Journal of Volcanology and Geothermal Research 165 (2007) 17 31

www.elsevier.com/locate/jvolgeores

Gas blowout from shallow boreholes at Fiumicino (Rome): Induced

hazard and evidence of deep CO2 degassing on the Tyrrhenian
margin of Central Italy
F. Barberi a,, M.L. Carapezza b , M. Ranaldi a , L. Tarchini a
a
b

Dipartimento Scienze Geologiche, Universit Roma Tre, L.go S. Leonardo Murialdo 1, 00146 Roma, Italy
Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma 1, Via Vigna Murata 605, 00143 Roma, Italy
Received 12 October 2006; accepted 23 April 2007
Available online 17 May 2007

Abstract
A borehole drilled at Fiumicino (Rome) down to only 27 m depth in a zone where no gas emission at the surface was known, caused a
gas blowout from a pressurized gas pocket confined beneath a clay cover. Gas slowly diffused from the borehole within superficial
permeable sand. Seven persons living in three ground floor flats of a near building had to be hospitalized due to CO2 exposure. All the
houses in the proximity were evacuated. At the request of the Fire Brigade two additional boreholes were drilled nearby, hoping that this
could rapidly exhaust the gas stored underground. To the contrary the soil gas flux near houses increased and indoor CO2 air concentration
rapidly rose to lethal values (15 to 30 vol.%). As a remediation we suggested to restore the continuity of the impervious gas cover by
squeezing quick-setting cement into the formation through new boreholes to be drilled near the existing ones. Although the first cement
squeeze reduced drastically the CO2 soil flux and indoor concentration, six additional squeezes had to be carried out in order to lower the gas
emission below the gas hazard threshold. The gas was mostly made of CO2 (98 vol.%) with minor N2 and CH4. Its chemical and isotopic
composition (13CCO2 = 1.55; 3He/4He =0.314 Ra) is similar to that of the gas manifestations of Mts. Sabatini and Alban Hills volcanic
areas. Though being somewhat contaminated by crustal and shallow organic volatiles, these gases likely have a component originated in the
mantle, that beneath the volcanic Roman Comagmatic Province is probably deeply contaminated with crustal material. The Fiumicino gas
blowout indicates that the area of Central Italy characterized by strong CO2 degassing extends westerly to include the Tyrrhenian coast.
2007 Published by Elsevier B.V.
Keywords: gas blowout; CO2 hazard; gas monitoring; remediation interventions; Earth CO2 degassing in Central Italy

1. Introduction
Central Italy, North and South of Rome, is characterized by the presence of two Quaternary volcanoes (e.g.
Mts. Sabatini and Alban Hills, Fig. 1). This area, as the
entire Tyrrhenian hinterland, has a thinned continental

Corresponding author. Tel.: +39 06 54888041.

E-mail address: barberi@uniroma3.it (F. Barberi).
0377-0273/$ - see front matter 2007 Published by Elsevier B.V.
doi:10.1016/j.jvolgeores.2007.04.009

crust (2025 km) and high heat fluxes (N 80 mW m 2)

(Gambardella et al., 2004). It is also characterized by a
very strong degassing of CO2 of deep provenance, as
estimated from the balance and the 13C of carbon
dissolved in regional aquifers (Chiodini et al., 2000; 2004;
Gambardella et al., 2004 and references therein). It has
also been suggested that the Apennine seismicity could be
driven by high pressure CO2 sources at depth, i.e. crustal
traps with overpressurized CO2 reservoirs (Miller et al.,
2004; Chiodini et al., 2004).

18

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

Fig. 1. Geological sketch-map of the Tyrrhenian margin of Central Italy (modified after De Rita et al., 1995). The extension of the volcanic rocks of
the Roman Comagmatic Province (RCP) and the location of the Fiumicino borehole are indicated.

The carbon dioxide rising from depth, whose origin is

controversial, dissolves into aquifers, geothermal or not,
hosted in buried Mesozoic limestones and in shallower
cold aquifers hosted either in Neogene clastic sediments
and in the Quaternary volcanic rocks. All these aquifers
release CO2-rich gases toward the surface mostly along
extensional fractures and faults, originating many
discrete gas manifestations or zones of high CO2 diffuse

emission from the soil. The quantity of CO2 released into

the atmosphere is locally so high to represent a serious
hazard to people and animals (Rogie et al., 2000;
Chiodini and Frondini, 2001; Carapezza et al., 2003;
Carapezza et al., 2005). In zones where there is no
significant CO2 release at the surface because of the
presence of an efficient cover of impervious rocks, the
existence at various depth of pressurized CO2 pockets has

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

been revealed by accidental and dangerous gas blowout

during drilling (Carapezza and Tarchini, 2007).
In February 2005 a similar accident, that only by chance
did not cause human casualties, occurred at Fiumicino, on
the delta of Tiber river (Fig. 1), in a zone where no surface
gas emission was known. The aim of this paper is twofold:
i) to describe this accident, its related hazard and the
activity carried out to control the gas emission and to
restore safety conditions and ii) to use chemical and
isotopic data of the emitted gas to show its endogenous
deep origin, so enlarging the surface of Central Italy with
evidence of strong CO2 Earth degassing.
2. Geological setting
Fiumicino is located at the mouth of Tiber river, SW
of Rome (Fig. 1). Since MiddleLate Pliocene active
extensional tectonic processes, with main fractures
oriented NWSE and NESW, controlled the evolution
of the Tyrrhenian margin of Central Italy, generating
structural highs and lows cut in Mesozoic limestones
and their allochthonous flysch cover (Di Filippo and
Toro, 1995). An extensive Quaternary alkalinepotassic
volcanism characterizes the region, forming the so-

19

called Roman Comagmatic Province originated by the

subduction of the Ionian Adriatic Plate (Peccerillo,
1985; Doglioni et al., 1999). Volcanic deposits of Mts.
Sabatini and Alban Hills crop out respectively N and E
of Fiumicino (Fig. 1). The Tiber delta plain is a late
PleistoceneHolocene succession which extends along
the coast of the Tyrrhenian sea for 30 km and inland for
13 km (Bellotti et al., 1995). In the zone of Fiumicino,
where the borehole was drilled, the outer delta succession, overlying a 0.8 My unconformity surface that cuts
Pleistocene marine sediments, is made up of 1516 m of
fine sands and silts of the present delta front, lying over
a N 25 m lithofacies of bluegrey marine shales with
intercalation of fine sands and sandy silts. The inner
delta sediments are typical of fluvial and lagoonal
depositional systems with levels of peat and peaty mud
(Bellotti et al., 1995).
3. The Fiumicino accidental gas blowout and the
remediation intervention
The borehole (F1 in Fig. 2) was located near an electric
cabin. Drilling started on 9 February 2005 at 9 a.m.,
to install a conductive ground electrode inside a clay level

Fig. 2. Map of the Fiumicino zone affected by the gas blowout of February 2005. The sites of the gas emitting boreholes (F) and of the wells where
cement squeezing was carried out (P) are indicated, as well as the limits of the investigated zones for CO2 flux.

20

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

Table 1
Health effects of respiratory exposure to carbon dioxide
Exposure limits Health effects
(% in air)
23
3
35
5
7.5
815
10
15
25+

Unnoticed at rest, but on exertion there may be marked shortness of breath

Breathing becomes noticeably deeper and more frequent at rest
Breathing rhythm accelerates. Repeated exposure provokes headaches
Breathing becomes extremely laboured, headaches, sweating and bounding pulse
Rapid breathing, increased heart rate, headaches, sweating, dizziness, shortness of breath, muscular weakness, loss of mental
abilities, drowsiness, and ringing in the ears
Headache, vertigo, vomiting, loss of consciousness and possibly death if the patient is not immediately given oxygen
Respiratory distress develops rapidly with loss of consciousness in 1015 min
Lethal concentration, exposure to levels above this are intolerable
Convulsions occur and rapid loss of consciousness ensues after a few breaths. Death will occur if level is maintained.

After www.ivhhn.org/gas/guidelines.htm.

them suffered loss of consciousness. These symptoms

are typical of exposure to CO2 air concentration from 5
to 15 vol.% (see Table 1). On the following morning the
Fire Brigade intervened. Carbon dioxide concentrations
exceeding the official safety threshold of 0.5% were
measured indoor and all the neighbouring houses were
evacuated. At the request of the Fire Brigade two new
boreholes (F2 and F3 in Fig. 2) were drilled at a depth
of 2728 m close to the first one, hoping that this could
rapidly exhaust the gas stored underground. On the
contrary, gas blowout (with no water) occurred also
from the new wells and indoor concentration continued
to increase. In the late afternoon of 11 February we were
called to give a scientific appraisal and to help in
solving the problem. At 8 p.m. we performed the first
CO2 flux measurements from the boreholes and from
the soil, and of CO2 indoor concentration. These were
extended in the following days and it was clear from the
results of the soil CO2 flux survey (see Section 3.1.2)
that the gas was preferentially diffusing from the F1F3

to be found underneath sandy superficial layers. At 2 p.m.,

at about 27 m depth, an abrupt decrease of the electrical
resistivity from 25 to 0.5 m was observed, and the
drillers presumed to have reached the clay. Drilling
continued for another meter, and then a 2 steel pipe
was installed inside the borehole. At 3 p.m. the gas
blowout occurred and a mixed jet of water, gas and clay
was emitted that reached the near houses (1, 2, 3 in
Fig. 2). The pipe was sticked for further 2 m and water
and clay was poured inside it. The cavity between
borehole and pipe was filled with sand, and the gas
emission apparently disappeared. The drillers presumed
to have closed the borehole and eliminated the gas
emission problem. On the morning of the following
day (10 February) the situation appeared similar and
the drilling rig was removed. During the night between
1011 February, the inhabitants of the near houses
(seven persons) requested medical assistance and were
hospitalized due to strong headache, sickness, dizziness, shortness of breath and pulse acceleration; two of

Table 2
Chronogram of the cement squeezing operations and cement characteristics
Operation

Date

Location

Depth

Cement

no.

(dd.mm, hh)

(see Fig. 2)

Borehole
(m)

Casing
(m)

Vol
(m3)

d(kg/l)

Pump P
(bar)

P1

21.02, 17

0.7 m from F2

30

28.5

1.60

1.52

P2
P3
P4
P5
P6
P7

23.02, 11
27.02, 10
28.02, 11
04.03, 08
19.03, 11
30.03, 17

Same as above
0.8 m from F3
F1
0.7 m from F1
0.8 m from F1
F1

30
27.2
30
30
27
27

28.5

3
6
3
7
5.3
1.4

1.70
1.70
1.70
1.77
1.77
1.70

23
23
1.53.5
22.5

22
21.3

Notes

Half of the injected cement poured out

from F2
Second squeezing into the same borehole
No cement lost; clay found at 1618 m depth
Squeezing through the 2 pipe
Clay found at 1618 m depth
Gas blocking additive introduced in cement
F1 redrilled; consolidated cement found from
6 to 12 m, pebbles at 15 m, clay at 16 m
depth

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

21

Fig. 3. Variations in the CO2 emission measured from 11 to 25 February 2005 from the three boreholes. F1 scale on the left axis; F2F3 scale on the
right axis. The date of the first cement squeeze P1 is indicated by the arrow.

boreholes to the houses, flowing within the superficial

high-permeability sands.
From 19 February to 30 March, at our request, four
new boreholes were drilled (see Fig. 2 for location) and
F1 was redrilled to perform cement squeezing inside the
formation to restore the continuity of the impervious
layer and thus to re-confine the gas underground. Seven
squeezing operations with different cement volumes and
characteristics (Table 2) were necessary from these
boreholes to restore the natural conditions and to eliminate the gas hazard inside the houses, allowing their
inhabitants to return after the evacuation.

During this period, in order to monitor the level of

the gas hazard and to check the effectiveness of the
cement squeezings, CO2 soil flux surveys were carried
out almost daily and a continuous automatic station for
the CO2 soil flux was installed in a flowerbed near
houses (see Fig. 2 for location). Indoor CO2 air concentration was measured frequently in the various flats
and in all available manholes and mancovers near
houses and on the road, in the electric cabin and in a
cellar. CO2 air concentration was also measured continuously for long periods within the evacuated ground
floor flats.

Fig. 4. Contour map of CO2 soil flux of the 15 February survey carried out before the first cement squeezing. The gas was flowing from the boreholes
towards the houses through the superficial porous sands; the anomalous degassing area had a surface of 2900m2.

22

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

3.1. Level of the gas emission before and after the

cement squeezing operations
3.1.1. CO2 emission from the boreholes
Field quick analyses by portable IR Drger (X-am
7000) revealed that the gas consisted mostly of CO2
with minor CH4 and a few ppm of H2S. The level of the
CO2 emission was estimated by the gradient of CO2
concentration increase within an accumulation chamber
placed above the open holes. This is a very gross
estimate of the gas flux, that served only to control the
changes in the gas emission level induced by the cement
squeezing. Results (in ppm/s) are shown in Fig. 3. The
strong decrease of the gas emission observed in
borehole F2 from 11 to 12 February was caused by
the partial collapse of the uncased well, that on 12
February resulted obstructed at 2.5 m depth; however, a
significant gas emission from F2 resumed in the
following days (Fig. 3). A strong fluctuation in the
level of CO2 emission from the boreholes was observed,
but without any recognizable decreasing trend (Fig. 3).
Minimum values of the CO2 emission were of about
1300 ppm/s at borehole F1, of 58,000 ppm/s at F2 and
of 44,000 ppm/s at F3. Gas emission from F1F3
boreholes completely ceased after the first cement
squeezing carried out on 21 February from well P1.
Actually about 3 m3 of the larry injected into P1 poured
out from the near F2 borehole and part of it was driven
to flow into well F3. In addition, as a consequence of the
partial loss of circulation during P1 drilling, no further
gas emission from F1 was observed.
3.1.2. CO2 soil flux
From 11 to 21 February, before the cement squeezing
operations, eleven CO2 soil flux surveys were carried out
with different point density and distribution, using a
portable accumulation chamber (see Chiodini et al.,
1998; Carapezza and Granieri, 2004 for the description
of the device). The most complete survey was performed
on 15 February. Starting from the boreholes, the CO2 soil
flux was surveyed in all possible directions until low
CO2 flux values were found. Unfortunately the presence
of many buildings, walking pavements and paved roads
(Fig. 2) prevented to cover the area with a regular grid
of measuring points. The CO2 soil flux was measured
on 61 points and a wide range of values was found from
2 g/m2 day to 31,370 g/m2 day with an average of
2026 g/m2 day. The statistical treatment of all data
collected in the various surveys indicates that the flux
background had a reasonable maximum threshold of
40 g/m2 day. In the survey of 15 February, 39% of the
values were within the background. All CO2 fluxes ex-

ceeding this threshold (58% of the measures) were considered as anomalous values generated by the gas
diffusion from the boreholes. The flux map pictured by
Kriging (Fig. 4) clearly indicates that the gas preferentially
Table 3
Main results of the CO2 soil flux surveys (in g/m2 day) before and after
cement squeezings a
Date

Meas. No.

11-Feb-05
12-Feb-05
13-Feb-05
14-Feb-05
15-Feb-05
16-Feb-05
17-Feb-05
18-Feb-05
19-Feb-05
20-Feb-05
21-Feb-05
21-Feb-05
22-Feb-05
23-Feb-05
23-Feb-05
24-Feb-05
25-Feb-05
26-Feb-05
27-Feb-05
27-Feb-05
28-Feb-05
28-Feb-05
01-Mar-05
02-Mar-05
03-Mar-05
4-Mar-05
05-Mar-05
06-Mar-05
07-Mar-05
08-Mar-05
09-Mar-05
10-Mar-05
14-Mar-05
15-Mar-05
16-Mar-05
17-Mar-05
18-Mar-05
19-Mar-05
19-Mar-05
21-Mar-05
22-Mar-05
23-Mar-05
25-Mar-05
29-Mar-05
30-Mar-05
30-Mar-05
01-Apr-05
19-Apr-05

16
3650
26
3128
26
3519
6
10,976
61
2027
26
4253
28
2968
30
4341
24
6497
13
6221
7
9465
1st cement squeezing in P1
29
306
29
269
2nd cement squeezing in P1
34
55
35
64
32
117
30
121
3rd cement squeezing in P2
30
108
4th squeezing in F1 borehole (P3)
54
85
32
90
32
86
5th cement squeezing in P4
32
69
22
36
32
51
32
53
32
55
32
52
32
54
32
52
31
55
29
44
29
50
30
44
6th cement squeezing in P5
30
42
30
40
18
65
18
46
30
24
13
52
7th squeezing in redrilled F1 (P6)
13
64
54
31

Mean value

Maximum
32,165
34,104
25,931
30,378
31,371
37,385
25,435
46,924
39,748
35,025
37,496
1174
2823
140
477
680
434
275
557
335
290
252
109
151
206
167
182
213
211
235
143
269
229
194
154
237
127
84
125
159
139

See Fig. 2 for the limits of the surveyed area and the location of
cement squeezing operations.

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

diffused within the superficial porous sands from the

boreholes eastward towards the houses; this was clearly a
direction with higher soil permeability, possibly also
because of the many earth-moving works that had been
carried out in that zone. The total CO2 flux was estimated
to be about 9 t/day from an anomalous degassing surface
of 2900 m2. After 15 February and before the cement
squeezing, a strong CO2 soil flux persisted on a wide
surface around the houses, with mean values always

23

higher than 2000 g/m2 day up to nearly 11,000 g/m2 day

and a maximum of 47,000 g/m2 day (Table 3).
On 22 February, a day after the first cement squeezing
from the newly drilled P1 well, the situation had completely changed (Table 3): no CO2 flux values higher
than 3000 g/m2 day were anymore found and the mean
flux had lowered from 2000 to 300 g/m2 day. This positive result and the persistence of anomalous, though reduced, CO2 flux values suggested to continue the cement

Fig. 5. Results of the CO2 soil flux surveys carried out in the fixed dense grid near the boreholes (see Fig. 2 for grid location).

24

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

squeezing operations. Data from Table 3 show that

following the various cement squeezings in new wells or
in the redrilled F1 borehole, the CO2 soil flux continued
progressively to lower, particularly after the 4 March
intervention, when mean values lowered below 70 g/m2
day. A survey carried out on 19 April, 20 days after the
last cement squeezing, showed that the CO2 soil flux had
lowered to a mean value of 31 g/m2 day (Table 3),
indicating that the natural degassing background of the
area had been nearly reestablished (74% of values within
the natural background).
In order to better evaluate the effectiveness of the
cement squeezing operations in reducing the gas release
at the surface, a dense network of fixed points for CO2
flux measurements was established in the zone between
the initial boreholes and the road (Fig. 2). This dense
network was surveyed several times from 6 March to 26
April. Also in this case a progressive reduction of the
CO2 soil flux was observed and only a few points around
50 g/m2 day (with average of 28 g/m2 day) persisted after
the completion of the cement squeezing (Fig. 5).
The CO2 soil flux was also continuously monitored
since 15 February by an automatic station placed in a
flowerbed in front of the houses, where high flux values
had been measured (see Fig. 2 for location). The CO2 soil
flux was measured by an accumulation chamber every
30, together with environmental parameters (soil T and

humidity, atmospheric P and T, wind speed and

direction). CO2 soil flux is actually affected by changes
in the environmental conditions such as atmospheric
pressure, wind velocity and soil moisture (Granieri et al.,
2003; Carapezza and Granieri, 2004). In the Fiumicino
case, the parameter that more strongly affected CO2 soil
flux was the soil moisture, also because several strong
rains occurred during the observation period that caused
a decrease of the soil permeability to gas, and hence of
the CO2 flux, because the soil pores were saturated with
water denser than the gas. In the insert of Fig. 6 the CO2
soil flux values recorded by the station are shown together with the soil moisture ones. The systematic strong
decrease of the gas flux when soil moisture increases can
be appreciated. For this reason, only flux values recorded
in dry conditions were considered indicative of the level
of actual degassing from the soil. The graph of Fig. 6
clearly indicates the strong reduction of the CO2 soil
flux produced by the first cement squeezing P1 of 21
February. Before the intervention the lowest flux values
were around 10,000 g/m2 day, the most frequent values
between 20,000 and 30,000 g/m2 day, with several
values exceeding 50,000 g/m2 day up to the maximum
of 110,000 g/m2 day. After the cement squeezing P1,
flux values lowered to below 550 g/m2 day, so that to
appreciate the results recorded in the following days, the
flux scale has to be modified (see insert in Fig. 6). The

Fig. 6. Temporal variation of CO2 soil flux monitored at the automatic station; note the strong reduction of CO2 flux after the first cement squeezing
P1 on 21 February (dashed vertical bar). Insert: variation from 21 February to 15 June of the 24h mobile average of CO2 flux and soil moisture.
Dashed vertical bars (P1P7) indicate the different cement squeezing operations. Note the different flux scale with respect to the main graph.

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

25

Fig. 7. Time variation of the daily maximum CO2 indoor concentration from 12 February to 31 May 2005. See Fig. 2 for the location of the
investigated sites. Note that safety conditions were reached once the cement squeezings restored the impervious cap.

CO2 soil flux values after the first squeezing were strongly
influenced by the rain; however, no further relevant
decrease of the maxima, that repeatedly approached
500 g/m2 day, was observed until the P4 cement squeezing of 4 March. This was carried out in the first well drilled
near F1 borehole (Fig. 2) and the CO2 soil flux maxima
recorded by the station lowered by more than half, to
about 200 g/m2 day. The following cement squeezing
operations furtherly decreased the CO2 soil flux that at the
end only rarely exceeded 50 g/m2 day (Fig. 6), with
dominant values around 3040 g/m2 day.
3.1.3. CO2 indoor air concentration
The indoor CO2 air concentration was initially
measured in all flats near the boreholes (Fig. 2) by
means of a portable IR Drger (X-am 7000). Successively the measurements were concentrated in three
ground floor flats and in other two rooms (the electric
cabin and an underground cellar), where [CO2] in air
was above dangerous values. Since 1 March (flat no. 1)
and 9 March (flats no. 2 and 3) indoor [CO2] was
measured every 10 at the ground level in the sleeping
rooms by three identical automatic Drger IR devices
(X-am 7000). In Fig. 7 the temporal variation of the
recorded indoor [CO2] maxima is reported. These values
do not correspond to actual potential maxima as they
depend on how long the spaces had remained closed,
without aeration, before the measurement. Data show
that in all the investigated sites, before the cement
squeezing operations, air [CO2] was often more than 10
vol.% and up to 30 vol.%. These were lethal concentrations (see Table 1) that justify the preventive evacuation

of these houses and rooms. After the first cement

squeezing of 23 February (P1) the indoor [CO2] values
lowered initially below 10 vol.% and, since 21 February,
below 5 vol.%. It is interesting to note that roughly 36
h were needed, since the first gas blowout from borehole
F1, for CO2 to diffuse laterally and reach dangerous
concentrations within the houses. Similarly indoor
[CO2] lowered below the hazard threshold of 5 vol.%,
from 36 to 48 h after the first P1 cement squeezing
(Fig. 7). The indoor [CO2] progressively lowered following the other cement squeezing operations. It went
back to normal air value (330 ppm) since early March in
the cellar and in the electric cabin. In flats 1 and 3 [CO2]
never exceeded 1.8 vol.% after 24 February, then it
progressively lowered to values around 0.5 vol.% after
the last cement squeezing (P7) of 30 March. In flat 2,
[CO2] remained somewhat higher, with maxima around
3.0 vol.%; here the floor of the monitored room was
badly constructed and displayed some evident cracks
and disjunctions from which a CO2 flux of 210 ppm/s
was measured. Once these cracks were sealed by silicon,
indoor [CO2] maxima fell to 1.01.5 vol.%. The families were then authorized to return home with the recommendation to install automatic efficient space aeration
systems and to rebuild the floor of flat 2 introducing an
impervious layer.
4. Gas composition and origin
Four gas samples were collected from the Fiumicino
boreholes. Gas mostly consisted of CO2 (9798 vol.%)
with minor N2 (1.6 vol.%) and CH4 (0.751.49 vol.%)

26

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

and only a few ppm of H2S. Such a composition falls

within the variation field of the cold gas manifestations
of Central Italy and particularly to that of the gases
emitted from the near volcanic areas of Alban Hills and
Mts. Sabatini (Table 4). The latters have higher H2S
content, frequently around 1 vol.%; hydrogen sulphide
was likely substantially lost in the Fiumicino gas by
oxidation to SO42 during interaction with the water of
the aquifers encountered by the rising gas.
In the ternary diagrams of the main components
(CO2N2CH4, Fig. 8a), the Fiumicino gas appears
enriched in CH4, as some Mts. Sabatini samples, with
respect to the dominant trend of Central Italy gas

manifestations, that is aligned along the CO2N2 side.

In North-Central Italy methane is particularly enriched
in the gas of the geothermal fields of Larderello and Mt.
Amiata and its origin is attributed by Minissale et al.
(1997) to thermogenic conversion of CO2 within the
geothermal reservoirs at 250300 C. This cannot be the
origin of methane in most of the surface gas manifestations of Central Italy where the temperature of the
carbonate aquifers is around 100150 C or less. The
Fiumicino gas composition likely reflects some addition
of low-temperature CH4 of organic origin to the deep
rising gas. In fact, a few kilometers inland, methane gas
emissions occur in the inner Tiber delta characterized by

Table 4
Chemical and isotopic analyses of Fiumicino gas and of gas emissions of Mts. Sabatini and Alban Hills volcanic areas
No. Locality

CH4
N2
(vol.%) (vol.%)

H2
He
CO
Ar
O2
(ppm) (ppm) (ppm) (vol.%) (vol.%)

3
He/4He
13CCO2
( vs. PBD) R/Ra

0.001
na
na
na

0.75
1.25
1.49
0.91

1.56
1.57
1.56
1.65

nd
nd
0.05
0.05

1.74
nd
2.3
2.4

nd
0.05
0.07
0.08

na
na
0.001
0.0013

0.017
0.015
0.007
0.022

1.55
na
1.50
0.75

0.314

43.87

94.49

0.0004

0.25

5.92

71

43

550

0.015

0.00103 0.53

0.89

194.40

93.9
93.7
98.54
94.2
97.6

1.30
0.35
0.8
4.65
1.07

0.56
0.048
0.042
0.149
0.015

4.22
5.25
0.5
0.98
1.28

12
b1
nd
b1
b1

14
3
1.9
2
9

na
nd
na
na

0.048
0.066
na
0.005
0.003

0.0004
0.547
0.05
b 0.001
b 0.001

1.34

42.51

na

0.0011

0.80

nd

5.3

nd

na

b 0.001 1.23

0.95

84.60

0.003
0.044

0.52
0.87

b1
7

2
1.6

na
nd

0.006
na

0.112
na

0.3
1.3

1.9

34.75

4.89
1.327
2.32

4
28
13

7
2
4

na
na
0.07

0.074 0.93
0.0154 0.0185
0.0152 0.204

0.51
2.15
2.4

1.23
0.34
0.24

14.7
0.95
0.85

14.289 5
2.6844 293
1.7599 b 1

17
7
7

na
0.05
0.1

0.2774 0.8242
0.0204 0.1276
0.0094 0.0947

1.93
2.71
0.69

2.02
na
na

6.23
na
na

H2 S
T
CO2
(C) (vol.%) (vol.%)

Fiumicino
1e IS A well F2
15.8 98.36
1e IS B well F3
12.7 96.00
1e IS n.2 well F3
96.95
97.42
1e IS n.1 well F2
Alban Hills
2e S. Maria delle
Mole well
3a Acqua Solfa
4a Cava dei Selci
4c Cava dei Selci
5a Lavinio
6a Solforata di
Pomezia
6c Solforata di
Pomezia
7a Trigoria
8d Valle Cupella
well
Mts. Sabatini
9b Baccano
10b B. go Pantano
11b Caldara
Manziana
12b Isola Farnese
13b Mt. Bischero
14b Parco della
Mola
15b P. re Annibaldi
16b P. gio
Capecchio
17b Puzzole
Zancona
18b Solfatara
Manziana

na
24

24.2 99.2

1.2
0.9
0.75
0.5
3.5

He/20Ne

na

98.1
98.25

1.25
0.5

22
40
15

93.52
96.91
98.04

b 0.0005 0.083
0.29
0.2892
0.42
0.0294

29
22
25

82.97
96.96
96.84

b 0.0005 0.0159
b 0.0005 1.0767
0.017
1.8668

10
26

97.60
98.03

0.125
0.193

0.0898
0.5602

1.0918 b 1
0.9747 b 1

2
2

0.04
na

0.0035 0.006
0.0083 0.0018

1.88
0.02

0.19
0.52

20
1.46

10

95.1

0.066

4.321

0.749

b1

0.0034 0.0025

4.64

0.456

34.5

11

96.48

0.98

0.1268

2.0973 1

0.05

0.0138 0.2284

4.14

na

na

na = not analysed; nd = below detection limit; measured on the field with Drger X-am 7000.
Data after: aGiggenbach et al. (1988); bMinissale et al. (1997); cCarapezza et al. (2003); dCarapezza and Tarchini, 2007; eThis work.

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

Fig. 8. CO2N2CH4 (a) and N2HeAr (b) triangular plots of the gas
emissions of Mts. Sabatini, Alban Hills and Fiumicino. Data after
Table 4.

lagoon organic deposits (Bellotti et al., 1995); here in

the gas manifestations of Ponte Galeria we measured a
CH4 content of 49.6 vol.%. The Fiumicino gas could be
the result of CH4 addition to a deep gas having a CO2/N2
ratio of 62, near to those of Solforata on Alban Hills and
Parco Mola on Mts. Sabatini (see dashed line in Fig. 8a).
In the triangular diagram N2HeAr (Fig. 8b) the
Fiumicino gas plots near the N2 corner, in a position
that was considered by Giggenbach (1991) as representative of gases from andesitic volcanoes of actively
subducting zones. Such an origin is very unlikely for
Central Italy cold gas manifestations, whose excess N2
(non-atmospheric nitrogen) has most probably a crustal
origin from metasedimentary rocks associated with deep
aquifers (Minissale et al., 1997). It is clear from Fig. 8b
that Fiumicino gas is the less contaminated by air or airsaturated water (ASW) of Mts. Sabatini and Alban Hills
gases, most of which plot on a FiumicinoASW potential mixing line. The Fiumicino gas could therefore
result from a significant addition of crustal N2 to a deep
mantle gas.

27

The Fiumicino gas has a 13C of CO2 ranging between

0.75 and 1.55 ( vs. PDB), values falling in the
variation range of the isotopic composition of the CO2
carbon of gas from both Mts. Sabatini and Alban Hills
(see Table 4). This might also reflect, as for CH4, a small
contribution of organic CO2 whose 13C being much
more negative, would lower the original isotope signature
of the deep provenance CO2. The origin of the CO2 of the
gas manifestations of Central Italy has been long debated.
A prevalent origin from the Mesozoic limestones was
proposed by Panichi and Tongiorgi (1976), whereas
Minissale (1991) and Chiodini et al. (2000) favoured a
prevalent mantle origin. The 13C values cannot help to
discriminate between these different sources. In fact a
simple contamination with atmospheric CO2 (13C from
6.0 to 8.0 vs. PDB; Faure, 1977), that could occur
also at depth in aquifers recharged by meteoric water, or
mixing with soil CO2 generated by plant respiration (13C
from 13.0 to 28.0 vs. PDB; Cerling et al., 1991),
could displace the 13C values of CO2 originated from
marine limestones (13C from 1.0 to + 2.0 vs. PDB;
Craig, 1963) toward the isotopic values considered typical
of the mantle CO2 (from 6.5 to 4.5 vs. PDB; Kyser,
1986). As a matter of fact the 13C of CO2 of the
manifestations of Central Italy covers a wide range from
7.0 to + 2.0 vs. PDB (Minissale et al., 1997 and
references therein) so that all the aforementioned sources
are possible and they may also likely coexist.
The helium isotopic composition of the CO2
dominant gases, i.e. the 3He/4He ratio (R) expressed
with respect to the same ratio in air (Ra), has been
frequently used to discriminate the role of mantle versus
crustal sources for these gases. The R/Ra values of the
gas manifestations of the volcanic areas of Central Italy
are much lower than in the typical mantle (R/Ra 8.0;
Marty and Jambon, 1987). They actually range from
0.89 to 1.90 in the Alban Hills gases and are somewhat
lower in the Mts. Sabatini gases (from 0.19 to 0.52) (see
Table 4 and data in Minissale et al., 1997 and references
therein). R/Ra of the Fiumicino gas (0.314) is again more
similar to the isotopic composition of Mts. Sabatini gas
manifestations than to those of Alban Hills, although
higher R/Ra values (1.232.02) have been found at
BaccanoCesano gas emissions in the Sabatini volcanic
complex (Minissale et al., 1997).
If a uncontaminated mantle with R/Ra around 8 exists
also beneath Central Italy, the low R/Ra values of the gas
emitted from this area would suggest only a minor mantle
contribution and a dominant crustal origin. Diagrams such
as that of Fig. 9 (O'Nions and Oxburgh, 1988) have been
used to show that these gases could actually be originated
by a large mixture of a mantle component with crustal

28

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

Fig. 9. R/Ralog(CO2/3He) plot after O'Nions and Oxburgh (1988). Full dots: Mts Sabatini samples; crosses: Alban Hills samples; open circle:
Fiumicino gas. Numbers refer to sample no. in Table 4.

He-enriched gases. In Fig. 9 most samples, including

Fiumicino gas, are aligned on a mixing line between a
crustal component and a mantle-rich component, more
evident for Cava dei Selci (4) and Solforata (8) gases in
Alban Hills. Only in the Larderello gas with R/Ra values
of 2.02.3, and possibly in Cesano gas (R/Ra = 2.02), the
volatile contribution from a typical mantle could be
around 2530% (Hooker et al., 1985). Consequently also
CO2 should be originated largely from crustal sources
such as the buried carbonate complexes (Minissale et al.,
1997). On the other hand, the mantle of this region could
be largely contaminated with crustal material and have R/
Ra significantly lower than that of the typical
suboceanic mantle (Tedesco et al., 1990). There is in
fact a strong geochemical evidence indicating that alkalipotassic magmas of the Roman Comagmatic Province
originate from mantle contaminated by subducted crustal
components or by slab derived crustal fluids (Peccerillo,
1999; Gasperini et al., 2002). The fluid inclusions in
olivine and clinopyroxene phenocrysts of the Quaternary
volcanic rocks of Central Italy have low R/Ra values very
near to those of the gas manifestations of the same area
(Martelli et al., 2004; Carapezza and Tarchini, 2007). The
possibility therefore exists that these gases largely
originate in a deep contaminated mantle or magmatic
source. Their isotopic signature, indicative of a marked
crustal component, could then be a geochemical character
inherited from the deep source, and only to a limited
extent be the result of contamination with crustal fluids
occurring during the gas upraise. In other words, a mixing

of mantle and crustal volatile components, like the one

indicated by Fig. 9, would still hold but with a mantle end
member having a significantly lower R/Ra and higher
CO2/3He.
5. Conclusions
The accidental gas blowout occurred at Fiumicino
boreholes in February 2005 significantly enlarges the
region of Central Italy known to be characterized by a
high level of deep CO2 degassing (Chiodini et al., 2004;
Gambardella et al., 2004). The gas surface manifestations
of this area, as well as the CO2 pressurized pockets
encountered at various depths by wells (Carapezza and
Tarchini, 2007), are mostly concentrated on structural
highs, i.e. on buried horsts of the Mesozoic limestones or
on they steep faulted flanks. These buried structures
produce gravity positive anomalies, as limestones are
denser than overlying rocks (Fig. 10). The Fiumicino zone
is somewhat anomalous to this respect, as it does not
correspond to any evident structural high, although just
near the Tiber mouth gravity residual anomaly values
increase from 2 to 0 mGal (Fig. 10). The top of the
limestones could be here at about 2000 m depth, whereas
it raises to only 500 m depth on the important Ciampino
structural high of Alban Hills (Di Filippo and Toro, 1995).
The preferential location of the gas manifestations on
carbonate structural highs can be explained because i) the
gas rising from depth encounters the first high-permeability formations in the Mesozoic limestones and

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

29

Fig. 10. Gravimetric map of the Roman Magmatic Province (modified after Di Filippo and Toro, 1995) with location of the main gas manifestations of
the area (numbers as in Table 4).

accumulates at their top, confined beneath an impervious

flysch cover; and ii) the gas easily escapes from these caps
towards the surface through the extensional faults and
fractures associated to the horst structures (Chiodini and
Frondini, 2001). On its way to the surface, the gas released
from the carbonate aquifer encounters, interacts and may
accumulate in perched aquifers that are present both in the
Neogene sedimentary deposits and in the overlying
Quaternary volcanics, where pockets of pressurized
CO2 have been encountered by wells (Carapezza and
Tarchini, 2007). At Fiumicino such a pressurized cap

formed at very shallow depth (ca. 25 m) within high

permeability sediments of the Tiber delta confined under
an efficient clay cover. Likely at Fiumicino the deep gas
reaches the surface through NE trending vertical faults
(Fig. 1; Funiciello and Parotto, 1978). The chemical and
isotopic composition is similar to that of the gas
manifestations of the volcanic area of Central Italy,
particularly to those of Mts. Sabatini volcanic complex.
Although the gas geochemistry has probably been
modified by the addition of crustal and shallow organic
components, it appears compatible with a deep mantle

30

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731

source confirming that the Tyrrhenian margin also

belongs to the zone of Central Italy affected by very
high CO2 Earth degassing (Chiodini et al., 2004). Finally,
because of the widespread occurrence of CO2 pressurized
pockets even at a very shallow depth, any well drilled in
the area is severely exposed to hazards from gas blowout
and, as a precautionary measure, dense drilling mud
should be employed and the rig should be equipped with
BOP (Blow Out Preventer). As a matter of fact, after the
February 2005 gas blowout we learnt that at least three
similar accidents had already occurred in the same zone,
all related to shallow boreholes.
Acknowledgements
The Civil Protection Office of the Province of Roma
provided the Draeger devises and the permanent CO2
flux station. The Fiumicino Municipality and ACEA
Company are thanked for the field assistance. An intense
technical collaboration was held with Ing. Alberto Bottai
who supervised the squeezing operations. This study has
been carried out with the support of the INGV-DPC
programme (Project V5-Diffuse degassing in Italy
Research units coordinated by F. Barberi and M.L.
Carapezza). The chemical and isotopic analyses of
Fiumicino gas were made in the labs of INGVOsservatorio Vesuviano, CNR-IGG Pisa and of INGVPalermo, through the courtesy of Giovanni Chiodini,
Roberto Cioni and Salvatore Inguaggiato, respectively.
References
Bellotti, P., Milli, S., Tortora, P., Valeri, P., 1995. Physical stratigraphy
and sedimentology of the Late PleistoceneHolocene Tiber Delta
depositional sequence. Sedimentology 42, 617634.
Carapezza, M.L., Granieri, D., 2004. CO2 soil flux at Vulcano (Italy):
comparison between active and passive methods. Appl. Geochem.
19, 7388.
Carapezza, M.L., Tarchini, L., 2007. Accidental gas emission from
shallow pressurized aquifers at Alban Hills volcano (Rome, Italy):
geochemical evidence of magmatic degassing? J. Volcanol.
Geotherm. Res. 165, 516. doi:10.1016/j.jvolgeores.2007.04.008.
Carapezza, M.L., Badalamenti, B., Cavarra, L., Scalzo, A., 2003. Gas
hazard assessment in a densely inhabited area of Colli Albani
volcano (Cava dei Selci). J. Volcanol. Geotherm. Res. 123, 8194.
Carapezza, M.L., Barberi, F., Tarchini, F.L., Cavarra, L., Granieri, D.,
2005. Le emissioni gassose dell'area vulcanica dei Colli Albani.
In: Carapezza, M.L., et al. (Ed.), Nuovi dati sull'attivit recente del
cratere del Lago Albano e sul degassamento dei Colli Albani. Atti
Accad. Naz. Lincei, 218, pp. 229242.
Cerling, T.E., Solomon, D.K., Quade, J.R., Bowman, J.R., 1991. On
the isotopic composition of carbon in soil carbon dioxide.
Geochim. Cosmochim. Acta 55, 34033405.
Chiodini, G., Frondini, F., 2001. Carbon dioxide degassing from the
Albani Hills volcanic region, Central Italy. Chem. Geol. 177, 6783.

Chiodini, G., Cioni, R., Guidi, M., Raco, B., Marini, L., 1998. Soil
CO2 flux measurements in volcanic and geothermal areas. Appl.
Geochem. 13, 135148.
Chiodini, G., Frondini, F., Cardellini, C., Parello, F., Peruzzi, L., 2000.
Rate of diffuse carbon dioxide Earth degassing estimated from
carbon balance of regional aquifers: the case of Central Apennine,
Italy. J. Geophys. Res. 105, 84238434.
Chiodini, G., Cardellini, C., Amato, A., Boschi, E., Caliro, S.,
Frondini, F., 2004. Carbon dioxide Earth degassing and seismogenesis in Central and southern Italy. Geophys. Res. Lett. 31.
doi:10.1029/2004GL019480.
Craig, H., 1963. The isotopic geochemistry of water and carbon in
geothermal areas. In: Tongiorgi, E. (Ed.), Nuclear Geology on
Geothermal Areas. Spoleto. C.N.R., pp. 1754 (Italian Council for
Research, Rome).
De Rita, D., Faccenna, C., Funiciello, R., Rosa, C., 1995. Stratigraphy
and volcano-tectonics. In: Trigila, R. (Ed.), The Volcano of the
Albani Hills. Tipografia SGS, Roma, pp. 3371.
Di Filippo, M., Toro, B., 1995. Gravity features. In: Trigila, R. (Ed.), The
Volcano of the Alban Hills. Tipografia SGS, Roma, pp. 213219.
Doglioni, C., Gueguen, E., Harabaglia, P., Mongelli, F., 1999. On the origin
of west-directed subduction zones and applications to the western
Mediterranean. In: Durand, B., Jolivet, L., Horvath, F., Seranne, F.
(Eds.), The Mediterranean Basins: Tertiary Extension Within the
Alpine Orogen. Geological Society, London, pp. 541561.
Faure, G., 1977. Principles of Isotope Geology. J. Wiley, New York.
589 pp.
Funiciello, R., Parotto, M., 1978. Il substrato sedimentario dei Colli
Albani: considerazioni geodinamiche e paleogeografiche sul margine
tirrenico dell'Appennino Centrale. Geol. Rom. 17, 233287.
Gambardella, B., Cardellini, C., Chiodini, G., Frondini, F., Marini, L.,
Ottonello, G., Vetuschi Zoccolino, M., 2004. Fluxes of deep CO2
in the volcanic areas of central- southern Italy. J. Volcanol.
Geotherm. Res. 136, 3152.
Gasperini, D., Blichert-Toft, J., Bosch, D., Del Moro, A., Macera, P.,
Albarde, F., 2002. Upwelling of deep mantle material through a
plate window: evidence from the geochemistry of Italian basaltic
volcanoes. J. Geophys. Res. 107, 2367.
Giggenbach, W.F., 1991. Chemical techniques in geothermal
exploration. In: D'Amore, F. (Ed.), Application of Geochemistry in Geothermal Reservoirs Development. UNITAR, Rome,
pp. 119144.
Giggenbach, W.F., Minissale, A.A., Scandiffio, G., 1988. Isotopic and
chemical assessment of geothermal potential of the Colli Albani
area, Latium, Italy. Appl. Geochem. 3, 475486.
Granieri, D., Chiodini, G., Marzocchi, W., Avino, R., 2003.
Continuous monitoring of CO2 soil diffuse degassing at Phlegraean Fields (Italy): influence of environmental and volcanic
parameters. Earth Planet. Sci. Lett. 212, 167179.
Hooker, P.J., Bertrami, R., Lombardi, S., O'Nions, R.K., Oxbourgh, E.R.,
1985. Helium-3 anomalies and crustmantle interaction in Italy.
Geochim. Cosmochim. Acta 49, 25052513.
Kyser, T.K., 1986. Stable isotope variations in the mantle. In: Valley, J.M.,
Taylor, H.P., O'Neil, J.R. (Eds.), Stable Isotope in High Temperature
Geological Processes. Rev. Miner., 16, pp. 141164.
Martelli, M., Nuccio, P.M., Stuart, F.M., Burgess, R., Ellam, R.M.,
Italiano, F., 2004. Heliumstrontium isotope constraints on mantle
evolution beneath the Roman Comagmatic Province, Italy. Earth
Planet. Sci. Lett. 224, 295308.
Marty, B., Jambon, A., 1987. C/3He in volatile fluxes from the solid
earth: implications for carbon geodynamics. Earth Planet. Sci. Lett.
83, 1626.

F. Barberi et al. / Journal of Volcanology and Geothermal Research 165 (2007) 1731
Miller, S.A., Collettini, C., Chiaraluce, L., Cocco, M., Barchi, M.,
Kaus, B.J.P., 2004. Aftershocks driven by a high pressure CO2
source at depth. Nature 427, 724727.
Minissale, A., 1991. Thermal springs in Italy: their relation to recent
tectonics. Appl. Geochem. 6, 201212.
Minissale, A., Evans, W.C., Magro, G., Vaselli, O., 1997. Multiple
source components in gas manifestations from north-central Italy.
Chem. Geol. 142, 175192.
O'Nions, R.K., Oxburgh, E.R., 1988. Helium, volatile fluxes and the
development of continental crust. Earth Planet. Sci. Lett. 90,
331347.
Panichi, C., Tongiorgi, E., 1976. Carbon isotopic composition of CO2
from springs, fumaroles, mofettes, and travertines of central and
southern Italy: a preliminary prospection method of geothermal

31

areas. Proc. 2nd U.N. Symp. Development and Use of Geothermal

Energy, San Francisco, 2029 May 1975, pp. 815825.
Peccerillo, A., 1985. Roman Comagmatic Province (Central Italy):
evidence for subduction-related magma genesis. Geology 13,
103106.
Peccerillo, A., 1999. Multiple mantle metasomatism in centralsouthern Italy: geochemical effects, timing and geodynamic
implications. Geology 27, 315318.
Rogie, J.D., Kerrick, D.M., Chiodini, G., Frondini, F., 2000. Flux
measurements of nonvolcanic CO2 emissions from some vents in
Central Italy. J. Geophys. Res. 105, 84358445.
Tedesco, D., Allard, P., Sano, Y., Wakita, H., Pece, R., 1990. Helium-3
in subaerial and submarine fumaroles of Campi Flegrei caldera,
Italy. Geochim. Cosmochim. Acta 54, 11051116.