at SciVerse ScienceDirect

Journal of Environmental Radioactivity 108 (2012) 50e59

Contents lists available

Journal of Environmental Radioactivity

journal homepage: www.elsevier .com/locate / jenvrad

In-situ radionuclide characterization of a submarine groundwater discharge siteat Kalogria Bay, Stoupa, Greece

Christos Tsabaris a,*, Dionisis L. Patiris a, Aristomenis P. Karageorgis a, George Eleftheriou b,Vassilis P. Papadopoulos a, Dimitris Georgopoulos a, Evangelos Papathanassiou a, Pavel P. Povinec c

aHellenic Centre for Marine Research, Institute of Oceanography, 19013 Anavyssos Attica, P.O. 712, GreecebNational Technical University of Athens, Department of Applied Mathematic and Physical Science, 15780 Athens, GreececComenius University, Faculty of Mathematics, Physics and Informatics, Mlynska dolina F-1, SK-84248 Bratislava, Slovakia

a r t i c l e i n f o

Article history:Received 19 January 2011Received in revised form27 July 2011Accepted 5 August 2011Available online 8 September 2011

Keywords:Marine environmentSubmarine groundwater dischargeRadonThoronIn-situ gamma-ray spectrometryKalogria Bay

* Corresponding author. Tel.: þ30 22910 76410.E-mail address: tsabaris@hcmr.gr (C. Tsabaris).

0265-931X/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.jenvrad.2011.08.005

a b s t r a c t

In-situ underwater gamma-ray spectrometer KATERINAwas used for continuous measurements of radonprogenies (214Pb, 214Bi), thoron progeny (208Tl) and 40K in submarine groundwater discharge (SGD) sitesat Kalogria Bay, SW Peloponnesus (Greece). The spectrometer was deployed attached on measuringplatform along with two conductivity - temperature data loggers while underwater battery packssupplied the system for acquisition periods up to 25 days. The radionuclide time series together withsalinity data were obtained for spring (wet) and summer (dry) seasons. The 40K activity concentrationscorrelated well with salinity of the emanating groundwater. Although the 214Bi and 208Tl activitiesshowed usually similar trends anticorrelating with salinity, in some cases 208Tl did not follow the 214Birecord due to changes in the dynamics of the groundwater aquifer. As the half-life of 220Rn is very short(55.6 s), its concentration in SGD may depend on the distance from its origin to the monitoring point. Theobserved temporal variations of 214Bi and 208Tl confirmed advantages of continuous in-situ monitoring ofSGD in coastal areas.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The marine environment and especially near shore environ-ments may be enriched in radionuclides as a result of varioushuman activities and natural phenomena. The primary sources ofanthropogenic radionuclides are global fallout from past atmo-spheric nuclear weapons tests (Livingston and Povinec, 2002), theChernobyl accident, and discharges of liquid radioactive wastesfrom nuclear reprocessing facilities (Livingston and Povinec, 2000;Matishov and Matishov, 2004).

On the other hand, transport of radionuclides enriching thecoastal zone may also be due to natural phenomena. The mostimportant are discharges of large water masses by rivers, as well asby submarine groundwater springs. The input of radionuclides tocoastal regions by rivers can take place by wash out of natural andglobal fallout radionuclides from soil, as well as technologicallyenhanced natural radioactive materials within watershed areas, orby disposal of such materials across rivers. On the other hand,

All rights reserved.

groundwater is rich in natural radionuclides which enter into thewater body from the subsoil mainly by weathering.

The submarine groundwater discharge (SGD) is an alternativepathway for such materials to be transported into a coastal marineenvironment (Burnett et al., 2006). The concentrations of radio-nuclides are usually low, so there is no threat to the environment.However, many studies on SGD showed spatial and temporalvariations of several radionuclides which could be useful for betterunderstanding of coastal processes (Burnett et al., 2006). Amongthem radon and radium isotopes, and the primordial thorium anduranium are the commonest radio-tracers. Radon is considered anexcellent indicator for the presence and the intensity of ground-water springs in coastal areas, since it diffuses from the subsoil intothe body of groundwater reaching high activity levels, and can beeasily detected by various techniques. The most frequently usedmethods are based on sampling water at specified time intervals.Radon is degassed from the sample, and it is transported bya suitable gas to measuring chambers where particles (alpha andbeta) which are emitted due to radon and/or its progeny decays arecounted. Although all themeasuring process may take place on-siteby mobile detection systems (e.g RAD-7), the SGD is not reallysurveyed continuously as there are gaps between two consequent

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e59 51

sample analyses. Moreover, grabbing sediment samples requiredfor calibration is a high cost and time consuming procedure.

It has been recognized during the last years that the utilizationof in-situ gamma-ray detection systems would be a useful tech-nique in SGD studies, since such systems are capable for continuousmonitoring of radionuclide levels in water (Povinec et al., 1996,1997; Osvath and Povinec, 2001a; Tsabaris and Ballas, 2005;Tsabaris, 2008), or for mapping the distribution of radionuclidesin sediments (Osvath et al., 1999, 2001b). In-situ systems could befully-integrated, capable to stand many demanding environmentalconditions (depth stresses, chemical corrosion, etc.), and operatingsubmerge in the aquatic environment. Also, they offer a full auto-mation of the measuring process, fast and cost efficient determi-nation of radionuclide activities since there is no need for sampling,transport and chemical treatment (Povinec et al., 2001). Addition-ally, another option could be the remote transmission of data whenthe monitoring systems are installed on permanent stations farfrom the laboratories (Wedekind et al., 1999; Debauche, 2004;Osvath et al., 2005; Tsabaris et al., 2005; Tsabaris, 2008).

Recently in-situ gamma-ray continuous monitoring systemshave been successfully applied in SGD studies (Povinec et al.,2006a; Tsabaris et al., 2010) since groundwater is rich in radonand its decay products are usually gamma-ray emitters. The in-situgamma-ray spectrometry can be applied for simultaneous quanti-tative determination of all radionuclides present in the emanatinggroundwater (Tsabaris et al., 2008; Bagatelas et al., 2010).

The aim of this study is to present results from the application ofthe in-situ underwater gamma-ray spectrometer named KATERINA(the abbreviation comes from the Greek words Innovative Sensorfor Artificial and Natural Radioactivity) in the study of submarinegroundwater springs at the Kalogria Bay, located southwest of thePeloponnesus Island (Greece).

2. Materials and methods

The underwater gamma-ray spectrometer KATERINA wasdeployed in the Kalogria Bay for several periods from July 2009 tillMay 2010. The system operated autonomously without computerconnection, pre-programmed to acquire gamma-ray spectra every12 h. The spectra were buffered into an internal memory, whichwere later recovered from the system. External battery packswere used as a power supply providing operation time up to 25days. The gamma-ray spectrometer was accompanied with twoconductivity-temperature logger (CT) providing data to calculatesalinity values. The results are presented as time series in order toinvestigate radionuclide activity variations with salinity records ofthe emanating groundwater.

2.1. Study area

The Kalogria Bay is located north of Stoupa town, in the south-west Peloponnesus (Messinia Prefecture). The area is characterizedby the presence of numerous minor SGDs, visible as small gyres atthe sea surface (small circles in Fig. 1a). The major SGD site islocated w100 m offshore and creates at the sea surface twoimpressive gyres, with diameter varying between 25 and 60 m(Fig. 1b). The SGD emanates groundwater as a vertical jet froma depth of 25 m. The spring is partly covered by a limestone blockwith dimensions approximately 5 � 4 � 1 m. Divers observedsmaller jets emanating groundwater from the sides of the rockblock, from surrounding rock fissures, as well as three minor SGDsin a distance of a fewmeters from themajor SGD. During thewinterperiod, the discharge was so strong that divers could hardlyapproach the core of the SGD and place the measuring platformsafely on the sea bed.

2.2. The in-situ underwater gamma-ray spectrometer

The underwater gamma-ray spectrometer KATERINA is a fully-integrated unit since there is no need of computer connection toacquire and store the data. All the modules are enclosed insidea special housing made by acetal, which allows deployment depthsup to 400 m, and at the same time exhibits the lowest gamma-rayattenuation. The detection module is based on a 3” � 3” NaI(Tl)scintillation crystal in contactwith a photomultiplier tubewithbuild-in high voltage controller. The initial signal is processed consecutivelyby a preamplifier, an amplifier, and then by an analog to digitalconverter. Electronic drifts of the amplification gain, which areusually observed in field installations are auto-compensated bya dedicated circuit on the amplifier module. The signals then passthrough amulti-channel analyzer, and the obtaineddata are stored ina non-volatilememorymodulewithmicrocontroller (after a preset ofthe acquisition time). A power unit is included to distribute the inputvoltage (12e15 V DC) from an external source (underwater batterypack) to the electronic modules.

In order to use the spectrometer for the determination of theradon 222Rn daughters (214Bi and 214Pb), the thoron 220Rn daughter(208Tl) and 40K, the system was energy calibrated and tested for itsstability with respect to temperature variations and energy reso-lution. The calibrations were done inside a laboratory tank of 5.5m3

filled with freshwater. At the bottom of the tank, an electric pumpcirculated the water to insure homogeneity. The water was spikedfor calibrations with standard solutions of mono-energetic radio-nuclides 137Cs, 40K 99mTc and 111In. The whole procedure of thecalibration is out of the scope of this work and it is described indetails in a previous work (Tsabaris et al., 2008).

After the analysis of the spectra the measured counts under anyphotopeak are converted to activity data (in Bq/m3) by thefollowing equation (Eq. (1)):

r�Bq=m3

�¼ cps

3mIg(1)

where cps denotes the counts per second recorded for each pho-topeak, 3m is the efficiency for an effective volume of water, and Ig isthe emission probability. The efficiency of the system, em, waspreviously determined experimentally at 141, 661 and 1461 keV(Bagatelas et al., 2010). Monte Carlo simulations were also carriedout for calculating the efficiency of the system at any gamma-rayenergy (Bagatelas et al., 2010). The minimum detectable activity(MDA) of the system was studied in both freshwater and seawaterenvironments (Bagatelas et al., 2010). As concerns radon progeny214Bi at 609 keV, the MDAwas calculated to be 0.03 and 0.05 Bq L�1

in freshwater and seawater, respectively. The MDA of the systemdepends on the 40K concentration in the water. However, theexpected activity of 214Bi in groundwater is at least by one order ofmagnitude above the MDA of the system.

Moreover, the background of the system due to potential radondaughter’s accumulation on the housingmaterial (acetal) as a resultof previous deployments into radon rich environments was exam-ined. Gamma-ray spectra were acquired in several deployments ina reference lake where 222Rn activities were below the MDA of thesystem. The spectrum analysis exhibited that the gamma-ray linesfrom radon progenies were not present, implying that any back-ground activity from the radon progenies into the material of theenclosure is below the MDA of the system.

The analysis of the spectra was performed using the SPECTRWsoftware (Kalfas, 2010). The quantification for single gamma-lines(concerning mainly the radon/thoron progenies 214Pb, 214Bi and208Tl) was carried out following themethodology already described(Bagatelas et al., 2010). The 208Tl is a decay product of 224Ra with

Fig. 1. a) A map of the study area: small circles e minor SGD sites, two large circles indicates the springs of interest. b) The gyres of the studied springs (two large circles), asobserved on the sea surface.

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e5952

a transition probability of 35.4% according to the decay chainproperties of 232Th. So, the measured activity of 208Tl wasnormalized to the parent nuclide (224Ra) assuming radioactivityequilibrium of all radionuclides between 224Ra to 208Tl.

However, for reliable calculation of 40K activity, an additionalstep of analysis is required. This step minimizes the contribution ofeight, low intensity, gamma-lines of 214Bi (1377 keV with 4%,1385 keV with 0.77%, 1401 keV with 1.38%, 1408 keV with 2.48%,1509 keV with 2.2%, 1538 keV with 0.4%, 1543 with 0.35% and 1583with 0.72%), which interfere with the single line of 40K at 1461 keV.The interference of the eight 214Bi lines may take place in theenergy range of 1360e1570 keV which is 2.5 times the FWHM ofthe system at 1461 keV. The correction is realized according to Eq.(1) by the following steps:

(a) calculating the activity concentration of 214Bi from the analysisof the single photopeak lines at 609 and 1764 keV;

(b) calculating the counts per second (cps) which the measuredactivity of 214Bi (step a) contributes to the eight gamma-raylines interfering with 40K;

(c) subtracting the cps of those 214Bi gamma-ray lines from the cpsof the 40K photopeak. Finally, the result (in cps) from thesubtraction is converted to activity in Bq L�1 using the cor-rected cps and the standard methodology described inBagatelas et al. (2010).

2.3. Conductivityetemperature data logger

To measure the salinity of the emanating water, a con-ductivityetemperature data logger (SeaBird SBE 37-InductiveModem MicroCAT) was used. This instrument was attached onthe frame with the underwater gamma-ray spectrometer, and ithad been programmed to measure salinity/temperature every

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e59 53

30 min. The internal battery pack provides power supply to theinstrument for continuous operation for several months. After therecovery of the platform the data were collected via a PC. The aimof monitoring salinity was to investigate the quality of theemanating water and to correlate the salinity values with 40K data,as well as with other radionuclides present in groundwater.

3. Results

The underwater gamma-ray spectrometer KATERINA anda conductivityetemperature data logger were deployed six timesduring July 2009 to May 2010. The KATERINA provided an overallspectrum and a number of 12-h spectra according to the total timeof its operation. Two of the overall spectra after the acquisitionperiods of 9 summer days and 21 spring days are depicted inFig. 2a and Fig. 2b, respectively. These spectra correspond to thehighest and lowest radon progeny activities observed in the wholeperiod of the experimental work. The spectrum that was acquiredin July 2009 exhibits non intense gamma-peaks of radon progeniesdue to the low concentration of radon in groundwater. In Fig. 2athe contribution of 214Pb is observed only via the photopeak of351 keV gamma-line. On the contrary, the spectrum which wasacquired during April 2010 showed several intense photopeaks ofradon progenies (Fig. 2b). This should be due to higher levels of

Fig. 2. Acquired gamma-ray spectra on the SGD site: a) during adry period (JulyeAugust2009), and b) during a wet period (MarcheApril 2010).

radon in the SGD, since a contribution from scavenging of radondaughters from the air by rain (Povinec et al., 2008a) could beneglected. The activity of 214Pb was determined using three typicalenergy lines which were clearly observed in the spectrum, andtheir analysis verified the quantification procedure of the method,even at the low energy range (<351 keV).

The results obtained from the analysis of the overall spectraduring summer 2009 and spring 2010 are listed in Table 1a and b,respectively. For 40K activity calculations corrections due to theinterference of gamma-lines emitted from to 214Bi were taken intoaccount. During the dry period of July 2009 (Fig. 2a), the radon/thoron progenies and 40K gamma-lines were clearly observed atgamma-ray energies of 351 keV (214Pb), 609 keV (214Bi), 2614 keV(208Tl) and 1461 keV (40K). The average value of the radon progenieswas 0.56 � 0.06 Bq L�1, of the thoron progeny (208Tl) it was(0.15 � 0.01) Bq L�1, and in the case of 40K activity it was9.2 � 0.7 Bq L�1 (Table 1a).

After the wet period of April 2010 the radon/thoron progeniesand 40K were also observed at many gamma-ray energies (Fig. 2b)at 242, 295, 351 keV (for 214Pb), 609,1120,1764, 2220 keV (for 214Bi),2614 keV (for 208Tl) and 1461 keV (for 40K). Although the quanti-tative analysis was not performed for the other low intensegamma-lines (such as at 765, 934, 1377 and 2447 keV), these werealso clearly observed. The average value of the radon progenies was3.2 � 0.3 Bq L�1 and of the thoron progeny (208Tl) it was0.24 � 0.02 Bq L�1. The average values of 214Pb activity concentra-tion are in good agreement within their uncertainties compared tothe 214Bi activity concentration (see Table 1b). The activity of 40Kwas found equal to 0.91 � 0.05 Bq L�1 which was much lower thanvalues observed at the open sea (Povinec et al., 1996; Tsabaris andBallas, 2005). The low 40K activity indicates that the water massesexhibited low salinity due to the intense freshwater outflow.

The 214Bi, 40K, 208Tl activities, derived by the 12-h spectraanalysis are presented as time series of deployments in Fig. 3aef.Salinity data are added to graphs for interpreting possible intrusionof seawater into the SGD site, as well as to perform an inter-calibration exercise of the salinity calculation via the 40K activitymeasurements (Tsabaris and Ballas, 2005).

In the July 2009 deployment (22/7e1/8) the 40K (around10 Bq L�1) and 208Tl (around 0.14 Bq L�1) activities were constantwithin standard deviations (Fig. 3a). On the contrary, the activity of214Bi exhibits an inverse behavior with salinity, the highest value(0.65 Bq L�1) appears at the beginning of the period (23e24/7), andthe lowest one (0.42 Bq L�1) at the end (28/7e1/8). At the same

Table 1The results obtained from the analysis of the overall spectra during the a) July 2009deployment and b) April 2010 deployment.

Isotope EnergykeV

Intensity%

Marine efficiency�10�4 m3

ActivityBq L�1

UncertaintyBq L�1

a)214Pb 351 37 2.08 0.55 0.06214Bi 609 46 1.74 0.56 0.05214Bi 1764 19 1.21 0.58 0.04208Tl (normalized) 2614 100 1.06 0.15 0.0140K 1461 11 1.29 9.2 0.7b)214Pb 242 7.5 2.30 3.3 0.3214Pb 295 19 2.19 3.3 0.3214Pb 351 37 2.08 3.2 0.3214Bi 609 46 1.73 3.0 0.3214Bi 1764 19 1.21 3.2 0.2214Bi 1120 17 1.41 3.0 0.2214Bi (3-lines) 2200 6.5 1.13 3.0 0.2208Tl (normalized) 2614 100 1.06 0.24 0.0240K 1461 11 1.29 0.91 0.05

Fig. 3. Activity records of 214Bi, 40K and 208Tl (normalized) and salinity data as time series during a) July 2009 b) October 2009 c) January 2010 d) March 2010 e) April 2010 and f)May 2010.

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e5954

Fig. 3. (continued).

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e59 55

Fig. 3. (continued).

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e5956

time the salinity data exhibit variations throughout the period(from 30 to 34). The lowest salinity values are observed after 28/7.

In the October 2009 deployment (8/10e16/10) the radionuclidedata (Fig. 3b) exhibit a constant trend till 14/10, followed by a rapid

variation from 14/10 till the end of the measurement (16/10). The40K and 208Tl activities exhibit constant trends of about 6.5 and0.22 Bq L�1, and a rapid decrease to 1.05 and 0.05 Bq L�1, respec-tively. On the contrary, the activity of 214Bi exhibits a similar

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e59 57

constant trend (1 Bq L�1) till 14/10, and then a rapid increase up to1.75 Bq L�1. At the same time salinity exhibits similar behavior as40K and 208Tl, starting with values of around 12 and ending atalmost 1.

In the January 2010 deployment (20/1e6/2) the radionuclidedata exhibit a constant trend from 20/1 till 31/1 followed by a rapidvariation till 6/2 (Fig. 3c). The 214Bi and 208Tl activities reveal similarbehavior starting with constant levels of about 2.3 and 0.2 Bq L�1

up to 1/2, and afterward their activities rapidly increased to 3.5 and0.8 Bq L�1, respectively. On the contrary, the activity of 40K exhibitsa constant trend at the beginning (7 Bq L�1), while after 1/2 a rapiddecrease to 3.5 Bq L�1. At the same time salinity reveals similarbehavior as 40K, beginning with values at around 22, and reachinga minimum of 1 (on 3/2). At the end of the period it was increasingagain up to 10, however, this increase was not followed bydecreases in 214Bi and 208Tl levels (or the 40K increase), indicating atleast 4 days delay in the radionuclide levels after salinity changes.

During the March 2010 deployment (3/3e23/3), the activities ofall radionuclides exhibit an almost constant trend throughout themeasuring period (Fig. 3d). More specifically, 208Tl activity isaround 0.35 Bq L�1, the activity of 214Bi slightly decreases from 5.3to 4.0 Bq L�1, however the activity of 40K remains surprisingly high,around 6.5 Bq L�1. At the same time salinity reveals low valuesindicating the emanation of freshwater from the spring throughoutthe period. The high values of 40K observed during March 2010cannot be attributed to the salinity of the water. We noticed thatdue to an intense discharging of groundwater, the deploymentposition of the spectrometer did not remain stable. This wasconfirmed by small damages found on the platform and on thehousing of the KATERINA spectrometer, suggesting that the plat-form was moving, hitting occasionally adjacent rocks. As the posi-tion of the platform had to be far enough from the rocks in order toavoid contributions from radionuclides found in the rock bodies,during the deployment of March 2010 this condition was probablynot fulfilled. Therefore, the observed 40K activity may be attributedto the rock environment around the spring instead of the springitself. As a result the radionuclide data during the March deploy-ment are not considered reliable and they were not utilized forfurther analysis.

During the April 2010 deployment (23/3e11/4), classified as thewet period, the 214Bi levels were constant at 3.3 Bq L�1, and the 208Tllevels varied between 0.17 and 0.25 Bq L�1 (Fig. 3e). The low 40Klevels (a mean value of 0.8 Bq L�1) were in accord with low salin-ities (around 2), indicating a presence of groundwater.

Finally, during the deployment of May 2010 (12/5e2/6) the 40Kactivity exhibits a slight decrement with time (Fig. 3f) varying from7.7 till 5.9 Bq L�1. The activity of 214Bi varied around 1.7 Bq L�1,showingminima on 25/5 and 30/5, both around 1.4 Bq L�1. The 208Tlrecord is opposite to the 214Bi one, showing a clear maximum(0.36 Bq L�1) on 30/5. At the same time the salinity data exhibitweak variations throughout the period (from 10.2 to 12).

4. Discussion

As radon (222Rn) decays with a half-life of 3.83 days, it is a usefultracer for investigation of groundwatereseawater interactions(Burnett et al., 2006). The 222Rn is transported into groundwater viadiffusion through rocks fractures of the aquifer (originating inminerals with high 238U content). It could also be supported from226Ra (its half-life is 1620 y) of groundwater itself or from sedi-ments. As 226Ra activity in groundwater is expected to be very low(0.01e0.05 Bq L�1; Povinec et al., 2008b), the activity of 226Ra maybe considered negligible in comparison to measured 214Bi activity.This fact indicates that the diffusion from the aquifer is the majorenrichment mechanism for radon in groundwater. An inverse

relation between the 214Bi (222Rn) activity concentration andsalinity (Povinec et al., 2006b) was confirmed in this study (Fig. 3band c), indicating that radon could be an ideal tracer for studyinggroundwatereseawater interactions in coastal areas.

Thoron (220Rn) as a short-lived isotope of radon (its half-life isonly 55.6 s) may be suitable for investigation of groundwa-tereseawater interactions in areas which are not too far from thethoron origin. The 208Tl (220Rn), similarly as 222Rn, could beenriched in coastal waters via two mechanisms:

(i) Thoron (originating in minerals with high 232Th content) isdiffused to groundwater through rock fractures, and then it isreleased to coastal waters by SGD. If the SGD release rates arehigh (usually during wet seasons), the thoron concentrationsin coastal waters would be also high, opposite to watersalinities which would be close to groundwater values (e.g.Fig. 3e).

(ii) Thoron is supported from its parent radionuclide 224Ra (itshalf-life is 3.66 d) found in sediment or directly in seawater.During the dry seasons the SGD fluxes are usually low, so theseawater with high salinity (and 40K content) dominates incoastal waters. As the 224Ra activity concentration in coastalwaters influenced by SGD is decreasing from the coast to theopen sea (contrary to the 226Ra content which remainsconstant; Povinec et al., 2008b), a similar trend should be seenfor thoron as well. The observed 208Tl concentrations as pre-sented in Fig. 3a (0.14 Bq L�1) may represent such a case whenthe monitoring station during the dry season was under theinfluence of seawater with high salinity and 40K contents.

However, in most cases we have to deal in coastal areas withmixed (groundwater plus seawater) waters, as documented inFig. 3. Generally, using the radon and thoron concentrations wecould define corresponding end-members representing ground-water (3.5 and 0.8 Bq L�1, respectively) and seawater (0.6 and0.14 Bq L�1, respectively). An interesting case, which is outside ofthis parameterization is presented in Fig. 3b, where 208Tl activitieswere decreasing with salinity. The measured 208Tl levels(0.05 Bq L�1) are at least 10 times lower than expected. This couldbe either due to the fact that a source of thoron in groundwater hasbeen shifted to a longer distance from themonitoring station, or thevelocity of groundwater flow has decreased. In both cases thethoron would decay before it could reach the monitoring station.

Another interesting radionuclide record is presented in Fig. 3f.Although the salinity record shows a minimum on 18/5 (10.0), anda slightly increasing afterward, the 40K record is decreasing withtime (reaching a minimum on 30/5, 5 Bq L�1). The 214Bi recordshowsminima on 25/5 and 30/5, opposite to the 208Tl record, whichshows a clear maximum on 30/5 (0.35 Bq L�1). These recordsconfirm that we have to deal with very dynamic groundwa-tereseawater interactions, which would not be possible to observewithout continuous in-situ measurements.

Moreover, the KATERINA system offers indirect salinitymeasurements based on the activity of 40K which can be convertedto salinity units taking into account radioactive and chemicalparameters (Tsabaris and Ballas, 2005). The correlation between 40Kand salinity is depicted in Fig. 4a. The agreement between thesalinity measurements obtained by the two systems is consideredsatisfactory although the measurement processes are totallydifferent. The CT logger acquires salinity of water which is in directcontact with the sensor. On the contrary, the KATERINA may detectgamma-rays emitted from 40K into an effective spherical volumearound it (w6.2m3),with a radius equal to themaximumdistance ofgamma-rays at 1461 keV that may transverse before being totallyattenuated by seawater (Bagatelas et al., 2010). The KATERINA

Fig. 4. a) Comparison of salinity values obtained directly from conductivity andtemperature data of the CT data logger versus salinity values calculated using the 40Kconcentration; b) The relation of the radon progeny 214Bi activity with salinity datafrom the CT data logger.

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e5958

offers therefore spatially averaged measurements of salinity. Ingeneral, when strong flow of freshwater is emanating from thespring, both instruments record almost the same values of salinitysince the effective volume of the system contains mainly theemanating freshwater. In other cases, the salinity values slightlydiffer due to the different mixing mechanisms between ground-water and seawater into the effective volume of the system.

Additionally, the relation between radon progeny (214Bi) andsalinity was studied (see Fig. 4b). At low salinity values (0e5) theconcentration of radon is enhanced (3.5 Bq L�1) and it is stronglyaffected by slight salinity variations. On the contrary at highsalinities (25e35) the concentration of radon is low (w0.5 Bq L�1),and only slightly changes with salinity. This study strengthensthe use of radon as a tracer of SGD, as well as an indicator of mixingprocesses between groundwater and seawater in coastal regions.

5. Conclusions

The KATERINA system, deployed at the Kalogria Bay, locatedsouthwest of the Peloponnesus Island (Greece), has proved to be

a valuable instrument for investigation of temporal variations ofradionuclides in submarine groundwater discharge (SGD). Themain observations made in this study may be summarized asfollows:

(i) Three natural gamma-ray emitters (40K, 214Bi representing222Rn, and 208Tl representing 220Rn) were continuouslymonitored in SGD. The 40K activity concentrations correlatedwell with salinity of the emanating groundwater providinginformation about the quality of the spring water, as well asa possibility of overexploitation of groundwater by humans.

(ii) Themonitoring of 214Bi in SGDmay help to understandmixingprocesses between groundwater and seawater, while themonitoring of 208Tl may provide information on short-termprocesses in coastal aquifers manifested by changes in thecomposition and fluxes of SGD.

(iii) Although the 214Bi and 208Tl activities had usually similartrends, in some cases 208Tl did not follow 214Bi due to changesin the dynamics of the groundwater system. As the half-life of220Rn is very short (55.6 s), its concentration in SGD maydepend on the distance from its origin to themonitoring point.

(iv) The observed temporal variations of 214Bi and 208Tl activities inSGD confirm that we have to deal with very dynamicgroundwatereseawater interactions, which would not bepossible to observewithout continuous in-situ measurements.

(v) This study strengthens the use of radon isotopes as tracers ofSGD, as well as an indicator of mixing processes betweengroundwater and seawater in coastal regions.

Acknowledgments

The authors thank the divers of the HCMR underwater team(Mr. K. Katsaros, V. Stasinos and Th. Photopoulos), and the Captainof the vessel (D. Eksarchouleas) for assistance. The local authoritiesof Messinia Prefecture are acknowledged for financial support. DLPacknowledges a post doctoral scholarship in the field of Soil andWater-Sources Science granted by the State Scholarship Founda-tion of Greece (IKY).

References

Bagatelas, C., Tsabaris, C., Kokkoris, M., Papadopoulos, C.T., Vlastou, R., 2010.Determination of marine gamma activity and study of the minimum detectableactivity (MDA) in 4pi geometry based on Monte Carlo simulation. Environ-mental Monitoring and Assessment 165, 159e168.

Burnett, W.C., Aggarwal, P.K., Aureli, A., Bokuniewicz, H., Cable, J.E., Charette, M.A.,Kontar, E., Krupa, S., Kulkarni, K.M., Loveless, A., et al., 2006. Quantifyingsubmarine groundwater discharge in the coastal zone via multiple methods.Science of the Total Environment 367, 498e543.

Debauche, A., 2004. Continuous radioactivity monitoring systems. From the pre-history of radioprotection to the future of radioecology. Journal of Environ-mental Radioactivity 72, 103e108.

Kalfas, C.A., 2010. SPECTRW, A software code for gamma-ray analysis. PersonalCommunication.

Livingston, H.D., Povinec, P.P., 2000. Anthropogenic marine radioactivity. Ocean &Coastal Management 43, 689e712.

Livingston, H.D., Povinec, P.P., 2002. AMillennium perspective on the contribution ofglobal fallout radionuclides to Ocean Science. Health Physics 82 (5), 656e668.

Matishov, D.G., Matishov, G.G., 2004. Radioecology in Northern European Seas. ISSN1431e6250. Springer-Verlag, Berlin, ISBN 3-540-20197-1.

Osvath, I., Povinec, P.P., Huynh-Ngoc, L., Comanducci, J.-F., 1999. Underwater gammasurveys of Mururoa and Fangataufa lagoons. Science of the Total Environment237/238, 277e286.

Osvath, I., Povinec, P.P., 2001a. Seabed g-ray spectrometry: applications at IAEA-MEL. Journal of Environmental Radioactivity 53, 335e349.

Osvath, I., Povinec, P.P., Baxter, M.S., Huynh-Ngoc, L., 2001b. Mapping of thedistribution of 137Cs in Irish Sea sediments. Journal of Radioanalytical andNuclear Chemistry 248, 735e739.

Osvath, I., Povinec, P.P., Livingston, H.D., Ryan, T.P., Mulsow, S., Comanducci, J.F.,2005. Monitoring of radioactivity in NW Irish Sea water using a stationaryunderwater gamma-ray spectrometer with satellite data transmission. Journalof Radioanalytical and Nuclear Chemistry 263, 437e440.

C. Tsabaris et al. / Journal of Environmental Radioactivity 108 (2012) 50e59 59

Povinec, P.P., Osvath, I., Baxter, M.S., 1996. Underwater gamma-spectrometry withHPGe and NaI (Tl) detectors. Applied Radiation and Isotopes 47, 1127e1133.

Povinec, P.P., Osvath, I., Baxter, M.S., Harms, I., Huynh-Ngoc, L., Liong Wee Kwong, L.,Pettersson, H.B.L., 1997. IAEA-MEL’s contribution to the investigation of Kara Searadioactivity and radiological assessment.Marine Pollution Bulletin 35, 235e241.

Povinec, P.P., La Rosa, J., Lee, S.-H., Mulsow, S., Osvath, I., Wyse, E., 2001. Recent devel-opments in radiometric and mass spectrometry methods for marine radioactivitymeasurements. Journal of Radioanalytical and Nuclear Chemistry 248, 713e718.

Povinec, P., Aggarwal, P., Aureli, A., Burnett,W.C., Kontar, E.A., Kulkarni, K.M.,Moore,W.S.,Rajar, R., Taniguchi, M., Comanducci, J.-F., Cusimano, G., Dulaiova, H., Gatto, L.,Hauser, S., Levy-Palomo, I., Ozorovich, Y.R., Privitera, A.M.G., Schiavo, M.A., 2006a.Characterisation of submarine groundwater discharge offshore south-eastern Sicilye SGD Collaboration. Journal of Environmental Radioactivity 89, 81e101.

Povinec, P.P., Comanducci, J.-F., Levy-Palomo, I., Oregioni, B., 2006b. Monitoring ofsubmarine groundwater discharge along the Donnalucata coast in the south-eastern Sicily using underwater gamma-ray spectrometry of radon decayproducts. Continental Shelf Research 26, 874e884.

Povinec, P.P., Osvath, I., Comanducci, J.-F., 2008a. Underwater gamma-ray spec-trometry. In: Povinec, P.P. (Ed.), Analysis of Environmental Radionuclides.Elsevier, Amsterdam, pp. 449e480.

Povinec, P.P., Bokuniewicz, H., Burnett, W.C., Cable, J., Charette, M., Comanducci, J.-F.,Kontar, E.A., Moore, W.S., Oberdorfer, J.A., de Oliveira, J., Peterson, R., Stieglitz, T.,Taniguchi, M., 2008b. Isotope tracing of submarine groundwater dischargeoffshore Ubatuba, Brazil: results of the IAEAeUNESCO SGD project. Journal ofEnvironmental Radioactivity 99, 1596e1610.

Tsabaris, C., Ballas, D., 2005. On-line gamma ray spectrometry at open sea. AppliedRadiation Isotopes 62, 83e89.

Tsabaris, C., Bagatelas, C., Dakladas, Th., Papadopoulos, C.T., Vlastou, R., Chronis, G.T.,2008. An autonomous in situ detection system for radioactivity measurementsin the marine environment. Applied Radiation Isotopes 66, 1419e1426.

Tsabaris, C., 2008. Monitoring natural and artificial radioactivity enhancement inthe Aegean Sea using floating measuring systems. Applied Radiation Isotopes66, 1599e1603.

Tsabaris, C., Scholten, J., Karageorgis, A.P., Commanducci, J.-F., Georgopoulos, D.,Kwong L-, L.W., Papathanassiou, E., 2010. Application of an in situ underwatergamma spectrometer as a marine radon progeny monitor: continuous moni-toring of groundwater discharges into the coastal zone. Radiation ProtectionDosimetry 142, 273e281.

Wedekind, Ch., Schilling, G., Grüttmüller, M., Becker, K., 1999. Gamma-radiationmonitoring network at sea. Applied Radiation Isotopes 50, 733e741.