at SciVerse ScienceDirect

Journal of South American Earth Sciences 46 (2013) 1e8

Contents lists available

Journal of South American Earth Sciences

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

Preliminary radon measurements at Villarrica volcano, Chile

C. Cigolini a,b,*, M. Laiolo a, D. Coppola a, G. Ulivieri c

aDipartimento di Scienze della Terra, University of Torino, Via Valperga Caluso 35, 10125 Torino, ItalybNatRisk, Centro Interdipartimentale sui Rischi Naturali (“NatRisk”), University of Torino, ItalycDipartimento di Scienze della Terra, University of Firenze, Via Giorgio La Pira 4, 50121 Firenze, Italy

a r t i c l e i n f o

Article history:Received 11 September 2012Accepted 17 April 2013

Keywords:Radon surveyVillarrica volcanoFault zonesDiffuse degassing

* Corresponding author. Dipartimento di Scienze deVia Valperga Caluso 35, 10125 Torino, Italy. Tel.: þ39 (6705128.

E-mail address: corrado.cigolini@unito.it (C. Cigoli

0895-9811/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jsames.2013.04.003

a b s t r a c t

We report data from a radon survey conducted at Villarrica volcano. Measurements have been obtainedat selected sites by E-PERM� electrets and two automatic stations utilizing DOSEman detectors (SARADGmbh). Mean values for Villarrica are 1600 (�1150) Bq/m3 are similar to values recorded at Cerro Negroand Arenal in Central America. Moderately higher emissions, at measurement sites, were recorded on theNNW sector of the volcano and the summit, ranging from 1800 to 2400 Bq/m3. These measurementsindicate that this area could potentially be a zone of flank weakness. In addition, the highest radonactivities, up to 4600 Bq/m3, were measured at a station located near the intersection of the Liquiñe-Ofqui Fault Zone with the Gastre Fault Zone.

To date, the Villarrica radon measurements reported here are, together with those collected at Galeras(Colombia), the sole radon data reported from South American volcanoes. This research may contributeto improving future geochemical monitoring and volcano surveillance.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The measurement of radon emissions from active volcanoesrepresents an additional tool to detect variations in volcanic activityand forecast eruptions. In-soil radon measurements at active vol-canoes can also provide useful information on the occurrence andthe magnitude of diffuse degassing along their flanks. Volcanicareas are generally affected by the release of gases along faults,fractures and fumaroles (cf. Chiodini et al., 1996; Giammanco et al.,2007). However, some volcanoes, generally characterized bypersistent out-gassing from active vents (i.e., open conduit vol-canoes), may show very low emissions at proximal and distal areas(Williams-Jones et al., 2000; Varley and Armienta, 2001), and gasesare essentially concentrated within the plume itself (due to thehigh permeability that develops at the conduitewallrock interfacewhen the magma is approaching the vent, cf. Cigolini et al., 1984).

Radon is a noble, chemically inert gas, constantly generated inrocks, soils and crustal materials. It is principally represented by theisotope 222Rn (with a half life of 3.82 days) and it easily enters therock pores and migrates to significant distances from the site oforigin before its decay. Measuring the variations of radon, induced

lla Terra, University of Torino,0)11 6705107; fax: þ39 (0)11

ni).

All rights reserved.

only by physical factors since it is not a reactive element, can pro-vide valuable information on volcanic degassing as well as on thedynamics of fluid transport processes (cf. Heiligmann et al., 1997;Trique et al., 1999; Cartagena et al., 2004).

Variations in radon concentration have been observed beforeand during the onset of regional seismic events with magnitude 4or higher (Scholtz et al., 1973; Fleischer and Mogro-Campero, 1985;Igarashi et al., 1995; Planici�c et al., 2004; Pulinets et al., 2009;Cigolini, 2010).

Chirkov (1975) was the first scientist to report an increasingtrend followed by an anomalous peak in radon concentrationprior to the 1971 eruption at Karimsky volcano, Kamchatka. Inlater years, radon anomalies related to changes in volcanic ac-tivity and the onset of volcanic eruptions, have been extensivelyreported (Cox, 1980, 1983; Thomas et al., 1986; Segovia andMena, 1999). In particular, the latter authors concentrated theirwork on four explosive American stratovolcanoes (El Chichónand Popocatépetl in Mexico, Poás in Costa Rica, and Cerro Negroin Nicaragua): they showed a positive correlation between theincrease in radon activity (related to the initial stages of volcaniceruptions) and the Volcanic Explosivity Index (VEI) of singleeruptions. According to their findings, the ratio: peak Rn-values/mean quiescence Rn-values, may be as high as 22.6 for eruptionswith VEI ¼ 5 (such as the eruption that occurred at El Chichón, onMarch 28, 1982). For minor eruptions observed at the citedCentral America Volcanoes (with VEI < 2), the above ratiosranged from 4 to 4.8.

Fig. 1. Location map of Villarrica volcano and surrounding area. Dotted white linesrepresent the main recognized regional structures: LOFZ e Liquiñe-Ofqui Fault Zone;GFZ e Gastre Fault Zone (modified after Volland et al., 2007; Lohmar et al., 2007). Insetshows the location of Villarrica volcano within the Southern Andean Volcanic Zone(SAVZ). The rectangle represents the limits of the area investigated in this work.

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e82

More recently, Cigolini et al. (2001) used a network for radonmonitoring at Mount Vesuvius to differentiate signals produced byregional earthquakes from those derived from the local volcanicseismicity. Burton et al. (2004), on the basis of radon measure-ments, were able to infer the geometry of a hidden fault at MountEtna during a period of marked flank seismicity (October 2002).Cigolini et al. (2007) detected earthquakeevolcano interactions atStromboli volcano: radon anomalies occurred with a time-delaywith respect to the onset of major regional seismic events. Radonanomalies regarded as precursors of volcanic eruptions have beenreported by several authors (e.g., Connors et al., 1996; Alparoneet al., 2005; Cigolini et al., 2005). However, degassing in volcanicareas (diffuse and/or concentrated) may be particularly efficientand its monitoring is crucial in volcano surveillance (e.g., Allardet al., 1991; Carapezza et al., 2004; Viveiros et al., 2008). In addi-tion, systematic radon measurements have been carried out atseveral active volcanoes in Central America (Varley and Armienta,2001; Williams-Jones et al., 2000) and Southern America(Heiligmann et al., 1997). Radon transport to the surface occursalong faults or fractures, and it is controlled by bulk porosity andpermeability. It essentially migrates by convection and advection ata larger scale, but at the site-scale diffusion may be considerablyeffective (cf. Dueñas et al., 1997). Moreover, the radon gas ispassively carried by water and carbon dioxide (e.g., Gauthier andCondomines, 1999). The role of environmental parameters hasbeen shown to be critical in modulating in-soil radon concentra-tions (e.g., Mogro-Campero and Fleischer, 1977; Pinault andBaubron, 1996; Zimmer and Erzinger, 2003; Pérez et al., 2007;Laiolo et al., 2012). Similarly, the effects of environmental param-eters on CO2 degassing have been reported by Viveiros et al. (2008)and Carapezza et al. (2008).

Automatic alpha particles detectors and real-time radon mea-surements considerably improve field surveys (cf. Siniscalchi et al.,2010) and monitoring strategies (cf. Neri et al., 2006; Cigolini et al.,2009). Nowadays, radon data can be automatically transferred andelaborated, enabling us to filter the effects of environmental pa-rameters on radon degassing (cf. Laiolo et al., 2012). This allows usto refine volcano surveillance and alert procedures.

The main purpose of this paper is to present the results of aradon survey at Villarrica volcano by using different methods andtechniques. We provide measurements on local emissions alsooutlining their spatial variations that seem to be related to specificstructural and volcanological conditions. In addition, we compareour results with soil radon measurements acquired at other Centraland Southern American volcanoes.

2. Villarrica volcano

Villarrica is an ice-capped composite volcano located in CentralChile (39.42�S, 71.95�W; 2847 m in altitude), near the town andlake of the same name (Fig. 1). The base area of the cone reaches110 km2whereas the glacier at its summit extends for about 30 km2

(Rivera et al., 2008). Climbing this volcano is challenging due tovariations in climate and the presence of several crevasses hiddenbelow the snow cover (Fig. 2). Villarrica volcano lies in the SouthernCentral Volcanic Zone (Lara and Clavero, 2004; Ortiz et al., 2003)and grows onto an NWeSE volcano-tectonic segment that includesthe edifices of Quetrupillán and Lanín (proceeding eastward towardthe structural axis of the orogen). This alignment runs parallel tothe Gastre Fault Zone (Bohm et al., 2002) that is, in turn, affectingbasement rocks and is located approximately 5 km North of Vil-larrica (Fig. 1). This fault zone displaces the Liquiñe-Ofqui FaultZone (LOFZ) that runs parallel to the structural axis of this portionof the orogen (Melnick et al., 2002). The early Villarrica edifice grewinside a minor caldera, approximately 2 km wide that was formed

3500 years ago within a larger caldera (w6 km in diameter) duringthe early Pleistocene (Moreno et al., 1994a; Clavero and Moreno,2004). Older lavas were essentially basalts to basaltic andesitesand their composition has not changed substantially over time.However, domes and tephra of dacitic composition have been re-ported in basal deposits related to themajor caldera (Moreno,1993;Moreno et al., 1994b; Clavero andMoreno,1994;Witter et al., 2004;Hickey-Vargas et al., 2004). Several parasitic cones and fissure ventsmay be observed along its flanks. Plinian ignimbrites and pyro-clastic flows were emplaced during the Holocene and may reacheddistances of about 20 km (Silva Parejas et al., 2010; Lohmar et al.,2007). Similarly, lava flows (up to 18 km in length) were eruptedfrom the summit vents and flank fissures and could easily reach thebase of the cone. Historical eruptions have been recorded since1558 and consist of strombolian-type eruptions with mild tomoderate explosivity occasionally accompanied by lava effusions.Lahars may generate during eruptive periods and could bedangerous: they caused more than 100 fatalities during the twen-tieth century alone (Naranjo and Moreno, 1991).

The regional tectonic setting has been described by López-Escobar et al. (1995), Lavenu and Cembrano (1999), and Ortizet al. (2003). The Central Southern Volcanic Zone (CSVZ) of theSouthern Andes runs NNE for approximately 1000 km along the so-called Liquiñe-Ofqui Fault Zone (with a dextral strike-slip motion)(Cembrano and Herve, 1993; Lavenu and Cembrano, 1994; López-Escobar and Moreno, 1994). Its northern sector is offset by theGastre Fault Zone, trending N60W and running parallel to the Vil-larrica-Quetrupillán-Lanín alignment (Moreno, 1974; Cembranoand Moreno, 1994; Ortiz et al., 2003). It is well known that mostof the eruptions of the Villarrica are triggered by regional earth-quakes located above fracture zone or to the north of it (Petit-Breuilh, 1994). This cause-effect link was first reported for severalChilean volcanoes (Casertano, 1963; Barrientos and Acevedo, 1992;Barrientos, 1994) but its occurrence has been recently investigatedin different volcano-tectonic domains (Hill et al., 2002; Cigoliniet al., 2007; Delle Donne et al., 2010).

Since 1558, several eruptions of Hawaiian, strombolian, and/orviolent strombolian type with explosivity index VEI < 3, have

Fig. 2. a) Degassing plume at Villarrica volcano (November 9, 2004); the volcano was covered with a thick snow cap; b) view of the intracrateric lava lake with a small lava fountain(the inside walls are 150 � 10 m in diameter (Photo: Valérie Vidal); c) location of the automatic radon summit station that was deployed on the crater terrace (VSS, see Table 1); d)detail of the automatic radon station with battery.

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e8 3

occurred at Villarrica. Although major explosive eruptions aremainly restricted to Pliocene and Holocene, a potential hazard isnot excluded since approximately 80,000 people live in the areasurrounding the volcano (within a 40 km radius) (cf. Lara, 2004).

Current activity is represented by fluctuations of the lava lakelevel inside the summit crater (Fig. 2b). This process is accompaniedby gas-induced lava puffing and persistent degassing that mayevolve into moderate strombolian explosions with the ejection ofscoria bombs. Recent work on current activity has been carried outby Palma et al. (2008) who integrated visual observations of thelava lake, analysis of the seismic tremor, together with COSPECmeasurements of SO2 flux. Seismic and infrasonic data (collected in2002, 2004 and 2009) were reported by Ripepe et al. (2010). BesideSO2 fluxes, geochemical data on Villarrica include FTIR measure-ments on plume gases (Shinohara and Witter, 2005). The atmo-spheric concentrations of acidic gases (SO2, HCl and HF) weremeasured by Witter and Delmelle (2004).

Fig. 3. Map of radon emissions at Villarrica for the data obtained by E-PERM� mea-surements (performed from November 11 to November 21, 2004) together with thelocation of the radon stations (squares). Isoemissive curves have been obtained byusing the IDW (Inverse Distance to a Power) approach. Dashed white lines refer to theposition of the Liquiñe-Ofqui Fault Zone and the Gastre Fault Zone (see Fig. 1). Thelocations of the automatic stations are also reported (dots).

3. Methods

In-soil radon measurements were performed at 13 samplingsites (Fig. 3), using E-PERM� electrets (Kotrappa et al., 1993). Sta-tions for E-PERM� measurements where named S1 to S13 (seeTable 1 and Fig. 3). In addition we deployed 2 automatic stationsutilizing DOSEman detectors (SARAD Gmbh). These are referred toas VSS (located at summit of Villarrica, 2840m a.s.l.) and VLS (at thebase of the cone, at 1380 m a.s.l.). Electrets were placed in tubes(1m long and 12 cm in diameter) isolated by a cap and inserted intothe soil to a depth of about 60 cm, to attenuate the effects ofenvironmental parameters (Cigolini et al., 2001, 2005). E-PERM�

detectors were exposed from 1 to 10 days, and will provide an in-tegrated measurement of radon activities. Average errors in mea-surements are estimated to be �7% (Gervino et al., 2004).

Radon concentrations measured by E-PERM� electrets wereobtained according to the following relationship (Kotrappa et al.,1990):

C ¼Vi � Vf

(1)

� �CF� t

where Vi and Vf are the initial and final voltages measured bymeans of RadElec (i.e. SPER-1E Electret Voltage Reader), and CF is acalibration factor which is a function of the “Midpoint ElectretVoltage” (i.e. (Vi þ Vf)/2) and t is the exposure time. We used LongTerm electrets positioned into “Ion Chambers” of 210 ml in volume.These detectors provide more accurate measurements since theyminimize interferences due to thoron and other isotopes of theradon progeny. Due to remarkable differences in altitude betweenlower and summit stations, we applied the correction reported byKotrappa and Stieff (1992)

CRn ¼ C þ�4:617� h

30;480

�(2)

Table 1Summary of the radon concentration measurements acquired at Villarrica volcanoduring November 2004. The sites from S1 to S13 refer to measurements effectedwith E-PERM� electretes VLS and VSS are acquired with automatic DOSEManradonmeter. See text for details.

Site Latitude Longitude Elevation(m)

Time of exposure(day)

222Rn conc.(Bq/m3)

S1 39.394 S 71.964 W 1381 6.8 794S2 39.410 S 71.969 W 1562 9.2 449S3 39.420 S 71.939 W 2854 8.0 320S4 39.420 S 71.941 W 2840 7.9 2450S5 39.409 S 71.944 W 2298 8.1 2274S6 39.379 S 71.946 W 1168 7.0 1867S7 39.388 S 71.938 W 1417 7.0 321S8 39.473 S 71.852 W 1154 10.1 2011S9 39.469 S 71.857 W 1171 0.8 1739S10 39.373 S 71.863 W 825 8.0 4629S11 39.472 S 71.897 W 1510 1.0 1630S12 39.493 S 71.902 W 1320 1.0 900S13 39.409 S 72.053 W 730 1.0 1460VLS 39.394 S 71.964 W 1381 6.8 175VSS 39.420 S 71.941 W 2840 7.9 680

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e84

where C is the radon concentration measured by means of Eq. (1)and h is the altitude in m.

As stated above, we installed two automatic monitoring stations(one at the base of the cone and one at the rim of the summit crater)by utilizing two DOSEman electronic radon dosimeters (producedby SARAD Gmbh). Radon diffuses through a leather membrane intoa cylindrical measurement chamber (12 cm3 in volume) where thecharged particles interact with a Si-doped semiconductor detectorand are counted by an automated alpha spectrometer. Counts arerecorded and processed by means of a multichannel analyzer thatsubdivides the counts into distinct energetic fields or Regions ofInterest (ROIs), and generates the spectrum of the radon gas. Thesensitivity of the detector is comprised between 10 Bq/m3 and4 millions of Bq/m3.

DOSEman detectors used at Villarrica volcano were singularlycalibrated in a “radon chamber” in order to measure particleswithin defined energy windows. In particular, VSS (at the summit)and VLS (at the base) stations measured within 3780e8925 keVand 4095e9240 keV, respectively (see Fig. 4). These energy win-dows include the peaks of 222Rn and its progeny (218Po, 214Po) aswell as 220Rn, thoron (due to the decay of the 232Th chain). Gründeland Postendörfer (2003; p. 290) showed that the counts for 214Pomust be corrected since the higher side of the 220Rn spectrum

Fig. 4. Spectra of the total counts at VSS station that was measured from 11 to 17November 2004. Counts were subdivided in Regions of Interest (ROIs). For computingradon concentrations the counts of ROI1 and ROI2 (222Rn and 218Po, respectively) wereconsidered.

overlaps the 214Po peak. They suggest that 7.5% of the counts of thelatter are ascribed to thoron.

Radon concentrations in Bq/m3 were computed from the totalcounts of single ROIs (within a given sample-time) by means of thefollowing relationship (Gründel and Postendörfer, 2003):

CihBq=m3

i¼

�Cfi=Cts� ð1=tsÞ

�� 1000 (3)

where Cfi is the instrumental calibration factor (related to the vol-ume of the chamber) Cts are the counts, ts is the sample time (inminutes) and 1000 is the conversion factor from kBq/m3 to Bq/m3.Radon can be obtained in fast mode merely counting 222Rn and218Po, or slow mode by also including the counts of 214Po. Our datawere obtained according to the “fast mode” option since 214Po mayaggregate with moisture particles, and thoron interferences (on the214Po peaks) do not have to be dealt with. The instrumental un-certainties for radon concentrations of 1000 Bq/m3 were observedto be �25%, drastically decreasing at higher emissions (Streil et al.,2002).

4. Radon survey

Although Villarrica Volcano was covered with snow, we decidedto perform the measurements due to the paucity of radon data onSouthern American volcanoes. Sensors were exposed fromNovember 8 to November 21, 2004. The results were stored in ourdata bank that includes the radon data collected on Vesuvius,Stromboli and La Soufriere volcanoes. The results are quite signif-icant essentially for two reasons: first, to our knowledge, no dataare available on radon emissions on Villarrica; secondly they referto periods when the activity of the volcano was increasing. How-ever, due to the difficulties in reaching the summit, and theextended surface of the cone, we could not survey the volcano ingreater detail during our survey.

We recall that the network consisted of 13 measurements siteswithin a 10 km radius from the summit crater. The network wasdeployed taking into account both morphological and structuralfeatures. Measurement sites at the base of the volcano were alsoselected relatively close to roads or tracks. We avoided placing thestations in the snow or within the ice cap, since these conditionsmay reduce radon release (such as at station S3; cf. Conen andRobertson, 2002). Therefore, the measurement obtained at thisstation, i.e., 320 Bq/m3 was not included in the map of Fig. 3. Allother stations were inserted into the outcropping soil or, alterna-tively, reached by digging a hole into the snow cover. In Table 1 wereport the data obtained by means of E-PERM� electrets: mea-surements show values comprised between 300 and 4600 Bq/m3.These emissions fit well with those detected at other Central andSouth American volcanoes and volcanic areas (cf. Table 2 and Fig. 5).It is significant to mention that, prior to this contribution, Galera’sdata were the sole radon measurements reported in literature onSouth American volcanoes. The mean value for Villarrica is 1600(�1150) Bq/m3, which is similar to the background values recordedat Stromboli at distal measurement sites (Cigolini et al., 2009). Inaddition, they are in close agreement with those reported for CerroNegro and Arenal in Central America (Connors et al., 1996;Williams-Jones, 2000).

The data acquired did not show any particular correlation withaltitude, or with the proximity of measurements sites to the sum-mit crater area. We projected the data onto a DEM image to decodethe possible relation between the structural framework and radonrelease. This approach is very useful since it provides a direct viewon the geometry of degassing in space and time (cf. Buttafuocoet al., 2007; Carapezza et al., 2008; Cigolini et al., 2009). In Fig. 3

Table 2In-soil radon concentration values measured in the main Central and South American active volcanoes; by using different techniques.

Volcano Description Period Mean Min Max Methods References

Galeras (Colombia) Network e 23measurement sites

1993e19994 2 1 50 E-Perm� Heiligmann et al., 1997

Popocatépetl (Mexico) Network e 667measurement sites

1997e1999 4.65 0.1 35 Alpha-Scintillation e PYLON AB-5 Varley and Armienta, 2001

Villarrica (Chile) Network e 13measurement sites

Nov 2004 1.6 0.33 4.6 E-Perm� This publication

El Chichon (Mexico) Monitoring Network 1982 0.6 0.02 13.5 LR115 e Track-Etches Film Segovia and Mena, 1999Poás (Costa Rica) 7 automatic stations 1982e1992 6 0.02 24 LR115 e Track-Etches Film Segovia and Mena, 1999Cerro Negro (Nicaragua) Network e 29

measurement sites1994e1996 0.35 0.03 26.7 E-Perm� Connors et al., 1996

Irazu (Costa Rica) Three automatic stations Nov 1994 4 0.1 35 Barasol (ALGADE) Seidel et al., 1999Arenal (Costa Rica) Network e 20

measurement sites1995 0.8 0.04 2.55 E-Perm� Williams-Jones, 1996;

Williams-Jones et al., 2000St. Miguel (El Salvador) Network e 205

measurement sites1999e2000 4.1 0.08 31 Alpha-Scintillation e PYLON AB-5 Cartagena et al., 2004

San Salvador (El Salvador) Network e 380measurement sites

1999 0.86 0.04 10.51 Alpha-Scintillation e PYLON AB-5 Pérez et al., 2004

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e8 5

we summarized the map of radon concentrations with the calcu-lated isoemissive curves identified by variable colors. These havebeen obtained by applying the IDW (“Inverse Distance to a Power”)method, preferred to the Kriging method since it better outlinessystematic variations in radon emissions (cf. Salih et al., 2002).

Noticeably, higher radon activities (>2000 Bq/m3) are alignedalong an NNWeSSE trend which includes the summit crater areaand its southeastern end. This geometry suggests a possible zone ofweakness distributed along the NNW sector of the cone. Thisstructure converges NNW with a sector of the Gastre Fault Zone(GVF) which crosscuts Villarrica Lake. It is interesting to notice thatall the areas surrounding the GFV and the LOFZ are characterized byhigher radon emissions. In particular, the station located in the NEsector of our survey is in proximity of the intersection of the citedfault zones and measuredw4600 Bq/m3 (see Table 1, S4). However,it cannot be excluded that the presence of highly vegetated clustersmay affect radon concentrations due to high moisture contents insoils (cf. Nazaroff, 1992; Hosoda et al., 2007; Papachristodoulouet al., 2007).

Automatic measurements obtained with DOSEman radon de-tectors were collected from November 8 to November 17, 2004.Dosimeters, provided with additional batteries, were installedwithin a polycarbonate case (permeable to radon, Fig. 2) andinserted into the soil down to a depth of about 50 cm. The location

Fig. 5. Summary of mean radon concentrations measured at Central and SouthAmerica volcanoes. The scale is logarithmic and the bars refer to the minima andmaxima measured radon concentrations. Our data on the Villarrica volcano refer to theE-PERM� measurements. The dotted line is the arithmetic mean for the whole data set(see Table 2 for details).

of the automatic stations coincides with the position of S1 and S4E-PERM� measurement sites (see Table 1 and Fig. 3). In these casesbulk radon emissions were not particularly high (Fig. 6): the sum-mit station reached values of about 1620 Bq/m3, averaging 680(�307) Bq/m3, whereas the station located at the NW base of thecone (station 1) recorded lower values, with a mean of 170(�54) Bq/m3. It is worth mentioning that higher radon signals, atboth stations, were recorded from November 12 to November 15(Fig. 6), but no significant changes have been noted, by direct ob-servations, in volcanic activity. During this period moderate lavapuffing within the intra-crater lava lake located at 70e100m belowthe crater rim (Fig. 2b) was noted. We observed several andesiticscoria bombs surrounding the crater rim that were ejected during

Fig. 6. Automatic radon measurements at Villarrica volcano (mean values for theabove stations are 160 and 670 Bq/m3, respectively); a) Time series for the lowerstation located at the base of the cone (VLS; cf. Table 1 for the location); b) Time seriesfor the summit station installed at Villarrica volcano. Grey and dark lines refer to the30-min sampling time and a 4-h moving average, respectively.

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e86

our monitoring survey. However, Palma et al. (2008) report a pro-gressive increase in volcanic activity since early November 2004. Inaddition, satellite thermal alerts were recorded by MODIS (Mod-erate Resolution Imaging Spectroradiometer), particularly onNovember 16 and 17 (cf. Smithsonian Institution, 2004).

5. Discussion and conclusions

It is generally acknowledged that in-soil gas measurements playa key-role in better understanding volcano dynamics and therelated degassing processes. Gas release from soil generally in-creases in sectors characterized by structural discontinuities and/orweakness, such as faults and fracture systems. Thus, in-soil radonsurveys in space and time have been used in detecting hidden faultsor identifying the extension at depth of hydrothermal systems andrelated diffuse degassing (e.g., Burton et al., 2004; Hernández et al.,2004; Cigolini et al., 2005). Moreover, continuous measurements ofradon activity in specific sites are regarded as a useful tool fortracking variations in volcanic activity.

Villarrica is one of the most active volcanoes in the Andean re-gion. It is an open system volcano and its current activity isessentially intracrateric with the presence of a lava lake where thelava normally resides and degasses. Although volcanic activity isrelatively mild and of strombolian type, it is not excluded thatmagma rise may lead to effusive events (such as that of November1984). The increase of heat flow in the upper part of the cone,during eruptive episodes, may melt the icy cover and trigger theonset of lahars that could threaten the lives of inhabitants live inthe area surrounding the volcano (particularly N and NW of thecone).

In addition, during several eruptions, failure of the volcanicedifice has followed the trend of regional fractures (Ortiz et al.,2003). These authors identify a zone of potential weakness alongthe NW sector of the cone (cf. p. 250 and Fig. 3 of their work). Thus,eruption styles may be the result of endogenous and morphotec-tonic factors that control the evolution of single eruptive episodes.Our contribution shows that the NNW sector of the cone exhibitshigher radon emissions and further supports the idea that this areacould be a zone of potential flank weakness. The survey also in-dicates that the magnitude of radon concentrations is not neces-sarily related to the altitude: i.e., the presence of a positivecorrelation between radon emissions and altitude. In turn, thisphenomenon has been shown to be effective at Stromboli andVesuvius (Cigolini et al., 2005; Cigolini, 2010). Normally, if the maindegassing occurs in the crater area at the top of the volcano expecthigher values might be expected at the summit. However, if there isan open conduit and a constant volcanic plume (such as Popoca-tépetl and Villarrica) most of the gas is released within the plumeitself (Varley and Armienta, 2001) and the gas fraction the mayreach the surrounding crater area is strongly controlled by thefracture network. This network could be subject to self-sealingprocesses (typical of volcanically related geothermal systems)that may limit and/or inhibit fluidmotion. The fracture network canpossibly be reopened during major eruptive cycles and/or seismicevents (Cigolini et al., 2005, 2007). It is therefore likely that thesemechanisms are operative at Villarrica volcano and the correlationbetween radon emissions and altitude was not detected during oursurvey. Another possible explanation is that the thick ice cover mayconsiderably inhibit radon emissions (Conen and Robertson, 2002).

We may conclude that the average 222Rn contents at Villarricaare in strong agreement with those recorded at other Central andSouth American volcanoes. Like Arenal, Villarrica does not showdramatic variations in radon from site to site, as observed atGaleras, Popocatépetl, and Poás where radon activities may behigher than 10,000 Bq/m3 (cf. Heiligmann et al., 1997; Varley and

Armienta, 2001; William-Jones et al., 2000). In our case, radonemissions are essentially controlled by tectonics: highest radonconcentrations (up to 4600 Bq/m3) are reached at the intersectionbetween the GVF and LOFZ that further supports the view thatradon is a key tool for identifying fractures and fault zones.We trustthat these data (together with the those collected at Galeras, thesole radon measurements reported regarding South Americanvolcanoes) could be useful for planning future monitoring activitiesat Villarrica in an effort to improve its volcano surveillance.Therefore, we suggest that a radon network dedicated to volcanomonitoring could be deployed by taking into account the locationsof our measurements sites and our preliminary results.

Acknowledgments

This research has been funded by the Italian Ministry for Uni-versity and Research (MIUR). We thank Valérie Vidal for providingsome images of Villarrica volcano. L. Wesley and A. Durkan, revisedthe English. The manuscript has been improved by the criticalcomments and suggestions of two anonymous reviewers.

References

Allard, P., Carbonelle, J.M., Faivre-Peirret, R., Martin, D., Sabroux, J.C., Zettwoog, P.,1991. Eruptive and diffusive emissions of CO2 from Mt. Etna. Nature 351,387e391.

Alparone, S., Behncke, B., Giammanco, S., Neri, M., Privitera, E., 2005. Paroxysmalsummit activity at Mt. Etna (Italy) monitored through continuous soil radonmeasurement. Geophysical Research Letters 2, L16307. http://dx.doi.org/10.1029/2005GL023352.

Barrientos, S., Acevedo, P.S., 1992. Seismological aspects of the 1988-1989 Lonqui-may (Chile) volcanic eruption. Journal of Volcanology and Geothermal Research53 (1e4), 73e87.

Barrientos, S., 1994. Large thrust earthquakes and volcanic eruptions. Pure andApplied Geophysics 142, 225e237.

Bohm, M., Lüth, S., Echtler, H., Asch, G., Bataille, K., Bruhn, C., Rietbrock, A.,Wigger, P., 2002. The Southern Andes between 36� and 40�S latitude: seismicityand average seismic velocities. Tectonophysics 356, 275e289.

Burton, M., Neri, M., Condarelli, D., 2004. High spatial resolution radon measure-ments reveal hidden active faults on Mt. Etna. Geophysical Research Letters 31,L07618.

Buttafuoco, G., Tallarico, A., Falcone, G., 2007. Mapping soil gas radon concentration:a comparative study of geostatistical methods. Environmental Monitoring andAssessment 313, 135e151.

Carapezza, M.L., Inguaggiato, S., Brusca, L., Longo, M., 2004. Geochemical precursorsof Stromboli 2002e2003 eruptive events. Geophysical Research Letters 31 (7).http://dx.doi.org/10.1029/2004GL019614.

Carapezza, M.L., Ricci, T., Ranaldi, M., Tarchini, L., 2008. Active degassing structuresof Stromboli and variations in diffuse CO2 output related to the volcanic activity.Journal of Volcanology and Geothermal Research 182 (3e4), 231e245.

Cartagena, R., Olmos, R., López, D.L., Soriano, T., Barabona, F., Hernández, P.A.,Pérez, N.M., 2004. Diffuse soil degassing of carbon dioxide, radon and mercuryat San Miguel volcano, El Salvador. In: Rose, W.I., Bommer, J.J., Lopez, D.L.,Carr, M.J., Major, J.J. (Eds.), Natural Hazards in El Salvador. Geological Society ofAmerica Special Paper, vol. 375, pp. 203e212. Boulder, Colorado.

Casertano, L., 1963. General characteristics of active Andean volcanoes and sum-mary of their activities during recent centuries. Bulletin Seismological Societyof America 53, 1415e1433.

Cembrano, J., Herve, F., 1993. The Liquiñe-Ofqui fault zone: a major Cenozoic strikeslip duplex in the southern Andes. In: Actas International Symposium on An-dean Geology, vol. 2, pp. 175e178.

Cembrano, J., Moreno, H., 1994. Geometría y naturaleza contrastante del vulcanismoCuaternario entre los 38� S y 46� S: ¿Dominios compresionales y tensionales enun regime transcurrente?. In: Septimo Congreso Geológico Chileno, Actas 1.Universidad de Concepción, pp. 240e244.

Chiodini, G., Frondini, F., Raco, B., 1996. Diffuse emission of CO2 from the Fossacrater, Vulcano Island Italy. Bulletin of Volcanology 58, 41e50.

Chirkov, A.M., 1975. Radon as a possible criterion for predicting eruptions asobserved at Karymsky volcano. Bulletin of Volcanology 39, 126e131.

Cigolini, C., Borgia, A., Casertano, L., 1984. Intra-crater activity, aa-block lava,viscosity and flow dynamics: Arenal Volcano, Costa Rica. Journal of Volcanologyand Geothermal Research 20, 155e176.

Cigolini, C., 2010. The dynamics of a double-cell hydrothermal system in triggeringseismicity at Somma-Vesuvius: results from a high-resolution radon survey(revisited). Bulletin of Volcanology 72, 693e704.

Cigolini, C., Salierno, G., Gervino, G., Bergese, P., Marino, C., Russo, M., Prati, P.,Ariola, V., Bonetti, R., Begnini, S., 2001. High-resolution radon monitoring and

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e8 7

hydrodynamics at Mount Vesuvius. Geophysical Research Letters 28 (21),4035e4039.

Cigolini, C., Gervino, G., Bonetti, R., Conte, F., Laiolo, M., Coppola, D., Manzoni, A.,2005. Tracking precursors and degassing by radon monitoring during majoreruptions at Stromboli Volcano (Aeolian Islands, Italy). Geophysical ResearchLetters 32, L12308. http://dx.doi.org/10.1029/2005GL022606.

Cigolini, C., Laiolo, M., Coppola, D., 2007. Earthquakeevolcano interactions detectedfrom radon degassing at Stromboli (Italy). Earth and Planetary Science Letters257, 511e525.

Cigolini, C., Poggi, P., Ripepe, M., Laiolo, M., Ciamberlini, C., Delle Donne, D.,Ulivieri, G., Coppola, D., Lacanna, G., Marchetti, E., Piscopo, D., Genco, R., 2009.Radon surveys and real-time monitoring at Stromboli Volcano: influence of soiltemperature, atmospheric pressure and tidal forces on 222Rn degassing. Journalof Volcanology and Geothermal Research 184 (3e4), 381e388.

Clavero, J., Moreno, H., 1994. Ignimbritas Licán y Pucón: Evidencias de erupcionesexplosivas andesítico- basálticas postglaciales del Volcán Villarrica, Andes delSur, 39�25’S. In: Septimo Congreso Geológico Chileno, Actas 1. Universidad deConcepción, pp. 250e254.

Clavero, J., Moreno, H., 2004. Evolution of Villarrica volcano. In: Lara, L., Clavero, J.(Eds.), 2004. Villarrica Volcano (39.5�S), Southern Andes, Chile, vol. 61. ServicioNacional de Geología y Minería, Boletín, pp. 17e27.

Conen, F., Robertson, L.B., 2002. Latitudinal distribution of radon-222 flux fromcontinents. Tellus 54B, 127e133.

Connors, C., Hill, B., La Femina, P., Navarro, M., Conway, M., 1996. Soil Rn-222 pulseduring the initial phase of the June August 1995 eruption of Cerro Negro,Nicaragua. Journal of Volcanology and Geothermal Research 73, 119e127.

Cox, M.E., 1980. Ground radon survey of a geothermal area in Hawaii. GeophysicalResearch Letters 7, 283e286.

Cox, M.E., 1983. Summit outgassing as indicated by radon, mercury and pH map-ping, Kilauea Volcano, Hawaii. Journal of Volcanology and Geothermal Research16, 131e151.

Delle Donne, D., Harris, A.J.L., Ripepe, M., Wright, R., 2010. Earthquake-inducedthermal anomalies at active volcanoes. Geology 38, 771e774. http://dx.doi.org/10.1130/G30984.1.

Dueñas, C., Fernández, M.C., Carretero, J., Liger, E., Pérez, M., 1997. Release of 222Rnfrom some soils. Annales Geophysicae 15, 124e133.

Fleischer, R.L., Mogro-Campero, A., 1985. Association with subsurface radon changesin Alaska and the Northeastern United States with earthquakes. Geochimicaand Cosmochimica Acta 49, 1061e1071.

Gauthier, P.J., Condomines, C., 1999. 210Pb-226Ra radioactive disequilibria in recentlavas and radon degassing: inferences on the magma chamber dynamics atStromboli and Merapi volcanoes. Earth and Planetary Science Letters 172, 111e126.

Gervino, G., Cigolini, C., Lavagno, A., Marino, C., Prati, P., Pruiti, L., Zangari, G., 2004.Modelling temperature distributions and radon emission at Stromboli Volcanousing a non-extensive statistical approach. Physica A 340 (1e3), 402e409.

Giammanco, S., Sims, K.W.W., Neri, M., 2007. Measurements of 220Rn and 222Rn andCO2 emissions in soil and fumarole gases on Mt. Etna volcano (Italy): implica-tions for gas transport and shallow ground fracture. Geochemistry, Geophysicsand Geosystems 8, Q10001.

Gründel, M., Postendörfer, J., 2003. Characterization of an electronic radon gaspersonal dosimeter. Radiation Protection Dosimetry 107 (4), 287e292.

Heiligmann, M., Stix, J., Williams-Jones, G., Sherwood Lollar, B., Garzón, V.,G., 1997.Distal degassing of radon and carbon dioxide on Galeras volcano, Colombia.Journal of Volcanology and Geothermal Research 77, 267e283.

Hernández, P., Pérez, N., Salazar, J., Reimer, M., Notsu, K., Wakita, H., 2004. Radonand helium in soil gases at Cañadas caldera, Tenerife, Canary Islands, Spain.Journal of Volcanology and Geothermal Research 131, 59e76.

Hickey-Vargas, R., López-Escobar, L., Moreno, H., Clavero, J., Lara, L., Sun, M., 2004.Magmatic evolution of the Villarrica Volcano. In: Lara, L., Clavero, J. (Eds.), 2004.Villarrica Volcano (39.5�S), Southern Andes, Chile, 61. Servicio Nacional deGeología y Minería Boletín, pp. 39e45.

Hill, D.P., Pollitz, F., Newhall, C., 2002. Earthquake-volcano interactions. PhysicsToday 55 (11), 41e47.

Hosoda, M., Shimo, M., Sugino, M., Furukawa, M., Fukushi, M., 2007. Effect of soilmoisture content on radon and thoron exhalation. Journal of Nuclear Scienceand Technology 44 (4), 664e672.

Igarashi, G., Saeki, S., Takahata, N., Sumikawa, K., Tasaka, S., Sasaki, Y., Takahashi, M.,Sano, Y., 1995. Ground-water radon anomaly before the Kobe earthquake inJapan. Science 269, 60e61.

Kotrappa, P., Dempsey, J.C., Stieff, L.R., 1990. A pratical E-PERM� (Electret PassiveEnvironmental Radon Monitor) system for indoor 222Rn measurement. HealthPhysics 58 (4), 461e467.

Kotrappa, P., Stieff, L.R., 1992. Elevation correction factors for E-PERM radon mon-itors. Health Physics 62 (1), 82e86.

Kotrappa, P., Dempsey, J.C., Stieff, L.R., 1993. Recent advances in electret ionchamber technology. Radiation Protection Dosimetry 47, 461e464.

Laiolo, M., Cigolini, C., Coppola, D., Piscopo, D., 2012. Developments in real-timeradon monitoring at Stromboli Volcano. Journal of Environmental Radioac-tivity 105, 21e29.

Lara, L., 2004. Villarrica-Lanín chain: tectonic constraints for volcanism in a transversalalignment. In: Lara, L., Clavero, J. (Eds.), 2004. Villarrica Volcano (39.5�S), SouthernAndes, Chile, 61. Servicio Nacional de Geología y Minería Boletín, pp. 13e16.

Lara, L., Clavero, J., 2004. Villarrica Volcano. Southern Andes, vol. 61. ServicioNacional de Geología y Minería Boletín, p. 73.

Lavenu, A., Cembrano, J., 1994. Neotectonica de rumbo dextral en la zona defalla Liquiñe-Ofqui: geometria, cinematica y tensor de esfuerzo. In: SéptimoCongreso Geológico Chileno, Actas 2. Universidad de Concepción, pp. 1080e1084.

Lavenu, A., Cembrano, J., 1999. Compressional and transpressional stress pattern forPliocene and Quaternary brittle deformation in fore-arc and intra-arc zones(Andes of Central and Southern Chile). Journal of Structural Geology 21, 1669e1691.

Lohmar, S., Robin, C., Gourgaud, A., Clavero, J., Parada, M.A., Moreno, H., Ersoy, O.,López-Escobar, L., Naranjo, J.A., 2007. Evidence of magma-water interactionduring the 13,800 years BP explosive cycle of the Licán Ignimbrite, Villarricavolcano (southern Chile). Revista Geológica de Chile 34 (2), 233e247.

López-Escobar, L., Moreno, H., 1994. Geochemical characteristics of the SouthernAndes basaltic volcanism associated with the Liquiñe-Ofqui fault zone between39.0� and 45.0�S. In: Séptimo Congreso Geológico Chileno, Actas 2. Universidadde Concepción, pp. 1388e1393.

López-Escobar, L., Cembrano, J., Moreno, H., 1995. Geochemistry and tectonics of theChilean Southern Andes basaltic Quaternary volcanism (37e46�S). RevistaGeológica de Chile 22 (2), 219e234.

Melnick, D., Folguera, A., Rosenau, M., Echtler, H., Potent, S., 2002. Tectonics fromthe northern segment of the Liquiñe-Ofqui fault system (37e39�S), PatagonianAndes. In: Papers Presented at the ISAG International Symposium on AndeanGeodynamics, Toulouse, pp. 413e416.

Mogro-Campero, A., Fleischer, R.L., 1977. Subterrestrial fluid convection: a hypoth-esis for long distance migration of radon within the earth. Earth and PlanetaryScience Letters 34, 321e325.

Moreno, H., 1974. Airplane flight over active volcanoes of central-south Chile. In:Guide Book D-3 of the IAVCEI International Symposium on Andean and Ant-arctic Volcanology Problems. Departamento de Geologia, Universidad de Chile,Santiago, p. 56.

Moreno, H., 1993. Volcán Villarica, geología y evaluación del riesgo volcánico,regiones IXyX, 39�25’ S. ServNacGeolMinChile,1:50,000GeologicalMap, p.112.

Moreno, H., Clavero, J., Lara, L., 1994a. Actividad explosiva postglacial del VolcánVillarrica, Andes del Sur. In: Séptimo Congreso Geológico Chileno, Actas 1.Universidad de Concepción, pp. 329e333.

Moreno, H., Lopez-Escobar, L., Cembrano, J., 1994b. The Villarrica-Quetrupillan-Lanin volcanic chain: a review and probable significance in the SouthernAndes, Chile. In: Séptimo Congreso Geológico Chileno, Actas 2. Universidad deConcepción, pp. 339e341.

Naranjo, J.A., Moreno, H., 1991. Actividad explosiva postglacial en el Volcán Llaima,Andes del Sur (38�450S). Revista Geológica de Chile 18 (1), 69e80.

Nazaroff, W.W., 1992. Radon transport from soil to air. Reviews of Geophysics 30,137e160. http://dx.doi.org/10.1029/92RG00055.

Neri, M., Behncke, B., Burton, M., Galli, G., Giammanco, S., Pecora, E., Privitera, E.,Reitano, D., 2006. Continuous soil radon monitoring during the July 2006 Etnaeruption. Geophysical Research Letters 33 (24). http://dx.doi.org/10.1029/2006GL028394.

Ortiz, R., Moreno, H., García, A., Fuentealba, G., Astiz, M., Peña, P., Sánchez, N.,Tárraga, M., 2003. Villarrica volcano (Chile): characteristics of the volcanictremor and forecasting of small explosions by means of a material failuremethod. Journal of Volcanology and Geothermal Research 128, 247e259.

Palma, J.L., Calder, E.S., Basualto, D., Blake, S., Rothery, D.A., 2008. Correlations be-tween SO2 flux, seismicity, and outgassing activity at the open vent of Villarricavolcano, Chile. Journal of Geophysical Research 113, B10201. http://dx.doi.org/10.1029/2008JB005577.

Papachristodoulou, C., Ioannides, K., Spathis, S., 2007. The effect of moisture contenton radon diffusion through soil: assessment in laboratory and field experi-ments. Health Physics 92 (3), 257e264.

Pérez, N.M., Salazar, J.M.L., Hernández, P.A., Soriano, T., López, D.L., Notsu, K., 2004.Diffuse CO2 and 222Rn degassing from San Salvador volcano, el Salvador, centralAmerica. In: Rose, W.I., Bommer, J.J., Lopez, D.L., Carr, M.J., Major, J.J. (Eds.),Natural Hazards in El Salvador. Geological Society of America Special Paper, vol.375, pp. 227e236. Boulder, Colorado.

Pérez, N.M., Hernández, P.A., Padrón, E., Melián, G., Marrero, R., Padilla, G.,Barrancos, J., Nolasco, D., 2007. Precursory subsurface 222Rn and 220Rndegassing signatures of the 2004 seismic crisis at Tenerife, Canary Islands. Pureand Applied Geophysics 164 (12), 2431e2448.

Petit-Breuilh, M.E., 1994. Contribución al conocimiento de la cronologia eruptivahistórica del Volcán Villarrica, 1558e1985. Revista Frontera 13, 71e99.

Pinault, J.L., Baubron, J.C., 1996. Signal processing of soil gas radon, atmosphericpressure and soil temperature data: a new approach for radon concentrationmodeling. Journal of Geophysical Research 101, 3157e3171.

Planici�c, J., Radoli�c, V., Vukovi�c, B., 2004. Radon as an earthquake precursor. NuclearInstruments and Methods in Physics Research A 530, 568e574.

Pulinets, S.A., Ouzounov, D.P., Giuliani, G.G., Ciraolo, L., Taylor, P.T., 2009. Atmo-sphere and radon activities observed prior to Abruzzo M6.3 earthquake of April6, 2009. EOS Transactions AGU 90 (52). Fall Meeting Suppl. Abstract, U14A-07.

Ripepe, M., Marchetti, E., Bonadonna, C., Harris, A.J.L., Pioli, L., Ulivieri, G., 2010.Monochromatic infrasonic tremor driven by persistent degassing and convec-tion at Villarrica Volcano, Chile. Geophysical Research Letters 37, L15303. http://dx.doi.org/10.1029/2010GL043516.

Rivera, A., Corripio, J.G., Brock, B., Clavero, J., Wendt, J., 2008. Monitoring ice-cappedactive Volcán Villarrica, southern Chile, using terrestrial photography combinedwith automatic weather stations and global positioning systems. Journal ofGlaciology 54 (188), 920e930.

C. Cigolini et al. / Journal of South American Earth Sciences 46 (2013) 1e88

Salih, M.I., Pettersson, H.B.L., Sivertun, A., Lund, E., 2002. Spatial correlation be-tween radon (222Rn) in groundwater and bedrock uranium (238U): GIS andgeostatistical analyses. Journal of Spatial Hydrology 2 (2), 1e10.

Scholtz, C.H., Sykes, L.R., Aggarval, Y.P., 1973. Earthquake prediction: a physical basis.Science 181, 803e810.

Segovia, N., Mena, M., 1999. Soil radon pulses related to the initial phases to volcaniceruptions. Il Nuovo Cimento 22C (3e4), 275e279.

Seidel, J.L., Monnin, M., Garcìa Vindas, J.R., Richard, L.P., Staudacher, T., 1999. Sys-tematic radon survey over active volcanoes. Il Nuovo Cimento 22C, 363e368.

Shinohara, H., Witter, J., 2005. Volcanic gases emitted during mild strombolianactivity of Villarrica volcano, Chile. Geophysical Research Letters 32, L20308.http://dx.doi.org/10.1029/2005GL024131.

Silva Parejas, C., Druitt, T.H., Robin, C., Moreno, H., Naranjo, J.A., 2010. The HolocenePucón eruption of Volcàn Villarrica, Chile: deposit architecture and eruptionchronology. Bulletin of Volcanology 72, 677e692. http://dx.doi.org/10.1007/s00445-010-0348-9.

Siniscalchi, A., Tripaldi, S., Neri, M., Giammanco, S., Piscitelli, S., Balasco, M.,Behncke, B., Magrì, C., Naudet, V., Rizzo, E., 2010. Insights into fluid circulationacross the Pernicana fault (Mt. Etna, Italy) and implications for flank instability.Journal of Volcanology and Geothermal Research 193, 137e142.

Smithsonian Institution, 2004. Villarrica. Bulletin of the Global Volcanism Network29 (12).

Streil, T., Oeser, V., Feige, S., 2002. An electronic radon dosimeter as a multipurposedevice-a bridge between dosimetry and monitoring. Geofisica Internacional 41,285e288.

Thomas, D.M., Cox, M.E., Cuff, K.E., 1986. The association between ground gas radonvariations and geologic activity in Hawaii. Journal of Geophysical Research 91,12186e12198.

Trique, M., Richon, P., Perrier, F., Avouac, J.P., Sabroux, J.C., 1999. Radon emanationand electric potential variations associated with transient deformation nearreservoir lakes. Nature 399, 137e141. http://dx.doi.org/10.1038/20161.

Varley, N.R., Armienta, M.A., 2001. The absence of diffuse degassing at Popocatepetlvolcano, Mexico. Chemical Geology 177, 157e173.

Volland, S., Sturm, M., Lukas, S., Pino, M., Muller, J., 2007. Geomorphological andsedimentological evolution of a lake basin under strong volcano-tectonic in-fluence: the seismic record of Lago Calafquén (south-central Chile). QuaternaryInternational 161, 32e45.

Viveiros, F., Ferreira, T., Cabral Vieira, J., Silva, C., Gaspar, J.L., 2008. Environmentalinfluences on soil CO2 degassing at Furnas and Fogo volcanoes (São MiguelIsland, Azores archipelago). Journal of Volcanology and Geothermal Research177 (4), 883e893.

Williams-Jones, G., Avril 1996. The Distribution and Origin of Radon, CO2and SO2 Gases at Arenal Volcano, Costa Rica. MSc thesis. Université deMontréal.

Williams-Jones, G., Stix, J., Heiligmann, M., Charland, A., Sherwood Lollar, B.,Arner, N., Garzón, G.V., Barquero, J., Fernandez, E., 2000. A model of diffusedegassing at three subduction-related volcanoes. Bulletin of Volcanology 62,130e142.

Witter, J.B., Delmelle, P., 2004. Acid gas hazards in the crater of Villarrica volcano(Chile). Revista Geológica de Chile 31 (2), 273e278.

Witter, J.B., Kress, V.C., Delmelle, P., Stix, J., 2004. Volatile degassing, petrology, andmagma dynamics of the Villarrica Lava Lake, Southern Chile. Journal of Volca-nology and Geothermal Research 134, 303e337.

Zimmer, M., Erzinger, J., 2003. Continuous H2O, CO2, 222Rn and temperature mea-surements on Merapi Volcano, Indonesia. Journal of Volcanology andGeothermal Research 125, 25e38.