Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system

STOTEN-21767; No of Pages 12 Science of the Total Environment xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

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

Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system F. Grimmeisen a,⁎, M.F. Lehmann b, T. Liesch a, N. Goeppert a, J. Klinger a, J. Zopfi b, N. Goldscheider a a b

Karlsruhe Institute of Technology (KIT), Institute of Applied Geosciences (AGW), Division of Hydrogeology, Kaiserstr. 12, D-76131 Karlsruhe, Germany University of Basel, Aquatic and Stable Isotope Biogeochemistry, Department of Environmental Sciences, Bernoullistr. 30, CH-4056 Basel, Switzerland

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

• A multiple tracer approach was used to study urban water source partitioning and mixing. • First nitrate δ15N and δ18O isotope data of a groundwater in Jordan are presented. • Distinct water δD and δ18O signatures allowed source identification in an urban water cycle. • Endmember mixing calculations revealed significant contributions of city effluents to groundwater. • Leaky networks and sewers contribute between 32% and 71% to polluted groundwater.

a r t i c l e

i n f o

Article history: Received 10 October 2016 Received in revised form 7 January 2017 Accepted 10 January 2017 Available online xxxx Editor: Jay Gan Keywords: Urban hydrogeology Karst aquifer Nitrate isotopes Semi-arid climate Jordan Pollution

a b s t r a c t Water supply in developing countries is prone to large water losses due to leaky distribution networks and defective sewers, which may affect groundwater quality and quantity in urban areas and result in complex subsurface mixing dynamics. In this study, a multi-stable isotope approach was used to investigate spatiotemporal fluctuations of surface and sub-surface water source partitioning and mixing, and to assess nitrogen (N) contamination in the urban water cycle of As-Salt, Jordan. Water import from the King Abdullah Canal (KAC), mains waters from the network, and wastewater are characterized by distinct isotopic signatures, which allowed us to quantify city effluents into the groundwater. Temporal variations in isotopic signatures of polluted groundwater are explained by seasonally fluctuating inflow, and dilution by water that originates from Lake Tiberias and enters the urban water cycle via the KAC. Isotopic analysis (N and O) and comparison between groundwater nitrate and nitrate from mains water, water imports and wastewater confirmed that septic waste from leaky sewers is the main contributor of nitrate contamination. The nitrate of strongly contaminated groundwater was characterized by highest δ15NNO3 values (13.3 ± 1.8‰), whereas lowest δ15NNO3 values were measured in unpolluted groundwater (6.9‰). Analogously, nitrate concentration and isotopic ratios were used for source partitioning and qualitatively confirmed δDH2O and δ18OH2O-based estimates. Dual water isotope endmember mixing calculations suggest that city effluents from leaky networks and sewers contribute 30–64% to the heavily polluted groundwater. Ternary mixing calculations including also chloride revealed that 5–18% of the polluted groundwater is wastewater. Up to two thirds of the groundwater originates from mains, indicating excessive water loss from the network, and calling for improved water supply management. © 2017 Elsevier B.V. All rights reserved.

⁎ Corresponding author. E-mail address: [email protected] (F. Grimmeisen).

http://dx.doi.org/10.1016/j.scitotenv.2017.01.054 0048-9697/© 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

2

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

1. Introduction Urban water management involves multiple aspects related to water supply, wastewater treatment and water resources management (Almeida et al., 2014). Consequently, urban hydrogeology encompasses an interdisciplinary understanding of the sources, transport, distribution and mixing of water and contaminants in the context of urban growth, societal changes, and climate variability (Howard, 2007; Lerner, 2002). Urban sprawl alters the environment, often with deteriorating effects on groundwater, both at different scales (Hibbs and Sharp, 2012). In order to safeguard against water shortages and to reduce contamination, it is crucial to control water losses in both supply and sewer networks. In this context, it is important to quantify recharge from leaky utility systems and to assess the relative contribution of individual sources and their impacts on groundwater quality (Schirmer et al., 2013). Effects of urbanization on aquifers are spatially and temporally multifaceted, rendering the identification of water sources and inputs intricate (Christian et al., 2011). Urban groundwater management requires indicators that allow the detection and tracing of man-made impacts (e.g. leakage from water pipelines, septic waste) (Strauch et al., 2007). The identification of artificial recharge processes from leaky utility systems is challenging because of its variability in time and space, often with multiple sources being involved (e.g. Barrett et al., 1999; Rutsch et al., 2006; Vazquez-Sune et al., 2010; Wolf et al., 2006). Densely populated areas in developing countries often suffer from poor water availability and/or defective infrastructure (Armstrong, 2009; Kumar et al., 2013), causing problems that are likely to be exacerbated with increasing water scarcity in the future (Vairavamoorthy et al., 2008). Furthermore, detailed information on urban groundwater

dynamics is often lacking, hampering the implementation of remediation measures. In Jordan, freshwater supplies rely primarily on groundwater (ElNaqa and Al-Shayeb, 2009), which is a limited resource. As a consequence of political crises in the Middle East, Jordan's population has grown strongly and intensified the stress on water resources. Most of the population (9.5 million in 2016) lives in urban centers located in the mountainous northwest of Jordan, such as the cities of Amman, Zarqa and As-Salt. Groundwater from nearby mountain aquifers is an important water source for these densely populated areas. To cope with their extensive freshwater demands, these urban centers import large amounts of surface water from the Yarmouk River and Lake Tiberias, which are transferred 110 km via the King Abdullah Canal (KAC) along the Lower Jordan Valley (LJV) (Alkhoury et al., 2010; Jiries et al., 2004) (Fig. 1A). Numerous studies highlight increased nitrate (NO− 3 ) concentrations in aquatic environments worldwide as widespread problem in urban and rural areas (e.g. Pasten-Zapata et al., 2014; Umezawa et al., 2008; Wakida and Lerner, 2005). Nitrate contamination can have a cascading set of negative consequences for aquatic environments, such as eutrophication, oxygen consumption and anoxia (e.g. Lehmann et al., 2015), and biodiversity loss. High nitrate concentrations in aquifers can lead to the deterioration of drinking water supplies. For several decades, nitrate N (and more recently O) isotopes have been used to trace nitrogen sources, including point sources (e.g., sewage plants), multipoint sources (e.g., leaky sewers), and diffuse sources (e.g., fertilizer inputs) (e.g. Aravena et al., 1993; Hale et al., 2014; Widory et al., 2005). Pinpointing the origin of N contamination is complicated, because nitrogen sources are partly not distinguishable based on their isotopic composition alone (Xue et al., 2009) (e.g. sewage and animal

Fig. 1. Location of the study area (A) and sampling sites (B). Number 1 to 4 in (A) refer to water resources contributing to King Abdullah Canal. Abbreviations refer to groundwater sites (GW) and mixed surface water sites (MW). The karst springs GW1–2 are used for drinking water supply, and GW3–4 for irrigation. GW5.1 and 5.2 are not used.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

waste have similar isotopic compositions), and processes such as denitrification lead to isotope fractionation altering the primary isotopic signatures (e.g. Fukada et al., 2004; McMahon and Bolke, 2006). In this study, we investigate the anthropogenic impacts on a karst aquifer below the city of As-Salt (Fig. 1B). This mid-size city (88,900 residents in 2011) has to cope with water scarcity, groundwater pollution, and water distribution concerns (Grimmeisen et al., 2016; Zemann et al., 2015). The situation of As-Salt, with its symptomatic water problems, can be considered representative for the whole area (Riepl, 2012). Therefore, the study serves as a test case for urban water management in similar semi-arid developing countries. Using a multi-tracer approach, our goal was to investigate dynamics and mixing of urban water sources to assess the impacts of contamination on groundwater quality, and to quantify water losses by leakage. Chloride (Cl−) was used as conservative tracer, unaffected by water treatment or biochemical processes. Stable isotope ratios of water (δDH2O and δ18OH2O) were determined to constrain mixing processes. As different water sources have different hydrological histories and, hence, distinct H and O isotopic signatures, their isotopic compositions are diagnostic and can be used for allocating specific natural and anthropogenic water reservoirs (e.g. Adar and Nativ, 2003; Lecuyer et al., 2012). Groundwater in the study area suffers from severe NO− 3 contamina15 tion. Using isotopic measurements of dissolved NO− 3 (δ NNO3 and δ18ONO3) – the first data of their kind for groundwater in Jordan – in combination with other isotopic/hydrochemical tracers, our goal was also to trace the sources and fate of contaminant nitrate, and to quantify the wastewater contribution to the karst aquifer underneath the city. 2. Study area 2.1. Hydrogeologic setting The study area is located on the western slopes of the Transjordanian Mountains in the upper part of Wadi Shueib, which is characterized by a dense network of steep valleys (Fig. 1A). The semi-arid Mediterranean climate causes short rainy winters and long dry summers. The drainage of the area is directed into one stream towards the southwestern outlet of the wadi. The main urban centers are the cities of As-Salt and Fuhais, where the majority of the local population lives. The major impact on the local environment originates from effluents in the city of As-Salt, which is located at 800–1100 m above sea level (asl; Fig. 1B). Groundwater quality is deteriorating due to untreated wastewater inputs from leaky sewers and cesspits infiltrating into karstic limestone, which is particularly vulnerable to anthropogenic pollution (Al-Kharabsheh et al., 2013; Margane et al., 2010). The wastewater treatment plants (WWTPs) of As-Salt and Fuhais release their effluents into the main Wadi Shueib stream, where they mix with surface runoff from the city and discharge from local springs. This study focuses on six springs: Shoreia = GW1, Baqqouria = GW2, Hazzir = GW3, Farkha = GW4, Jadour lower = GW5.1 and Jadour upper = GW5.2 (Fig. 1B). GW2 to GW5 are situated downstream of As-Salt, near the main course of Wadi Shueib, and GW1 in Wadi Azraq, a side valley. All six springs drain one larger catchment, comprising two Upper Cretaceous karst aquifers, the Es Sir Limestone and the Hummar formation, which are hydrogeologically interconnected by complex tectonics (Hahne et al., 2008). The area comprises several sub-catchments. GW2 emerges at the lowest elevation (390 m asl) and has the highest average discharge (116 L s− 1; Supplementary Table S1). The local hydrogeology is described by Werz and Hoetzl (2007) and Zemann et al. (2015). Water supplies within the urban hydrological cycle and potential urban sources with impact on the groundwater within the sub-catchments of the springs are schematically illustrated in Fig. 2 (right panel). The degree of anthropogenic influence on groundwater quality increases with the hydrographic proximity to the urban area and degree

3

of urbanization. Contamination arises from sporadic application of animal manure in rural areas (e.g. Wadi Azraq, GW1), leaky cesspits or septic tanks in small settlements, inflow from WWTPs (e.g. GW2), and, in the suburban communities from leaks in piped sewers and cesspits (e.g. GW3 and GW4). In the city center of As-Salt contamination and highest water losses, both from the supply network and from sewer systems can be expected (e.g. GW5). Zemann et al. (2015) also observed that levels of NO− 3 , E. coli and pharmaceutical compounds in GW2–4 are indicative of the urban influence by sewer effluents. Mains water for the As-Salt water district is supplied by approximately equal shares of local groundwater (44% to 54% between 2011 and 2013) and water import from KAC (46% to 56%; Fig. 2 and Supplementary Table S2). Groundwater is mainly extracted from GW2 and local wells located up-stream of As-Salt. GW1 contributes marginally to the public network. Overall, 53% to 59% of the yearly public water supply is classified as non-revenue water (i.e. lost before it reaches the end user), attributed to network losses and partly to unauthorized consumption.

2.2. Water transfer through the King Abdullah Canal The KAC, an artificial open canal, diverts water for irrigation to the LJV, and ultimately feeds the Zai water treatment plant (WTP) for freshwater distribution in Amman and As-Salt, among other urban centers (Fig. 1A). Water from KAC represents a mixture of waters from the Yarmouk River, Lake Tiberias, the Mukheiba wells (Fig. 2 and Supplementary Table S3), and to a minor extent, from several side wadis. Total KAC inputs sum up to 100 to 200 million cubic meters per year (MCM/a). Approximately 50 to 70 MCM/a of the drinking water leaving the Zai WTP are channeled to Amman and 5 MCM/a to As-Salt (unpublished data, Water Authority of Jordan). Lake Tiberias supplies 30–50% to the annual discharge of KAC, with strong seasonal fluctuations. Between April and November, lake water makes up 50% to 60% of the water in the canal; from December to March, the lake contributes b20% (Alkhoury et al., 2010; Supplementary Fig. S1 and Table S4). The other inputs (side wadis and Yarmouk River) also undergo seasonal variations, particularly during winter when discharge rates are high due to flood events. Then, reduced inflows from the lake are compensated by riverine inputs to KAC.

3. Materials & methods 3.1. Sampling campaigns and site selection Between 2012 and 2013, we conducted five water-sampling campaigns (Supplementary Table S5 and Fig. S2), which were planned to account for hydrological differences between summer and winter. Samplings took place twice at the end of winter in February/March 2012 (1.) and 2013 (3.), and twice at the end of summer in September 2012 (2.) and November 2013 (5.). One sampling campaign took place at the beginning of the dry season in May/June 2013 (4.). Groundwater samples were collected from the six karst springs (GW). We also sampled wastewater at the As-Salt WWTP, rainwater in As-Salt, treated import water from the KAC, and mains water in As-Salt. Three wastewater samples were collected in June 2013 (MW3). Samples of water imported from the KAC were taken at the Zai WTP (MW1) in June and November 2013. The mains water (MW2) ultimately used for consumption at As-Salt is MW1 water mixed with treated groundwater from GW2, GW1 and local wells. Representative MW2 samples from As-Salt were collected from the tap in a household near GW3. Precipitation samples were collected in January 2013 using a rain gauge placed on the roof of a building to GW3. In total, 103 groundwater samples (GW) and 13 mixed water samples (MW) were collected and analyzed. Sampling locations are given in Fig. 1.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

4

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

Fig. 2. Schematic illustration of the water supply and the urban water cycle of As-Salt. Sampling locations GW1–GW5 are karst springs; MW1–MW3 are mixed water resources. Numbers are based on unpublished data from the Water Authority of Jordan; MCM/a stands for million cubic meters per year.

3.2. Sampling and analyses

spring and sampling campaign. MDS provides a visual representation of a geometric configuration of points in a map to compare patterns of proximities (e.g. similarities or distances between points or groups) among a set of objects (Borg et al., 2012; Mito et al., 2011). Statistical analyses were conducted using OriginPro and R software (R Development Core Team 2008; stats package).

Water samples for the analysis of dissolved anions (Cl−, NO− 3 ) and stable water isotope ratios were filtered using 0.45-μm micropore membrane filters. Prior to analysis, all samples were stored at 4 °C. Anion concentrations were determined by ion-chromatography (DIONEX ICS-2100), isotope ratios by a spectroscopic water isotope analyzer (Los Gatos Research, U.S.A.). Water isotope signatures are reported in the delta notation (δDH2O, δ18OH2O) in ‰, relative to V-SMOW (Vienna-Standard Mean Ocean Water): δ = (Rsample / Rstandard − 1) × 1000‰, where R = D / H or 18O/16O in the sample and standard, respectively. Replicate reproducibility was generally better than ±0.5‰ for δDH2O and ±0.08‰ for δ18OH2O. For the NO− 3 isotopic ratio and concentration measurements, separate aliquots were collected and kept + frozen until analysis. Nitrite (NO− 2 ) and ammonium (NH4 ) were determined using photometric tests (Merck Spectroquant). The detection −1 limits for NO− , and 0.26 mg L−1 for 2 concentrations was 0.03 mg L + NH4 . The nitrogen (N) and oxygen (O) isotopic compositions of NO− 3 were determined using the denitrifier method (Casciotti et al., 2002; Sigman et al., 2001). N and O isotope ratios are reported as δ-values in ‰ relative to air N2 and V-SMOW, respectively. For isotope value calibration, internal (δ15N = 14.15‰) and international KNO3 reference materials with reported δ15N and δ18O values of 4.7‰ and 25.6‰ (IAEA-N3), and − 1.8‰ and 27.9‰ (USGS 34), respectively, were used. Replicate reproducibility was generally better than 0.3‰ for δ15N and 0.4‰ for δ18O.

For the quantitative assessment of different fractions (f) contributing to groundwater at locations GW2–5.2 down-gradient of As-Salt, three endmembers were defined that significantly differed in their combined δ18OH2O and Cl− concentration signature: (i) groundwater, (ii) wastewater from sewers, (iii) mains water from the distribution network. For pure-groundwater (fGW = 1; GW: δ18OH2O = − 5.79 and Cl− = 38.6 mg L−1) hydrochemical signatures, average values for GW1 were used. For wastewater (fwaste = 1; Waste: δ18OH2O = − 4.13 and Cl− = 252.0 mg L−1), we assumed water sampled at the As-Salt WWTP (MW3) to be most representative, and for mains water (fmains = 1; Mains: δ18OH2O = − 4.24 and Cl− = 136.4 mg L− 1) the hydrochemical signatures of MW2 were used in the endmember calculations. Contributions of the different endmembers to the mixed groundwater at springs down-stream of As-Salt were calculated using a threecomponent mixing analysis:

3.3. Statistical analyses

1 ¼ f GW þ f mains þ f waste

ð1Þ

δ18 OSpring ¼ f GW  δ18 OGW þ f mains  δ18 Omains þ f waste  δ18 Owaste

ð2Þ

We performed ANOVAs on the four parameters Cl−, NO− 3 , δDH2O and δ18OH2O to compare the effects of sampling location and time for each parameter separately. Since assumptions of normality (KolmogorovSmirnoff test; significance level N 0.05) or homogeneity of variances (Levene test; sig. b0.01) were partially not reached (Supplementary Table S6), non-parametric Kruskal-Wallis ANOVAs were applied. Twoway ANOVAs were performed with a reduced set of data groups (without GW5.1–5.2). Additionally, multidimensional scaling (MDS) was carried out on the metric values of the same four parameters for each

3.4. Endmember mixing analysis

−

−

−

−

½Cl Spring ¼ f GW  ½Cl GW þ f mains  ½Cl mains þ f waste  ½Cl waste

ð3Þ

f refers to the fraction of groundwater, mains water and wastewater, respectively.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

4. Results 4.1. Solute concentrations (Cl− and NO− 3 ) and isotopic compositions (H2O and NO− 3 ) The concentrations of analyzed anions (Cl− and NO− 3 ) and the water and nitrate isotopic compositions of spring water samples collected during all samplings are summarized in Fig. 3. Table 1 presents mean values from spring samples and urban water (MW-samples). The five springs emerging in the city of As-Salt (GW5.1 and GW5.2), in the Wadi Shueib (GW4 to GW2) and Wadi Azraq (GW1) showed a considerable spatial variation of Cl−, NO− 3 (Fig. 3A and C) and δDH2O

5

and δ18OH2O (Fig. 3B and D). The observed concentration and isotopic changes appear to be linked to the degree of urban influence, with elevated anion concentrations and higher δDH2O and δ18OH2O at sampling sites proximal to the city (Fig. 3B and D). Groundwater at GW3–5 contained Cl− at concentrations between 80 and 135 mg L−1, and −1 NO− , which is the maximum admissible concentration ac3 N 50 mg L cording to the Jordanian drinking water standard. Groundwater and −1 springs with NO− are considered severely polluted. 3 N 50 mg L Groundwater in the more rural sites displayed lower Cl− concentrations (mean ± SD) of 43.4 ± 6.7 mg L−1 (GW1) and 59.3 ± 9.0 mg L−1 −1 (GW2) and NO− and 32.9 ± 3 concentrations of 28.3 ± 3.4 mg L 3.4 mg L− 1, respectively, indicating a comparatively low pollution

18 15 18 Fig. 3. Variations of mean Cl− (A) and NO− 3 concentrations (C), δDH2O (B) and δ OH2O (D), and δ NNO3 (E) and δ ONO3 (F) of groundwater (GW). Error bars show minimum and maximum values during one sampling campaign.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

6

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

Table 1 18 15 18 Mean values (±standard deviation) of Cl− and NO− 3 , δDH2O and δ OH2O, and δ NNO3 and δ ONO3 of groundwater (GW), import water from the KAC, local mains water, and wastewater. n stands for the number of samples. Samples GW1 GW2 GW3 GW4 GW5.1 GW5.2 Imported water (MW1) Mains water (MW2) Wastewater (MW3)

Mean n Mean n Mean n Mean n Mean n Mean n Mean n Mean n Mean n

± SD ± SD ± SD ± SD ± SD ± SD ± SD ± SD ± SD

Cl− mg L−1

NO− 3 mg L−1

δDH2O ‰

43.4 ± 6.7 18 59.3 ± 9.0 18 103.5 ± 15.4 18 98.1 ± 8.5 17 114.4 ± 8.9 15 107.5 ± 10.1 16 171.6 ± 3.3 3 132.5 ± 7.8 7 236.6 ± 36.2 4

28.3 ± 3.4 18 32.9 ± 3.4 18 52.7 ± 3.6 18 57.2 ± 5.2 17 70.5 ± 7.3 15 68.2 ± 7.0 16 7.2 ± 3.4 3 15.7 ± 2.3 7 4.1 ± 2.1 4

−26.8 14 −26.4 14 −24.4 17 −25.1 13 −24.1 13 −24.2 13 −15.5 3 −18.8 7 −19.8 3

level. NO− 2 contributed only negligible amounts to NOx in groundwater samples, with concentrations between 0.06 and 0.11 mg L−1 (data not shown). δDH2O and δ18OH2O of polluted springs showed greater temporal fluctuations than those observed at less polluted sites, with lower values in February to March, and higher values during the dry season. Considerable spatial variation was also observed in the N and O isotope composition of NO− 3 , but without any clear spatial trend (Fig. 3E, F). Lowest δ15NNO3 values (~6.5‰) were observed at the least polluted groundwater site (GW1), and highest values were found for GW3 (12.4‰) (Fig. 3E). The δ18ONO3 variability was lower, with values ranging from 2.4‰ to 4.8‰ (Fig. 3F). Cl− concentrations in samples from local inputs, such as mains water (MW2: 132.5 ± 7.8 mg L−1) and wastewater (MW3: 236.6 ± 36.2 mg L−1), were considerably higher than in groundwater (86.4 ± 28.6 mg L−1) (Table 1). The relatively high Cl− concentrations in the mains water are attributed to a mix of local groundwater and water imported from the KAC, which contains considerably more Cl− (MW1: 171.6 ± 3.3 mg L−1; Supplementary Table S6). The observed elevated Cl− concentrations of KAC water are consistent with the weighted mean concentrations of the main inputs to the canal, i.e., water from Lake Tiberias with Cl− between 200 and 340 mg L−1 (Hötzl et al., 2008) and from Yarmouk River (133 mg L−1) (Farber et al., 2004). The maximum Cl− concentrations in the untreated wastewater samples (MW3) −1 corresponded to comparatively low levels of NO− ), 3 (4.1 ± 2.1 mg L −1 + despite high levels of NH+ (≫10.3 mg L ). At the other sites, NH con4 4 centrations were below detection limit. Dissolved NO− 2 concentrations were also below detection limit in mains water (MW2) and import water (MW1), and slightly elevated in wastewater (0.1–0.8 mg L−1; data not shown).

4.2. Statistical assessment of spatiotemporal variability The pronounced seasonal effects with regards to the groundwater quality (Fig. 3) is confirmed through Kruskal-Wallis ANOVAs of Cl− and NO− 3 concentrations, as well as stable isotope ratios of water (δDH2O and δ18OH2O) (Supplementary Table S8). The ANOVAs confirmed significant seasonal differences in the analyzed hydrochemical/isotope parameters, with p-values between 0.024 and 0.041. Generally, mean Cl− concentrations were higher in summer, with differences of N10 mg L−1 compared to winter. Similarly, δDH2O and δ18OH2O values in summer were systematically higher than winter values, whereas NO− 3 concentrations did not vary significantly (p = 0.077).

± 0.5 ± 0.6 ± 1.2 ± 0.6 ± 1.3 ± 1.1 ± 0.3 ± 0.9 ± 0.2

δ18OH2O ‰

δ15NNO3 ‰

δ18ONO3 ‰

−5.9 14 −5.8 14 −5.3 17 −5.5 13 −5.0 13 −5.2 13 −3.0 3 −4.0 7 −4.1 3

6.9 2 8.6 2 11.5 ± 0.7 6 9.6 2 9.7 2 9.6 1 13.3 ± 1.8 3 9.7 ± 0.5 4 9.3 2

3.0 2 3.9 2 3.9 ± 0.5 6 3.4 2 2.5 2 2.6 1 6.1 ± 3.1 3 4.8 ± 1.8 4 7.4 2

± 0.1 ± 0.1 ± 0.3 ± 0.3 ± 0.4 ± 0.3 ± 0.1 ± 0.2 ± 0.1

We also performed two-way ANOVAs on a reduced data set (GW1– 4) to evaluate the effects of space and time (Supplementary Table S9). All tests showed significant variations between sampling locations and sampling periods (p b 0.05). Spatial effects were generally stronger than temporal effects. The most significant differences were observed for Cl− concentrations. No significant differences in space or time were observed for NO− 3 concentrations (p = 0.065). The MDS map allows the classification of the data set into two “spatial” clusters reflecting unpolluted (GW1–2) and polluted springs (GW3–5.2) (Supplementary Fig. S3). The MDS analysis also revealed clear hydrochemical parameter grouping in the context of temporal variability: samples taken during winter (February/March) are distinct from summer samples (September and November). Seasonal differences seem to be particularly pronounced for samples from polluted springs, whereas early-summer samples (May/June) from the unpolluted springs display a hydrochemical and isotopic signature similar to those collected during winter. 5. Discussion 5.1. Water source identification within the urban water cycle 5.1.1. Chloride and nitrate concentration-based constraints − Positive correlation (R2 = 0.684) between NO− was ob3 and Cl − served in groundwater from GW1–5.2. Moreover, NO3 and Cl− concentration signatures allow the categorization of the springs according to their groundwater regimes: the relatively unpolluted springs GW1–2 and the polluted springs GW3–5 (Fig. 4A). Springs GW1 and GW2 are used for local supply; hence, their hydrochemical signatures plot in the same space as groundwater from the two local supply wells. − The NO− 3 /Cl ratio has successfully been used to gain information on N dynamics and sewage inputs, because Cl− concentration changes only by mixing (i.e., dilution), and remains usually unaffected by wastewater treatment (Hinkle et al., 2008; Koba et al., 1997). Hence, plotting NO− 3 / Cl− molar ratios versus Cl− concentration helps distinguishing different input sources. In Fig. 4B, various potential sources to the As-Salt ground− water cycle are indicated. NO− 3 /Cl signatures not only permit the identification of sources affecting local groundwater directly or indirectly, they can also be used to elucidate mixing. The three most important processes are: (1) mixing of import water with groundwater for mains water distribution, (2) groundwater pollution due to sewer leak− age, (3) groundwater dilution due to network losses. The NO− 3 /Cl ratios of mains water lie between the values of import water (MW1) and groundwater used for water supply (GW1, GW2 and local wells),

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

7

− − − Fig. 4. Relationship between (A) Cl− and NO− 3 concentrations, and (B) between Cl molar concentrations and NO3 /Cl molar ratios of different water sources from the local water system. Means of two local wells are based on Zemann et al. (2015). The urban water cycle comprises three main processes/sources affecting local groundwater body: groundwater is mixed with import water for mains water distribution and consumption (1), mains water percolation (2), and sewage percolation (3).

which indicate mixing of the two waters (process 1). Sewage effluents − (MW3) display low NO− 3 /Cl ratios. Whereas we did not find clear evidence for nitrate regeneration by nitrification in the wastewater sam− ples, it is reasonable to assume that their NO− 3 /Cl is likely to increase after sewage percolation and rapid oxidation of the wastewater NH+ 4 to NO− 3 in the aquifer. Samples from the polluted springs (GW3– GW5) show elevated Cl− and NO− 3 concentrations, suggesting mixing of unpolluted groundwater (used for supply) with urban effluents, i.e. with nitrogen-contaminated sewage (process 2) and mains water (pro− cess 3) (Fig. 4). Thus the NO− 3 /Cl results underscore that local groundwater is significantly influenced by human impacts and consequently highly vulnerable to microbial contamination, as recently documented − (Grimmeisen et al., 2016). However, the NO− data do not allow 3 /Cl the quantitative disentanglement of network-loss dilution from pollution effects. 5.1.2. Water H and O isotopic constraints The isotopic composition of the water samples can provide additional constraints of groundwater sources and indicate recharge patterns (Clark and Fritz, 1997). Fig. 5 compares groundwater isotope ratios with respect to the eastern (EMWL) and local (LMWL: δD = 6.27 × δ18O + 11.4) meteoric water lines. Unpolluted spring water displays relatively low δD and δ18OH2O values (GW1/GW2: − 27.1 to − 26.0‰, and − 6.0 to − 5.7‰), close to the LMWL. All samples plot below both meteoric water lines. Samples from polluted springs (GW3–GW5: −25.8 to −22.4‰, and −5.8 to −4.5‰) showed higher δD and δ18OH2O, plotting below the LMWL. Linear regression of δD and δ18OH2O for the combined spring samples data set resulted in a δD/ δ18O slope of 3.11, which is distinctively lower than that of the LMWL (6.27) and indicates stronger influence by kinetic versus equilibrium isotope effects (Gat, 2010). The δD/δ18O slope depends on local conditions of humidity, and the fraction of water evaporated. Evaporation

results in increasing δD and δ18OH2O values and in lower regression slopes (e.g., Koh et al., 2010; Pasten-Zapata et al., 2014). The observed δD/δ18O trends between GW1 and GW5 support a direct link of unpolluted spring water to meteoric water originating under current climatological conditions that prevail in the Eastern Mediterranean basin (GW1 and GW2), and the replenishment from local recharge with water that has undergone significant evaporative loss. Local evaporative enrichment of the heavy water isotopes is also indicated when looking at δ18OH2O alone. Mains water δD and δ18OH2O signatures are reflective of local water management practices including mixing of import water and local groundwater. δD and δ18OH2O of supply wells are expected to reflect the isotopic composition of unpolluted springs (given their similar anion concentrations (Fig. 4A) and based on earlier studies in the same area, e.g. Bajjali (2006), Abu-Jaber and Kharabsheh (2008)). The import water contributing to the network is itself a mix of different water sources contributing to KAC water. Its isotopic signatures (MW1) are consistent with mixing of water from Lake Tiberias, the Yarmouk River and Mukheiba wells, since the δD and δ18OH2O signature of imported water plots on a mixing line between the different sources (Fig. 5). This nicely corroborates the validity of water isotopes to assess mixing of water sources, particularly in an area where the isotopic signatures of the respective sources are quite distinct. It becomes obvious that a large part of import water from KAC originates from Lake Tiberias with considerably higher δD and δ18OH2O (Gat and Dansgaard, 1972; Siebert, 2006). Since lake water contribution to KAC varies seasonally (Fig. 2 and Supplementary Fig. S1), the water isotopic composition of KAC water is expected to vary accordingly. In turn, the seasonal fluctuations in the H and O isotope signatures of import water likely modulate the isotopic composition of the mains water in As-Salt. The H and O isotopic data confirm that during the time when import water samples were collected (June and November 2013),

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

8

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

Fig. 5. Isotopic composition of regional water sources and groundwater. Spring water δDH2O and δ18OH2O mean values are each based on five sampling periods and six springs. Stable isotopic compositions are compared to regional water sources documented in the literature, and put in perspective to the eastern (EMWL; Gat and Carmi, 1970) and local meteoric water line (LMWL; Bajjali, 2012). The LMWL is based on weighted-average precipitation data collected from 11 rainfall stations in Jordan between 1965 and 2005. Data for Yarmouk River (unpublished) were provided by the Water Authority of Jordan.

contribution of Lake Tiberias water to KAC was about 50%, which in turn contributed about 50% to the water mix that is distributed in As Salt for consumption (Fig. 4 and Supplementary Fig. S1). 5.2. Nitrate N and O isotopic constraints on N contamination and mixing The dual isotope approach to trace nitrate contamination is based on the fact that nitrate from different origins have distinct isotopic signatures (Kendall et al. (2007) for a review). Artificial nitrate fertilizer, for example, displays significantly higher δ18O values compared to most other nitrate sources, whereas its δ15N is comparatively low. In contrast, nitrate derived from organic sources and septic waste tends to exhibit elevated δ15N values, but comparatively low δ18O values (Amberger and Schmidt, 1987; Thibodeau et al., 2013). Hence, for simple pointsource analyses, and in the absence of isotope-signal altering processes in the aquifer (e.g., denitrification under O2-deficient conditions), coupled nitrate N and O isotope fingerprinting can be quite helpful. Fig. 6A illustrates the isotopic compositions of NO− 3 in the study area. The dual isotope plot indicates that NO− 3 in the spring water samples derives mainly from manure and septic waste. All groundwater values fall also within the range indicative for nitrate that has undergone microbial nitrification. Synthetic fertilizers seem to be a negligible N source in the study area. No clear relationship was observed between δ15NNO3 and δ18ONO3 when all spring water samples were considered (slope = 0.15, R2 = 0.11, n = 15). However, using samples only from the polluted springs (GW3–5) for regression analysis, a weak positive relationship (R2 = 0.36, n = 11) was observed. The slope of the regression line (i.e., Δδ15NNO3:Δδ18ONO3) was approximately 2:1, which in other groundwater studies has been taken as indication for microbial denitrification (e.g., Aravena and Mayer (2010) and references therein). The δ15N of nitrate in the groundwater samples were generally similar to the δ15NNO3 of local mains water (9.7 ± 0.5‰), as well as wastewater (9.3‰). δ18ONO3 values of the mains water (4.8 ± 1.8‰) were also similar, but considerably higher in wastewater (7.4‰) than in the groundwater and mains water. Import water showed large variations and generally higher values of δ15NNO3 (13.3 ± 1.8‰) and δ18ONO3 (6.1 ± 3.1‰). Based on the generally elevated levels of δ15NNO3 (at low δ18ONO3) in all components of the urban water cycle, we may cautiously state that N contamination in the groundwater of the study area, and at

low levels even in the supply water and the urban network, is broadly due to septic/wastewater-derived nitrate (which may partially be denitrified in the aquifer). However, the data remain ambiguous with regards to the transport vectors of the wastewater N and mixing between the water sources. Additional insight with regards to urban mixing processes may be gained from the δ15N versus NO− 3 comparison (Fig. 6B). Again, the mains water geochemical signatures reflect the practice of mixing local groundwater (GW1–2) and import water. That is, they fall between the two mixing lines in between the mean values for groundwater and import water, indicating an approximately equivalent share between these two endmember sources, and confirming the water isotope data. Similarly, the combined nitrate concentration and δ15NNO3 data observed at GW2 suggest admixing of waters from of GW1 and GW3 water, explaining the somewhat elevated δ15NNO3 values of GW2 with respect to GW3. Hence our data confirm the hydrogeological connectivity between groundwater in the sub-catchments of GW1 and GW3, which both seem to recharge the sub-catchment of GW2. Simple mixing of more or less contaminated water sources in the local aquifer system, however, cannot explain the complete data set. Unclear, for example, are the high δ15NNO3 values and high NO− 3 concentrations in particular at GW3. Whereas we already argued that the whole urban water cycle seems affected by contamination with wastewater (usually leading to elevated δ15NNO3), any direct link between the wastewater N source and elevated nitrate concentrations and δ15NNO3 values in the urban groundwater remains hidden (Fig. 6B) as the fate of contaminant ammonium-N from leaky wastewater/septic systems is unknown. It can generally be assumed that the addition of NH+ 4 from sewer systems ultimately leads to increased NO− 3 concentrations and higher δ15NNO3 values in groundwater. NH+ 4 is rapidly and completely convert− ed to NO− 3 , hence the isotopic composition of NO3 in the groundwater tends to reflect the N isotopic signature of the NH+ 4 source (Aravena and Mayer, 2010). We speculate, that the δ15N values found for GW3 (and to a minor extent also for GW4 and GW5) can best be explained by the infiltration of ammonium-rich (and NO3-poor) wastewater, and the nitrification of the strongly 15N-enriched NH+ 4 in the groundwater aquifer. This

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

9

Fig. 6. (A) δ18ONO3 versus δ15N in the Wadi Shueib catchment. The pool-δ15N values are based on Aravena and Mayer (2010) and Xue et al. (2009), and references therein. (B) δ15NNO3 values versus NO3− concentrations.

“cryptic” high-δ15N wastewater nitrate may become even more enriched in 15N due to isotope fractionation by microbial denitrification in anoxic parts of the aquifer, as indicated by the concomitant and characteristic increase of both nitrate δ15N and δ18O (e.g. Lu et al., 2015) in the more polluted portions of the aquifer. While we can exclude that extensive anoxia prevails anywhere in the groundwaters of the study area, it is plausible to assume that particularly in the polluted urban areas, anoxic conditions may occur locally due to improper disposal of sewage and/or in anoxic microenvironments of the polluted aquifer (Pedersen et al., 1991). There, besides denitrification also anaerobic ammonium oxidation (anammox) may contribute to nitrogen attenuation and the enrichment of 15N-NOx in groundwater (Brunner et al., 2013; Robertson et al., 2012). Recent studies suggest that anammox is widespread in terrestrial aquatic environments, including aquifers contaminated by ammonium (Humbert et al., 2010; Moore et al., 2011). 5.3. Endmember mixing analysis on artificial recharge of polluted springs In Fig. 7A the δDH2O and δ18OH2O of all water samples are summarized to highlight spatiotemporal variations and distinct characteristics: (i) samples of GW1 plot close to the LMWL and do not show pronounced seasonal dynamics. (ii) Particularly GW3 and GW5 were more enriched in the heavy isotopes, and isotopic enrichment is not caused by local evaporation only, but rather by mixing with D and 18 O-enriched water, i.e. urban seepage water with short residence times (few days). During the wet season, the isotopic composition of polluted springs seems to resemble that of unpolluted springs (Fig. 7A), reflecting the dominant influence of natural recharge. During summer, urban waters were significantly more enriched in the heavy water isotopes than the groundwater (Fig. 7B). The elevated δD and δ18O signatures of urban waters (MW-samples) result from water refinement practices prior to network distribution. The fact that wastewater samples show a similar isotopic signature as mains water,

indicates that no significant mixing or fractionation processes occurred after distribution. This allows us to treat wastewater and network-loss water as one single source of artificial recharge, and to quantify the percolation of combined city effluents to the groundwater aquifer. In order to do so, a two-endmember water-isotope based mixing model using δDH2O and δ18OH2O was applied. These calculations are analogous to the three-component mixing approach, except that mains water and wastewater endmembers were combined. This combined endmember can be considered as “city effluent”, including wastewater and leakage water from the network. Isotope values of GW1 (GW: δ18OH2O = − 5.79 and δDH2O = − 26.8) on the one end and of city effluents (Mains + Waste: δ18OH2O = − 4.19 and δDH2O = − 19.9) on the other were used as natural and polluted endmembers, respectively. GW1 is likely to reflect natural background conditions given that its δD and δ18O values are close the LMWL. Mains water and mixing practices for the water distribution in As-Salt undergo significant seasonal variations, e.g. due to the variable contribution of water from Lake Tiberias (see section 5.1.2). Nevertheless, in the δD/δ18O space, the combined GW data (including all GW data) follow the mixing line between unpolluted water (GW1) and effluent water (mains and wastewater) with a slightly lower slope of the GW regression line suggesting input from another seasonally variable source of water. Mixing calculation of the artificial recharge in polluted springs was based on samples taken in May and June only, since the wastewater samples that constrain the polluted endmember were taken during that period (Fig. 7B). The May/June sampling corresponded to a period without precipitation and hence, is most representative for base flow conditions. The reduced May and June isotope data set is particularly well represented by the mixing line (Fig. 7B), confirming the validity of the twoendmember approach. The fraction of urban seepage water (fmains + waste) in the groundwater ranged between 3% and 64% (Supplementary Table S10). Particularly at the polluted springs (GW5), artificial recharge was very high,

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

10

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

Fig. 7. δDH2O and δ18OH2O for all samplings (A) and mean values (B) between May to June 2013. The LMWL is indicated according to Bajjali (2012). (C) δ18OH2O versus NO− 3 concentrations for all samplings; (D) δ18OH2O versus Cl− concentrations (means) for May to June 2013. Error bars indicate minimum and maximum values. The black dashed lines connect the two endmembers in (B) and the three endmembers in (D), which are used for the mixing analysis.

indicating extreme contamination with effluent waters. At GW2, the contribution from urban seepage was rather small. Nonetheless, hydrochemical analyses confirmed some urban influence also at GW2, as indicated by the nitrate isotope data (section 3.2). In order to further constrain the relative importance of the two main components of urban effluent water (i.e., mains versus wastewater), we investigated ternary mixing using δ18OH2O versus NO− 3 (Fig. 7C), and δ18OH2O versus Cl− (Fig. 7D; see endmember values in section 3.4).

Fig. 7C depicts δ18OH2O versus NO− 3 and illustrates temporal and spatial differences of the groundwater and the chosen endmembers. The natural endmember (GW1) contains already significant NO− 3 concentrations that additionally vary throughout the year. This strongly reduces the sensitivity of any approach based on nitrate concentrations. Cl− concentrations in the unpolluted endmember (GW1) are in good agreement with historical data, which supports our assumption that they represent the natural background. However, seasonal

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

fluctuations were also observed for Cl− concentrations in mains and wastewater, blurring the endmember signatures of the two effluent water sources. During summer, Cl− in groundwater increases due to increased Cl− in import water and also increased soil water salinity caused by local transpiration effects (Zagana et al., 2007). Additional variability arises from local water masses mixing (= groundwater + import water) and (variable) mixing in the KAC, both affecting the water isotope ratios and Cl− concentrations of mains water and the wastewater. We conducted ternary mixing calculation using the mean values for the May/June 2013 samplings (Supplementary Table S10). Our calculations revealed that mains water (fmains) contributed between 10% and 66% to water at the polluted springs, whereas the fractions of wastewater (fwaste) ranged between 5% and 26%. The estimated range for total artificial recharge (fmains + waste) was between 32% and 71%, consistent with the estimates based on the binary approach. Generally, the partitioning between the urban sources of effluent waters seems to vary significantly. The two springs in the city area GW5.1 (fmains = 66%) and GW5.2 (fmains = 36%) contained the highest fractions of mains water, highlighting the importance of network loss particularly in the central urban area. Mains water contribution was significantly lower at GW3 and GW4, whereas wastewater contamination was rather higher. At GW2, the fmains + waste was 9.3%, an estimate that stands in excellent agreement with the two-endmember mixing calculation (9.7%). Yet with the additional constraint from Cl− we can also state that the urban effluent is almost equally partitioned between wastewater and mains water (fwaste = 5.1%, fmains = 4.2%). Both, the ternary and binary mixing calculations clearly indicate increasing city effluent recharge with proximity to the urban area, where up to two thirds of the groundwater originates from mains and water pipelines. While our quantitative assessment is based on early summer data only (May/June), the isotope and hydrochemical data for the other sampling periods provide qualitative evidence that significant network losses and waste-water leakage occurs year-round. 6. Conclusions In this study, environmental isotopes were used to identify and quantify surface and sub-surface mixing processes in an urban groundwater system in Jordan. The isotope-hydrochemical analysis allowed us to disentangle pre-consumption water source mixing, and highlighted distinct seasonal changes in the relative importance of the water sources used for water import through the KAC and supply in the urban network. We could demonstrate that the groundwater resources underneath the city are strongly affected by enormous water infrastructure losses, with up to 70% of the groundwater in the urban area originating from city effluents. Our data clearly highlight an increasing effluent contamination trend with proximity to the urban center, and a spatially variable partitioning between the effluent components. The nitrate isotope data suggest that N contamination in the urban aquifer most likely originates from septic waste ammonium, which is oxidized after release to the urban aquifer. These are the first nitrate isotope data for a groundwater aquifer in Jordan. While the limited data set is still ambiguous with regards to the mixing and exact fate of contaminant nitrogen sources (nitrification/denitrification), it provides evidence for city-effluent contamination/artificial recharge that is qualitatively consistent with the presented mixing models. Wastewater contributes up to 25% to the groundwater, but contribution from leaky sewer systems is still much smaller than the contribution from mains water, indicating the relevance of artificial recharge by network losses. Our study demonstrates that environmental isotopes are a useful tool to monitor mixing processes in urban settings, to provide substantial knowledge on urban groundwater systems and their contamination, and to gain constraints on urban network loss. Such knowledge is a prerequisite for planning effective measures to preserve groundwater quality and to prevent waste of water. The approach presented here should

11

also be applicable to other water supply systems in semi-arid regions that depend on water imports from distant sources, allowing constraints on water source mixing, network leakage and contamination in urban groundwaters. Acknowledgements We are grateful to Eng. Susan Kilani from of the Jordanian Ministry of Water and Irrigation (MWI) for providing access to unpublished data. We thank Dr. Ali Sawarieh for local help. Laboratory analyses were conducted with the help of Dr. Thomas Kuhn (Aquatic and Stable Isotope Biogeochemistry, University of Basel), Gesine Preuß, Daniela Blank, Chris Buschhaus and Christine Roske-Stegemann (AGW, KIT). The German Federal Ministry of Education and Research (BMBF) is acknowledged for funding the SMART project through grants 02WM10791086 and 02WM1211-1212. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2017.01.054. References Abu-Jaber, N., Kharabsheh, A., 2008. Ground water origin and movement in the upper Yarmouk Basin, Northern Jordan. Environ. Geol. 54, 1355–1365. Adar, E., Nativ, R., 2003. Isotopes as tracers in a contaminated fractured chalk aquitard. J. Contam. Hydrol. 65, 19–39. Al-Kharabsheh, N.M., Al-Kharabsheh, A.A., Ghnaim, O.M., 2013. Effect of septic tanks and agricultural wastes on springs' water quality deterioration in Wadi Shu'eib catchment area - Jordan. Jordan Journal of Agricultural Sciences 9 (1) (2013). Alkhoury, W., Ziegmann, M., Frimmel, F.H., Abbt-Braun, G., Salameh, E., 2010. Water quality of the King Abdullah Canal/Jordan–impact on eutrophication and water disinfection. Toxicol. Environ. Chem. 92, 855–877. Almeida, M.C., Vieira, P., Smeets, P., 2014. Extending the water safety plan concept to the urban water cycle. Water Policy 16, 298–322. Amberger, A., Schmidt, H.L., 1987. Natürliche Isotopengehalte von Nitrat als Indikatoren für dessen Herkunft. Geochim. Cosmochim. Acta 51, 2699–2705. Aravena, R., Mayer, B., 2010. Isotopes and Processes in the Nitrogen and Sulfur Cycles. Environmental Isotopes in Biodegradation and Bioremediation. pp. 203–246. Aravena, R., Evans, M.L., Cherry, J.A., 1993. Stable isotopes of oxygen and nitrogen in source identification of nitrate from septic systems. Ground Water 31, 180–186. Armstrong, A., 2009. Water pollution urban waste. Nat. Geosci. 2, 748. Bajjali, W., 2006. Recharge mechanism and hydrochemistry evaluation of groundwater in the Nuaimeh area, Jordan, using environmental isotope techniques. Hydrogeol. J. 14, 180–191. Bajjali, W., 2012. Spatial variability of environmental isotope and chemical content of precipitation in Jordan and evidence of slight change in climate. Appl Water Sci 2, 271–283. Barrett, M.H., Hiscock, K.M., Pedley, S., Lerner, D.N., Tellam, J.H., French, M.J., 1999. Marker species for identifying urban groundwater recharge sources: a review and case study in Nottingham, UK. Water Res. 33, 3083–3097. Borg, I., Groenen, P.J., Mair, P., 2012. Applied Multidimensional Scaling. Springer Science & Business Media. Brunner, B., Contreras, S., Lehmann, M.F., Matantseva, O., Rollog, M., Kalvelage, T., Klockgether, G., Lavik, G., Jetten, M.S.M., Kartal, B., Kuypers, M.M.M., 2013. Nitrogen isotope effects induced by anammox bacteria. Proc. Natl. Acad. Sci. U. S. A. 110, 18994–18999. Casciotti, K.L., Sigman, D.M., Hastings, M.G., Bohlke, J.K., Hilkert, A., 2002. Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal. Chem. 74, 4905–4912. Christian, L.N., Banner, J.L., Mack, L.E., 2011. Sr isotopes as tracers of anthropogenic influences on stream water in the Austin, Texas, area. Chem. Geol. 282, 84–97. Clark, I., Fritz, P., 1997. Environmental Isotopes in Hydrogeology. CRC Press. El-Naqa, A., Al-Shayeb, A., 2009. Groundwater protection and management strategy in Jordan. Water Resour. Manag. 23, 2379–2394. Farber, E., Vengosh, A., Gavrieli, I., Marie, A., Bullen, T.D., Mayer, B., Holtzman, R., Segal, M., Shavit, U., 2004. The origin and mechanisms of salinization of the Lower Jordan River. Geochim. Cosmochim. Acta 68, 1989–2006. Fukada, T., Hiscock, K.M., Dennis, P.F., 2004. A dual-isotope approach to the nitrogen hydrochemistry of an urban aquifer. Appl. Geochem. 19, 709–719. Gat, J.R., 2010. Isotope hydrology: a study of the water cycle. Vol 6: World Scientific. Gat, J.R., Carmi, I., 1970. Evolution of isotopic composition of atmospheric waters in Mediterranean-Sea Area. J. Geophys. Res. 75 (15), 3039–3048. Gat, J.R., Dansgaard, W., 1972. Stable isotope survey of the fresh water occurrences in Israel and the northern Jordan Rift Valley. J. Hydrol. 16, 177–211. Grimmeisen, F., Zemann, M., Goeppert, N., Goldscheider, N., 2016. Weekly variations of discharge and groundwater quality caused by intermittent water supply in an urbanized karst catchment. J. Hydrol. 537, 157–170.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054

12

F. Grimmeisen et al. / Science of the Total Environment xxx (2017) xxx–xxx

Hahne, K., Margane, A., Borgstedt, A., 2008. Geological and tectonical setting of the upper Wadi Shuayb catchment area, Jordan. Technical Cooperation Project ‘Groundwater Resources Management’, Special Report No. 7, Hannover. Hale, R.L., Turnbull, L., Earl, S., Grimm, N., Riha, K., Michalski, G., Lohse, K.A., Childers, D., 2014. Sources and transport of nitrogen in arid urban watersheds. Environ. Sci. Technol. 48, 6211–6219. Hibbs, B.J., Sharp, J.M., 2012. Hydrogeological impacts of urbanization. Environ. Eng. Geosci. 18, 3–24. Hinkle, S.R., Bohlke, J.K., Fisher, L.H., 2008. Mass balance and isotope effects during nitrogen transport through septic tank systems with packed-bed (sand) filters. Sci. Total Environ. 407, 324–332. Hötzl, H., Möller, P., Rosenthal, E., 2008. The Water of the Jordan Valley: Scarcity and Deterioration of Groundwater and Its Impact on the Regional Development. Springer. Howard, K.W., 2007. Urban groundwater, meeting the challenge. IAH Selected Papers on Hydrogeology 8. CRC Press. Humbert, S., Tarnawski, S., Fromin, N., Mallet, M.P., Aragno, M., Zopfi, J., 2010. Molecular detection of anammox bacteria in terrestrial ecosystems: distribution and diversity. ISME J. 4, 450–454. Jiries, A., Ziadat, A., Lintelmann, J., 2004. The quality of drinking water at source of west Amman, Jordan. Water Int. 29, 392–397. Kendall, C., Elliott, E.M., Wankel, S.D., 2007. Tracing anthropogenic inputs of nitrogen to ecosystems. Stable Isotopes in Ecology and Environmental Science 2, 375–449. Koba, K., Tokuchi, N., Wada, E., Nakajima, T., Iwatsubo, G., 1997. Intermittent denitrification: the application of a N-15 natural abundance method to a forested ecosystem. Geochim. Cosmochim. Acta 61, 5043–5050. Koh, D.C., Mayer, B., Lee, K.S., Ko, K.S., 2010. Land-use controls on sources and fate of nitrate in shallow groundwater of an agricultural area revealed by multiple environmental tracers. J. Contam. Hydrol. 118, 62–78. Kumar, M.S.M., Manohar, U., Pallavi, M.R.M., Anjana, G.R., 2013. Urban water supply and management. Journal of the Indian Institute of Science 93, 295–317. Lecuyer, C., Bodergat, A.M., Martineau, F., Fourel, F., Gurbuz, K., Nazik, A., 2012. Water sources, mixing and evaporation in the Akyatan Lagoon, Turkey. Estuar. Coast. Shelf Sci. 115, 200–209. Lehmann, M.F., Simona, M., Wyss, S., Blees, J., Frame, C.H., Niemann, H., Veronesi, M., Zopfi, J., 2015. Powering up the “biogeochemical engine”: the impact of exceptional ventilation of a deep meromictic lake on the lacustrine redox, nutrient, and methane balances. Front. Earth Sci. 3. Lerner, D.N., 2002. Identifying and quantifying urban recharge: a review. Hydrogeol. J. 10, 143–152. Lu, L., Cheng, H., Pu, X., Liu, X., Cheng, Q., 2015. Nitrate behaviors and source apportionment in an aquatic system from a watershed with intensive agricultural activities. Environmental Science: Processes and Impacts 17, 131–144. Margane, A., Subah, A., Hamdan, I., Hajali, Z., Almomani, T., 2010. Delineation of groundwater protection zones for the Wadi Shuayb springs. Technical Cooperation Project ‘Groundwater Resources Management’, Technical Report No. 14, Amman. McMahon, P.B., Bolke, J.K., 2006. Regional patterns in the isotopic composition of natural and anthropogenic nitrate in groundwater, high plains, USA. Environ. Sci. Technol. 40, 2965–2970. Mito, Y., Ismail, M.A.M., Yamamoto, T., 2011. Multidimensional scaling and inverse distance weighting transform for image processing of hydrogeological structure in rock mass. J. Hydrol. 411, 25–36. Moore, T.A., Xing, Y., Lazenby, B., Lynch, M.D., Schiff, S., Robertson, W.D., Timlin, R., Lanza, S., Ryan, M.C., Aravena, R., Fortin, D., Clark, I.D., Neufeld, J.D., 2011. Prevalence of anaerobic ammonium-oxidizing bacteria in contaminated groundwater. Environ. Sci. Technol. 45, 7217–7225.

Pasten-Zapata, E., Ledesma-Ruiz, R., Harter, T., Ramirez, A.I., Mahlknecht, J., 2014. Assessment of sources and fate of nitrate in shallow groundwater of an agricultural area by using a multi-tracer approach. Sci. Total Environ. 470-471, 855–864. Pedersen, J.K., Bjerg, P.L., Christensen, T.H., 1991. Correlation of nitrate profiles with groundwater and sediment characteristics in a shallow sandy aquifer. J. Hydrol. 124, 263–277. Riepl, D., 2012. Knowledge-based Decision Support for Integrated Water Resources Management With an Application for Wadi Shueib, Jordan. KIT Scientific Publishing. Robertson, W.D., Moore, T.A., Spoelstra, J., Li, L., Elgood, R.J., Clark, I.D., Schiff, S.L., Aravena, R., Neufeld, J.D., 2012. Natural attenuation of septic system nitrogen by anammox. Ground Water 50, 541–553. Rutsch, M., Rieckermann, J., Krebs, P., 2006. Quantification of sewer leakage: a review. Water Sci. Technol. 54, 135. Schirmer, M., Leschik, S., Musolff, A., 2013. Current research in urban hydrogeology – a review. Adv. Water Resour. 51, 280–291. Siebert, C., 2006. Saisonale chemische Variationen des See Genezareth, seiner Zuflüsse und deren Ursachen. Freie Universität Berlin, Germany. Sigman, D.M., Casciotti, K.L., Andreani, M., Barford, C., Galanter, M., Bohlke, J.K., 2001. A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 73, 4145–4153. Strauch, G., Möder, M., Wennrich, R., Osenbrück, K., Gläser, H.-R., Schladitz, T., Müller, C., Schirmer, K., Reinstorf, F., Schirmer, M., 2007. Indicators for assessing anthropogenic impact on urban surface and groundwater. J. Soils Sediments 8, 23–33. Thibodeau, B., Helie, J.F., Lehmann, M.F., 2013. Variations of the nitrate isotopic composition in the St. Lawrence River caused by seasonal changes in atmospheric nitrogen inputs. Biogeochemistry 115, 287–298. Umezawa, Y., Hosono, T., Onodera, S., Siringan, F., Buapeng, S., Delinom, R., Yoshimizu, C., Tayasu, I., Nagata, T., Taniguchi, M., 2008. Sources of nitrate and ammonium contamination in groundwater under developing Asian megacities. Sci. Total Environ. 404, 361–376. Vairavamoorthy, K., Gorantiwar, S.D., Pathirana, A., 2008. Managing urban water supplies in developing countries - climate change and water scarcity scenarios. Phys. Chem. Earth 33, 330–339. Vazquez-Sune, E., Carrera, J., Tubau, I., Sanchez-Vila, X., Soler, A., 2010. An approach to identify urban groundwater recharge. Hydrol. Earth Syst. Sci. 14, 2085–2097. Wakida, F.T., Lerner, D.N., 2005. Non-agricultural sources of groundwater nitrate: a review and case study. Water Res. 39, 3–16. Werz, H., Hoetzl, H., 2007. Groundwater risk intensity mapping in semi-arid regions using optical remote sensing data as an additional tool. Hydrogeol. J. 15, 1031–1049. Widory, D., Petelet-Giraud, E., Négrel, P., Ladouche, B., 2005. Tracking the sources of nitrate in groundwater using coupled nitrogen and boron isotopes: a synthesis. Environ. Sci. Technol. 39, 539–548. Wolf, L., Klinger, J., Held, I., Hotzl, H., 2006. Integrating groundwater into urban water management. Water Sci. Technol. 54, 395–403. Xue, D.M., Botte, J., De Baets, B., Accoe, F., Nestler, A., Taylor, P., Van Cleemput, O., Berglund, M., Boeckx, P., 2009. Present limitations and future prospects of stable isotope methods for nitrate source identification in surface- and groundwater. Water Res. 43, 1159–1170. Zagana, E., Obeidat, M., Kuells, C., Udluft, P., 2007. Chloride, hydrochemical and isotope methods of groundwater recharge estimation in eastern Mediterranean areas: a case study in Jordan. Hydrol. Process. 21, 2112–2123. Zemann, M., Wolf, L., Grimmeisen, F., Tiehm, A., Klinger, J., Hötzl, H., Goldscheider, N., 2015. Tracking changing X-ray contrast media application to an urban-influenced karst aquifer in the Wadi Shueib, Jordan. Environ. Pollut. 198, 133–143.

Please cite this article as: Grimmeisen, F., et al., Isotopic constraints on water source mixing, network leakage and contamination in an urban groundwater system, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.01.054