Title : Natural radioactivity content in groundwater of Mt . Etna ' s eastern flank and gamma background of surrounding rocks

Waters of Mt. Etna are the main source of drinking water for the local population and are also distributed in municipal supply systems to neighbouring areas. Radioactivity in underground waters and surrounding rocks from the eastern flank of Mt. Etna was investigated on the basis of 9 water and 8 rock samples from 12 localities altogether. Three samples were from water drainage galleries and six from water wells. All water intakes are used for consumption. Activity concentration of uranium isotopes 234,238U, radium isotopes 226,228Ra and radon 222Rn were determined with the use different nuclear spectrometry techniques. The determination of uranium isotopes was carried out with the use of alpha spectrometry. The measurements of radium and radon activity concentration in water were performed with the use of a liquid scintillation technique. Additionally, rocks surrounding the intakes were examined with gamma spectrometry. All water samples showed uranium concentration above Minimum Detectable Activity (MDA), with the highest total uranium (234U + 238U) activity concentration equal to 149.2±6 mBq/L. Conversely, all samples showed radium isotopes activity concentrations below MDA. Radon activity concentration was within the range from 2.91±0.36 to 21.21±1.10 Bq/L, hence these waters can be classified as low-radon waters. Gamma natural background of the rocks surrounding the water sampling sites was found on the same levels as other volcanic


Introduction
There is a growing concern about the health effects produced by natural radioactivity in the environment.Natural radionuclides are present in water, air, soil, plants and animals and in consequence in the human diet.In groundwater, their presence is determined by their activity concentration in soil and bedrock.Groundwater reacts with the surrounding rocks and re-leases elements which can be dissolved in it.Long-lived isotopes of uranium, radium and potassium are mainly responsible for natural radioactivity content.The most extensively studied natural radionuclide in all types of natural environments is radon since this element was found to be directly or indirectly responsible of about 2% of all death from cancer in Europe [Darby et al. 2005].In particular, indoor radon levels derived from groundwater use in the USA are thought to cause 1-7% of fatal lung cancers [Cothern et al. 1986].
Concerning an active volcanic environment such as Mt.Etna still few detailed studies exist on the levels of radionuclides other than radon in groundwater.Waters of Mt.Etna are the main source of drinking water for the local population and are also distributed in municipal supply systems to neighbouring areas.Moreover, they are used for both agricultural and industrial purposes.
A previous study on the determination of the activity concentration of 234,238 U, 226,228  Ra and 222 Rn in Etna's groundwater intakes was recently carried out [Kozłowska et al. 2009] showing values of uranium isotopes generally higher than the Minimum Detectable Activity (MDA), radium isotopes generally below MDA and low radon activity concentrations.The latter were in the same range as those found by D'Alessandro and Vita [2003], who performed an extensive survey in Mt.Etna's groundwater for the specific determination of dissolved radon contents.
In this work we present geochemical data on the activity concentration of uranium isotopes 234,238 U, radium isotopes 226,228  Ra and 222 Rn in collected un-derground waters from the aquifers of the eastern flank of Mt.Etna volcano.This area was selected as a good case study because it is characterised by large volumes of groundwater, strong interaction between groundwater, volcanic rocks, volcanic gases and some hydrothermal fluids [Aiuppa et al. 2000, Brusca et al. 2001, Parello et al. 2001, Aiuppa et al. 2004]; it is also the area of Mt.Etna with the highest values of dissolved radon [D'Alessandro and Vita 2003] and the only area where radium isotopes were detected [Kozłowska et al. 2009].All these facts stimulated the authors to perform detailed studies on natural radioactivity content in groundwater from this area.Moreover, gamma spectrometry measurements were performed on five samples from volcanic rocks surrounding water intakes and three samples from volcanic rocks taken close to summit craters, in order to investigate the natural background radioactivity of the area.

Sampling
Nine groundwater samples from aquifers of the eastern flank of Mt.Etna were collected for the purposes of this study.We chose sampling sites whose waters are continuously used for drinking purposes by the local population and whose sampling point was not too far from the water well/drainage gallery.There were other waters available for collection but they were either not used for drinking or their closest collection point was too far from the well/drainage gallery, so there was a high chance of losing radon by radioactive decay before collection.
Sampling was performed three times, in 2011, 2012 and 2013 year.Figure 1 presents the investigated area and the sampling sites.Three samples were taken from water drainage galleries and six from water wells.In the vicinity of five of these sites rocks samples were collected for the analysis of radioactive isotopes.Besides, three more rock samples were collected from the old-  Ferrara [1990]), delimited by thick blue lines, are also shown together with the three main rift zones (NE Rift, W Rift and S Rift, respectively).VB = Valle del Bove morphological depression.Volcano-tectonic data of Mt.Etna modified from Acocella and Neri [2003].
est lavas outcropping in the uppermost slopes of the volcano, thus representing the rock types present in the recharge areas of the aquifers that contain the groundwater collected downslope.CO 2 (pCO 2 ) measured in the corresponding sites during previous hydrogeochemical surveys [Giammanco et al. 1998, Parello et al. 2001, Giammanco et al. 2007a].
For radium and uranium determination the samples were collected in 2 and 0.5 volume polyethylene bottles, respectively.Samples were acidified to prevent the precipitation of iron as ferric hydroxide in the collecting containers.The initial chemical treatment for radium isotopes was performed at the Department of Physics and Astronomy of the University of Catania, Italy.Further treatment of the samples for radium and uranium determination was performed at the Department of Nuclear Physics and Its Applications, Institute of Physics, University of Silesia in Katowice, Poland.
For radon determination, samples of the volume of 10 ml were collected by disposable syringes directly at the intakes and immediately put into scintillation vials under 10 ml of scintillation cocktail [Suomela 1993a].The measurements were performed at the Department of Physics and Astronomy of the University of Catania.

Hydrogeology and hydrochemistry of the area
The investigated area is the eastern flank of Mt.Etna volcano (Figure 1).The volcanic edifice of Mt.Etna has grown up during the last 500,000 years on the eastern coast of Sicily (southern Italy) with the alternate superimposition of lava flows and pyroclastic deposits [Romano 1982, Chester et al. 1985].During the last 5000 years Etna has been almost continuously erupting and has emitted volcanic products essentially made of lavas, ranging in composition from alkali-basalt to hawaiite [Chester et al. 1985].The volcano built up its edifice over a sedimentary substratum with an estimated thickness greater than 15 km, according to Cristofolini et al. [1979].This basement is mostly composed of clays of Messinian age to the East and South and of clay, marl and quartz-arenite units, aging from Triassic to upper Pliocene, to the West and North.
In general, Etna presents a hydrogeological situation common to many other basaltic volcanoes: fissured and highly permeable lava layers are interbedded with discontinuous layers of scarcely permeable pyroclastics.These volcanic products are superimposed on an impermeable sedimentary substratum, with permeability coefficients ranging from 10 -7 to 10 -13 cm/s [Schilirò 1988].The base of the main aquifers of Etna thus lies at the contact between the volcanic rocks and these sediments [Ogniben 1966, Ferrara and Pappalardo 2008, Branca and Ferrara 2013].The volcanic rocks hosting the aquifer under study basically include most of the lithotypes found on Mt.Etna.They mainly belong to more or less differentiated products of the alkaline (sodic) series [Cristofolini and Romano 1982, Ro-mano 1982, Chester et al. 1985, Branca and Ferrara 2013].However, the concurrent presence of alkali olivine basalts ranging from "transitional" to sub-alkaline, belonging to the deepest basal lavas, is highly probable [Branca and Ferrara 2013].
Both of these magmatic series have a chemical composition ranging from basic to intermediate.The main mineralogical phases present in such products consist, in order of abundance, of plagioclase (~65%, with a variable composition), pyroxene (~28%, having an average augitic composition), olivine (~3%) and oxides (~3%) such as magnetite and sometimes titanomagnetite and ilmenite.Other minerals present to a lesser extent are apatite, alkali feldspars and in some cases nepheline, leucite and biotite [Cristofolini and Romano 1982, Romano 1982, Chester et al. 1985].
According to the model first proposed by Anzà et al. [1989], Etna's groundwaters are of meteoric origin, based on their Deuterium/Hydrogen and 18 Oxigen/ 16 Oxigen isotopic ratios.When rain waters enter the aquifer, they become enriched in CO 2 which, according to the d 13 C values already available in literature, is mostly of magmatic origin [Allard et al. 1997, Aiuppa et al. 2004].This process is ubiquitous on Mt.Etna, although it is enhanced in the most fractured and seismically active zones of the volcano (i.e. the SW and the E flanks) and it seems to mask any evolution of water chemistry from recharge to discharge areas [Aiuppa et al. 2004, and literature therein cited].
On the basis of structural, geological and geophysical data, three main hydrogeological basins were identified [Ogniben 1966, Ferrara 1990] (Figure 1).They roughly correspond to three sectors of the volcano: I) the eastern one, tributary of the Ionian Sea; II) the southwestern one, tributary of the Simeto River; III) the northern one, whose waters feed the Alcantara River.
The general chemical composition of Etna's groundwater ranges from bicarbonate alkaline-earth to bicarbonate alkaline [Anzà et al. 1989, Dongarrà et al. 1993, Giammanco et al. 1996, Aiuppa et al. 2004].As regards the major species in solution, the relative abundance is generally (Na + , Mg ++ )>Ca ++ >K + among the cations, with magnesium prevailing on sodium in the waters with higher salinity, whereas among the anions bicarbonate (HCO 3 − ) is always the prevailing dissolved species.Only very few samples show a chloride-sulphate alkaline-earth composition, normally due to the high concentrations of sulphate ions.
Values of Total Dissolved Solids (TDS) are normally well correlated with the corresponding values of calculated partial pressure of CO 2 (pCO 2 ) in equilibrium with water.This indicates a strong chemical interaction among meteoric water, the host rocks, from which the ions in solution were extracted by leaching, and a gaseous phase mainly composed of CO 2 .This gas is very abundant both in the ground waters and in the soils of Etna, besides its massive emission through the volcanic plume [Allard et al. 1991, Giammanco et al. 1995, D'Alessandro et al. 1997], and its origin is mostly magmatic [Anzà et al. 1989, Giammanco et al. 1995].Dongarrà et al. [1993] also showed that the concentration of the major species in solution and particularly the calculated values of partial pressure of CO 2 with which waters equilibrated can be subject to marked variations in time.Such variations were found to be caused both by seasonal effects and, more significantly, by changes in the volcanic activity of Etna.Some minor chemical species are brought into water directly from the rocks or from the volcanic gases that carry them as volatile species.This is the case of radon, that is present both in the volcanic rocks of Mt.Etna and in the volcanic gases that leak through the flanks of the volcano along tectonic faults [e.g., Neri et al. 2011].
The surface of the area investigated by our samplings was about 120 km 2 .The geological features of this area mainly consist of Etna's volcanic products.The volcanic rocks that outcrop in the study area are essentially represented by lava flows belonging to the so-called Stratovolcano Supersynthem, whose activity started from 57 ka [Branca et al. 2011], although some minor outcrops belong to the Timpe Supersynthem, that started some 220 ka [Branca et al. 2011].All of the samples of volcanic rocks collected belong to the former.

Water samples analysis
Since this project is a continuation of the research started in year 2009 the measurement techniques used for water samples studies were already described in the previous publication [Kozłowska et al. 2009].There were two nuclear spectrometry techniques involved: liquid scintillation counting and alpha spectrometry.Determination of radon activity concentration was performed at the Department of Physics and Astronomy of University of Catania with the use of Beckman LS6500 liquid scintillation counter.The radium content in water samples was measured with the use of the WinSpectral 1414 a/b liquid scintillation counter from Wallac at the Institute of Physics, University of Silesia, Katowice, Poland.The measurements of uranium 234,238 U isotopes were carried out with the use of alpha spectrometers 7401VR from Camberra-Packard also at the Institute of Physics in the University of Silesia.
Radon activity concentration was determined using a calibrated window for alpha counting.The Min-imum Detectable Activity (MDA) was calculated according to Currie's publication [Currie 1968] and was equal to 1 Bq/L at 1800 s counting time.Radium was co-precipitated with barium into Ba(Ra)SO 4 according to radiochemical procedure based on the Polish Norm [1989].The precipitate with radium was transferred into a scintillation vial, mixed with water and gellation scintillator.Theoretical calculations were based on Bateman's equations [Bateman 1910].MDA was equal to 10 mBq/L for 226 Ra and 20 mBq/L for 228 Ra at 3600 s counting time and 3 L sample volume.The separation of uranium from other radionuclides present in water was performed based on the procedure worked out by Suomela [1993b].The a-spectrometry source was prepared by uranium co-precipitation with NdF 3 and deposition on polypropylene discs (0.1 µm) according to the procedure worked out by Sill [1987].MDA was equal to 0.4 and 0.6 mBq/L for 234,238 U isotopes, respectively, for a 2 day counting time.

Volcanic rocks samples analysis
The studies of volcanic rocks samples were carried out both in the Department of Physics and Astronomy of University of Catania in Italy and in the Department of Nuclear Physics and Its Applications of the Institute of Physics at the University of Silesia in Poland.Apart from scientific purposes these studies served as intercomparison measurement in the two laboratories.Collected samples were transferred to the laboratory in the Department of Physics and Astronomy of University of Catania where they were characterized, dried, crushed and ground to a fraction smaller than 0.2 mm for homogenization.
The measurements in both laboratories were carried out in Marinelli geometry, in 0.6 dm 3 containers, where the rock samples were sealed for 30 days in order to allow secular equilibrium in the uranium and thorium series.The natural radioactivity of 226 Ra, 228 Ra( 228 Ac) and 40 K in collected samples were determined using HPGe gamma spectrometers produced by ORTEC Company.The spectra were analyzed with the use of GENNIE gamma acquisition program and GANNAS spectra analysis program and MAESTRO-32 software.The energy calibration and the detector efficiency curves were obtained on the basis of standards prepared from certified materials from the International Atomic Energy Agency from Vienna, Austria and Central Laboratory for Radiological Protection from Warsaw, Poland.The activity of 226 Ra was calculated as the weighted mean of the values obtained directly from the 186.2 keV line and the activities of 214 Pb (241.9, 295.2, 351.9 keV) and 214 Bi (609.3, 1120.3 keV) isotopes.The determination of the presence of 228 Ra and calculation of its activity was based on the following gamma-ray transitions: 212 Pb (238.6, 300.1 keV), 228 Ac (338.3, 911.1, 968.9 keV). 40K was calculated from one 1460 keV line.The detection limit for radium measurements was 1 Bq/kg and 10 Bq/kg for potassium 40 K.

Results and discussion
The results of the analyses of groundwater and volcanic rocks collected in the present survey for the detection of radionuclides are described in the following sections relatively to uranium isotopes, radium and radon and reported in Table 2 and corresponding figures.

Uranium
The activity concentrations for uranium isotopes varied from 1.5±0.5 mBq/L to 71.0±5.7 mBq/L for 238 U and from 2.0±0.6 to 78.6±5.7 mBq/L for 234 U. A strong positive correlation (Pearson's correlation coefficient r=+0.99) was observed between the concentration of 238 U and its daughter 234 U.Moreover, a highly significant positive correlation (r=+0.85) between the total uranium content in water and TDS values was also noted.The highest total uranium ( 234 U + 238 U) activity concentration was equal to 149.2±6 mBq/L, that is slightly higher than usual European values [Kozłowska et al. 2008, Walencik et al. 2012 andreferences herein].Moreover, comparing the present results with uranium data published for 17 brands of drinking waters collected in the whole Italy one can state that they are comparable to the values obtained by Jia et al. [2009].Our results confirm those of previous geochemical surveys of Mt.Etna groundwaters, underlining the important role both of uranium enrichment in water through rock leaching and of the redox state of water in controlling the dissolution of uranium [Aiuppa et al. 2000].This element is actually characterized by high mobility in the aqueous system under oxidized conditions, because it forms soluble oxo-hydroxo complexes in its higher oxidation state (VI).Among the most common complexes in equilibrium with Mt.Etna waters are the uranyl-carbonates [Aiuppa et al. 2000].This process would explain the highest uranium values found in this study.However, under more reduced conditions of waters, uranium can be scavenged by iron minerals and ferric oxyhydroxide, thus resulting to be depleted in those waters.This second process would explain the lowest uranium values found in our samples.
Interesting results were obtained for the 234 U/ 238 U activity ratios in the studied waters (Table 2, column 9), as they can be interpreted in terms of possible interference between groundwater and local precipitations (rain/melted snow) that easily percolate through highly permeable lava layers along major faults in the eastern flank of the volcano.The values varied in the range from 0.9±0.2 to 1.7±0.4,with values of arithmetic mean and median equal to 1.2 and 1.1, respectively.The results suggest achievement of secular equilibrium state for uranium isotopes (both arithmetic mean and median values are close to 1), rather than preferential leaching of 234 U (VI+) atoms from the mineral lattice of the host rocks, apparent for the disequilibrium state.
Our results confirm a previous study performed for the Mt.Etna aquifers, where the highest value of 234 U/ 238 U was equal to 1.4±0.1 [Kozłowska et al. 2009].As regards waters studied in the rest of Italy [Jia et al. 2009], the authors observed a slight disequilibrium between 234 U and 238 U activity concentration in most of the water samples, with mean isotopic activity ratios of 1.35±0.33.
In the literature, groundwater flowing through rocks characterized by heterogeneous structure (e.g.presence of several deep faults and fractures) and where different weather conditions may also play a role, show 234 U/ 238 U values up to 30 [Osmond et al. 1983].
Plotting the 234 U/ 238 U ratio versus the inverse of total uranium content 1/U (expressed as l/µg) it is possible to infer processes that occur in waters in time.Groundwater characterized by constant 234 U/ 238 U values but variable contents of reversed uranium (1/U) in time is likely affected by mixing with other water devoid of uranium, such as, for example, fresh rainwater.Figure 2 presents the relation between 234 U/ 238 U and 1/U for some of our studied waters (i.e., those for which analyses were repeated twice or three times).The plot indicates only minor mixing between the selected groundwater samples and shallow infiltration waters, probably rainwater or surface runoff produced from melting of snow during early spring.Moreover, the variation of 1/U with time seems to be particularly evident only for the waters sampled at higher altitude (i.e., samples E_5 and E_7), thus confirming such hypothesis.
Besides activity concentrations, Table 2 presents also the results of uranium mass expressed in [µg/L], since the consumption limits are given in these units.In 2011, the World Health Organization (WHO) proposed a limit for 238 U in groundwater equal to 30 µg/L (0.19 Bq/L), based on the chemical toxicity of uranium as recommended by EPA [2000].Our investigated waters are used for drinking purposes, so the calculation of the effective doses due to uranium absorption seems to be reasonable.Assuming a consumption of 2 L of water per day per person, based on the highest uranium activity con-centration measured in our samples one can calculate a yearly effective dose of 5.1 µSv, which is much below the proposed limit of 100 µSv/year [WHO 2011].

Radium and gamma background
All water samples analysed in this study showed activity concentrations of radium isotopes 226,228 Ra below MDA.This is in apparent contrast with the general rule that high TDS values, such as those found in the investigated groundwater, correspond to high radium activity concentrations [Martín Sánchez et al. 1999, Dávila Rangel et al. 2002, Kralik et al. 2003, Godoy and Godoy 2006, Kozłowska et al. 2008].On the other hand, activity concentrations of radium isotopes in other Italian waters [Jia et al. 2009] were also mostly below 10 mBq for both isotopes.Our previous hypothesis [Kozłowska et al. 2009] was that these results may point to a low radium content in the host rocks of Etna's aquifers.
We performed gamma spectrometry measurements on rock samples collected in the vicinity of the water intakes sampled in this study.Activity concentrations of 226 Ra varied from 45.5±2.2Bq/kg to 85.2±4.4Bq/kg, with average value of 61.8 Bq/kg.This value is almost identical to the median (61.4 Bq/kg), which confirms the normal distribution of the data.The activity concentration of 228 Ra( 228 Ac) varied from 28.8±0.9Bq/kg to 67.7±2.6 Bq/kg, with average 53.3Bq/kg and median 47.2 Bq/kg.The content of 40 K in the studied samples varied from 331±15 Bq/kg to 725±35 Bq/kg.Our measured values, together with average values from other Italian magmatic rocks taken from literature for comparison, are presented in Figure 3.
Our study showed that the activity concentrations of 226 Ra and 228 Ra ( 228 Ac) found in the rocks of Mt.Etna are on the same levels as the average European values found in bedrock.IARC [1988] comprehensive studies concluded that 226 Ra concentrations in soils span several orders of magnitude, but they are generally comprised between 10 and 50 Bq/kg, with an estimated average concentration of 25 Bq/kg.Similar values, from 10 to 100 Bq/kg, can be found in other publications [Cothern and Smith 1987, Nevissi and Bodansky 1987, Polański 1988].An average of 25.3 Bq/kg for 226 Ra and of 24.4 Bq/kg for 228 Ra can be found in soils of Poland [Radiological Atlas of Poland 2012].By comparison, the Agency for Toxic Substances and Disease Registry reported on its web site concentrations of 226 Ra in soils of northern Italy in the range 3-140 Bq/kg, excluding regions with extremely high levels of natural radioactivity.
Recently, Plant et al. [2003] stated that recent volcanic rocks such as those on Mt.Etna slopes, may be rich in uranium, which should result also in high radium contents.Chiozzi et al. [2001] published values of uranium levels in Lipari and Vulcano islands and Chiozzi et al. [2003] did the same for Stromboli, Salina, Filicuidi and Panarea islands.These data were based on gamma spectrometric measurements of 214 Bi line (1.76 MeV) or 208 Tl line (2.62 MeV) so these results are representative of 226 Ra or 228 Ra radioactive decay, rather than that of uranium or thorium, for which radioactive equilibrium had to be assumed.Converting the results of Chiozzi et al. [2001Chiozzi et al. [ , 2003] ] from ppm to Bq/kg, we obtain 226 Ra values from 32 to 240 Bq/kg for Lipari and Vulcano and from 14 to 117 Bq/kg in the other volcanic islands of the Aeolian arc.Following the same procedure, we obtain 228 Ra values in the range 37-268 Bq/kg for Lipari and Vulcano and 15-113 Bq/kg for the other Aeolian volcanoes.Brai et al. [2002], using gamma spectrometry, investigated rock samples from Stromboli island obtaining similar results as those of Chiozzi et al. [2003] ( 226 Ra values in the range 31-112 Bq/kg and 228 Ra (from 228 Ac line) values from 30 to 106 Bq/kg).It seems that the activity concentrations of 226,228 Ra obtained for Mt.Etna rock samples in our study are very similar to the average results obtained for Stromboli, but slightly lower than those of Lipari and Vulcano.Volcanic products from the latter islands are richer in a wide variety of chemical elements, mostly silica, and hence they may be richer also in radioactive elements from the uranium chain, thus justifying the observed differences with Mt.Etna.This hypothesis is also supported by the results of D'Alessandro and Vita [2003], who found a significant positive correlation between median radon content in volcanic groundwater from several active volcanic areas of southern Italy (Mt.Etna, Vulcano, Pantelleria, Ischia) and the median total uranium content in the respective local reservoir rocks, with uranium being in turn more enriched in silica-rich volcanic rocks.As regards the apparent similarities between Mt.Etna and Stromboli, despite the two volcanoes are quite different in terms both of chemical composition of eruptive products and of styles of activity, these could be explained by a similar behaviour of Ra en-richment in the erupted fraction of the magmas from the two volcanoes by processes of magma formation, differentiation and eruption, as modelled by Capaldi et al. [1976], thus indicating similar pre-eruption histories for the magmas at both volcanoes.
In any case, the low radium activity in the groundwater of Mt.Etna would be explained mostly by the low affinity of this element for the aqueous environment, as radium tends instead to remain in the host rocks of the aquifers [Walencik et al. 2010].Furthermore, as for uranium, the behaviour of radium in the aqueous environment is strongly controlled by redox conditions.Actually, under the typical Eh and pH conditions of groundwater of Mt.Etna, radium tends to form soluble Ra 2+ ions.However, in the presence of significant amounts of sulphate and/or carbonate ions in solution, insoluble complexes such as RaSO 4 or RaCO 3 are likely to form, thus removing radium from water.This process could be very efficient in the sampled waters, thus contributing to lower the contents of this element in solution.

Radon
The activity concentration of 222   observed within the presently analyzed waters cannot be simply explained by the different content of parent radionuclides in the Etna volcanic rocks, due to the rather homogeneous chemical composition of the latter, but rather they can be a result of different natural factors related to interaction between volcanic gases, water and volcanic rocks.The first is the enhanced water-rock interaction in the eastern flank of Etna that causes a strong release of major ions (described by TDS values) and hence of radon as a by-product of radium decay.This factor, however, seems to be effective only for some of the sampled waters.Figure 4 shows the correlation between values of dissolved radon and values of TDS in all of our samples.As evident from the plot, our samples show two distinct behaviors.In fact, samples E_1, E_5 and E_9 apparently show a much poorer correlation between radon activity in water and the corresponding TDS values than the rest of our samples.These three water samples come from the same aquifer, so the plot suggests an additional input of radon in the aquifer represented by these samples and therefore a radon source other than a simple dissolution of volcanic rocks.To study this, we plotted radon activity values versus the corresponding pCO 2 values in our water samples (Figure 5).Generally, the figure shows a correlation between the two parameters, indicating that radon activity in our sampled waters increases with increasing contents of dissolved CO 2 .This evidence, therefore, supports the idea that radon is carried by magmatic CO 2 into the aquifers of the eastern flank of Etna, thus enriching water in both of these chemical species, and rules out the possibility of different radon/CO 2 ratios between the two group of samples highlighted in the plot of Figure 4. Therefore, a possible explanation for the higher-than-normal radon content in samples E_1, E_5 and E_9 is to assume that radon in these waters derives, if not solely at least in large part, from direct and strong input of magmatic gas in the aquifer in the proximity of these sampling sites.Actually, sites E_1 and E_5 are located very close to the Fiandaca Fracture Zone [Azzaro et al. 2012, De Guidi et al. 2012] and site E_9 is located very close to the Santa Venerina Fracture Zone [Azzaro et al. 2012, De Guidi et al. 2012], both being very active seismogenic faults.Therefore, it is reasonable to infer a large, CO 2 -driven, radon release into local groundwater through active shallow rock fracturing along those major structural lines.
A specific discussion must also be made for site E_4, due to its peculiar characteristics.According to the above indications, the high amount of radon previously measured by other authors in the waters from this site ( 222 Rn=43 Bq/L) [D'Alessandro and Vita 2003] would stem both from enhanced dissolution of volcanic rocks and from strong direct input of CO 2 gas into the aquifer, the latter testified by the very high pCO 2 values calculated for this water.However, it is worth noting from Figure 5 that at site E_4 the radon activity value that we measured is much lower than that expected according to the observed general correlation with pCO 2 values.Actually, if we remove the point corresponding to this site, the general correlation is greatly improved (r=+0.91),thus suggesting some physical mechanism acting in the water from site E_4 that either leads to a removal of radon from solution or that hampers radon dissolution into the ground water at very high values of dissolved CO 2 .The most reasonable explanation for this change in behavior is that the influx of the carrier gas (CO 2 ) in the ground water becomes so high that it overwhelms the shallow source of radon and thus dilutes it, so the gas entering the aquifer and dissolving into it has a much lower radon/CO 2 ratio.Actually, a similar process of radon dilution from high CO 2 flux was already observed by Giammanco et al. [2007b] in soil and fumarole gas emissions from Mt. Etna and by Bonfanti et al. [1995] in soil gas samples from South East Sicily.In our case, the "dilution" effect would start at values of pCO 2 higher than 0.2 atm (Figure 5).
Apart from the fact that the Mt.Etna waters can be treated as low radon activity waters, the values of the activity ratios 222 Rn/ 238 U in waters (Table 2) varied in a wide range from 50±11 to 7856±1511.The obtained results indicated that 222 Rn isotope was much better transported than 238 U isotope.

Summary and conclusions
This paper presents the first full description of radionuclides ( 234 U, 238 U, 226 Ra, 228 Ra, 222 Rn) contents in groundwaters of the eastern flank of Mt.Etna volcano and of radioactivity background ( 226 Ra, 228 Ra) in sur-rounding rocks.We studied nine groundwater samples from different local aquifers and eight rock samples from twelve localities altogether.Three groundwater samples were taken from water drainage galleries and six from water wells.All water intakes are used for drinkable purposes.Rock samples were assumed to represent the host rocks of local aquifers.
The present results confirmed those of a previous study performed on Mt.Etna [Kozłowska et al. 2009], taking into account a large number of aquifers, but with a much lower spatial density of sampling point in each aquifer.Activity concentrations of uranium isotopes 234,238 U, radium isotopes 226,228 Ra and radon 222 Rn were determined with the use of different nuclear spectrometry techniques.Rocks surrounding the intakes were examined with gamma spectrometry.
Our new study showed that radionuclides levels in waters localized in the eastern flank of Mt.Etna are not different than the values obtained for the whole area of the volcano [Kozłowska et al. 2009].Radionuclides are weakly leached by groundwater from the surrounding reservoir rocks, especially 226,228 Ra that were at levels below MDA.The average 238 U and 234 U activity concentrations were 21.3 mBq/L and 22.7 mBq/L, respectively, whereas the average 234 U/ 238 U activity ratio was equal to 1.2, thus proving achievement of secular equilibrium for uranium isotopes.Furthermore, all waters may be classified as low-radon waters, with maximum activity concentrations of 19.2 ± 0.8 Bq/L.
Gamma spectrometry studies of the rock samples  collected in the vicinity of water intakes showed 226 Ra and 228 Ra mean values equal to 61.8 and 53.3 Bq/kg, respectively.These values are slightly lower than those of volcanic rocks from other active volcanoes in this geographic area.

Figure 1 .
Figure 1.Map of study area and sampling sites.Green squares represent water drainage galleries; blue open circles represent water wells; yellow open diamonds indicate the location of the three lava rock samples taken near the summit area of the volcano.The three main hydrogeological basins of the volcano (I = eastern basin; II = western basin; III = northern basin, according toFerrara [1990]), delimited by thick blue lines, are also shown together with the three main rift zones (NE Rift, W Rift and S Rift, respectively).VB = Valle del Bove morphological depression.Volcano-tectonic data of Mt.Etna modified fromAcocella and Neri [2003].

Figure 2 .
Figure 2. Graphical correlation between 234 U/ 238 U activity ratios and the inverse of uranium concentration values (1/U) in investigated waters from the eastern flank of Mt.Etna in three different years of sampling (data for 2007 survey are taken from Kozłowska et al. [2009]; data for 2011 and 2012 surveys are from this work).

Figure 3 .
Figure 3. Activity concentrations of 226 Ra (open circles) and 228 Ra( 228 Ac) (filled triangles) in [Bq/kg] in the rocks/soil samples surrounding the water intakes sampled in this work.Activity values were obtained from c-spectrometry analyses.For comparison, the range of the results for Stromboli is presented [Chiozzi et al. 2003], together with average values from other Italian magmatic rocks [ANPA 2000, ARPA Piemonte 2006].

Figure 4 .
Figure 4. Correlation between average radon activity concentration in solution and TDS values in the sampled waters of Mt.Etna.The bestfit lines highlight two populations of samples related to two different mechanisms of radon input into solution (see text for explanation).Values of R 2 correlation coefficient are also shown.
RADIOACTIVITY IN WATERS AND ROCKS OF MT.ETNA

Figure 5 .
Figure 5. Correlation between radon activity in solution and pCO 2 values in the sampled waters of Mt.Etna.The two lines show the main mechanism of radon enrichment and dilution produced by increasing input of volcanic CO 2 into the aquifers (see text for explanation).The value of the R 2 correlation coefficient for the best-fit line relevant to samples affected by radon enrichment is also shown.

Table 1 .
Table 1 presents a description of the intakes and the physical-chemical properties of the investigated waters, i.e. total dissolved solids (TDS), pH, electrical conductivity and temperature.The table also shows the values of partial pressure of dissolved RADIOACTIVITY IN WATERS AND ROCKS OF MT.ETNA Physical-chemical properties of investigated waters and samples description.NA=not available; MPS=municipal power supply; * pCO 2 values from Parello et al. [2001]; ** pCO 2 values from Giammanco et al. [1998]; § average pCO 2 value for this site, from Giammanco et al. [2007a].
Rn varied from 2.91±0.36Bq/L to 21.21±1.1 Bq/L.Obtained radon values are comparable to those of previous ground water surveys at Mt. Etna [D'Alessandro and Vita 2003, Kozłowska et al. 2009].The variations in radon activity RADIOACTIVITY IN WATERS AND ROCKS OF MT.ETNA