Gas geochemistry of natural analogues for the studies of geological CO 2 sequestration

Applied Geochemistry 24 (2009) 1339–1346

Contents lists available at ScienceDirect

Applied Geochemistry

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

Gas geochemistry of natural analogues for the studies of geologicalCO2 sequestration

N. Voltattorni a,*, A. Sciarra a, G. Caramanna b, D. Cinti a, L. Pizzino a, F. Quattrocchi a

a Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata n� 605, 00143 Rome, Italyb Earth Science Dep., University ‘‘La Sapienza”, Piazzale A. Moro n� 5, 00185 Rome, Italy

a r t i c l e i n f o

Article history:Received 29 November 2007Accepted 17 April 2009Available online 3 May 2009

Editorial handling by R. Fuge

0883-2927/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.apgeochem.2009.04.026

* Corresponding author.E-mail address: [email protected] (N. Volt

a b s t r a c t

Geological sequestration of anthropogenic CO2 appears to be a promising method for reducing theamount of greenhouse gases released to the atmosphere. Geochemical modelling of the storage capacityfor CO2 in saline aquifers, sandstones and/or carbonates should be based on natural analogues both in situand in the laboratory. The main focus of this paper has been to study natural gas emissions representingextremely attractive surrogates for the study and prediction of the possible consequences of leakage fromgeological sequestration sites of anthropogenic CO2 (i.e., the return to surface, potentially causing local-ised environmental problems). These include a comparison among three different Italian case histories:(i) the Solfatara crater (Phlegraean Fields caldera, southern Italy) is an ancient Roman spa. The area ischaracterised by intense and diffuse hydrothermal activity, testified by hot acidic mud pools, thermalsprings and a large fumarolic field. Soil gas flux measurements show that the entire area dischargesbetween 1200 and 1500 tons of CO2 per day; (ii) the Panarea Island (Aeolian Islands, southern Italy)where a huge submarine volcanic-hydrothermal gas burst occurred in November, 2002. The submarinegas emissions chemically modified seawater causing a strong modification of the marine ecosystem.All of the collected gases are CO2-dominant (maximum value: 98.43 vol.%); (iii) the Tor Caldara area (Cen-tral Italy), located in a peripheral sector of the quiescent Alban Hills volcano, along the faults of the ArdeaBasin transfer structure. The area is characterised by huge CO2 degassing both from water and soil.Although the above mentioned areas do not represent a storage scenario, these sites do provide manyopportunities to study near-surface processes and to test monitoring methodologies.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Natural degassing phenomena can be studied as ‘‘natural ana-logues” in the framework of geological storage and sequestrationof anthropogenic CO2 emissions, especially when the risk of possi-ble leakage at the surface is taken into account with potential con-sequences for the biosphere (Voltattorni et al., 2006). Muchinternational effort has been expended to investigate CO2 storagewithin deep geological formations (e.g. Holloway, 1996), withindeep oceans (Marchetti, 1977; Golomb et al., 1992; Austvik andLoken, 1992; Wilson, 1992), and in deep-water sediments (Koideet al., 1997). Among the many solutions posed to reduce furtheremissions and to remediate the existing problem, capture and stor-age of CO2 in geological formations is one of the most promisingapproaches. Differing sequestration techniques have been pro-posed: through injection of CO2 associated with enhanced oilrecovery, by storage in depleted oil and gas reservoirs, throughreplacement of absorbed CH4 by CO2 in deep coal-beds, by

ll rights reserved.

attorni).

injection into saline aquifers, through storage in salt caverns andby replacement of CH4 in seafloor clathrate hydrates (Yezdimeret al., 2002). Beside the problems of minimizing the costs of CO2

injection and gaining public acceptance, one of the main topics isto reduce the risk of CO2 leakage through fault zones and furtherensuring long-term CO2 retention in the geological formations. Asalready known, the long-term effects and stability of a man-madeCO2 geological storage facility is very difficult to predict with lab-oratory or modelling experiments due to the size and long timescales involved. An useful approach which can address the difficul-ties of laboratory and modelling experiments, is to study ‘‘naturalanalogues”, that is, natural sites where CO2 is produced at depthwith the resulting gas either being trapped in deep reservoirs orleaking to the surface along preferential migration pathways likefaults or fractures (Lombardi et al., 2006; Pearce, 2006; Pizzinoet al., 2002). The study of these sites can be a valid tool for the bet-ter understanding of the physical and chemical processes involvedas well as for the development of new tools for site assessment andmonitoring of any potential CO2 geological storage site.

This paper summarises some of the results obtained from theinvestigation of three natural analogue sites in Italy including theSolfatara crater (Phlegraean Fields caldera, southern Italy), Panarea

1340 N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346

Island (Aeolian Archipelago, Sicily) and Tor Caldara (Alban Hill vol-canic complex, Central Italy). All these sites are CO2-leaking andthey represent natural laboratories for the direct study of the microand macro scale migration of gas, as well as the possible effects ofCO2 leakage on the shallow environment. Soil-gas surveys andanalysis of gas leakage rates can define how CO2 migrates throughthe near surface environment as well as identify migration path-ways (using tracer gases such as Rn and He) that may be potentialroutes for CO2 escape in the near-surface.

2. Sampling and analysis

Many techniques are used to sample soil-gases, gas from sub-marine gaseous emission points and to measure gas fluxes. Sam-pling was performed using well tested methods (Lombardi andReimer; 1990; Bertrami et al., 1990; Hutchinson and Livingston,1993; Matthias et al., 1980; Chiodini et al., 1995, 1998, 2000).

2.1. Soil-gases

Soil gas samples were collected from shallow point sources. Thesoil-gas survey was generally carried out in the summer during aperiod of stable meteorological conditions. A 1 m hollow steelprobe with a conical point at the bottom and a sampling port ontop is inserted to a depth of 0.5 m below the ground surface. Two50 cc samples of soil gas are extracted with a syringe for cleaningthe probe, then a soil-gas sample is extracted and stored in anevacuated 25 mL steel cylinder for laboratory analysis (He, H2,O2, N2, CO2, CH4 and H2S) by means of a Perkin–Elmer AutoSystemXL gas chromatograph.

A RAD7 Durridge� alpha spectrometer was used for Rn surveys.Radon particles generate positively charged 218Po ions after enter-ing the ionization chamber and they are collected on the detectorby electrical high-voltage field sources. Radon calculation is basedon the sum of 218Po and 214Po peaks. Radon values were deter-mined after 15 min (necessary time for Po and Rn nuclei equilib-rium, that is about 5 times the half-life of 218Po) pumping from asteel probe at 0.5 m at depth. A desiccant (drierite) trap and an in-let filter protect the detector ionization chamber from soil humid-ity (>10%).

2.2. Gas fluxes

Gas flux measurements have been performed in situ using thechamber technique by means two different methods:

(i) in the Solfatara crater, characterised by high fluxes, an accu-mulation chamber with a volume of 50 L was used in orderto cover a very large range of soil fluxes with good sensitivityand linearity. The chamber was set on the ground in such away as to eliminate the input of atmospheric air. Every5 min a gas sample was extracted from an external septumlocated on the top of the chamber and analysed in loco usinga portable gas chromatograph and, at the same time, theRAD7 Durridge� alpha spectrometer for Rn flux measure-ments. The gas measurements result in an accumulationcurve which can be used to calculate soil gas flux accordingto Hutchinson and Livingstone (1993) and Tuccimei and Sol-igo (2008):

Ugas ¼ ðCfinal � CinitialÞV=DtA

where U is the gas flux (kgm�2 d�1), V is the volume of thechamber, A is the surface area measured and Dt is the timedifference between the first and the last measurement ofthe gas concentrations in accordance with the slope of thegrowth curve over the course of time.

(ii) The soil CO2 flux at Tor Caldara was measured by a West Sys-temTM instrument as the fluxes were very low (in comparisonwith the previous studied area) and the use of a 50 L chamberis not advisable since it takes a long time to achieve a growthcurve. The West SystemTM chamber has an assured low rate ofmixing, pressure equilibration between the inside and theoutside of the chamber and real time (PDA memorization)measurements with a portable on-line Li-COR�, modelLI820. The error caused by the interference of the H2O signal(generated by the humidity, normally cut by a MagnesiumPerchlorate drier), has been evaluated to be lower than 1%.

A different method was developed to measure the gas flow atPanarea area because of the submarine degassing processes. A fun-nel was inverted and placed on the vent to be measured and wasthen connected to a tank of known volume. By counting the fillingtime, it was possible to determine the flow rate (L/min) at any gi-ven depth. The flow rate of the vent was then determined byreporting this measure at SPT.

By knowing the degassing surface covered by the funnel, it waspossible to infer the flow in L/min/m2. Because of the limitations inthe dimensions of the measuring tank and the funnel, this systemmay be used only with small or medium rate gas flows (up to 1–1.5 L/s). The sampling system must have ballast to counter itsbuoyancy once it is filled with the gas. For a 5-L tank linked to a30 cm. diameter funnel, a ballast of 8 kg was used.

2.3. Submarine fluid sampling

Due to the presence of water, sampling methods adapted to theundersea environment were developed (Caramanna et al., 2005). Inorder to collect free/dry gas samples, a plastic funnel was inverted(30 cm diameter with 12 kg ballast around the lower ring) andplaced precisely on the gas vent to be sampled. All of the samplerswere stored in a plastic box that was carried underwater by the di-vers. The funnel was connected, through a silicon hose, to a Pyrexglass flask with twin valves. This flask was pre-filled with air at apressure above that of the hydrostatic pressure expected at thesampling depth in order to stop seawater from entering the sam-pler. Afterwards, collected gas samples were analysed by meansof gas chromatography (He, H2, O2, N2, CO2, CH4 and H2S) and c-spectrometry (Rn) at the laboratory.

3. Results and discussion

Three Italian sites have been considered to show different geo-logical scenarios with natural CO2 gas accumulations and emis-sions in order to study processes controlling how leaks mayoccur, their potential impacts on near-surface ecosystems and tosuggest methodologies for monitoring gas leakage in the frame-work of CO2 geological sequestration. The studied areas are tecton-ically and volcanically active and there is natural production of CO2

that migrates to the surface via faults and fractures. All three inves-tigated areas are located along the Tyrrhenian margin that is a geo-logically-active region due to the African plate subducting belowthe European plate with a northeastern trend. Deep subduction ba-sins and volcanic arcs lie in the Tyrrhenian Margin and serve asclues of tectonic and volcanic activity. Near the basins and the vol-canic regions there is an increase of heat flow (up to 200 mW/m2 inthe Tyrrhenian abyssal plain) that coincides with gravity and mag-netic anomalies (Dando et al., 1999).

3.1. Panarea Island (Aeolian Islands, southern Italy)

Panarea Island belongs to the Aeolian Arc, a volcanic structureextending for about 200 km along the north-western side of the

N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346 1341

Calabro-Peloritano block. This is a fragment of the Hercynian-Al-pine orogenic belt (consisting of various types of metamorphic,sedimentary and intrusive rocks) that detached from the Corsica-Sardinia block and migrated south-eastward to its present positionduring the opening of the Tyrrhenian Sea. The distribution of vol-canoes is strongly controlled by regional fault systems which areoriented E–W, NW–SE and NE–SW. Seismic studies reveal a crustof about 20–25 km beneath the Aeolian Arc (Gasparini et al.,1982; Barca and Ventura, 1991, and references therein) which indi-cates mantle upwelling. The dynamics of the arc (located alongsome regional N–S, E–W and NE–SW oriented fault systems) iscontrolled by tectonics through still active faults (Gasparini et al.,1982; Lanzafame and Rossi, 1984; Capaccioni et al., 2007).

In the proximity of Panarea Island (Aeolian Islands, southernItaly) a huge submarine volcanic-hydrothermal gas burst occurredduring November, 2002. The high-pressure gas leakage createdsinkholes with the collapse of the seafloor. From November 2002to the present, five gas emission points (named Vent 1, Vent 2, Vent8, Sink and Black point) have been studied with the aim of investi-gating possible variations in the geochemical parameters overtime. Gas emission points are located at the sea-bottom amongsome islets and reefs which represent the subaerial remnant ofan old volcanic centre (Gabbianelli et al., 1986, 1990).

The temperatures of leaking fluids are variable at the differentgas emission points (Fig. 1): highest temperature measurementsrefer to Black point ranging between 110 �C and 137 �C, exceptingthe first measurement (a few days after the gas burst, on Novem-ber, 13, 2002) being 30 �C. Lower temperatures (mean value:86.7 �C) have been measured at the Sink point and at Vent 8 (meanvalue: 52.11 �C). According to Capaccioni et al. (2007), a possibleexplanation for the temperature variability at the different gasemission points is related to an inferred magmatic system centredon or closer to Black point and whose diameter probably does notexceed a few hundred meters. Vents 1 and 2 have recorded thelowest values (mean values, respectively: 39.75 and 33.8 �C). Beinglocated at the margins of the inferred magmatic system, the lattervents have probably been affected by thermal cooling in the declin-ing stage, because of a rapid inflow of cold seawater from thesurroundings.

Flow measurements refer to three surveys that were conducted(in December 2006, February 2007 and May 2007) over four vents(Vent 8, Vent 1, Vent 2 and Black point). The strongest gas flow wasfrom Vent 8 (4.14 � 104 m3/y at SPT); in this vent the gas rises froma small spot at about 9 m in depth on the gravel seafloor originat-

10

30

50

70

90

110

130

150

13/11/2

002

20/11/2

002

28/11

/200

2

30/11

/200

2

09/12

/200

2

10/12

/200

2

14/12

/2002

17/12

/2002

18/12

/2002

11/02

/2003

12/02

/2003

25/04

/2003

28/05

/2003

03/06

/2003

04/06

/2

T (°C

)

Sea Vent 1Vent 2SinkholeVent 8Black Point

Fig. 1. T measurements at Panarea vents. Fluids from vents are very hot (especially from Saffect temperatures of the surrounding seawater.

ing as a plume that reaches the surface. The second strongest flowcomes from the Black point (1.46 � 104 m3/y at SPT) at 23 m depth;in this vent the gas exits from a small orifice in an agglomerate thatwas created from the deposition of metal sulphosalts. The flow car-ries a cloud of sulphosalts particulate creating a vent similar to theoceanic ‘‘black smokers” (but on a much smaller scale) from whichthe name ‘‘Black point” derives. The third greatest flow is from theVent 2 area. In this area, the gas originates from some lineamentson the gravel seafloor creating a wide degassing area(1.17 � 104 m3/y at SPT). In Vent 1 (3.58 � 103 m3/y at SPT) thegas flow is an order of magnitude less than the previous one; thisvent represents the oldest gas emission point since the 2002 burst.

The total estimated emitted volume of gas from all the mea-sured vents is about 8 � 104 m3/y at SPT, or about 1.6 � 102 t ofCO2/y. This value represents only a small amount of the total CO2

emitted from the entire degassing area.Gases have been collected from the seafloor at variable depths

(depending on gas emission point depth, Fig. 2) and analysed inaccordance with standard methods (Giggenbach, 1975; Capassoand Inguaggiato, 1998).

The chemical compositions of the submarine gas emissionshave displayed a complex combination of temporal and spatial var-iability from November 2002 to December 2006.

All of the collected gases (Table 1) are CO2-dominant (the con-tent varies from a minimum of 83.64 vol.% to a maximum of 98.43vol.%). Fig. 3 shows a comparison of the CO2 values from the fivemonitored vents through a statistical distribution (box plots). Theboxes show the upper and lower quartiles, median inside thebox, the vertical lines show the normally-distributed minimumand maximum while the horizontal line shows outliers. The CO2

leakage varies at the different vents being higher at the Black pointand lowest at the Sink point. However, median values are very sim-ilar for each vent suggesting a common degassing input linked tolocal tectonic features. In fact, all the gas emission points are lo-cated along N–S, E–W and NE–SW oriented active faults control-ling the Aeolian Volcanic District.

The H2S content has only been determined since July 2003 andranges from 0.4 to 4 vol.%. Lowest values were found at Black point(mean value: 0.64 vol.%) and values are quite constant during thewhole sampling period. At this gas emission point, also the temper-ature trend is quite constant (mean value: 127.1 �C) and Fig. 4shows a slight correlation between H2S concentrations and Tvariations. The same trend has been highlighted for Sink point,Vent 8 and Vent 1 and, for this reason they are not reported in

003

21/07

/2003

22/07

/2003

23/07

/2003

03/09

/2003

14/05

/2004

15/05/2

004

16/05/2

004

07/07

/2004

18/11

/2004

16/12

/2004

19/02

/2005

10/06

/2006

22/09

/2006

15/12

/2006

ink and Black point with temperatures >90 �C) but due to their low flow, they do not

Fig. 2. Panarea gas vent locations. The gas emissions are among the Dattilo, Lisca Nera, Bottaro and Lisca Bianca islets. They follow local tectonic faults, in particular the N40E-trending fault that links the Panarea and Stromboli volcanic structures.

Table 1Mean values of gases emitted at Panarea Island.

Gas species Mean value

CO2 (%, v/v) 98CH4 (ppm, v/v) 10N2 (%, v/v) 0.4He (ppm, v/v) 11H2 (ppm, v/v) 1100H2S (%, v/v) 2.2

2

82

84

86

88

90

92

94

96

98

100

CO

2 con

cent

ratio

n (%

, v/v

)

Black Point

Vent 8

Vent 2

Vent 1

Sink

Fig. 3. Box plots of soil gas CO2 data from the Panarea vents. The boxes show theupper and lower quartiles, median inside the box, the vertical lines show thenormally-distributed minimum and maximum while the horizontal line showsoutliers. Note that the median values are very similar for each vent suggesting acommon degassing input linked to local tectonics.

Black Point

0.4

0.5

0.6

0.7

0.8

0.9

Jul_0

3

Sep_0

3

May_04

Jul_0

4

Nov_04

Dec_0

4

Feb_0

5

Jun_0

6

Sep_0

6

Dec_0

6

H2S

(%. v

/v)

0

20

40

60

80

100

120

140

160

T (°C

)

Vent 2

0

1

2

3

4

5

Jul_0

3

Sep_0

3

Feb_0

4

May_04

Jul_0

4

Nov_04

Dec_0

4

Feb_0

5

Jun_0

6

Sep_0

6

Dec_0

6

H2S

(%. v

/v)

0

5

10

15

20

25

T (°

C)

H2S (%)T (°C)

Fig. 4. Temporal diagrams of temperature versus H2S content. The variation oftemperatures at different gas emission points (Black point, Vents 1 and 8, Sinkpoint) do not influence the H2S concentrations. At Vent 2, in contrast, a sharpdecrease in temperature corresponds to a slight increase in the gas speciessuggesting an abrupt thermal cooling of the feeding system from July 2004 up toFebruary 2005.

5060708090

100110120130140150

march03

apr03

may03

july03

sept03

may04

july04

nov04

dec04

feb05

june06

sept06

dec06

T (°C

)

300350400450500550600650700750800

CH

4 (v

ol. p

pm)

T (°C)CH4 (vol. ppm)

Fig. 5. Temporal diagram of the temperatures and CH4 content of Black Point. Thehigh CH4 concentrations likely suggest a thermogenic origin of the gas and therelationship between temperature and gas concentration variations stronglysupports this hypothesis.

1342 N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346

Fig. 4. Different behaviour is displayed by Vent 2. The H2S concen-trations and temperature are inversely correlated: a sharp decreasein temperature corresponds with a slight increase in the gas spe-cies suggesting an abrupt thermal cooling of the feeding systemfrom July 2004 up to February 2005.

Mean CH4 content is about 20 vol. ppm except in the Blackpoint (BP) where the mean value is 588.6 vol. ppm. The hypoth-esis for the thermogenic origin of CH4 in the Black point is con-firmed by the relationship between CH4 and temperature (Fig. 5).Helium is always present, however from November 2002 to Sep-tember 2003, its content was very high (>2000 vol. ppm for Vent1 and Vent 2). After that period, all the values have seemed tobe more stable with concentrations of approximately 10 vol.ppm (Fig. 6). Only the Sink sample shows constant He values,but this vent has only been studied since its May 2003 discoveryand no records are available before this period. Hydrogen is al-ways present. With the exception of Vent 1 and Sink point,which have constant values, the trend of H2 concentration is

H2 (vol . ppm)

0

200

400

600

800

1000

1200ve nt 1

ve nt 2

ve nt 8

s ink

BP

0

1000

2000

3000

4000

5000

6000

nov-02

dec-02

jan-03

feb-0

3

march

-03

apr-0

3

may-03

july-

03

sept-0

3feb

-04

may-04

july-

04

nov-04

dec-04

feb-0

5

june-0

6

sept-0

6

dec-06

nov-02

dec-02

jan-03

feb-0

3

march

-03

apr-0

3

may-03

july-

03

sept-0

3feb

-04

may-04

july-

04

nov-04

dec-04

feb-0

5

june-0

6

sept-0

6

dec-06

He (vol . ppm) ve nt 1

ve nt 2

ve nt 8

s ink

BP

Fig. 6. Temporal diagrams of He and H2 contents from the Panarea vents. Bothgases are always present in all samples although after July 2003, He undergoes arapid decrease while, in contrast, H2 increases, especially at Vent 2 and Sink.

N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346 1343

completely inverse to that of He (Fig. 6): the lowest values (<10vol. ppm) were detected during the period from November 2002to September 2003 and the highest values (up to 1000 vol. ppmfor Vent 2) have been measured since February 2004. Data col-lected before November 2, 2002 have values typical of a hydro-thermal environment (Italiano and Nuccio, 1991), with lower H2/CH4 ratios which are likely due to H2 removal caused by surfaceoxidative processes. The data collected after the recent gaseousemissions, however, show consistent contents of H2 and CH4 thatare more typical of high-T volcanic fumaroles (Caliro et al.,2004). A comparison between variations of gas concentrationsand temperatures over time has shown no relationship exceptingfor CH4, as discussed above. Nevertheless, the temporal evolutionof the He and H2 composition after February 2004 would,according to Capaccioni et al. (2007), be due to thermal cool-ing/heating-up phases of the feeding system.

Thus, the overall gas compositions and their variation over timesuggest a continuing evolution of the hydrothermal/volcanic sys-tem controlling the Panarea emissions.

3.2. The Phlegraean Fields Caldera (southern Italy)

The Phlegraean Field magmatic system is active, as the lasteruption occurred in 1538 A.D. at Monte Nuovo. Faults affectingthe Phlegraean Field caldera follow two preferred strikes, NW–SEand NE–SW, that also affect the Campanian Plain and the inner sec-tors of the Apennine belt (Hyppolite et al., 1994; Orsi et al., 1996).The most impressive thermal manifestations (including fumaroles,mud pools and vigorously boiling pools) are located in the Solfataraarea, a subcircular depression with a 12 km diameter. It originatedabout 35 ka BP after the eruption of the Campanian Ignimbrite Rosiand Santacroce, 1984). The Bocca Grande (Large Mouth) is the nameof the main fumarole and it has water vapour temperatures reach-ing approximately 160 �C. Within the mouth, the vapour conden-sate contains compounds such as realgar (AsS), cinnabar (HgS)and As sulfide (As2S3) which give a yellow-reddish colour to thesurrounding rocks. In the middle of the Solfatara area a bubblingmud pool (called ‘‘La Fangaia”) is made up of rainwater and vapourcondensate.

The Phlegraean Field Caldera was investigated, during Novem-ber 2006, by means of a detailed soil-gas survey in the inter cratersector, during which 54 soil gas flux measurements (1 sample/50–100 m) were performed through the accumulation chamber meth-od. Table 2 shows some statistical parameters for the all measuredgas species both as fluxes and concentrations. However, in thiscontext, only results of CO2 flux and concentration will bediscussed.

The contour map (Fig. 7) shows an area of about 0.5 km2 of highUCO2 values representing an important diffuse degassing structurecrossed by a NW–SE band of low fluxes. The UCO2 values rangefrom 0 to 5500 g/m2 � d with an average of 1127.32 g/m2 � d. Thehighest fluxes were found in the ‘‘La Fangaia” and near the ‘‘BoccaGrande” and ‘‘Bocca Nuova” fumaroles. These fumaroles have thehighest outlet temperatures (145–165 �C) among the several fuma-roles present in the area (mean discharge temperature = 100 �C).Fumarole effluents have similar chemistry, with H2O as the maincomponent, followed by CO2 and H2S (Chiodini et al., 2001).

The same area was investigated during June 2002 (Voltattorniet al., 2006) and a comparison of the two surveys (Fig. 7) performedin different years and seasons, has highlighted that the highestUCO2 values are always within an area bordered by faults and frac-tures, confirming that the degassing process is strictly related to lo-cal tectonic structures.

3.3. The Tor Caldara area

The Natural Reserve of Tor Caldara belongs to the Alban Hill vol-canic complex formed about 700 ka ago, during different eruptivesteps within the extensive tectonics linked to the opening of the Tyr-rhenian Sea (Chiarabba et al., 1997). The Meso-Cenozoic carbonatebasement, underlying the volcanic edifice and enclosing a deeplow-enthalpy reservoir of around 110–140 �C (Giggenbach et al.,1988; Quattrocchi et al., 2001; Mariucci et al., 2008), suggests thatthe volcanic region has been involved in recent tectonics, forminghorst-graben sequences inside the basement itself (Funicielloet al., 1987; Toro, 1976; Chiarabba et al., 1994; Di Filippo and Toro,1995; De Rita et al., 1995): the boundaries of these structures, wherethe thickness of the Neogene clay deposits is minimum, coincidewith many diffusive degassing structures (DDS). One of them is thestudied Tor Caldara natural reserve pertaining to the ‘‘Ardea Basin”fault system, as defined by Faccenna et al. (1994). In the Alban Hillquiescent volcanic structure many other DDSs are located: the Zolfo-rata-Pomezia and the Cava dei Selci sites are the most important (i.e.,6.1 � 108 [moles y�1] of CO2 for both structures), with noteworthyNatural Gas Hazard inside (Chiodini and Frondini, 2001; Pizzinoet al., 2002). Minor DDSs are the Trigoria-Vallerano, located near asmall seismogenic structure activated during the 1995 Rome earth-quake (Mw 3.8, Marra, 1999), the Tivoli-Bagni di Tivoli DDS, re-acti-vated during 2001 and the Ardea-Fossignano DDS, wheregeochemical and thermal anomalies in groundwater occur (Quatt-rocchi et al., 2001). Most of the earthquakes and related degassingstructures on the Alban Hills occur along a well known N–S regionalfault (De Rita et al., 1992; Marra, 2001).

Historic episodes of increase in the gas-discharges from DDSand/or strong gas exhalations over large areas have been observedin many sectors of the Alban Hill Volcano corresponding with localand/or regional earthquakes (Pizzino et al., 2002, and referencestherein). These phenomena suggest a strong relation between gas-eous releases/bursts and stress–strain at depth (Pizzino et al., 2002,and references therein). Many authors (Chiodini and Frondini,2001; Annunziatellis et al., 2003; Carapezza et al., 2003; Tuccimeiet al., 2006; Mariucci et al., 2008) have studied the chemistry ofthese diffuse gaseous exhalations; they are CO2-dominated (90–98% vol./vol.) and contain minor amounts of N2, CH4, H2S, He andRn. In particular, CO2 emitted by the degassing structures of the Al-

Table 2Some statistical parameters for all the measured gas species fluxes and concentrations at the Phlegraean Fields Caldera. U is expressed in KBq �m�2 � day�1 for Rn and g/m2 dayfor all the other gas species. He, H2 and CH4 concentrations are expressed in ppm, v/v; O2, N2, CO2 and H2S in %, v/v and Rn in Bq/m3.

He H2 O2 N2 CO2 CH4 H2S Rn

Flux (U)Minimum 4.86E�04 0.00 0.07 0.45 83.34 0.00 0.00 0.00Maximum 1.76E�03 0.65 2.04 9.01 5287.21 1524.96 390.24 92.76Mean 1.18E�03 0.09 0.85 3.85 1127.32 361.49 28.34 18.23Stand. dev. 3.79E�04 0.15 0.69 2.92 1395.00 481.30 89.84 21.37

Gas concentrationMinimum 0 0.09 16.49 61.74 3.80E�03 0 0 0Maximum 9.05 572.67 20.45 78.32 7.26 165.51 2.62 33,767Mean 3.52 150.32 20.54 75.63 3.89 85.10 0.52 5504.44Stand. dev. 2.22 987.04 0.96 2.87 2.25 54.65 0.73 8608.65

Fig. 7. Contour maps of CO2 fluxes at the Phlegraean Fields Caldera. The results from two surveys performed in different years (June 2002 on the left and November 2006 onthe right) and seasons highlights that the highest UCO2 values are always within an area bordered by faults and fractures.

1344 N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346

ban Hills is high temperature-derived, being a mix to different ex-tents between a magmatic source and the decarbonation processesaffecting the deep carbonate basement (Giggenbach et al., 1988;Chiodini and Frondini, 2001; Annunziatellis et al., 2003; Mariucciet al., 2008).

The Tor Caldara area was investigated during June, 2007 in theframework of ‘‘Diffuse Degassing in Italy” risk assessment project(funded by the Civil Protection Department) by means of a detailedsoil-gas survey and, simultaneously, UCO2 measurements.

Some statistical parameters of the analysed soil gas species arereported in Table 3. The studied gases generally show a uniformdistribution and low concentrations although the high CO2, CH4

and H2 standard deviation and the strong difference between meanand median highlight the ‘‘spot” distribution of some anomalies inthe data set. Since the highest CH4 values have been found in themain degassing sector where UCO2 is also extremely high, C isoto-pic analysis of the CH4 (in the dry gas phase) has been carried outin order to understand its origin. The result (�22.66 ‰ vs. PDB) iscomparable with those characterising the CH4 discharging in themain gas vents of central Italy (northern Latium and Tuscany, Min-issale et al., 1997). In particular, data point to a high temperaturebiogenic origin of CH4 (i.e. thermogenic, as defined by Welhan,

1988), ruling out both magmatic and/or biological/bacterial (i.e.at low temperature) production. Hydrogen anomalies (only 1% ofthe whole values) are strictly related to the highest CH4 valuesshowing a similar spatial and relative concentration distribution.

Most of CO2 soil-gas concentration and flux anomalies are in themain degassing sector of the investigated area but they do notoverlap. Although CO2 soil-gas concentration and flux have thesame statistical distribution (Fig. 8) differing just in the anomalousvalues, spatially they have a different distribution: Fig. 9 showstwo maps representing UCO2 and CO2 soil-gas concentrations.The local trends are very similar but soil-gas concentration showsa more diffusive distribution. Furthermore, flux distribution isshifted towards the south suggesting the presence of preferentialmigration pathways such as faults and fracture networks.

The d13C value of CO2 in the anomalous degassing areas isaround �0.10‰ vs. PDB (Mariucci et al., 2008), strongly supportingthe involving of carbonates in the production of the observed CO2

degassing.The Tor Caldara chemical composition of emitted gases reflects

the deep origin degassing characterising the Alban Hill volcanic areaalthough different tectonic structures/gas traps along the buriedfault systems affect the final distribution observed at the surface.

Table 3Some statistical parameters for all the analysed soil gas species and CO2 flux at Tor Caldara natural reserve.

He (ppm, v/v) H2 (ppm, v/v) O2 (%, v/v) N2 (%, v/v) CH4 (ppm, v/v) CO2 (%, v/v) Rn (Bq/m3) UCO2 (g/m2 day)

Minimum 4.44 0.11 15.29 61.83 0.14 0.10 100 1.39Maximum 15.20 6582.77 20.45 78.24 197.44 23.46 12,900 31746.66Mean 5.39 178.44 19.63 76.53 8.68 2.21 4097 1079.26Median 5.03 0.80 19.90 77.28 2.41 0.88 3640 16.33Stand. dev. 1.72 1027.04 0.99 2.88 30.98 4.19 3450 4547.03

-3 -2 -1 0 1 2 3 4Expected Normal Value

-5

0

5

10

15

20

25

30

35

CO

2 (%

, v/v

)

ΦC

O2

(g/m

2da

y)

ΦCO2 (g/m2day) CO2(%, v/v)

Outliers

Background values

Anomalous values

0

5000

10000

15000

20000

25000

30000

35000

Fig. 8. Normal probability plot for both CO2 concentration and flux. Note that CO2

concentration and flux have almost the same statistical distribution. For both, it ispossible to distinguish three populations: background, anomalous values andoutliers (the latter are more numerous for UCO2) but from an analysis of the dataset, the highest soil gas CO2 anomalous values do not correspond to the highestUCO2 values.

N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346 1345

4. Conclusions

The potential risks of geological CO2 storage must be under-stood and geologists are required to predict how CO2 may behaveonce stored underground. As natural geological accumulations ofCO2 occur in many basins in Italy, considering that volcanic andseismically active areas allow CO2-rich fluids to migrate to the nearsurface, three of these areas have been investigated. They representinteresting natural analogues for the study and prediction of thepossible consequences of gas leakage from geological sequestra-tion sites (i.e., the return of gas to surface, potentially causing local-ised environmental problems).

Fig. 9. Contour maps of CO2 soil-gas concentration (on the right) and flux (on the left).distribution and flux distribution is shifted towards the south suggesting the presence o

Furthermore, the study of the use of leaking sites could help todevelop, test and optimize various monitoring technologies. In thisframework, the use of soil gas and flux techniques has been shownto be an useful and very simple method that can be well adaptedboth to site-assessment prior to installation of a CO2 sequestrationsite (in order to look for gas permeable structures) and for eventualmonitoring of the site once injection begins.

The three studied Italian sites, representing different geologicalscenarios, are linked to migration of significant quantities of CO2

towards the surface. At present the above mentioned areas, in par-ticular the Phlegraean Field Caldera, are monitored in order tostudy the temporal evolution of different phenomena. The studiedareas show evident consequences of the CO2 release at the surface:both the Phlegraean Field and Tor Caldara areas are characterisedby the absence of vegetation in the degassing zones while the mar-ine environment at Panarea Island underwent a drastic ecosystemchange (that caused the death of mainly benthonic life forms andserious damage to the sea-grass Posidonia oceanica) during thegas burst which occurred in November, 2002.

The aim of the research was to study CO2 leaks and their behav-iour in time and space. To date, the obtained results suggest thatgas uprising is much localised around restricted areas, often con-trolled by local tectonics (faults and/or fractures). Generally speak-ing, the studied temporal gas concentration variations can beexplained by considering different phenomena: a long period ofaccumulation with high fluxing of the gas from the buried reser-voir/magmatic chamber, a change in the confining pressure inthe fault domain, the sudden gas release triggered by seismic activ-ity, the rapid rise of gas to the surface along more permeable path-ways. All this implies that, in the framework of geological CO2

sequestration, it is necessary to carefully study gas permeabilityzones taking also into account faults and/or fractures that can beenhanced by seismic activity.

Furthermore, the results suggest that surface monitoring of aclosed geological repository may be carried out for the risk

The local trends are very similar but soil-gas concentrations show a more diffusivef preferential migration pathways such as faults and fracture networks.

1346 N. Voltattorni et al. / Applied Geochemistry 24 (2009) 1339–1346

assessment of the potential and nature of CO2 migration alongpreferential pathways.

Acknowledgments

Funding provided by the Civil Protection Department (‘‘DiffuseDegassing in Italy” risk assessment project, V5-UR 11 and ‘‘Researchon active volcanoes, precursors, scenario, hazard and risks”, pro-ject, V3-1-Colli Albani – Task2), is gratefully acknowledged. Com-ments and suggestions of the two reviewers (Dr. J. Klerks and Dr.J. Lewicki) are greatly appreciated. The authors sincerely thankDr Cantucci B., Dr Procesi M. and Mr. Piccolini L. for their help dur-ing field work: without their ‘‘sacrifices” this paper would not havebeen written.

References

Annunziatellis, A., Ciotoli, G., Lombardi, S., Nolasco, F., 2003. Short- and long-termgas hazard: the release of toxic gases in the Alban Hills volcanic area (centralItaly). J. Geochem. Explor. 77, 93–108.

Austvik, T., Loken, K.P., 1992. Deposition of CO2 on the seabed in the form ofhydrates. Energy Convers. Manage. 33, 659–666.

Barca, D., Ventura, G., 1991. Evoluzione vulcano-tettonica dell’isola di Salina(Arcipelago delle Eolie). Mem. Soc. Geol. Ital. 47, 401–415.

Bertrami, R., Buonasorte, G., Ceccarelli, A., Lombardi, S., Pieri, S., Scandiffio, G., 1990.Soil gases in geothermal prospecting: two case histories (Sabatini Volcanoesand Alban Hills, Latium, Central Italy). J. Geophys. Res. 95, 21475–21481.

Caliro, S., Caracausi, A., Chiodini, G., Ditta, M., Italiano, F., Longo, M., Monopoli, C.,Nuccio, P.M., Paonita, A., Rizzo, A., 2004. Evidence of a recent input of magmaticgases into quiescient volcanic edifice of PANAREA, Aeolian Islands, Italy.Geophys. Res. Lett. 31, L07619. doi:10.1029/2003GL019359.

Capaccioni, B., Tassi, F., Vaselli, O., Tedesco, D., 2007. Submarine gas burst at PanareaIsland (southern Italy) on 3 November 2002: A magmatic versus hydrothermalepisode. J. Geophys. Res. 112, B05201. doi:10.1029/2006JB0044359.

Capasso, G., Inguaggiato, S., 1998. A simple method for the determination ofdissolved gases in natural waters: an application to thermal waters fromVolcano Island. Appl. Geochem. 13, 631–642.

Caramanna, G., Voltattorni, N., Caramanna, L., Cinti, D., Galli, G., Pizzino, L.,Quattrocchi, F., 2005. Scientific diving techniques applied to thegeomorphological and geochemical study of some submarine volcanic gasvents (Aeolian Islands, southern Tyrrhenian Sea, Italy). In: Proceedings of 24thDiving for Science Symposium at American Academy of Underwater sciences11–12 March 2005, Mystic, Connecticut (USA).

Carapezza, M.L., Badalamenti, B., Cavarra, L., Scalzo, A., 2003. Gas hazard assessmentin a densely inhabitated area of Colli Albano volcano (Cava dei Selci, Roma). J.Volc. Geotherm. Res. 123, 81–94.

Chiarabba, C., Malagnini, L., Amato, A., 1994. Three-dimentional velocity structureand earthquake relocation in the Alban Hills Volcano, Central Italy. Bull. Seism.Soc. Am. 84, 295–306.

Chiarabba, C., Amato, A., Delaney, P.T., 1997. Crustal structure, evolution, andvolcanic unrest of the Alban Hills, Central Italy. Bull. Volcanol. 59, 161–170.

Chiodini, G., Frondini, F., 2001. Carbon dioxide degassing from the Albani Hillsvolcanic region, Central Italy. Chem. Geol. 177, 67–83.

Chiodini, G., Frondini, F., Ponziani, F., 1995. Deep structures and carbon dioxidedegassing in Central Italy. Geothermics 24, 81–94.

Chiodini, G., Cioni, R., Guidi, M., Raco, B., Marini, L., 1998. Soil CO2 fluxmeasurements in volcanic and geothermal areas. Appl. Geochem. 13, 543–552.

Chiodini, G., Frondini, F., Cardellini, C., Parello, C., Peruzzi, L., 2000. Rate of diffusecarbon dioxide Earth degassing estimated from carbon balance of regionalaquifers: the case of central Apennines, Italy. J. Geophys. Res. 105, 8423–8434.

Chiodini, G., Frondini, F., Cardellini, C., Granirei, D., Marini, L., Ventura, G., 2001. CO2

degassing and energy release at Solfatara volcano, Campi Flegrei, Italy. J.Geophys. Res. 106, 16216–16221.

Dando, P.R., Stuben, D., Varnavas, S.P., 1999. Hydrothermalism in the MediterraneanSea. Prog. Oceanogr. 44, 333–367.

De Rita, D., Funiciello, R., Rosa, C., 1992. Volcanic activity and drainage networkevolution of the Colli Albani area. Acta Vulcanologica 2, 185–198.

De Rita, D., Faccenna, C., Funiciello, R., Rosa, C., 1995. Stratigraphy and volcano-tectonics. In: Trigila, R. (Ed.), The Volcano of the Alban Hills, Rome, Italy, pp. 33–71.

Di Filippo, M., Toro, B., 1995. Gravity features. In: Trigila, R. (Ed.), The Volcano of theAlban Hills, Rome, Italy, pp. 213–219.

Faccenna, C., Funiciello, R., Mattei, M., 1994. Late Pleistocene N–S shear zone alongLatium Thyrrenian margin: structural characters and volcanologicalimplications. Boll. Geof. Teor. Appl. 36, 141–144.

Funiciello, R., Gasparini, C., Montone, P., Parotto, M., Salvini, F., 1987. Il campo difatturazione superficiale dei Colli Albani (Roma): implicazioni per un modellosismotettonico dell’area. Bollettino GNV, 373–384.

Gabbianelli, G., Gillot, P.V., Lanzafame, G., Ligi, M., Postpischl, D., Rossi, P.L., 1986.Controllo strutturale nell’evoluzione vulcanica di Panarea (Isole Eolie). ItalianNational Res. Council Internal Report, pubbl. n� 239, Catania, Italy.

Gabbianelli, G., Gillot, P.V., Lanzafame, G., Romagnoli, C., Rossi, P.L., 1990. Tectonicand volcanic evolution of Panarea (Aeolian Islands, Italy). Mar. Geol. 92, 313–326.

Gasparini, P., Iannaccone, G., Scandone, P., Scarpa, R., 1982. The seismotectonic ofthe Calabrian Arc. Tectonophysics 84, 267–287.

Giggenbach, W.F., 1975. A simple method for the collection and analysis of volcanicgas samples. Bull. Volcanol. 39, 15–27.

Giggenbach, W.F., Minissale, A.A., Scandiffio, G., 1988. Isotopic and chemicalassessment of geothermal potential of the Colli Albani area, Latium region, Italy.Appl. Geochem. 3, 475–486.

Golomb, D.S., Zemba, S.G., Dacey, J.W.H., Michaels, A.F., 1992. The fate of CO2

sequestered in deep ocean. Energy Convers. Manage. 33, 675–683.Holloway, S. 1996. The Underground Disposal of Carbon Dioxide Final Report for the

CEC of Contract Number JOU2-CT92-0031. British Geological Survey.Hutchinson, G.L., Livingston, G.P., 1993. Use of chamber systems to measure trace

gas fluxes. American Society of Agronomy, 55 (special publication).Hyppolite, J., Angelier, J., Roure, F., 1994. A major change revealed by quaternary

stress patterns in the southern Apennines. Tectonophysics 230, 199–210.Italiano, F., Nuccio, P.M., 1991. Geochemical investigations of submarine volcanic

exhalations to the east of Panarea, Aeolian Islands, Italy. J. Volcan. Geotherm.Res. 46, 1–185.

Koide, H., Takahashi, M., Shindo, Y., Tazaki, Y., Iijima, M., Ito, K., Kimura, N., Omata,K., 1997. Hydrate formation in sediments in the sub-seabed disposal of CO2.Energy 22, 279–283.

Lanzafame, G., Rossi, P.L., 1984. Evidenze di attivita‘ tettonica recente a Panarea(Eolie): implicazioni vulcanologiche. Geologica Romana 23, 131–139.

Lombardi, S., Reimer, G.M., 1990. Radon and Helium in soil gases in the Phlegreanfields, central Italy. Geophys. Res. Lett. 17, 849–852.

Lombardi, S., Annunziatellis, A., Beaubien, S.E., Ciotoli, G., 2006. Near-surface gasgeochemistry techniques to assess and monitor CO2 geological sequestrationsites. In: Lombardi, S., Altunina, L.K., Beaubien, S.E. (Eds.), Advances in theGeological Storage of Carbon Dioxide NATO Science Series, vol. 65. SpringerPublishing, Berlin, pp. 141–156.

Marchetti, C., 1977. On geoengineering and the CO2 problem. Climate Change 1, 59–68.

Mariucci, M.T., Pierdominici, S., Pizzino, L., Marra, F., Montone, P., 2008. Looking intoa volcanic area: an overview on the 350 m scientific drilling at Colli Albani(Rome, Italy). J. Volcanol. Geotherm. Res. 176, 225–240.

Marra, F., 1999. Low-magnitude earthquakes in Rome, structural interpretation andimplications for the local stress-field. Geoph. J. Internat. 138, 231–243.

Marra, F., 2001. Strike-slip faulting and block rotation: a possible triggeringmechanism for lava flows in the Alban Hills? J. Structural Geol. 23, 127–141.

Matthias, A.D., Blackmer, A.M., Bremner, J.M., 1980. A simple chamber technique forfield measurements of emissions of nitrous oxide from soils. J. Environ. Qual. 9,251–256.

Minissale, A., Ewans, W.C., Magro, G., Vaselli, O., 1997. Multiple source componentsin gas manifestations from north-central Italy. Chem. Geol. 142, 175–192.

Orsi, G., De Vita, S., Di Vito, M., 1996. The restless, resurgent Campi Flegrei nestedcaldera (Italy): constrains on its evolution and configuration. J. Volcanol.Geotherm. Res. 74, 179–214.

Pearce, J.M., 2006. What can we learn from natural analogues? An overview of howanalogues can benefit the geological storage of CO2. In: Lombardi, S., Altunina,L.K., Beaubien, S.E. (Eds.), Advances in the Geological Storage of Carbon Dioxide.NATO Science Series, vol. 65. Springer Publishing, Berlin, pp. 129–138.

Pizzino, L., Galli, G., Mancini, C., Quattrocchi, F., Scarlato, P., 2002. Natural gas hazard(CO2, 222 Rn) within a quiescent volcanic region and its relations with tectonics:the case of the Ciampino-Marino area. Nat. Hazards 27, 257–287.

Quattrocchi, F., Capelli, G., De Rita, D., Faccenna, C., Funiciello, R., Galli, G., Giordano,G., Goletto, D., Mancini, C., Mazza, R., Pizzino, L., 2001. The Ardea Basin fluidgeochemistry, hydrogeology, and structural patterns: new insights about thegeothermal unrest activity of the Alban Hills quiescent volcano (Rome, Italy)and its geochemical hazard surveillance. In: Cidu, R. (Ed.), Proceedings of 10thInternational Conference (WRI-10). Water–Rock Interaction, June 2001.Balkema, Villasimius, Italy, pp. 111–114.

Rosi, M., Santacroce, R., 1984. Volcanic hazard assessment in the Phlegrean Fields: acontribution based on stratigraphic and hysterical data. Bull. Volcanol. 47, 359–370.

Toro, B., 1976. Gravimetry and deep structure of the Sabatinian and Alban volcanicareas (Latium). Geologica Romana 15, 301–310.

Tuccimei, P., Soligo, M., 2008. Correcting for CO2 interference in soil radon fluxmeasurements. Radiat. Meas. 43, 102–105.

Tuccimei, P., Giordano, G., Tedeschi, M., 2006. CO2 release variations during the last2000 years at the Colli Albani volcano (Roma, Italy) from speleothems studies.Earth Planet. Sci. Lett. 243, 449–462.

Voltattorni, N., Caramanna, G., Cinti, D., Galli, G., Pizzino, L., Quattrocchi, F., 2006.Study of natural CO2emissions in different Italian geological scenarios: arefinement of natural hazard and risk assessment. In: Lombardi, S., Altunina,L.K., Beaubien, S.E. (Eds.), Advances in the Geological Storage of Carbon DioxideNATO Science Series, vol. 65. Springer Publishing, Berlin, pp. 75–190.

Welhan, J.A., 1988. Origins of methane in hydrothermal systems. Chem. Geol. 71,183–198.

Wilson, T.R.S., 1992. The deep ocean disposal of carbon dioxide. Energy Convers.Manage. 33, 627–633.

Yezdimer, E.M., Cummings, P.T., Chialvo, A.A., 2002. Determination of the Gibbs freeenergy of gas replacement in sI clathrate hydrates by molecular simulation. J.Phys. Chem. A 106, 7982–7987.