ARTICLE IN PRESS
Journal of Hydrology xxx (2009) xxx–xxx
Contents lists available at ScienceDirect
Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol
Natural versus anthropogenic sources in the surface- and groundwater dissolved load of the Dommel river (Meuse basin): Constraints by boron and strontium isotopes and gadolinium anomaly
Emmanuelle Petelet-Giraud a,*, Gerard Klaver b, Philippe Negrel a a BRGM, 3 Avenue C. Guillemin, BP6009, 45060 Orléans Cedex 2, France b DELTARES, Budapestlaan 4, Postbus 80015, 3508 TA Utrecht, The Netherlands article info summary
Available online xxxx The river Dommel, a tributary of the Meuse River, drains an area of intensive agriculture (livestock farm- ing, maize and grassland over 50% of the basin), and a dense population of about 600,000 people repre- Keywords: senting 20% of the total area. The combined human activities in the Dommel catchment lead to a large Sr isotopes amount of dissolved elements and compounds released in surface- and groundwaters. The aim of this B isotopes study was to discriminate the natural (including infiltration of Meuse water) versus anthropogenic Gd anomaly sources of the dissolved load, and to identify the various pollution sources such as agriculture, industrial Dommel basin activity, and wastewater treatment plants, using geochemical tools including major- and trace elements, Sr and B isotopes, and rare earth elements (REE). For that purpose, a same-day geochemical ‘‘Snapshot” picture of the entire basin was combined with monthly monitoring in strategic points. The major- and trace elements analyses allowed discriminating the main pollution sources affecting the basin, i.e. point versus diffuse sources. Strontium isotopes helped to identify each tributary and to calculate mixing proportions. Combining these calculations with the Sr- isotopic data obtained from the ‘‘Snapshot” sampling campaign during a low-flow period, shows that Meuse water infiltration represents 25% of the total Dommel discharge. Boron isotopes used for assessing the amount of water affected by anthropogenic input cannot discriminate between the two main anthro- pogenic inputs, i.e. urban wastewater and the zinc-smelter effluent, as they have similar d11B values. Finally, the REE, and especially the use of Gd anomalies (Gd*), demonstrated the generalized impact of urban wastewater on the streams of the Dommel Basin. The coupled use of different geochemical tracers (Sr and B isotopes together with Gd*) in addition to the standard major-element analyses, led to discriminating the various anthropogenic components influenc- ing the Dommel Basin water quality. With these tools it also became possible to assess the complex water circulation and exchanges between water compartments, including the major role of Meuse water through the Bocholt–Herenthals canal. Ó 2009 Elsevier B.V. All rights reserved.
Introduction The Dommel is a small river that has its source in north-eastern Belgium and runs through the southern part of the Netherlands Application of multiple geochemical tracers makes it possible to where it joins the river Meuse close to ‘s-Hertogenbosch (Fig. 1). differentiate between point and diffuse pollution sources influenc- The upstream water quality in the Dommel is strongly influenced ing river-water quality. An ideal tracer for distinguishing between by discharge from point- and diffuse sources. In all tributaries the different sources should be conservative and thus neither react flowing through densely populated areas, the water quality is af- with solids and other waterborne elements, nor be biodegradable. fected by industrial and urban waste-water inflow. In the upstream Its content in the (waste) water of point- and diffuse sources part of the Dommel, the wastewater treatment plants (WWTP) of should differ markedly from that in natural water to allow detec- Peer, Eksel, Overpelt and Lommel in Belgium discharge into the tion of low discharge water contents (Rabiet et al., 2005; Kulaksiz river. A suitable tracer, successfully used in previous wastewater and Bau, 2007). studies, is boron and its isotopes. Boron is used in detergents and boron concentrations in urban wastewater can be as high as 5mgL 1 (Lazarova et al., 2003); as it is not removed in traditional * Corresponding author. Tel.: +33 2 38 64 37 75; fax: +33 2 38 64 34 46. WWTPs it also appears to be conservative in rivers (Chetelat and E-mail address: [email protected] (E. Petelet-Giraud). Gaillardet, 2005). Boron isotopes are even more sensitive than
0022-1694/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2009.02.029
Please cite this article in press as: Petelet-Giraud, E., et al. Natural versus anthropogenic sources in the surface- and groundwater ... J. Hy- drol. (2009), doi:10.1016/j.jhydrol.2009.02.029 ARTICLE IN PRESS
2 E. Petelet-Giraud et al. / Journal of Hydrology xxx (2009) xxx–xxx
Fig. 1. Dommel catchment and simplified geological map. The letters A and B represent respectively the Feldbiss and the Peel Faults.
boron concentrations for tracing urban wastewater in surface some of these geochemical tracers will be used in combination water (Barth, 2000; Chetelat and Gaillardet, 2005). However, this with strontium isotopes as tracers for the discharge water of holds only when no other substances with the same isotopic signa- the zinc smelter. ture are discharged into the river and if the background value of In addition, since the 1970s agricultural activity has changed the surface water differs sufficiently from that of the urban waste- from dairy farming to intensive livestock farming; the manure of water (Barth, 2000). the latter is spread on land and leads to a large load of nitrogen, Another suitable tracer for wastewater is anthropogenic gado- phosphate and metals. Thus, the anthropogenic pollution in the linium (Gd). In water, the rare earth element (REE) signature is lar- Dommel and its tributaries not only derives from the point sources gely inherited from the rocks or sediments with which this water outlined above, but also from the diffuse input from historical interacts. In urban areas, their natural distribution is modified in industrial and recent agricultural activities. This diffuse input of relationship with anthropogenic influences and the REE signature anthropogenic substances in the surface water of the Dommel is can be used for tracing the provenance of water in rivers (Smedley, not constant, but varies depending upon the groundwater com- 1991; Sholkovitz et al., 1999; Elbaz-Poulichet et al., 2002). These partment discharging to the surface water (Rozemeijer and Broers, changes consist mainly in a pronounced positive gadolinium 2007). anomaly. Gd is used in the form of gadopentetic acid (Gd(DTPA)2 ) The geochemical and isotopic variations caused by both point as a contrast agent in magnetic-resonance imaging (MRI). This sources and variably diffuse input into the Dommel were regularly complex, which remains stable in water, displays excess concen- monitored from December 2004 to September 2006. The purpose trations above natural levels in Europe (Bau and Dulski, 1996; of this paper is to identify the abovementioned sources. In addition Kummerer and Helmers, 2000; Elbaz-Poulichet et al., 2002; Rabiet to the fixed long-term monitoring stations, a ‘‘snapshot” sampling et al., 2005; Kulaksiz and Bau, 2007). Effluents from wastewater (samples taken the same day all over the catchment) was done to plants will contain positive Gd anomalies and Gd is therefore a follow the geochemical tracers through the upper Dommel from potentially good tracer for WWTPs in urban areas where almost their point sources towards the downstream monitoring station. all households are connected to a WWTP. In this manner an ‘‘instantaneous view” of the behavior of the ap- Upstream in the Dommel, the most important industrial dis- plied tracers was obtained and could be used as input for interpret- charge is the Umicore zinc smelter in Overpelt (Belgium). Re- ing the monitoring results. The main goal of this study was to trace leased zinc and cadmium cannot be used as effective tracers as the transfer of the identified pollutants from their respective they adsorb onto particles in the river. The discharge water of sources through the hydrological compartments, together with a the smelter also contains exceptionally high concentrations of, better characterization of the functioning of the Dommel Basin, among others, Cl, Cs, Tl, Rb, K, Mo, Li and also B. As the behavior especially for the relations between surface water and of Cl, Cs, Tl, Rb K, Mo and Li in surface water is conservative, groundwater.
Please cite this article in press as: Petelet-Giraud, E., et al. Natural versus anthropogenic sources in the surface- and groundwater ... J. Hy- drol. (2009), doi:10.1016/j.jhydrol.2009.02.029 ARTICLE IN PRESS
E. Petelet-Giraud et al. / Journal of Hydrology xxx (2009) xxx–xxx 3
Sampling and analytical methods ment consists mainly of sand. Groundwater levels are usually within 1–3 m below surface. Water is pumped from the middle of the river at a depth of 10 Geologically, the upper Dommel catchment is divided into two cm below surface with a peristaltic pump (Eikelkamp 12.25.01). It parts. The south-western part is the Campine High and the rest of is then filtered through a 0.45 lm polyethersulfon filter in a the catchment lies in the Roer Valley Graben, which is bounded 142 mm diameter filtration unit (Eikelkamp, 12.31.01). For each by the Feldbiss and Peel faults (Schokker, 2003, Fig. 1). The Roer sample, four bottles are filled: a brown PE bottle for anions, a Valley Graben is covered by the Boxtel Formation, a sandy deposit 500 ml HDPE bottle for major cations and trace elements, a with small amounts of micas and feldspars, and by heterogeneous 100 ml HDPE bottle for Sr isotopes and a 1000 ml HDPE bottle loam and peat layers; maximum thickness 35 m). The cover for B isotopes. Each bottle was carefully rinsed with the filtered sands are underlain by the Sterksel Formation consisting of Pleisto- water before being completely filled. The samples for major cations cene sand and gravel deposits from the Meuse River. In the Cam- and trace elements and for Sr-isotope analyses were acidified with pine High, the Sterksel Formation is the top unit. In the Kempen ultrapure HNO3 to pH < or = 2. All samples were kept cool before region the sandy soils of the Boxtel and Sterksel formations are analysis. vulnerable to leaching due to the acidifying conditions caused by The pH and electrical conductivity (EC, standardized to 20 °C) an absence of carbonates, a small amount of clay minerals (<1%), were systematically measured on site, after calibration with stan- and low contents of organic matter (Wilkens and Loch, 1997). dard buffers, with a handheld WTW multi 340i meter equipped Agriculture occupies 56% of the Dommel catchment and con- with a Sentrix 81 WTW pH-electrode and a Tetra Con 325 WTW sists mainly of intensive livestock farming. Crops are mainly maize Ec-electrode. and grassland with lesser amounts of arable farming and tree cul- The water samples were analyzed by ICP-AES (Ca, Na, K, Mg; tivation. About 20% is urbanized and the remaining part is occu- uncertainty less than 10%) ion chromatography (Cl, SO4,NO3; pied by a dense network of nature reserves (Pieterse et al., 2003). uncertainty less than 10%) and ICP-MS (B, Cd, Co, Cs, K, Li, Ni, The mean annual rainfall in the Dommel area is approximately REE, Rb, Sr, Sb, Tl, Zn; uncertainty 10–15%) and potentiometric 740 mm/yr, the reference evapotranspiration for grassland is 560 2 method according to N EN ISO 9963-1(HCO3 ; CO3 , uncertainty mm/yr and the average annual runoff is 190 mm/yr. Groundwater 5%). seepage is estimated to contribute up to 70% to the annual runoff Chemical separation of Sr was done with an ion-exchange col- for the Dommel catchment (Pieterse et al., 2003; Rozemeijer and umn (Sr-Spec), with total blank <0.5 ng for the entire chemical pro- Broers, 2007). cedure. After chemical separation, 1/5th of the sample was loaded The flow chart of the basin is supplied in Fig. 2, the long-term onto a tungsten filament and analyzed with a Finnigan MAT average discharge figures are from Waterboard de Dommel 262 multiple collector mass spectrometer. The 87Sr/86Sr ratios (www.Dommel.nl) and the Vlaamse Milieu Maatschappij (VMM, were normalized to a 86Sr/88Sr ratio of 0.1194. An average internal www.VMM.be). The four WWTPs have a treatment capacity of precision of ±10 10 6 (2r) was obtained during this study. The 55,000 equivalent habitants. According to the Waterboard de Dom- reproducibility of 87Sr/86Sr ratio measurements was tested through mel, about 25% of the WWTP effluent water is household water, the duplicate analyses of the NBS 987 standard and the mean value rest being drainage and overflow water from urban areas. The con- was close to 0.710227 ± 17 10 6 (2r; n = 70). tribution of the WWTPs to the Dommel discharge at Neerpelt (1.3 Boron isotopes were determined on a Finnigan MAT 261 solid- m3 s 1) is over 30%. Inputs from the Lommel WWTP are about 1.6 source mass spectrometer after removing the major ions from the times higher than the discharge of the Eindergatloop stream. water sample using a cationic resin IR120, and boron separation Upstream from the confluence with the Keersop and Run streams, using Amberlite IRA-743 selective resin according to the procedure the Dommel also receives water from the Hageven nature reserve, of Mossadik, 1997. The boron sample was then loaded onto a single which is partly fed by seepage from the Bocholt–Herenthals canal Ta filament with mannitol and Cs, and the B isotopes determined ( 0.3 m3 s 1), the latter containing water from the Meuse River. þ by measuring the Cs2BO2 ion. The values are plotted on the d scale In addition to urban wastewater, the basin receives contami- nated waters from industrial and agricultural activities. The (expressed in ‰) relative to the NBS951 boric acidnh . standard, where anthropogenic impact of past and present zinc smelting activities the d value is defined as: d11B(‰)= 11B 10B sample can be considered firstly as diffuse source through past atmo- 11 10 3 11 10 B BÞstandardg 1: 10 . The B/ B value obtained for the spheric Zn and Cd emissions, now concentrated in the soil and NBS951 boric acid standard after oxygen correction was clearly higher than the natural background in the rivers, and sulfur 11 4.0431 ± 0.0023 (2rm, n = 31). The reproducibility of the d B dioxide (SO2) emissions. The zinc smelting can also be considered determination, based on replicate analysis, is ± 0.5‰ and internal as a point source, as the treated wastewater from the Umicore errors are often better than 0.3‰. smelter discharges into the Eindergatloop stream. These effluents The REE concentrations (Table 2) were directly measured by are the final products from the physico-chemical treatment of ICP-MS, without the generally used pre-concentration step to en- smelter wastewater through several pH adjustments (using hance REE concentrations (Hennebruder} et al., 2004). amongst others chalk hydrate) where the metals, fluorides and sul- fates are removed (OVAM, 2006). Recent improvements in the The study area treatment processes (OVAM, 2006) have led to a reduction of more than 90% for Cd and 60% for Zn in the Eindergatloop water in 2006– The Dommel catchment 2007, compared to the 2002–2005 period. Since the 1970s, agricultural activity changed from dairy farm- The Dommel catchment, 1.800 km2 large, is a riverine system in ing to intensive livestock farming, the manure of which leads to a northern Belgium (380 km2) and the southern part of the Nether- heavy load of nitrogen, phosphates, metals and sulfate, in addition lands. The Dommel discharges into the Meuse River downstream to the sulfate originating from industrial sulfur-dioxide emissions. of Eindhoven (Fig. 1). The main tributaries are the Eindergatloop, Wastewater of the urban areas is collected and treated in wastewa- Keersop and Run (Figs. 1 and 2). From Wouberg to Eindhoven, ter treatment plants in the upper Dommel, four of them (Peer, Ek- the Dommel streambed is approximately 2 m below the surface, sel, Overpelt and Lommel in Belgium) discharging their effluents in the flow velocities are 0.5–1.0 m s 1 (Waterboard de Dommel, streams. These effluents are major pollution point sources and also 2007) and because of these high flow velocities the bottom sedi- an appreciable water input in the hydrological balance.
Please cite this article in press as: Petelet-Giraud, E., et al. Natural versus anthropogenic sources in the surface- and groundwater ... J. Hy- drol. (2009), doi:10.1016/j.jhydrol.2009.02.029 ARTICLE IN PRESS
4 E. Petelet-Giraud et al. / Journal of Hydrology xxx (2009) xxx–xxx
Fig. 2. Flow chart with the relationships between sampling points, tributaries, Bocholt–Herenthals canal and point and diffuse sources. Long-term average discharges are taken from Waterboard de Dommel (www.Dommel.nl) and the Vlaamse Milieu Maatschappij (VMM, www.VMM.be).
Monitoring stations (K2FeO4) is used as an environmentally friendly oxidant. The high boron contents also suggest the use of borate. The sampling locations are schematically shown on Fig. 2. Sta- All major-element data (Table 1) are plotted in a Piper diagram tions 240013, 240013AV, 240013Run and 240013DBR were regu- (Fig. 3) representing the ion proportions. The cation diagram larly sampled from December 2004 to September 2006. Stations clearly distinguishes the Run and Keersop samples (surface and 241015, 241012, 240011, 240013Run, 240013AV and 240013 were groundwater) from the smelter effluents (mainly composed of also sampled for the ‘‘Snapshot” campaign. In addition to the Na + K). The Eindergatloop-Up, whose discharge mainly consists Waterboard de Dommel monitoring, additional monitoring sta- of the Lommel WWTP effluents, is also very concentrated in tions were selected around the confluence of the Dommel with Na + K. The Dommel-Up sample presents lower Ca and higher Na its tributary Run (Figs. 1 and 2) as the latter flows only through contents than the Run and Keersop, and is also partly fed by WWTP an agricultural area and is assumed to be representative of diffuse effluents. The Dommel River samples taken downstream of the source input in the upper Dommel catchment. Water was also ta- Eindergatloop plot between these two groups of samples, confirm- ken from an artificial pond (240013AV) isolated from the streams ing that it is a mixture between the Eindergatloop and the Dom- and fed only by groundwater and rainwater. mel-Up. In the anion diagram, the Run waters show the highest
SO4 proportion together with the lowest Cl + NO3 proportions. At Results and discussion the opposite, the smelter effluents have the lowest SO4 proportion together with the highest Cl + NO3 proportions. Thus, in the gen- Anthropogenic sources: diffuse and point sources eral diagram, the water from the Run and Keersop tributaries are of the CaSO4 type, mostly reflecting the diffuse atmospheric sulfur The monitoring during this study shows that the wastewater fallout from past and present smelting activities, agriculture, and from Umicore has high contents, among others, of B, Ca, K, Na, likely also from denitrification processes by pyrite oxidation (Roze-
Rb, SO4 and occasionally Sr. These elements are related to the meijer and Broers, 2007). The above-mentioned SO4 sources are chemicals used in effluents treatment, e.g. Ca from chalk water superimposed on each other in these agricultural catchments. On (OVAM, 2006) in the WWTP, or produced as a secondary product, top of the large amount of SO4 in these tributaries, chlorine and ni- e.g. SO4 (OVAM, 2006). Although it is unknown which chemical trates are also good indicators of agricultural activities with maize, products are used by Umicore for removing the metals, the high al- the only crop to which an unlimited amount of manure and fertil- kali contents of the wastewater suggests that potassium ferrate izers can be applied.
Please cite this article in press as: Petelet-Giraud, E., et al. Natural versus anthropogenic sources in the surface- and groundwater ... J. Hy- drol. (2009), doi:10.1016/j.jhydrol.2009.02.029 laect hsatcei rs s eee-iad . ta.Ntrlvru nhooei ore ntesrae n rudae . .Hy- J. ... groundwater and surface- the in sources anthropogenic versus Natural al. et E., doi:10.1016/j.jhydrol.2009.02.029 Petelet-Giraud, (2009), as: drol. press in article this cite Please Table 1 Major elements (mg L 1), 87 Sr/86 Sr and d11 B(‰) and associated Sr and B concentrations (lgL 1) of the main stations in the Dommel Basin monitored for this study. Complementary data used in this paper are available upon request from the authors. 1 1 1 1 1 1 1 1 11 87 86 1 1 Name/date Na (mg L ) Mg (mg L ) K (mg L ) Ca (mg L ) HCO 3 (mg L ) Cl (mg L )NO3 (mg L )SO4 (mg L ) d B(‰) Sr/ Sr B ( lgL )Sr(lgL ) Dommel river Dommel-Up 17/07/07 23.7 5.6 11.3 26.9 48.2 39.8 20.8 50.1 2.9 0.71137 43 98 Dommel-DBR 09/09/05 96.4 6.6 69.2 44.9 – – – – 1.6 0.71029 136 161 30/09/05 76.6 6.3 52.0 40.7 22.1 143.4 14.3 134.5 2.1 0.71036 123 152 25/11/05 71.5 6.8 45.0 42.3 74.5 116.8 14.1 115.5 2.8 0.71042 68 125 20/12/05 75.7 7.2 35.5 41.3 102.1 108.1 16.4 101.8 2.0 0.71053 57 126 23/01/06 60.6 6.9 34.9 41.7 72.6 93.5 15.2 104.8 2.6 0.71055 52 130 27/02/06 54.7 7.3 32.6 42.0 70.4 89.1 17.4 97.8 5.2 0.71058 55 134 06/04/06 53.4 7.6 29.2 44.3 71.9 82.4 15.3 91.5 4.6 0.71061 35 110 01/06/06 – – – – 73.1 81.6 16.5 93.2 7.1 0.71049 45 125 17/07/07 47.9 6.4 31.9 38.3 77.6 84.3 12.9 86.3 3.3 0.71050 55 127 Dommel-Down 20/07/05 54.8 4.8 44.9 32.6 0.0 136.1 12.6 139.0 1.2 0.71031 94 144 xxx–xxx (2009) xxx Hydrology of Journal / al. et Petelet-Giraud E. 09/09/05 91.5 6.9 65.4 43.5 21.8 172.2 12.2 152.0 2.6 0.71038 130 159 30/09/05 75.5 6.9 50.5 40.5 30.9 135.6 13.9 135.7 2.6 0.71048 120 155 25/11/05 65.6 7.0 40.3 40.7 72.0 105.3 14.1 112.1 3.9 0.71057 62 123 20/12/05 61.6 7.6 29.7 40.5 84.3 93.0 16.0 99.7 3.5 0.71068 49 130 PRESS IN ARTICLE 23/01/06 56.6 7.3 32.5 41.8 68.0 87.1 15.0 106.0 3.3 0.71066 38 133 27/02/06 50.0 7.6 30.1 41.7 67.4 82.0 16.5 98.8 6.9 0.71069 51 134 08/03/06 63.8 93.6 17.8 98.7 – – – 145 06/04/06 50.0 7.7 28.6 44.6 65.6 80.3 15.4 95.1 4.7 0.71072 30 106 03/05/06 54.0 6.9 23.3 37.9 74.5 70.5 13.6 95.3 2.7 0.71067 – 137 01/06/06 67.9 80.5 16.7 96.9 8.1 0.71062 44 102 17/07/07 48.0 6.8 30.5 38.6 73.6 79.8 12.3 87.5 – – 51 128
Tributaries Eindergatloop-Up 30/01/06 190.7 8.3 146.3 62.9 175.6 287.0 34.6 254.4 – – 110 156 17/07/07 106.6 6.0 90.4 26.6 171.2 111.6 13.1 155.0 2.6 0.70972 68 64 Eindergatloop-Down 30/01/06 395.4 7.9 338.8 113.8 158.6 698.6 31.2 472.3 – – 188 201 17/07/07 265.8 6.0 182.8 43.2 152.7 362.7 14.2 265.5 1.0 0.70954 100 90 Keersop 30/01/06 25.1 9.1 9.6 44.0 59.2 38.2 12.0 84.2 – – 15 162 17/07/07 20.2 7.7 7.2 42.8 88.2 34.4 7.7 65.1 13.9 0.71034 29 158 Run 20/07/05 17.8 9.7 7.6 31.2 17.9 29.2 1.4 104.7 24.3 0.71161 27 168 09/09/05 18.8 9.5 8.2 32.8 33.5 28.3 3.4 97.6 22.4 0.71162 35 170 30/09/05 15.0 10.0 8.5 33.0 5.1 28.4 5.7 112.9 27.3 0.71184 23 184 25/11/05 17.5 8.5 10.2 33.0 29.4 30.6 8.1 92.0 25.9 0.71156 18 135 20/12/05 19.2 10.9 14.2 42.6 48.8 30.2 15.8 110.5 24.7 0.71149 16 157 23/01/06 19.2 11.0 12.3 42.7 41.2 31.3 13.9 112.4 25.4 0.71150 15 161 27/02/06 18.4 10.6 12.6 40.9 38.7 32.0 17.7 112.4 25.4 0.71154 20 161 08/03/06 – – – – 39.3 32.6 16.1 111.5 – – – 164 06/04/06 20.1 11.3 13.3 43.4 37.9 31.3 15.2 108.2 25.9 0.71150 13 143 03/05/06 17.7 9.8 10.3 37.8 36.5 29.0 9.2 106.3 26.1 0.71144 – 154 01/06/06 – – – – 43.1 30.2 16.0 104.5 23.9 0.71147 23 146
Groundwater(near Run) 20/07/05 17.0 8.9 8.7 29.0 46.0 26.0 B.D. 79.2 22.9 0.71149 25 147 09/09/05 16.8 9.2 8.6 32.6 48.9 27.4 B.D. 84.6 22.8 0.71154 28 160 (continued on next page) 5 ARTICLE IN PRESS
6 E. Petelet-Giraud et al. / Journal of Hydrology xxx (2009) xxx–xxx
In an NO3 vs. Cl diagram (Fig. 4), the groundwater from the Run ) 1