Interactive comment on “ Tracing groundwater salinization processes in coastal aquifers : a hydrogeochemical and isotopic approach in NaCl brackish waters of north-western Sardinia , Italy ”

As suggested by the Referee #1 we have calculated the correct sea water-rain water mixing line based on oxygen-18 and chloride. We are going to correct Figure N-1 (corresponding to our Figure 6) and to change accordingly the related comment in the revised version of the manuscript. The mixing line supports the assumption that chloride and oxygen-18 in water may derive by mixing of rain water and sea water. Further, in the sulfate vs chloride Figure N-2 (corresponding to our Figure 4) the reviewer observes that only few sulfate values match the model of sulfate coming from mixing of


Introduction
In the Mediterranean area the demand of good quality water is rapidly increasing and the processes of salinization (e.g.Petalas and Lambrakis, 2006;El Yaouti et al., 2009;Ghiglieri et al., 2012;Sdao et al., 2012) often threaten the challenge for exploitation of water additional resources such as groundwaters.Salinization of aquifers in coastal areas is the result of concomitant processes due to both marine water intrusion and rock-water interaction, which in some cases are hardly distinguishable.In Sardinia, the Nurra area, which is located in the northwestern part of the island, has a coastline that stretchs up to 80 km (Fig. 1), and salinization due to marine water intrusion has been recently evidenced as consequence of bore hole exploitation (Ghiglieri et al., 2012).The geology of the Nurra records a long history from Paleozoic to Quaternary, resulting in relative structural complexity and in a wide variety of lithologies, including Variscan low-grade metamorphic basement consisting of phyllites, quartzites, and metabasites, lower-middle Permian continental sediments and volcanites, middle Triassic to Cretaceous red beds, evaporites and shallow-marine carbonate, lower Miocene ignimbrites, alluvial deposits of Messinian age, and alluvial and eolian Quaternary deposits (Mameli et al., 2007;Mongelli et al., 2012).
In the Nurra, notwithstanding the importance of local groundwater as the main source of good quality water, exploitation has been uncontrolled and, due to intensive human activities and recent climatic changes, the area has become vulnerable Introduction

Conclusions References
Tables Figures

Back Close
Full to desertification (Ghiglieri et al., 2006).As a consequence, since the water demand is relevant and, similarly to other Mediterranean islands, surface-water resources can periodically suffer from drastic shortage (Ghiglieri et al., 2009).Chemical data available for the Nurra aquifers (Ghiglieri et al., 2009) show these groundwaters are affected by a large chemical variability, including, for instance, TDS values (from 600 up to 4000 mg L −1 ), chloride concentrations (from 3 up to 76 mg L −1 ), and sulphate concentrations (from 0.2 up to 40 mg L −1 ).This variability involves that various geochemical processes may affect the composition of the resource.Ghiglieri et al. (2009) suggested that the initial chemical composition of source water was conditioned by groundwater-rock interaction, including ion exchange with hydrothermal minerals and clays, incongruent solution of dolomite, and sulphate reduction.These statements, the relevance of the water resource and its role as strategic reserve in a climate evolving toward semiarid conditions, claim for a detailed study focusing on the processes determining the hydrogeochemistry of the Nurra groundwater and its quality, by tracking the sources of ions responsible for high salinity, aiming to assume the Nurra study case as a model for coastal aquifer hosted in Mesozoic carbonate-evaporite platform.In fact, the origin of the saline component of groundwaters is difficult to assess using only chemical data whereas coupling chemical and isotopic composition enhances our comprehension of the processes causing salinization of continental waters (e.g.Faye et al., 2005;Bouchaou et al., 2008;Gattacceca et al., 2009).With this in mind we report new data, regarding brackish waters of Na-Cl type of the Nurra, including major ions and selected trace elements (B, Br, I and Sr) and isotopic data, including δ 18 O, δD in water, and δ 34 S and δ 18 O in dissolved sulphate.To better depict the origin of the salinity we also analyzed a set of Nurra Triassic evaporites for mineralogical and isotopic composition.Introduction

Conclusions References
Tables Figures

Back Close
Full The structural framework of north-western Sardinia mainly derives from its Mesozoic and Tertiary tectonic evolution (Combes et al., 1993;Mameli et al., 2007).This has to be related to the Bedoulian Movements, to the Pyrenean Phase and to the North Apennine collision followed by the opening of the Ligure-Provencal back arc basin (Carmignani et al., 2004;Mameli et al., 2007;Oggiano et al., 2009).The cover rocks are affected by NE-SW-trending folds and thrusts; evaporites commonly occur as d écollement horizons and are exposed in the cores of anticlines and/or dec òllement surfaces.Since the Burdigalian, the area was subjected to an extensional tectonic related to the opening of the Liguro-Provencal Basin, followed by moderate uplift during the Pliocene.As a whole Nurra, consists of a structural high that represents the uplifted part of a wide block, tilted to the east.To the west, the Nurra area borders the eastern passive margin of the Liguro-Provencal backarc basin; to the east, it abuts the edge of a N-S trending Miocene half-graben, the Porto Torres Basin (Thomas and Gennessaux, 1986;Funedda et al., 2000).The Mesozoic and Cenozoic structural evolution of the region resulted in a thin-skinned tectonics with the Mesozoic cover represented by a sequence made up of limestone, dolostone and, at a lesser extent, marlstones and evaporites that deformed independently from the Palaeozoic basement, which outcrops in the westernmost part of the region: in fact, as a whole, the older rock sequences are progressively exposed westward.
The middle Triassic succession at Nurra, rests on red beds of Permian-Triassic age and consists mainly of pure dolostones and limestones, with clay-rich beds occurring within the Triassic deposits as marly limestones, and clayey gypsum deposits.Marls also occur in the early and Late Jurassic strata, the former associated with dark Liassic limestone with euxinic facies and the latter with typical lagoonal-lacustrine "Purbeckian" facies (Pecorini, 1969).The most of Jurassic succession consists of limestones and dolostones with a thickness exceeding 700 m.The Jurassic beds host the most relevant aquifer of the area (Ghiglieri et al., 2009).Lower Cretaceous is represented by Introduction

Conclusions References
Tables Figures

Back Close
Full pure Urgonian limestones, upper Cretaceous, lies unconformably on the Urgonian calcarenites along a bauxite level; it consists of Hippurites-bearing, limestones and marls of Late Cretaceous age (Coniacian to Maastrichtian).The whole Cretaceous beds have maximum thickness of about 400 m and host some perched aquifers out of the study area.In the study area the Mesozoic rocks are locally capped by Tertiary pyroclastic flows and by alluvial deposits of Messinian age, consisting of 30 to 80 m thick alluvial sequence mostly made up of clays and matrix-supported conglomerates.This deposit constitutes an important hydrogeologic unit for the north western part of the Nurra region since it acts as an aquitard that seals the confined aquifers hosted in the Mesozoic succession (Fig. 2).
The sampled area mostly pertains to the Porto Torres hydrogeological basin where the Jurassic aquifer has either reduced thickness in comparison to the Calich basin (Ghiglieri et al., 2009) or is absent toward the west.The groundwater flow in this basin is toward the northern shore (Asinara Gulf) whereas in the Calich basin is toward the south.The two hydrogeological systems are separated by a structural high toward which the axis of the main structures converge (B-B section in Fig. 2).
A detailed geological mapping of the area allow to recognize another structural high between Mount Zirra and Rocca della Bagassa which acts as geological watershed between the Calich basin and a small hydrogeological basin (Baratz Lake basin) flowing toward the western coast (Porto Ferro Gulf).The sampled waters pertain to the western part of the Porto Torres basin and to the Baratz Lake basin, within an aquifer hosted in Triassic carbonate rocks, cataclastic evaporite and sandstone (red beds).These saturated deposits are recharged by the Palaeozoic basement to the west and by the Jurassic carbonate hills to the east.

Sampling and analysis
Water samples from 19 springs and wells and two samples from the Baratz Lake were collected in September and October of 2011 in the coastal areas of the Nurra (Fig. 1).Introduction

Conclusions References
Tables Figures

Back Close
Full In addition a seawater sample was collected from a site 0.1 km away from the Porto Ferro coastline and a rainwater sample was collected in September 2011 nearby the Baratz Lake site.Many of the sampled springs and wells are sources for drinking and irrigation waters.We used a high-resolution multiparametric probe (Hach HQ 30d) to measure the pH, temperature, and electrical conductivity (E.C.) of each sample.
All water samples were filtered through 0.45 µm MF-Millipore membrane filters in the field and stored in high-density polyethylene bottles (50 and 100 m L −1 ).Prior to their use, these bottles were cleaned with nitric acid (HNO 3 ) and then rinsed with deionized water.The bottles were filled to the top with water, capped without leaving any head space, stored in a refrigerated container (∼ 4 • C) during transportation to the laboratory, and kept cool until analysis.At each sampling site, two water samples (for cation analyses) were collected and acidified with Suprapur ® HNO 3 (1 % v/v), after filtration, in order to prevent metal precipitation.For anion analysis, an un-acidified 100 m L −1 sample was collected.Alkalinity was determined in the field by titration with HCl (0.1 M).Cation concentrations (Ca, Mg, Na, K and Sr) were analyzed using Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP/OES) at the Activation laboratory of Actlabs (Canada) with precision better than ±5 %.Anion concentrations were determined for Cl, SO 4 , NO 3 , and Br using ion chromatography (Dionex CX-100), and minor element (I and B) were determined using an inductively coupled plasma-mass spectrometry (ICP/MS).Ionic balance was computed for each sample taking into account major species.All samples exhibited imbalances lower than 5 %.Several certified reference materials (NIST 1643e, NIST 1640E and SLRS-5) were processed and analyzed along with the samples to assess the accuracy of our method.Our concentration measurements for these certified reference materials agree with the certified values.
For oxygen isotopic analysis about 2 mL of each groundwater sample was equilibrated with CO 2 by shaking for 6 h at 25 • C (Epstein and Mayeda, 1953).For the hydrogen isotopic analysis, metallic zinc was used to produce hydrogen gas in the zinc reduction method (Coleman et al., 1982).Stable isotope ratios were measured on a dual inlet Finnigan Delta Plus IRMS with analytical precision of better than 0.2 ‰ for oxygen Introduction

Conclusions References
Tables Figures

Back Close
Full and 1 ‰ for hydrogen.Five water samples calibrated with respect to V-SMOW and GISP International Standards were used as working standards.For the sulfur isotopic analysis, dissolved SO 4 was precipitated as BaSO 4 by addition of BaCl 2 .The sample was then acidified to pH < 2 in order to dissolve any precipitated BaCO 3 .For δ 34 S analysis, SO 2 gas was prepared using the method of Yanagisawa and Sakai (1983).
The isotopic composition of sulfur was determined using Continuous Flow-Isotope Ratio Mass Spectrometry (CF-EA-IRMS) at Isotope Science Laboratory of University of Calgary (ISL-UofC).The analytical precision is 0.3 ‰ for δ 34 S-(SO 4 ) and 0.5 ‰ for ). Isotopic results were expressed as ‰ deviation (δ notation) relative to the international standards (V-SMOW for 18 O and 2 H, and V-CDT for 34 S and 18 O in dissolved SO 4 ; Gonfiantini et al., 1995).Finally, to properly evaluate the water-rock interaction processes a set of three evaporites was sampled and analyzed for mineralogical and isotopic composition.The mineralogy of bulk samples was obtained by X-ray powder diffraction (XRPD) using a Rigaku Rint 2200 diffractometer with CuKα radiation at 40 kV and 30 mA.

Mineralogical and isotopic features of the Nurra evaporites
Three samples of evaporites of upper Triassic age have been analyzed for mineralogy and isotopic composition (Table 2; GR, GG and GB samples in Fig. 1).The rocks have been collected at the transition from Muschelkalk carbonates to Keuper evaporites where the alternance of grayish, whitish, and reddish evaporite levels, from older to younger, occur.All the evaporites are composed by gypsum; in the case of the grayish level XRD analyses reveal the presence of halite and quartz as minor components marine evaporites of upper Triassic age from +10.9 to +18.3 ‰ (Krouse and Grinenko, 1991 and references therein).

Water chemistry
Temperature, pH, EC (25 • C) values and the chemical compositions of water samples are provided in Table 1.The pH values range between 6.2 and 8.5, with the exception of the Baratz Lake samples (hereafter LB1 and LB2) which have higher values (9.2 and 8.2).Ambient water temperature is between 16.5 and 20.9 • C with the exception of the sample SP4 which show a value of 26.7 • C. The electrical conductivity ranges from 1240 to 7046 µS cm −1 .The amount of total dissolved solids (TDS) is usually in the 1-20 g L −1 range, with the exceptions of the samples SP6 (TDS = 0.92 g L −1 ) and CS2 (TDS = 0.98 g L −1 ), and the water samples can be classified as brackish waters according to the classification of Drever (1997).The measured concentrations of major ions in the water samples are plotted on a Piper diagram (Fig. 3), which identifies the chemical compositions of the water samples as the Na-Cl type.In the anion concentration plot the samples are roughly distributed along the HCO 3 -Cl edge, between the rain water and seawater points and fall close to the Cl apex.
In the diagrams ions vs. chlorine (Fig. 4) both the data of the Na-Cl water samples and the rainwater-seawater mixing line (hereafter RSML) are plotted to evaluate possible seawater intrusion.The Ca/Cl, SO 4 /Cl, and Sr/Cl ratios in the water samples are much higher than expected on the basis of a simple mixing between rain water and seawater, thus suggesting that sulphate dissolution contributes to increasing the dissolved component.The Na/Cl is generally higher than that of the RMSL suggesting that dissolution of mineral phase(s) may contribute to add Na + to the solutions.The K/Cl is generally lower than that depicted by the RSML suggesting K + derives from silicate dissolution only.The variation in the boron contents are not correlated to the variation in the chlorine contents and the B/Cl ratio of the water samples is generally from lower to much lower than that of the RSML.Low values of the B/Cl ratio are associated to water-rock reaction, since Cl is preferentially leached with respect to B 1049 Introduction

Conclusions References
Tables Figures

Back Close
Full and since B is adsorbed on clays (Leybourne et al., 2007).Halogens are retained as particularly useful to investigate the features of the saline component of groundwater (e.g.Boschetti et al., 2011).In the iodine vs. chlorine binary diagram the distribution of the water samples is scattered and most of them, characterized by a high I/Cl ratio, depart significantly from the RSML.Only in the bromine vs. chlorine binary diagram the water samples fit the RSML thus supporting the hyphothesis of a seawater intrusion.However, it has been stressed that the Br/Cl systematics of groundwaters is complex and that the Br/Cl ratio may not be a useful discriminator of marine and non-marine sources of salinity (Leybourne et al., 2007).Finally, the NO 3 concentration of three water samples (PZ24 = 127 mg L −1 , PZ26 = 74 mg L −1 , CS5 = 91 mg L −1 ) exceeds the maximum admissible concentration of 50 mg L −1 defined under Italian law (D.L. 31/2001).This claims for future and more detailed studies concerning environmental aspects.More in general, the lack of any significant (and positive) correlation of NO 3 with Cl (r = −0.26)and SO 4 (r = −0.14)exclude a nitrate origin associated to the salinization processes.

Isotopic composition of water and dissolved sulphate
The results of δ 18 O, δD, and oxygen and sulphur isotopes of dissolved sulphate analyses are presented in Table 2. Isotopic compositions of water samples range from −6.6 to −2.1 ‰ for δ Most of the waters, in the δ 18 O vs. δD diagram, plot in a relatively tight cluster between the Regional Meteoric Water Line (RMWL, Chery, 1988;Celle et al., 2004) and the Global Meteoric Water Line (GMWL, Craig, 1961) suggesting they are meteoric in origin (Fig. 5).Lake water (LB1 and LB2 samples) and crop out waters (SP6, SP4, PZ18, PZ22) are enriched in the heavy O isotope forming a distinct subset (hereafter 1050 Introduction

Conclusions References
Tables Figures

Back Close
Full LCO waters).These samples fall on a line with a slope of 4.96, which is considerably lower than that of the RMWL (about 8).Such low slope can be produced by evaporation effects (Rozansky and Frohlich, 2001) or mixing of groundwater and seawater.However the hypothesis of any mixing between groundwater and seawater (either due to recent seawater intrusion or to addiction of interstitial waters) is clearly ruled out by the Clδ 18 O relationship (Fig. 6).The interstitial waters can be considered palaeofluids that had been trapped below the Miocene-Pleistocene sediments.In Fig. 6  O values together to chlorine contents show that there is no need to invoke an seawater source.This assumption is also supported by the lack of any correlation between distance from coastline and chlorine contents (Fig. 7).

The origin of salinity
As previously stated, both the elemental chemistry and the isotopic composition of groundwaters point toward an origin for the saline component not related to any seawater contribution.The concentration and isotopic composition of dissolved SO 4 in groundwater is related to both its source and mechanism of formation and the sulphur isotopes of dissolved sulphate can be used to identify the mixing of groundwaters of different origin (e.g.Schwarcz and Cortecci, 1974;Taylor and Wheeler, 1994;Ayora et al., 1995;Schulte et al., 1996).In addition, in an area where largely soluble salts, and especially gypsum, occur, the δ 18 O-SO 4 values can be an useful tool for identifying sources of sulfate since the oxygen exchange between sulfate and water is extremely slow at low temperatures (Chiba and Sakai, 1985).Therefore, the δ 18 O-SO 4 value usually remains unaltered after sulfate formation and hence reveals information about the formation process (Mayer et al., 1995;Van Donkelaar et al., 1995;Mitchell et al., 1998).Dual-isotope approach has been used with considerable success in both Introduction

Conclusions References
Tables Figures
In Fig. 8 (δ 34 S-SO 4 vs. δ 18 O-SO 4 binary diagram), all investigated samples, including the LCO subset, fall in the field of marine evaporites (Clark and Fritz, 1997).A few samples are characterized by isotopic values consistent with those of the Upper Triassic Nurra evaporites (from +14.4 to 15.4 ‰ for δ 34 S-SO 4 values and from +10.4 to +11.6 ‰ for δ 18 O-SO 4 values) while others samples show higher values.The enrichment in measured 18 O and 34 S can either be induced by fractionation due to bacterial SO 4 reduction (Clark and Fritz, 1997) or derive from an isotopic heavier source.The lack of H 2 S, the high Eh values, and the presence of dissolved O 2 (Ghiglieri et al., 2009 and references therein) exclude that these waters are affected by microbial SO 4 reduction.The isotopic heavier supply thus likely derives from the interaction with marine sediments of different age with respect to the upper Triassic (Keuper) Nurra evaporites (Fig. 8).Mass-balance curves between rainwater and evaporites of different Triassic age (Keuper, Muschelkalk, and Buntsandstein) support this hypothesis (Fig. 9) also in agreement with the occurrence of sediments of Triassic age in the investigated area.
Furthermore the presence of halite within the gypsum levels, as demonstrated by the XRD analysis, suggests that dissolution of evaporitic levels is responsible for the high chlorine abundances in the Nurra waters, also in agreement with the positive linear relationship existing between chlorine and sulphate contents (r = 0.60).In addition the saturation indexes for gypsum, anhydrite, halite and sylvite are below the unit (Table 2) suggesting the dissolution of soluble salts is an ongoing process.Finally this, in turn, involves that the salinization of the Nurra waters, in a climatic regime which evolves toward drier conditions, is a phenomenon that could be dramatically accentuated in the near future.Introduction

Conclusions References
Tables Figures

Back Close
Full

Summary
In the Nurra area occur brackish waters of the Na-Cl type composition which have Cl contents up to 2025 mg L −1 .The ratios between dissolved ions and chlorine, with the exception of the Br/Cl ratio, are not those expected on the basis of a simple mixing between rain water and seawater.
The δ 18 O and δD data indicate that most of the waters are within the Regional Meteoric Water Line and the Global Meteoric Water Line supporting the idea that they are meteoric in origin.Due to evaporation few waters are 18 O-enriched.
A relevant consequence of the meteoric origin of the Nurra Na-Cl type water is that the Br/Cl ratio, extensively used to assess the origin of salinity in fresh water, and that in the present case is compatible with a seawater-rainwater mixing, thus erroneously supporting the hypothesis of a marine intrusion, should be used with care also in near coastal carbonate aquifers.
A dual-isotope approach based on δ 34 S and δ 18 O in dissolved sulphate proved to be useful in assess the origin of salinity in the Nurra Na-Cl brackish water.All investigated samples have isotopic composition within the isotopic range of marine evaporites.A few samples are characterized by isotopic values consistent with those of the upper Triassic (Keuper) Nurra evaporites that, in this study, were analyzed for the first time for isotopic and mineralogical composition.Others samples have heavier isotopic composition consistent with interaction with the isotopic composition of sediments of older Triassic age (Muschelkalk and Buntsandstein) also occurring in the area.Overall, and consistent with the geology and the lithological features of the study area, δ 34 S and δ 18 O in dissolved sulphate suggest that water-rock interaction is the responsible for the Nurra Na-Cl brackish water composition.Evaporites dissolution also explain the high chlorine contents since halite has been detected in the gypsum levels.
Finally, the Nurra Na-Cl brackish water are undersaturated with respect to the more soluble salts involving, in a climate evolving toward semi-arid conditions, that the salinization process could dramatically intensify in the near future.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . The evaporites have δ 34 S-SO 4 values between +14.4 and +15.4 ‰ and δ 18 O-SO 4 values between +10.4 and +11.6 ‰ .These values are in the range of isotopic composition of Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 18 O and from −39 to −16 ‰ for δD.The seawater sample gave a value of +1.1 ‰ for δ 18 O and +5 ‰ for δD.Rainwater sample shows values of −5.5 ‰ for δ 18 O and −29 ‰ for δD.Sulfate in the investigated groundwater samples is characterized by positive δ 34 S and δ 18 O values ranging between +15 and +21.2 ‰ and between +9 and +14.1, respectively.

Fig. 1 .Fig. 3 .
Fig. 1.DTM-base geological map of investigated area.The localization of sampling sites and code of the analyzed water and rock sample are shown.The limit of hydrogeological basins are from Ghiglieri et al. (2009).The blue arrows indicate the direction of prevailing groundwater flow.See text for further details.
mass-balance theoretical curves between values of rainwater (RW) and modern seawater (SW -data from this work) and rainwater and interstitial waters (IW) trapped in Miocene sediments (data from B öttcher et al., 1999) have been calculated and plotted together with the measured isotopic data.All samples fall away from the theoretical curves.In short, the δ 18

Table 1 .
Location of sampling points (water and rock) and chemical composition of the investigated waters.