthalate C16H22O4 Plastic additive 8.91 / n.a. 24 + 5.3 Oleamide C18H35NO Surfactant 10.42 / n.a. 59 + 6.6 Oxindole C8H7NO Natural product 4.22 / n.a. 29 + 1.2 Pentaethylene Glycol C10H22O6 Petroleum derivate 2.67 / n.a. 18 + n.a. Pentapropylene glycol (PPG n5) C15H32O6 Moisturizer 6.05 / n.a. 59 + 0.2 Perfluoro-1-butanesulfonamide C4H2F9NO2S PFAS n.a. / 4.92 71 - 2.2 Perfluorobutanesulfonic acid (PFBS) C4HF9O3S PFAS n.a. / 4.16 88 - 2.3 Perfluorodecanoic acid (PFDA) C10HF19O2 PFAS n.a. / 6.15 53 - 6.3 Perfluoroheptanesulfonic acid (PFHpS) C7HF15O3S PFAS n.a. / 5.56 71 - 4.3 Perfluoroheptanoic acid (PFHpA) C7HF13O2 PFAS n.a. / 5.17 88 - 4.3 Perfluorohexanesulfonate (PFHxS) C6F13O3S- PFAS n.a. / 5.19 100 - 3.6 Perfluorohexanoic acid (PFHxA) C6HF11O2 PFAS n.a. / 4.68 100 - 3.6 Perfluorononanoic acid (PFNA) C9HF17O2 PFAS n.a. / 5.88 65 - 5.6 Perfluorooctanesulfonic acid (PFOS) C8HF17O3S PFAS n.a. / 5.87 76 - 5 (continued on next page) 5 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 Table 1 (continued ) Compound Molecular formula Type of product/industrial category RT (min) DR%b IMc LogPd ESI (+) / ESI (-) Perfluorooctanoic acid (PFOA) C8HF15O2 PFAS n.a. / 5.56 88 - 4.9 Perfluoropentanesulfonic acid (PFPeS) C5HF11O3S PFAS n.a. / 4.74 71 - 3 Perfluorovaleric acid (PFPeA) C5HF9O2 PFAS n.a. / 4.00 88 - 2.9 Phendimetrazine C12H17NO Pharmaceutical 7.24 / n.a. 53 + 1.9 Pilocarpine C11H16N2O2 Pharmaceutical 0.94 / n.a. 82 + 1.1 Ritalinic acid C13H17NO2 Pharmaceutical (metabolite) 3.37 / n.a. 88 + -2.4 Simazine C7H12ClN5 Biocide 6.16 / n.a. 82 + 2.2 Stearic acid C18H36O2 Surfactant n.a. / 8.28 35 - 7.4 Tapentadol C14H23NO Pharmaceutical 8.32 / n.a. 76 + 3.5 Tetradecylsulfate C14H29O4S- Surfactant n.a. / 6.81 6 - 5.8 Tetrapropylene glycol (PPG n4) C12H26O5 Moisturizer 5.10 / n.a. 47 + -0.1 Topiramate C12H21NO8S Pharmaceutical n.a. / 4.15 76 - -0.8 Tributyl phosphate C12H27O4P Flame retardant 8.98 / n.a. 71 + 2.9 Triethyl citrate C12H20O7 Coatings, additive, PCPs 6.02 / n.a. 18 + 0.1 Triethyl phosphate C6H15O4P Industrial use 5.38 / n.a. 82 + 0.8 Triphenylphosphine oxide C18H15OP Rubber additive 7.37 / n.a. 100 + 2.8 Tris(1-chloro-2-propyl) phosphate C9H18Cl3O4P Flame retardant 7.54 / n.a. 35 + 3.3 Valproic acid C8H16O2 Pharmaceutical n.a. / 4.29 100 - 2.8 Xylenesulfonate C8H9O3S- Surfactant n.a. / 2.79 76 - 1.3 a n.a.: non-ionizable in this mode. bDetection rate (DR%) accounting for the chemicals detected in the groundwater samples (detected chemicals in the groundwater samples divided by the total groundwater samples). cIonization mode (IM) in ESI (positive=+ and/or negative=-). dLogarithm of the octanol-water partition coefficient (LogP) computed by XLogP3 3.0 (PubChem release 2021.05.07). eN/A = non-available. Table 2 Chemicals which followed a trend (Coefficient of Determination (R2) > 0.7) with salinity values. R2 for a linear regression between salinity and corrected peak area (C) and a linear regression between lnsalinity and C, is included. Family Chemicals R2 R2 (Salinity VS C) (lnSalinity VS C) PFAS PFHpA 0.875 0.857 PFBS 0.872 0.794 PFHxA 0.821 0.777 PFOS 0.876 0.882 PFOA 0.831 0.825 PFPeA 0.810 0.818 Pharmaceuticals Topiramate 0.717 0.721 Carbamazepine 0.741 0.590 could also be potentially useful for tracing SGD. However, these chem­ icals are consistently found at low concentrations in seawater [56,67], even in remote environments such as the Antarctic region or the Arctic Ocean [9,36]. Therefore, the low levels determined in the coastal sea, close to the detection limits, prevented the presence of PFAS from being attributed to groundwater input. This also led to the suggestion that PFAS might not be a good SGD tracer in the hydrogeological and coastal context. Its potential application should be limited to highly PFAS-contaminated groundwater systems, where the concentration of these chemicals could be easily traced in the seawater above the basal levels, which are usually found at pg⋅L− 1 level [26,36]. Another context in which they might feasibly function as trackers is in marine zones with a high proportion of freshwater, such as karstic or volcanic systems or Fig. 2. Screening and prioritization target workflow. semi-closed coastal areas (e.g., coastal lagoons or coves). The pharmaceuticals carbamazepine (Fig. 3c) and topiramate has been previously reported in groundwater from different regions in (Fig. A.3, SI-4), both employed as anticonvulsants, followed the same concentrations ranging from low pg⋅L− 1 to low mg⋅L− 1 [56]. In this reversal trend in concentration with salinity as PFAS. These chemicals study, PFOA and PFOS were the most abundant within the PFAS family were also present in the fresh-water samples. Their abundance was of chemicals, which is consistent with other studies in surface and much lower in the deep aquifer samples rich in SW and absent (below groundwater [34,56]. Due to their persistent behavior and high ioni­ the limit of detection) in the seawater sites. These results were in line zation efficiency in negative electrospray ionization mode (which leads with previous studies that proposed carbamazepine as an anthropogenic to low limits of detection), these substances may be useful tracers for the marker during groundwater infiltration from wastewater treatment mobility of organic chemicals within the aquifer. PFAS were present at processes because of its persistence, as it was assumed that it is not higher levels in the shallowest sites, especially in N2–10 and N1–10, subjected to degradation or adsorption under these conditions [12,31]. which were highly influenced by surface water infiltration, but also at Regarding topiramate, the available information in the literature is deep samples, which suggests origination and transport from further much scarcer, but it is expected to persist in aquatic environments. afield. According to these results and their intrinsic properties, PFAS Therefore, it could be proposed as another potential infiltration tracer in 6 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 Fig. 3. Distribution of PFAS, PFOS and carbamazepine along the aquifer. Fresh, brackish, and saline water were represented among with the silt layers which (semi) confined the system [40]. a) Distribution of the overall amount (normalized) of PFAS, including PFHpA, PFBS, PFHxA, PFOS, PFOA, PFPeA and its linear regression versus salinity. b) Distribution of PFOS and its linear regression versus salinity. c) Distribution of carbamazepine and its linear regression versus salinity. aquatic environments, including SGD in systems highly affected by fraction of seawater included sampling sites N2–20, N3–20, PP10, anthropogenic contamination, and thus, complement the use of carba­ NMar-10 and NMar-15. Finally, a saline zone (salinity greater than 20.0 mazepine. Other pharmaceuticals were also detected in the aquifer at PSU) with a high fraction of seawater, included the sites N1–20, NMar- different levels. Nonetheless, none of them, including those detected 20, N1–25, N2–25, N3–25 and NMar-25. A representative example of a with a high DR% and intensity (e.g., valproic acid, desvenlafaxine), distribution of chemicals not related to salinity can be found in benzo­ followed the same inverse trend relative to salinity as carbamazepine triazoles, which are widely used in the chemical industry as corrosion and topiramate, probably due to their higher biodegradability during inhibitors, and were detected in all the aquifer samples (Fig. 4). Methyl- the different transformation processes taking place under the aquifer benzotriazole and 5-chlorobenzotriazole were present at high levels in conditions. the freshwater end of the mixing zone (with a similar abundance in The abundance of the rest of the highly detected chemicals (DR% both) but with much smaller intensities in the shallow freshwater zone. >75%) did not follow a reverse trend relative to salinity. Nevertheless, This behavior could be explained by the different degradation rates as mentioned above, their distribution throughout the different zones of expected for the different aquifer conditions in those areas, especially the aquifer may be used to understand some of the processes taking the redox state which could vary from oxidizing/aerobic to reducing/ place. In this regard, different zones within the aquifer were identified in anaerobic. In addition, different transformation pathways have already terms of their salinity levels (Table A.2, SI-2). The freshwater zone been proposed for benzotriazoles and its methylated or chlorinated de­ (salinity less than 1.5 PSU) was characterized by low seawater influence rivatives, which could also contribute to this scenario [37,55]. Results and included the sampling sites N2–10, N2–15, N3–10, N1–10, N1–15 suggested that these chemicals were degraded under oxidizing condi­ and PP18. The brackish zone (salinity ranged from 1.5 to 5.0 PSU) tions, which are predominant in the shallow freshwater zone. In the represents a zone where water contains a significant, but minoritarian, mixing zone, where silty layers semi-confined the aquifer and the 7 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 Fig. 4. Distribution of chemicals along the aquifer, including: a) methyl-benzotriazole and b) 5-chlorobenzotriazole. environment is not that oxidizing (or is even reducing), the degradation 3 features (expressed as m/z, along with their detection rate (DR%), was lower. In addition, the large seawater intrusion produced in the retention time (RT), ionization mode (IM) and R2) were prioritized for salty zone diluted the freshwater and, therefore, decreased the abun­ further identification (Table 3). dance of these two chemicals. However, previous studies suggested the The non-target workflow followed in the identifications (Fig. 5) existence of additional sources for these chemicals in the aquifer, such as provided 3 features. One chemical was tentatively identified as probable runoff infiltration or input from wastewater treatment plants (WWTPs) structure (Level 3) while another identification remained as exact m/z of [28], which would complicate the system dynamics. interest (Level 5). One of the prioritized features was identified as in- source fragment associated to one of the aforementioned substances. 3.2. Non-target prioritization and identification of unknowns The first feature of interest was m/z 168.9883. This m/z value corre­ sponded to one of the main common fragments resulting from PFAS and it co- A total of 5543 features were detected in the aquifer after blank eluted with several target chemicals of this family (e.g., PFOS, PFOA or subtraction, and 1721 of them were present in more than 75% of the PFNA). Additionally, this fragment coeluted with the prioritized feature m/z: groundwater samples. In a similar manner as carried out in the target 524.9597, which also presented a mass defect and an isotopic pattern char­ analysis, their peak area in the chromatograms was compared with the acteristic of the PFAS family (Table 3). The feature presented a low intensity salinity values, to identify those with a linear correlation (R2 >0.70). and coeluted with PFOS and PFNA, thus, a DDA injection with 50 µL injection After eliminating matches already found in the target analysis, a total of volume was performed (Fig. 6). The fragments 418.9744, 218.9865, Table 3 Non-target features detected in the aquifer samples whose peak area in the chromatograms followed a reverse trend with salinity values. m/z RTa DR%b R2c IMd Molecular Tentative structure Levele (mDa) (min) Formula C vs Salinity C vs lnSalinity 524.9597 5.87 82 0.74 0.89 - C11HF19O2 3 168.9883 5.7 82 0.58 0.73 - [C3F7]- - 439.2931 7.12 100 0.70 0.54 - n.a.f n.a. 5 a Retention time (RT). b Detection rate (DR) expressed as a percentage, accounting for the groundwater samples. c Coefficient of determination (R2) for the linear regression between salinity and corrected peak area (C) as well as a linear regression between lnSalinity and C. d Ionization mode (IM), + or -, in ESI. e Level of identification according with Schymanski et al. Schymanski et al., (2014) [54]. f n.a.: information non-available. 8 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 remained as m/z of interest (level 5). 4. Conclusions This study presents the largest monitoring of CECs in the groundwater-seawater interface ever performed to-date. It took place in a well characterized Mediterranean coastal aquifer near Barcelona and included wide-scope target analysis for over 2000 chemicals. It resulted in 92 anthropogenic chemical positive identifications spread out over the different zones of the aquifer, including industrial chemicals, phar­ maceuticals, biocides, or personal care products. This high number of anthropogenic chemicals indicated the significant influence human ac­ tivities have on the chemical composition of environmental water at all levels, including barely accessible coastal aquifers. Additionally, these chemicals may reach the marine ecosystem through SGD along with other sea inputs. Hence, 44 chemicals out of the 92 detected in the aquifer, were also found in the sea samples. Nonetheless, their abun­ dance was estimated to be far lower than in the aquifer, in most cases. Additionally, eight of the identified chemicals, all of them known for their persistent behavior, were proposed as SGD tracers based on its inverse correlation with salinity. PFAS stood out among the selected chemicals, as six members of this family followed this trend, as well as two pharmaceuticals: carbamazepine and topiramate. Non-target analytical strategies were also applied in this study and resulted in the prioritization of three additional features, also candidates for tracing SGD. Among them, one was tentatively identified as 2- (difluoromethylene)− 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadeca­ fluoro-decanoic acid. This chemical consisted of a PFAS member, rein­ forcing the previous results. To the best of our knowledge, this is the first Fig. 5. Prioritization non-target workflow. evidence of presence for this perfluorinated chemical in the environ­ ment. While recognizing the successful use of those prioritized chem­ 168.9897 and 118.9920 were finally assigned. Only one molecular formula icals as anthropogenic markers during groundwater infiltration, their (C11HF19O2) presented a plausible structure according to MetFrag and use as SGD tracers might entail more significant challenges. Hence, the Sirius. Using PubChem database and a mass error threshold of 5 ppm, three use of PFAS, carbamazepine or topiramate as tracers should be limited to structural isomers were proposed as possible: 2-(difluoromethylene)− highly contaminated groundwater systems and/or in the context of sea 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-decanoic acid, (Z)− environments with a large proportion of freshwater, such as karstic 3,4,4,5,5,6,6,7,7,8,8,9,9,9-tetradecafluoro-2-(1,1,2,2,2-penta-fluoroethyl) systems or with a higher resident time in the sea (e.g., semi-closed non-2-enoic acid and 2-(difluoromethylene)− 3,3,4,4,5,5,6,6,7,7,8,9,9,9- coastal areas). The determination of naturally occurring substances tetradecafluoro-8-(trifluoromethyl)nonanoic acid. The first structure can from the aquifer with very low or even null presence in the sea was explain all the proposed fragments, while the second and the third one could among our objectives. The conducted prioritization of the non-target not explain the fragments m/z 418.9744 and 118.9920, respectively. features did not provide natural solid candidates however. Therefore, this chemical was considered tentatively identified (confidence All the provided information enriches the current knowledge on level 3). Submarine Groundwater Discharge and the reaction processes occurring The identification of this chemical is consistent with the results ob­ at groundwater-seawater interface. While much remains to be done, we tained in the target analysis and reinforce the potential use of PFAS as conclude that the ability of non-target analyses to identify unsuspected tracers for fresh SGD in groundwater-polluted environments. Also, to the chemicals might contribute to distinguishing the sources of water in best of the author’s knowledge, this is the first time this chemical has coastal aquifers, as well as their organic pollution and degradation been identified in the environment. processes. Lastly, for m/z 439.2931, no clear MS2 spectra could be obtained due to its low intensity. An unequivocal molecular formula could not be reached due to the absence of fragmentation. Thus, this feature Fig. 6. DDA injection (50 µL injection volume) for the identification of m/z: 524.9597. 9 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 CRediT authorship contribution statement [6] Arpin-Pont, L., Martínez-Bueno, M.J., Gomez, E., Fenet, H., 2016. Occurrence of PPCPs in the marine environment: a review. Environ Sci Pollut Res 23 (6), 4978–4991. https://doi.org/10.1007/s11356-014-3617-x. D. Gutiérrez-Martín: Formal analysis, Investigation, Data curation, [7] Avellan, A., Duarte, A., Rocha-Santos, T., 2022. Organic contaminants in marine Writing – original draft. R. Gil-Solsona: Conceptualization, Methodol­ sediments and seawater: A review for drawing environmental diagnostics and ogy, Writing – review & editing, Supervision. M.W. Saaltink: Writing - searching for informative predictors. In: Science of the Total Environment, 808. Elsevier B.V,, 152012. https://doi.org/10.1016/j.scitotenv.2021.152012. review & editing, Funding acquisition. V. Rodellas: Writing– review & [8] Bakker, M., Miller, A.D., Morgan, L.K., Werner, A.D., 2017. Evaluation of analytic editing. R. López-Serna: Writing – review & editing. A. Folch: solutions for steady interface flow where the aquifer extends below the sea. Conceptualization, Writing – review & editing, Funding acquisition. J. J Hydrol 660–664. https://doi.org/10.1016/j.jhydrol.2017.04.009. [9] Bengtson Nash, S., Rintoul, S.R., Kawaguchi, S., Staniland, I., Van Den Hoff, J., Carrera: Conceptualization, Writing – review & editing, Funding Tierney, M., Bossi, R., 2010. Perfluorinated compounds in the Antarctic region: acquisition. P. Gago-Ferrero: Conceptualization, Methodology, Writing Ocean circulation provides prolonged protection from distant sources. In: – review & editing, Project administration, Supervision, Funding Environmental Pollution, 158. Elsevier Ltd, pp. 2985–2991. https://doi.org/ 10.1016/j.envpol.2010.05.024. acquisition. [10] Brumovský, M., Bečanová, J., Kohoutek, J., Borghini, M., Nizzetto, L., 2017. Contaminants of emerging concern in the open sea waters of the Western Declaration of Competing Interest Mediterranean. Environ Pollut 229, 976–983. https://doi.org/10.1016/j. envpol.2017.07.082. [11] Chidichimo, F., De Biase, M., Rizzo, E., Masi, S., Straface, S., 2015. Hydrodynamic The authors declare that they have no known competing financial parameters estimation from self-potential data in a controlled full scale site. In: interests or personal relationships that could have appeared to influence J Hydrol (Amst), 522. Elsevier B.V, pp. 572–581. https://doi.org/10.1016/j. jhydrol.2015.01.022. the work reported in this paper. [12] Clara, M., Strenn, B., Kreuzinger, N., 2004. Carbamazepine as a possible anthropogenic marker in the aquatic environment: Investigations on the behaviour Data Availability of Carbamazepine in wastewater treatment and during groundwater infiltration. Water Res 38 (4), 947–954. https://doi.org/10.1016/j.watres.2003.10.058. [13] Diego-Feliu, M., Rodellas, V., Alorda-Kleinglass, A., Saaltink, M., Folch, A., Garcia- Data will be made available on request. Orellana, J., 2022. Extreme precipitation events induce high fluxes of groundwater and associated nutrients to coastal ocean. Hydrol Earth Syst Sci 26 (18), Acknowledgments 4619–4635. https://doi.org/10.5194/hess-26-4619-2022. [14] Diego-Feliu, M., Rodellas, V., Saaltink, M.W., Alorda-Kleinglass, A., Goyetche, T., Martínez-Pérez, L., Folch, A., Garcia-Orellana, J., 2021. New perspectives on the This work was funded by the project “Mixing and dispersion in the use of 224Ra/228Ra and 222Rn/226Ra activity ratios in groundwater studies. transport of energy and solutes (MEDISTRAES)” coded as PID2019- J Hydrol (Amst) January, 596. https://doi.org/10.1016/j.jhydrol.2021.126043. [15] Dührkop, K., Fleischauer, M., Ludwig, M., Aksenov, A.A., Melnik, A.V., Meusel, M., 110212RB-C21 of the Spanish Government (I+D+I Retos Investigación). Dorrestein, P.C., Rousu, J., Böcker, S., 2019. SIRIUS 4: a rapid tool for turning We acknowledge the contribution of all the members of the MEDI­ tandem mass spectra into metabolite structure information. In: Nat Methods, 16. STRAES I, II and III project, from the Groundwater Hydrology Group of Springer, US, pp. 299–302. https://doi.org/10.1038/s41592-019-0344-8. [16] Duque, C., Calvache, M.L., Pedrera, A., Martín-Rosales, W., López-Chicano, M., the Technical University of Catalonia (UPC) and the Spanish National 2008. Combined time domain electromagnetic soundings and gravimetry to Research Council (CSIC), the Autonomous University of Barcelona determine marine intrusion in a detrital coastal aquifer (Southern Spain). J Hydrol (UAB), the University of Barcelona (UB), the Geosciences Montpellier (Amst) 349 (3–4), 536–547. https://doi.org/10.1016/j.jhydrol.2007.11.031. [17] Duque, C., Knee, K.L., Russoniello, C.J., Sherif, M., Abu Risha, U.A., Sturchio, N.C., Laboratory, CNRS and the University of Rennes, for their support during Michael, H.A., 2019. Hydrogeological processes and near shore spatial variability experimental laboratory tests, the Argentona site creation, the setup of of radium and radon isotopes for the characterization of submarine groundwater the electrodes during piezometer installation and for ensuring site discharge. J Hydrol (Amst), 579 (September): 124192. Elsevier,. https://doi.org/ maintenance. IDAEA-CSIC authors would like to express their gratitude 10.1016/j.jhydrol.2019.124192. [18] EEA. 2016. “Seawater intrusion in coastal aquifers.” Accessed July 6, 2022. 〈htt to The Spanish Ministry of Science and Innovation through the support ps://www.eea.europa.eu/publications/92–9167-056–1/page010.html〉. received as “Centro de Excelencia Severo Ochoa 2019–2023′′ . P. Gago- [19] Falgàs Parra, E., 2007. Hydrogeophysics as a multidisciplinary tool on aquifer Ferrero acknowledges his Ramón y Cajal fellowship (RYC2019- appraisal: Focus on AMT capabilities. University of Barcelona, (UB). [20] Folch, A., del Val, L., Luquot, L., Martínez-Pérez, L., Bellmunt, F., Le Lay, H., 027913-I) from the AEI-MICI. V. Rodellas acknowledges financial sup­ Rodellas, V., Ferrer, N., Palacios, A., Fernández, S., Marazuela, M.A., Diego- port from the Beatriu de Pinós postdoctoral programme of the Catalan Feliu, M., Pool, M., Goyetche, T., Ledo, J., Pezard, P., Bour, O., Queralt, P., Government (2019-BP-00241). Marcuello, A., Garcia-Orellana, J., Saaltink, M.W., Vázquez-Suñé, E., Carrera, J., 2020. Combining fiber optic DTS, cross-hole ERT and time-lapse induction logging to characterize and monitor a coastal aquifer (May). In: J Hydrol (Amst), 588. Appendix A. Supporting information Elsevier,, 125050. https://doi.org/10.1016/j.jhydrol.2020.125050 (May). [21] Gago-Ferrero, P., Schymanski, E.L., Bletsou, A.A., Aalizadeh, R., Hollender, J., Thomaidis, N.S., 2015. Extended suspect and non-target strategies to characterize Supplementary data associated with this article can be found in the emerging polar organic contaminants in raw wastewater with LC-HRMS/MS. online version at doi:10.1016/j.jhazmat.2023.130876. Environ Sci Technol 49 (20), 12333–12341. https://doi.org/10.1021/acs. est.5b03454. [22] Gago-Ferrero, P., Bletsou, A.A., Damalas, D.E., Aalizadeh, R., Alygizakis, N.A., References Singer, H.P., Hollender, J., Thomaidis, N.S., 2020. Wide-scope target screening of >2000 emerging contaminants in wastewater samples with UPLC-Q-ToF-HRMS/ [1] Abarca, E., 2006. Seawater intrusion in complex geological environments. UPC. MS and smart evaluation of its performance through the validation of 195 selected Technical University of Catalonia. representative analytes (July 2019). In: J Hazard Mater, 387. Elsevier, 121712. [2] Adyasari, D., Pratama, M.A., Teguh, N.A., Sabdaningsih, A., Kusumaningtyas, M.A., https://doi.org/10.1016/j.jhazmat.2019.121712 (July 2019). Dimova, N., 2021. Anthropogenic impact on Indonesian coastal water and [23] Garcia-Orellana, J., Rodellas, V., Tamborski, J., Diego-Feliu, M., van Beek, P., ecosystems: current status and future opportunities. In: Mar Pollut Bull, 171. Weinstein, Y., Charette, M., Alorda-Kleinglass, A., Michael, H.A., Stieglitz, T., Pergamon, 112689. https://doi.org/10.1016/J.MARPOLBUL.2021.112689. Scholten, J., 2021. Radium isotopes as submarine groundwater discharge (SGD) [3] Alygizakis, N.A., Gago-Ferrero, P., Borova, V.L., Pavlidou, A., Hatzianestis, I., tracers: review and recommendations (April). In: Earth Sci Rev, 220. Elsevier B.V,, Thomaidis, N.S., 2016. Occurrence and spatial distribution of 158 pharmaceuticals, 103681. https://doi.org/10.1016/j.earscirev.2021.103681 (April). drugs of abuse and related metabolites in offshore seawater. In: Science of The [24] Gaw, S., Thomas, K.V., Hutchinson, T.H., 2014. “Sources, impacts and trends of Total Environment, 541. Elsevier,, pp. 1097–1105. https://doi.org/10.1016/J. pharmaceuticals in the marine and coastal environment.”. Philos Trans R Soc B: SCITOTENV.2015.09.145. Biol Sci. [4] Andrisoa, A., Stieglitz, T.C., Rodellas, V., Raimbault, P., 2019. Primary production [25] Goldman, M., Kafri, U., 2007. Hydrogeophysical applications in coastal aquifers. in coastal lagoons supported by groundwater discharge and porewater fluxes Appl Hydrogeophys 233–254. inferred from nitrogen and carbon isotope signatures. In: Mar Chem, 210. Elsevier [26] González-Gaya, B., Dachs, J., Roscales, J.L., Caballero, G., Jiménez, B., 2014. B.V,, pp. 48–60. https://doi.org/10.1016/j.marchem.2019.03.003. “Perfluoroalkylated substances in the global tropical and subtropical surface [5] Arp, H.P.H., Hale, S.E., 2019. . REACH: Improvement of guidance and methods for the oceans.”. Environ Sci Technol 48 (22), 13076–13084. https://doi.org/10.1021/ identification and assessment of PMT/vPvM substances. Dessau-Roßlau, Germany. es503490z. [27] Goyetche, T., Pool, M., Carrera, J., Luquot, L., 2022. Hydromechanical characterization of tide-induced head fluctuations in coastal aquifers: the role of 10 D. Gutiérrez-Martín et al. Journal of Hazardous Materials 448 (2023) 130876 delayed yield and minor permeable layers. J Hydrol (Amst). Elsevier B.V, p. 612. [48] Pool, M., Carrera, J., 2011. A correction factor to account for mixing in Ghyben- https://doi.org/10.1016/j.jhydrol.2022.128128. Herzberg and critical pumping rate approximations of seawater intrusion in coastal [28] Goyetche, T., Luquot, L., Carrera, J., Martínez-Pérez, L., Folch, A., 2022. aquifers. Water Resour Res 47 (5). https://doi.org/10.1029/2010WR010256. Identification and quantification of chemical reactions in a coastal aquifer to assess [49] Robinson, C., Gibbes, B., Carey, H., Li, L., 2007. Salt-freshwater dynamics in a submarine groundwater discharge composition. In: Science of The Total subterranean estuary over a spring-neap tidal cycle. J Geophys Res 112 (C9), Environment, 838. Elsevier,, 155978. https://doi.org/10.1016/J. C09007. https://doi.org/10.1029/2006JC003888. SCITOTENV.2022.155978. [50] Ruttkies, C., Neumann, S., Posch, S., 2019. Improving MetFrag with statistical [29] Grau-Martínez, A., Folch, A., Torrentó, C., Valhondo, C., Barba, C., Domènech, C., learning of fragment annotations. In: BMC Bioinformatics, 20. BMC Bioinformatics, Soler, A., Otero, N., 2018. “Monitoring induced denitrification during managed pp. 1–14. https://doi.org/10.1186/s12859-019-2954-7. aquifer recharge in an infiltration pond. J Hydrol (Amst) 561, 123–135. https:// [51] Ruttkies, C., Schymanski, E.L., Wolf, S., Hollender, J., Neumann, S., 2016. MetFrag doi.org/10.1016/j.jhydrol.2018.03.044. relaunched: Incorporating strategies beyond in silico fragmentation. In: [30] Guo, Q., Li, H., Boufadel, M.C., Sharifi, Y., 2010. Hydrodynamics in a gravel beach J Cheminform, 8. Springer International Publishing, pp. 1–16. https://doi.org/ and its impact on the Exxon Valdez oil. J Geophys Res Oceans 115 (12), 1–21. 10.1186/s13321-016-0115-9. https://doi.org/10.1029/2010JC006169. [52] Santos, I.R., Chen, X., Lecher, A.L., Sawyer, A.H., Moosdorf, N., Rodellas, V., [31] Jurado, A., López-Serna, R., Vázquez-Suné, E., Carrera, J., Pujades, E., Petrovic, M., Tamborski, J., Cho, H.M., Dimova, N., Sugimoto, R., Bonaglia, S., Li, H., Hajati, M. Barceló, D., 2014. Occurrence of carbamazepine and five metabolites in an urban C., Li, L., 2021. Submarine groundwater discharge impacts on coastal nutrient aquifer. Chemosphere 115 (1), 47–53. https://doi.org/10.1016/j. biogeochemistry. Nat Rev Earth Environ 2 (5), 307–323. https://doi.org/10.1038/ chemosphere.2014.01.014. s43017-021-00152-0. [32] Kanwischer, M., Asker, N., Wernersson, A.S., Wirth, M.A., Fisch, K., Dahlgren, E., [53] Savatier, M., Rocha, C., 2021. Rethinking tracer-based (Ra, Rn, salinity) Osterholz, H., Habedank, F., Naumann, M., Mannio, J., Schulz-Bull, D.E., 2022. approaches to estimate point-source submarine groundwater discharge (SGD) into Substances of emerging concern in Baltic Sea water: Review on methodological coastal systems. In: J Hydrol (Amst), 598. Elsevier B.V, 126247. https://doi.org/ advances for the environmental assessment and proposal for future monitoring. In: 10.1016/j.jhydrol.2021.126247. Ambio, 51. Springer, Netherlands, pp. 1588–1608. https://doi.org/10.1007/ [54] Schymanski, E.L., Jeon, J., Gulde, R., Fenner, K., Ruff, M., Singer, H.P., s13280-021-01627-6. Hollender, J., 2014. Identifying small molecules via high resolution mass [33] Koussis, A.D., Mazi, K., Riou, F., Destouni, G., 2015. A correction for Dupuit- spectrometry: Communicating confidence. Environ Sci Technol 48 (4), 2097–2098. Forchheimer interface flow models of seawater intrusion in unconfined coastal https://doi.org/10.1021/es5002105. aquifers. J Hydrol 525, 277–285. https://doi.org/10.1016/j.jhydrol.2015.03.047. [55] Shi, Z.-Q., Liu, Y.-S., Xiong, Q., Cai, W.-W., Ying, G.-G., 2019. Occurrence, toxicity [34] Kurwadkar, S., Dane, J., Kanel, S.R., Nadagouda, M.N., Cawdrey, R.W., and transformation of six typical benzotriazoles in the environment: A review. The Ambade, B., Struckhoff, G.C., Wilkin, R., 2022. Per- and polyfluoroalkyl substances Science of the Total Environment 661, 407–421. https://doi.org/10.1016/j. in water and wastewater: a critical review of their global occurrence and scitotenv.2019.01.138. distribution. In: Science of the Total Environment, 809. Elsevier B.V,, 151003. [56] Sims, J.L., Stroski, K.M., Kim, S., Killeen, G., Ehalt, R., Simcik, M.F., Brooks, B.W., https://doi.org/10.1016/j.scitotenv.2021.151003. 2022. Global occurrence and probabilistic environmental health hazard assessment [35] Lall, U., L. Josset, and T. Russo. 2020. “A Snapshot of the World’s Groundwater of per- and polyfluoroalkyl substances (PFASs) in groundwater and surface waters. Challenges.” https://doi.org/10.1146/annurev-environ-102017. In: Science of the Total Environment, 816. Elsevier B.V, 151535. https://doi.org/ [36] Li, L., Zheng, H., Wang, T., Cai, M., Wang, P., 2018. Perfluoroalkyl acids in surface 10.1016/j.scitotenv.2021.151535. seawater from the North Pacific to the Arctic Ocean: Contamination, distribution [57] Spiteri, C., Slomp, C.P., Tuncay, K., Meile, C., 2008. Modeling biogeochemical and transportation. In: Environmental Pollution, 238. Elsevier Ltd,, pp. 168–176. processes in subterranean estuaries: Effect of flow dynamics and redox conditions https://doi.org/10.1016/j.envpol.2018.03.018. on submarine groundwater discharge of nutrients. Water Resour Res 44 (2). [37] Liu, Y.S., Ying, G.G., Shareef, A., Kookana, R.S., 2011. Biodegradation of three https://doi.org/10.1029/2007WR006071. selected benzotriazoles under aerobic and anaerobic conditions. Water Research 45 [58] Szymczycha, B., Borecka, M., Białk-Bielińska, A., Siedlewicz, G., Pazdro, K., 2020. (16), 5005–5014. https://doi.org/10.1016/j.watres.2011.07.001. Submarine groundwater discharge as a source of pharmaceutical and caffeine [38] Llopis-Albert, C., and D. Pulido-Velazquez. 2013. “Discussion about the validity of residues in coastal ecosystem: bay of Puck, southern Baltic Sea case study. Sci Total sharp-interface models to deal with seawater intrusion in coastal aquifers.” htt Environ 713. https://doi.org/10.1016/j.scitotenv.2020.136522. ps://doi.org/10.1002/hyp.9908. [59] Tamez-Meléndez, C., Hernández-Antonio, A., Gaona-Zanella, P.C., Ornelas- [39] Maréchal, J.C., Wyns, R., Lachassagne, P., Subrahmanyam, K., Touchard, F., 2003. Soto, N., Mahlknecht, J., 2016. Isotope signatures and hydrochemistry as tools in Anisotropie verticale de la perméabilité de l′ horizon fissuré des aquifères de socle: assessing groundwater occurrence and dynamics in a coastal arid aquifer. Environ Concordance avec la structure géologique des profils d′ altération. In: Comptes Earth Sci 75 (9). https://doi.org/10.1007/s12665-016-5617-2. Rendus - Geoscience, 335. Elsevier Masson SAS, pp. 451–460. https://doi.org/ [60] Taniguchi, M., Dulai, H., Burnett, K.M., Santos, I.R., Sugimoto, R., Stieglitz, T., 10.1016/S1631-0713(03)00082-8. Kim, G., Moosdorf, N., Burnett, W.C., 2019. Submarine groundwater discharge: [40] Martínez-Pérez, L., Luquot, L., Carrera, J., Marazuela, M.A., Goyetche, T., Pool, M., updates on its measurement techniques, geophysical drivers, magnitudes, and Palacios, A., Bellmunt, F., Ledo, J., Ferrer, N., del Val, L., Pezard, P.A., García- effects. Front Environ Sci. Frontiers Media S.A. Orellana, J., Diego-Feliu, M., Rodellas, V., Saaltink, M.W., Vázquez-Suñé, E., [61] Trezzi, G., Garcia-Orellana, J., Rodellas, V., Santos-Echeandia, J., Tovar- Folch, A., 2022. A multidisciplinary approach to characterizing coastal alluvial Sánchez, A., Garcia-Solsona, E., Masqué, P., 2016. Submarine groundwater aquifers to improve understanding of seawater intrusion and submarine discharge: a significant source of dissolved trace metals to the North Western groundwater discharge. J Hydrol (Amst) 607 (January), 127510. https://doi.org/ Mediterranean Sea. In: Mar Chem, 186. Elsevier B.V, pp. 90–100. https://doi.org/ 10.1016/j.jhydrol.2022.127510. 10.1016/j.marchem.2016.08.004. [41] McKenzie, T., Habel, S., Dulai, H., 2021. Sea-level rise drives wastewater leakage to [62] U.S. EPA 1993. “Wellhead Protection: A Guide for Small Communities. EPA/625/ coastal waters and storm drains. Limnol Oceano Lett 6 (3), 154–163. https://doi. R-93/002.” org/10.1002/lol2.10186. [63] U.S. EPA 2022. “Contaminants of Emerging Concern including Pharmaceuticals [42] McKenzie, T., Holloway, C., Dulai, H., Tucker, J.P., Sugimoto, R., Nakajima, T., and Personal Care Products.” Accessed August 16, 2022. 〈https://www.epa.gov Harada, K., Santos, I.R., 2020. Submarine groundwater discharge: A previously /wqc/contaminants-emerging-concern-including-pharmaceuticals-and-personal-ca undocumented source of contaminants of emerging concern to the coastal ocean re-products〉. (Sydney, Australia). In: Mar Pollut Bull, 160. Elsevier Ltd, 111519. https://doi.org/ [64] Vallejos, A., Sola, F., Pulido-Bosch, A., 2015. Processes influencing groundwater 10.1016/j.marpolbul.2020.111519. level and the freshwater-saltwater interface in a coastal aquifer. Water Resour [43] Menger, F., Ahrens, L., Wiberg, K., Gago-Ferrero, P., 2021. Suspect screening based Manag 29 (3), 679–697. https://doi.org/10.1007/s11269-014-0621-3. on market data of polar halogenated micropollutants in river water affected by [65] Yu, X., Xin, P., Lu, C., 2019. Seawater intrusion and retreat in tidally-affected wastewater. J Hazard Mater. Elsevier, p. 401. https://doi.org/10.1016/j. unconfined aquifers: laboratory experiments and numerical simulations. In: Adv jhazmat.2020.123377. Water Resour, 132. Elsevier Ltd,. https://doi.org/10.1016/j. [44] Moore, W.S., 2010. The effect of submarine groundwater discharge on the ocean. advwatres.2019.103393. Ann Rev Mar Sci 2 (1), 59–88. https://doi.org/10.1146/annurev-marine-120308- [66] Zamrsky, D., Karssenberg, M.E., Cohen, K.M., Bierkens, M.F.P., Oude Essink, G.H. 081019. P., 2020. Geological Heterogeneity of Coastal Unconsolidated Groundwater [45] Ozler, H.M., 2003. Hydrochemistry and salt-water intrusion in the Van aquifer, Systems Worldwide and Its Influence on Offshore Fresh Groundwater Occurrence. east Turkey. Environ Geol 43 (7), 759–775. https://doi.org/10.1007/s00254-002- Front Earth Sci (Lausanne) 7 (January), 1–23. https://doi.org/10.3389/ 0690-0. feart.2019.00339. [46] Palacios, A., José Ledo, J., Linde, N., Luquot, L., Bellmunt, F., Folch, A., [67] Zhang, X., Lohmann, R., Sunderland, E.M., 2019. Poly- And Perfluoroalkyl Marcuello, A., Queralt, P., Pezard, P.A., Martínez, L., del Val, L., Bosch, D., Substances in Seawater and Plankton from the Northwestern Atlantic Margin Carrera, J., 2020. Time-lapse cross-hole electrical resistivity tomography (CHERT) (research-article). In: Environ Sci Technol, 53. American Chemical Society, for monitoring seawater intrusion dynamics in a Mediterranean aquifer. Hydrol pp. 12348–12356. https://doi.org/10.1021/acs.est.9b03230. Earth Syst Sci 24 (4), 2121–2139. https://doi.org/10.5194/hess-24-2121-2020. [68] Zhou, Y., Sawyer, A.H., David, C.H., Famiglietti, J.S., 2019. Fresh Submarine [47] Pluskal, T., S. Castillo, A. Villar-Briones, and M. Orešič. 2010. MZmine 2: Modular Groundwater Discharge to the Near-Global Coast. Geophys Res Lett 46 (11), framework for processing, visualizing, and analyzing mass spectrometry-based molecular 5855–5863. https://doi.org/10.1029/2019GL082749. profile data. 11