Journal of Volcanology and Geothermal Research 176 (2008) 82–93

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

j o u rn al ho m ep ag e: www.elsevier. com / l o ca t e/ jv o lge or e s

Pre-eruptive physical conditions of El Reventador volcano (Ecuador) inferred from

the petrology of the 2002 and 2004–05 eruptions
Pablo Samaniego a,b,⁎, Jean-Philippe Eissen b,1, Jean-Luc Le Pennec b, Claude Robin a,b, Minard L. Hall a,
Patricia Mothes a, Deborah Chavrit b, Joseph Cotten c
a
Instituto Geofísico, Escuela Politécnica Nacional, Ap. 17-01-2759, Quito, Ecuador
b
IRD, UMR 163, Laboratoire Magmas et Volcans, 5 rue Kessler, 63038 Clermont–Ferrand, France
c
Université de Bretagne Occidentale, CNRS-UMR 6538, 6 avenue Le Gorgeu, BP 809, 29283 Brest, France

ARTICLE INFO ABSTRACT

Article history: A petrological study of the eruptive products of El Reventador allowed us to infer the magmatic processes
Received 25 January 2007 related to the 2002 and 2004–05 eruptions of this andesitic stratovolcano. On November 3, 2002, El
Accepted 10 March 2008 Reventador experienced a highly explosive event, which was followed by emplacement of two lava flows in
Available online 19 March 2008
November–December 2002. Silica contents range from 62 to 58 wt.% SiO 2 for the November 3 pyroclastic
deposits to 58–56 and 54–53 wt.% SiO2 for the successive lava flows. In November 2004 eruptive activity
Keywords:
resumed supplying four new lava flows (56–54 wt.% SiO 2) between November 2004 and August 2005.
El Reventador
Ecuador
Volatile contents in matrix glasses and glass inclusions from the November 3 pyroclastic deposits allow us to
magmatic episodes estimate the total amount of SO2 and HCl released into the atmosphere during the paroxysmal phase (i.e. 80 kT of
recharge SO2 and 280 kT of HCl). Pre-eruptive pressure-temperature conditions of the magmas range from 300 to 150 MPa
petrology and ~ 1000 °C with high water contents (~ 5 wt.%). We propose the existence of an andesitic magma body located at
thermobarometry ~7–12 km depth that is frequently intruded by more primitive, hydrous magmas from a deeper source. The initial
degassing crystallization of amphibole from the hydrous primitive magma seems typical of El Reventador, as well as the
historically recurrent and regular periods of eruptive activity lasting several years. This eruptive behaviour coupled
with the fractionation and mixing processes inferred from the 2002 and 2004–05 petrologic data suggest that deep
magmatic recharge at El Reventador is frequent, and is probably responsible for the high frequency of eruptions.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction left a prominent and well-preserved 4 km-wide horseshoe-shaped

Improved understanding of volcanic eruptions is facilitated by the
increasing number of geophysical and geochemical techniques (e.g.
Scarpa and Tilling, 1996) coupled with phenomenological observation
of volcanic activity. These studies can be complemented by conven-
tional petrological methods focused on the mechanisms of magma
differentiation as well as by measurements and inferences of the vola-
tile contents and degassing modalities of ascending magmas. Applied
to magmas erupted sequentially, such tools help to constrain the pre-
eruptive P–T conditions and the geometry of the magma plumbing
system. These petrological studies provide an important tool for the
understanding of the dynamics of andesitic magma processes at depth.
El Reventador is a large stratovolcano located 90 km east of Quito in the
eastern subandean zone of Ecuador (Fig. 1a), which provides an ideal
laboratory for petrologic analyses of a sequence of eruptive episodes. El
Reventador experienced a large sector collapse around 19,000 yr BP that
⁎ Corresponding author. Instituto Geofísico, Escuela Politécnica Nacional, Ap. 17-01-
2759, Quito, Ecuador. fax: +593 2 256 7847.
E-mail address: psamaniego@igepn.edu.ec (P. Samaniego).
1
Deceased.
caldera (Aguilera et al., 1988). A young, symmetrical, 1 km-high edifice
has grown in this amphitheatre and is one of the most active volcanoes
in the Ecuadorian Andes. Based on sporadic historical documentation,
Hall (1980) and Simkin and Siebert (1994) identified at least 16 eruptive
periods since AD 1541. During the 20th century, eruptive episodes were
separated by regular time intervals and often were characterized by
durations of several years (e.g. 1898, 1912, 1926–29, 1944, 1958–60, and
1972–76). All eruptive episodes were moderate in scale (VEI 2–3), with
strombolian to vulcanian eruptive styles, producing ash falls, small-
volume pyroclastic and debris flows, and blocky lava flows.
Following 26 years of quiescence, El Reventador erupted again in
2002. On November 3, a sub-Plinian eruption occurred without sig-
nificant precursory signs (Hall et al., 2004), and was characterized by a
rapid increase in eruption intensity, from complete rest prior to the
night of November 2, to the paroxysmal event in the early morning of
November 3. The 2002 eruption onset was notably stronger (VEI 4)
than the preceding eruptions and was followed by the effusive em-
placement of two lava flows within the following month. A pause in
2003–04 ended with renewed effusive activity in November 2004 and
with a series of lava flows that formed during the following months on
the southeastern flanks of the active cone. The activity continued

0377-0273/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jvolgeores.2008.03.004
 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93 83

Fig. 1. (a) Geodynamical setting of the Ecuadorian volcanic arc (modified from Gutscher et al., 1999). Andes Cordillera is defined by the 2000 m contour. Trench is shown as toothed
line and main volcanoes indicated by open triangles. GSC: Galápagos Spreading Centre; GFZ: Grijalva Fracture Zone. El Reventador volcano (represented by a black triangle) is located
in the eastern border of the Ecuadorian arc. (b) Aerial photo of El Reventador caldera taken by Instituto Geográfico Militar (Quito, Ecuador) in 1983, showing historical lava flows, as
well as the extent of the 2002 pyroclastic flow deposits (PFs) and lava flows (LF1 and LF2) of 2002; (c) and (d) lava flows (LF3, LF4, LF5 & LF6) corresponding to 2004–05 activity.

mildly explosive in June–September 2005 coincident with the
termination of significant lava effusion and finally tapered off and
ended in December 2005. Intermittent effusive activity resumed in
March–April 2007 and continues at this time (September 2007). This
new eruptive episode is not addressed in the present study, although
the eruptive styles of the 2007 episode is similar to that of 2005.
Overall, the 2002 and 2004–05 eruption sequence is characterized
by distinct eruption styles of unequal size and duration (see below), as
well as periods of relative quiet. To gain insights into the root causes of
varied eruption processes at El Reventador, we studied the geochemical
and mineralogical characteristics of the 2002 and 2004–05 eruptive
products and assessed volatile contents in juvenile matrix glasses and
glass inclusions. The data are presented together with a petrogenetic
model that accounts for the observed magmatic suite and eruption
styles over the 3-yr-long period of activity. Repeated recharge of the
inferred magma reservoir and the significant role of amphibole in both
the petrogenesis and eruption styles are presented in this model.
2. Chronology of the 2002 and 2004–05 activity
2.1. The 2002 eruption
At around 02h00 (07h00 GMT) on November 3, the Instituto
Geofísico's National Seismic Network recorded VT and hybrid events
as well as a low-amplitude, almost continuous tremor. At dawn
(05h30–06h00), inhabitants in the area witnessed a steam column
rising above the crater, signalling the onset of a new eruption. The
height of the column and its ash content increased considerably
during the next few hours, reaching 16–17 km asl (above sea level) at
09h12, after a 45 min-long paroxysmal phase. The plume drifted to the
west towards the Interandean valley, depositing a layer of ash with a
bulk uncompacted volume estimated at ~ 3 × 108 m3 (Le Pennec et al.,
2003). At least five pyroclastic flows swept the caldera floor and two
reached the Quijos River ~ 9 km eastward (Fig. 1b). Lithic-rich deposits,
5–20 m thick, bearing vesiculated bombs were emplaced over a 12–
13 km2 area and correspond to a bulk volume of ~ 4–6× 107 m3 (Hall
et al., 2004). The overall tephra volume ranks the eruption at VEI 4,
making it the most explosive event during historical times at El
Reventador and in Ecuador in more than a century. Phenomenological
observations and geological investigations point to a strong phreato-
magmatic interaction in the eruption process (Hall et al., 2004). This is
evidenced by (1) a high degree of magma fragmentation leading to
fine-grained ash-fall deposits, (2) the presence of angular fragments of
basement rocks and older lavas commonly incorporated as xenoclasts
in the bombs, and (3) large amounts of vapour seen in the eruption
column. The summit crater was significantly enlarged during the
eruption (~ 40 × 106 m3, Hall et al., 2004) and notches opened in the
crater's northern and southern margins. Explosive activity progres-
sively waned after the paroxysmal phase and ceased a few hours later.
On November 6 the first blocky lava flow (LF1) spilled out of the
84 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93

southern notch and moved down the cone over the pyroclastic flow
deposits (Fig. 1b). Two weeks later, a second lava (LF2) issued from a
lateral vent located at 2600 m asl on the southeastern flank of the cone
~ 1 km distant from the main crater (Fig. 1b). LF2 bears abundant
gabbroic xenoliths. Hall et al. (2004) estimated the overall volume of
both lava flows at 3.6 × 107 m3.
2.2. The 2004–05 eruption
Gas and vapour were continuously released by El Reventador from
early 2003 to mid-2004. In August 2004, new seismic swarms and
increasing thermal anomalies within the crater preceded a second
eruptive episode (Johnson et al., 2006; Ramon et al., 2006, Troncoso
et al., 2006). A third lava (LF3) filled the crater and poured out of the
southern notch in November–December 2004, locally overriding LF1
(Fig. 1c). LF4 lava lobes descended 4.5 km down the cone during
March–April 2005. Two months later, from June to July 2005, LF5
accumulated on top of LF4 (Fig. 1d), lava flow effusion tapered, and the
eruptive activity evolved to strombolian–vulcanian styles in August–
September 2005. At least one small additional lava (LF6) flowed 1 km
down the southern flank of the cone in August 2005.
3. Petrology of the 2002 and 2004–05 eruptive products
for 10 samples from the 2004–05 lava flows (Samaniego et al., 2006).
The silica content of the entire 2002 and 2004–05 magmatic series
ranges from 53.5 to 62.1 wt.% normalized to an anhydrous basis
(Table 1). Most pyroclastic products of the paroxysmal phase display
homogeneous compositions (57.8–59.1 wt.% SiO2), however scarce,
rounded to subangular, light-coloured magmatic enclaves found in
bombs bear more siliceous compositions (61.5–62.1 wt.% SiO2). The
2002 lava compositions show a wider silica range, decreasing from
58.3–56.5 for LF1 to 54.6–53.5 wt.% for LF2. In the 2004–05 lava flow
sequence the silica content also decreases from 56.4–55.4 for LF3, and
LF4 to 54.8 for LF5 and 53.8 wt.% for a ballistic block contemporaneous
to LF6 (Fig. 2a). The magmatic suite plots near the limit between
medium- and high-K in the basic andesite to andesite fields of Gill
(1981) (Fig. 2b).
Major oxides (except Na2O and K2O) as well as Ni, Cr, and V are
negatively correlated with silica content, although some scattering is
shown for Al2O3, and P2O5. Large-Ion Lithophile Elements (LILE), High
Field Strength Elements (HFSE), and Light Rare Earth Elements (LREE)
are positively correlated to silica content (Fig. 3). Conversely, Sr, Y, and
Middle and Heavy Rare Earth Elements (MREE and HREE) show a
slight decrease with increasing silica content or no correlation at all.
Apart from Sr, HREE and Y, most elements from the 2002 to 2004–05
rocks define a single trend in the variations diagrams (Fig. 3).

3.1. Whole-rock geochemistry 3.2. Mineral compositions

Whole-rock major and trace element concentrations were ob- The 2002 rocks are porphyritic (~ 25–30 vol.% phenocrysts in pyro-
tained for 38 samples of the 2002 pyroclastic deposits and lavas and clastic blocks and ~ 40 vol.% in lava flows) with a mineral assemblage

Table 1
Selected whole rock major and trace elements of the 2002 and 2004–05 eruptions of El Reventador volcano. Major element (wt.%) and trace element (ppm) analyses were performed
at the Université de Bretagne Occidentale, Brest (France) by inductively coupled plasma-atomic emission spectroscopy (ICP-AES), see Cotten et al. (1995), except Rb determined by
flame atomic emission spectroscopy. Relative standard deviation is ca.1% for SiO2 and 2% for the other major elements except for low values (b 0.5% oxide). For trace elements, relative
standard deviation is ca. 5% except for Er (ca. 20%). ⁎ Total iron as Fe2O3

Year 2002 2002 2002 2002 2002 2002 2002 2002 2002 2004 2004 2005 2005 2005

Unit PFs PFs PFs PFs LF1 LF1 LF1 LF2 LF2 LF3 LF3 LF4 LF5 LF6

Sample REV 15BP REV 15BG REV 15F REV 16A REV 20 REV 19 REV 47 REV 31B REV 54 REV 42A REV 50 REV 49 REV 58 REV 62

SiO2 61.25 58.6 59.1 58.6 58 56.3 57.3 54.6 53.4 56.2 55.2 55.25 54.9 53.8
TiO2 0.61 0.74 0.72 0.75 0.76 0.83 0.76 0.9 0.95 0.85 0.89 0.88 0.95 0.98
Al2O3 17.2 17.55 17.6 17.65 17.6 17.7 17.7 17.72 17.42 17.74 17.92 18.1 17.6 17.7
Fe O ⁎
2 3 5.88 6.7 6.6 6.92 6.93 7.5 7.07 8.24 8.76 7.53 7.82 7.67 8.2 8.68
MnO 0.11 0.11 0.11 0.12 0.12 0.13 0.12 0.14 0.14 0.13 0.13 0.12 0.13 0.13
MgO 2.35 2.97 2.89 2.98 3.2 3.73 3.37 4.63 5.26 3.89 4.13 4.04 4.3 4.84
CaO 5.15 6.07 6.08 6.15 6.35 6.82 6.39 7.6 7.82 7 7.04 7.1 7.4 7.7
Na2O 4.35 4.47 4.38 4.48 4.4 4.39 4.38 4.16 4.04 4.39 4.46 4.49 4.28 4.2
K2O 2.43 2.14 2.16 2.09 2.02 1.83 2.01 1.6 1.53 1.75 1.74 1.73 1.66 1.58
P2O5 0.33 0.38 0.38 0.37 0.37 0.37 0.36 0.36 0.41 0.4 0.4 0.39 0.39 0.39
LOI 0.29 0.06 − 0.01 − 0.05 − 0.11 − 0.13 − 0.09 − 0.16 − 0.09 − 0.24 − 0.21 − 0.22 − 0.07 − 0.33
Total 99.95 99.79 100.01 100.06 99.64 99.47 99.37 99.79 99.64 99.64 99.52 99.55 99.74 99.67
Sc 8.3 10.8 10.3 10.7 11.6 13.5 12 16.2 18 13.5 14 13.5 14.5 15.5
V 110 136 132 144 142 158 143 179 196 159 166 165 184 195
Cr 16 24 19.5 26 30 40 30 69 99 46 52 50 62 69
Co 16 19 19 21 20 23 20 25 29 24 24 24 27 30
Ni 15 19 18.5 25 22 28 23 40 57 35 37 35 42 50
Rb 65.5 54 54 51.8 50.5 42.5 46 34 31 37 35.5 34.5 32.5 31.5
Sr 820 910 915 890 878 902 870 902 890 958 968 985 935 944
Y 13 14 13.8 14 14.7 15.4 14.5 16 16.3 14 14 14 14.3 14.5
Zr 160 154 159 155 152 143 147 131 129 144 138 136 134 128
Nb 9 9.2 8.9 9 8.5 8.4 8.7 8.3 8.3 8.4 8.3 8.4 8 7.8
Ba 1390 1210 1250 1230 1160 1050 1120 910 855 1072 990 1000 960 880
La 32 32 31.5 31.5 30.5 30.5 30 28 28 30 28 28 29 28
Ce 59 60 60 59 58 58 59 55 52 56 56 55 54 52.5
Nd 26 28 27.5 26.5 28 28.5 27 28 27 28 23 26.5 28 28
Sm 4.6 5 5.15 5.1 5.2 5 5.05 5.4 5.2 1.9 5.4 5.1 5 5.1
Eu 1.25 1.4 1.37 1.39 1.43 1.51 1.43 1.46 1.48 1.45 1.48 1.51 1.48 1.51
Gd 3.4 3.8 3.6 3.9 4.2 4.1 4 4.4 4.3 4.05 3.9 4.2 4.2 4.2
Dy 2.2 2.5 2.4 2.5 2.75 2.75 2.75 2.95 2.95 2.65 2.65 2.65 2.7 2.75
Er 1.15 1.2 1.15 1.25 1.45 1.5 1.35 1.5 1.5 1.3 1.25 1.3 1.4 1.4
Yb 1.14 1.2 1.14 1.19 1.28 1.36 1.3 1.37 1.37 1.18 1.16 1.16 1.2 1.16
Th 6.7 5.65 5.95 5.65 5.5 4.9 5.45 4.45 4 4.6 4.3 4.3 4.15 3.85
 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
Fig. 2. (a) Time versus wt.% SiO2 diagram for historic and recent rocks of El Reventador
volcano. Inset shows the progressive diminution in silica content during 2002 and
2004–05 eruptions. (b) K2O versus SiO2 classification diagram ( Gill, 1981) for 2002 and
2004–05 eruptive products as well as for historical lava flows (1912, 1944, 1972, 1973
and 1976). “Inc” signifies basic xenoliths in pyroclastic bombs. See text for abbreviations
significance.
composed of Pl+ Cpx + Opx+ Amph+ Mag with scarce olivine. The
2004–05 lavas are also highly porphyritic (40–50 vol.%) with Pl + Cpx +
Ol + Mag± Amph± Opx. In these rocks, olivine is common while ortho-
pyroxene is scarce. Amphibole is not abundant in the 2004–05 lavas as
compared to the 2002 products, and all crystals exhibit reaction rims
and disequilibrium features. Groundmass texture is mostly glassy in
2002 juvenile bombs and intersertal or cryptocrystalline in subse-
quent lavas. The gabbroic xenoliths found in LF2 show almost the
same mineral assemblage, Pl+ Cpx + Ol+ Mag± Opx± Amph, pointing to
a cumulate origin. In general, phenocryst from both eruptive episodes
share similar compositions (Fig. 4).
Plagioclase is the most abundant phenocryst phase (5–20 vol.%).
Crystals are generally euhedral and show a wide compositional range
from An 89 to An 37 (Fig. 4a and b; Table 2). High-Na and -K compo-
sitions (An27–34, Or6–10) were obtained in some phenocryst cores. Most
phenocrysts are normally zoned in the 2002 pyroclastic products (e.g.
An77–46), although dusty Ca-rich concentric zones and cores (up to
An80–88), as well as reverse (e.g. An 48–64) or oscillatory zonings are
commonly observed. Microlites from 2002 pyroclastic products show
restricted compositions (An 42–60). The compositional range of the
plagioclase phenocrysts is similar in the pyroclastic products and LF1,
but narrows in LF2 (An59–37). All plagioclase phenocrysts in the 2004–
05 lavas are normally zoned and bear similar compositional ranges
(e.g. An67–43).
85
Clinopyroxene occurs up to 5 vol.% in samples of both eruptive
episodes and exhibits diopsidic to augitic compositions. Euhedral and
subhedral phenocrysts have Mg#, which vary from 84 to 73, with
normal (e.g., Mg# 82–75) or slightly reversed (e.g., Mg# 74–78)
zonation (Table 2), where Mg#= 100⁎Mg/(Mg+ Fet) in mol% and Fet is
total iron as Fe2+. In a Mg# versus Al2O3 plot (Fig. 4c and d), the
dominant clinopyroxene population (Mg# 73–79; 1–6 wt.% Al2O3)
shows a crude negative correlation. However, a second group is re-
presented by high Mg# (79–84) and high-Al (3–6 wt.% Al2O3) and
corresponds to some phenocryst cores analysed in the 2002 and
2004–05 products.
Orthopyroxene is an accessory mineral in the basaltic andesite LF2,
as well as in those of the 2004–05 effusive phase. This contrasts with
the siliceous andesites of the 2002 pyroclastic deposits and the first
LF1 lava flow where it is more abundant (2–4 vol.%). In these rocks,
it occurs as euhedral enstatite phenocrysts (En 66–72, Mg# 70–76,
Table 2) with slight normal or reverse zonings (2–3% Mg# variation).
Amphibole is widespread in the 2002 products (≤ 4 vol.%), but rare
in 2004–05 lavas and consists of magnesium–hastingsite and scarce
tschermakite phenocrysts (Leake et al., 1997; Table 2). Most pheno-
crysts exhibit euhedral habits, but reaction rims are also common. This
mineral was also observed surrounding clinopyroxene and olivine
crystals. Two amphibole alteration types are identified here and are
similar to those of many other calc-alkaline suites. The dominant type
consists of reaction rims composed of Opx+ Cpx+ Pl+ Mag (“gabbroic-
type” alteration). The other is widespread in the 2002 lavas and results
from replacement of amphibole by cryptocrystalline aggregates of Fe–
Ti oxides and pyroxenes (“opacite” or black-type alteration).
Variations in aluminium content (10–14 wt.% Al2O3) allow us to
distinguish two amphibole populations. Amphibole compositions in
the 2002 pyroclastic rocks cover the whole compositional range, while
only Al-rich crystals are found in the 2002 lavas (Fig. 4e and f). It is
worth noting that the compositions of amphibole overgrowths on Cpx
and Ol phenocrysts correspond to a low-Al group. All amphibole
crystals in 2004–05 lavas exhibit reaction rims and correspond to both
high- and low-Al compositions.
Olivine is an accessory phase in the 2002 pyroclastic products
(b 1 vol.%). It has subhedral to anhedral forms with common dis-
equilibrium features (reabsorbed crystals and reaction rims) and nar-
row compositional ranges (Fo 71–81, Table 2). In the subsequent 2002
lavas, olivine (1–2 vol.%) shows a larger compositional range and most
crystals are normally zoned (Fo 87–67). In the 2004–05 lavas, olivine
crystals (3–4 vol.%) are widespread and have similar compositions
(e.g. Fo81–66) with frequent normal zoning.
Fe–Ti oxides appear as microphenocrysts and microlites (b1–2 vol.%),
as well as inclusions in pyroxene, amphibole and plagioclase pheno-
crysts. They correspond to magnetite and Ti-magnetite (5–14 wt.% TiO2
Table 2), but scarce exsolved ilmenite bands (~ 40 wt.% TiO2) were also
found in magnetite phenocrysts.
3.3. Composition of matrix glasses and glass inclusions
Major elements and the S, Cl, and F contents of matrix glasses (68
analyses) and glass inclusions in plagioclase (40), clinopyroxene (32),
orthopyroxene (14), and olivine (5) from the 2002 pyroclastic blocks
have been measured at Blaise Pascal University (Clermont–Ferrand,
France) using a CAMECA SX100 microprobe. All matrix glasses yielded
homogeneous dacitic compositions (64.8 ± 0.9 wt.% SiO 2, Table 3),
similar to those obtained for the glass inclusions in plagioclase (66.2 ±
2.0 wt.% SiO2), clinopyroxene (66.2 ± 1.0 wt.% SiO2), and orthopyroxene
(65.4 ± 1.0 wt.% SiO2). Less siliceous compositions have been obtained
for inclusions from olivine (63.0 ± 1.4 wt.% SiO2) and some plagioclase
crystals (57.8 ± 2.3 wt.% SiO2). In contrast, matrix glasses in the light-
coloured enclaves are rhyolitic (72.0 ± 1.2 wt.% SiO2). The water
contents estimated “by the difference” method (Devine et al., 1984)
are 1–2 wt.% in matrix glasses and 1–6 wt.% in glass inclusions.
86 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93

Fig. 3. Variation diagrams for selected major and trace elements for El Reventador eruptive products.

As the composition of the glass inclusions is similar to that of the
matrix glass, these inclusions likely represent the magmatic liquid
during the last evolutionary stages, prior to the eruptive episode. As
such, the difference in volatile contents between glass inclusions and
matrix glass can be used as an important degassing indicator during
the eruption. The matrix glasses contains 116 ± 66 ppm S, 1522
± 395 ppm Cl, and 449 ± 153 ppm F (Table 3). In contrast, in the dacitic
glass inclusions, S ranges from 180 to 300 ppm, Cl from 2800 to
3000 ppm, and F from 400 to 800 ppm (Table 3). Moreover, several
andesitic inclusions analysed in olivine and plagioclase point to a
much more volatile-rich liquid with 500–600 ppm S, 3100–3200 ppm
Cl and 550–1600 ppm F. The sulphur content in glass inclusions is as
expected for andesitic magmas (b 1000 ppm), whereas the chlorine
and fluorine contents of glass inclusions in the El Reventador magma
are high. Commonly, andesitic magmas contain b 2000 ppm Cl and
b 500 ppm F (Wallace, 2005; Oppenheimer, 2003). The low S and Cl
 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93 87

Fig. 4. Chemical composition of minerals of El Reventador eruptive products. Plagioclase composition plotted as An (mol%) versus Or (mol%) for (a) 2002 and (b) 2004–05.
Clinopyroxene composition plotted as Mg# versus Al2O3 wt.% for (c) 2002 and (d) 2004–05. Amphibole composition plotted as Mg# versus Al2O3 wt.% for (e) 2002 and (f) 2004–05.

concentrations of the matrix (compared with glass inclusions) indi-
cate that both elements degassed together with water during the
paroxysmal phase of the eruption. In contrast, F concentrations in
both matrix glasses and inclusions are similar.
4. Discussion
4.1. SO2 and HCl released during the 2002 eruption
The analysed glass inclusions do not show evidence of degassing,
fracturing or secondary mineral crystallization. Nevertheless, minor
differences in the major element concentrations between matrix
glasses and glass inclusions, -especially a minor SiO 2 enrichment in
the glass inclusions as compared to the matrix glass-, can be explained
by post-entrapment crystallization, as suggested in some cases by a 1–
2 μm-thick overgrowth of the host mineral. This also explains why
plagioclase glass inclusions are more Al2O3, CaO and Na2O-poor rela-
tive to the groundmass, as well as pyroxene inclusions are MgO, CaO
and FeO-poor relative to the groundmass. We estimated the relevance
of post-entrapment crystallization by adding an amount of the host
mineral to the initial melt with a composition similar to that of the
groundmass (cf. Saito et al., 2005). Our tests yield a maximum of
10 vol.% of plagioclase and 5 vol.% pyroxene overgrowth. This process
may partly explain the compositional evolution in the glass inclusions,
but fails to account for the cause of large water and volatile differences
observed between matrix and inclusion glasses.
The measured S and Cl contents in matrix and glass inclusions can
be used to estimate the amount of volatiles released to the atmo-
sphere during the paroxysmal eruption. In the case of total S emis-
sions, the relative concentrations of SO 2 and H2S in the exsolved
volatile phase is controlled by temperature, pressure, fugacities of O2
and H2O, and initial gas content of the melt. It is known that H 2S
usually prevails over SO 2 in deep pre-eruptive magma bodies with
reducing conditions, whereas SO2 is the dominant species in oxidized
magmas (Wallace, 2001). In addition, the relative proportion of these
two species changes towards a predominance of SO 2 during magma
ascent (Burgisser and Scaillet, 2007). In the case of El Reventador
volcano, the Cl-rich signature measured in glass inclusions and matrix
glasses suggests that low-pressure interaction with the hydrothermal
systems is not an important process, given the high solubility of HCl in
hydrothermal water. This suggests that the volatiles released during
the paroxysm seem to have been transported quickly from the re-
servoir. On the other hand, the presence of sulphide globules as
inclusions in Fe–Ti oxides, as well as the composition of these oxides
(see below), point to moderately oxidizing conditions. This evi-
dence implies that H2S concentrations are relatively inconsequential,
 88
88 Table 2 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
Selected analyses of phenocryst (in wt.%) for El Reventador lavas. Mineral composition were analyzed at Université Blaise Pascal (Clermont–Ferrand, France) using a CAMECA SX100 microprobe. The operating conditions were 15 kV
accelerating voltage, 15 nA beam current and 10 s counting time

Plagioclase Clinopyroxene

Year 2002 2002 2002 2002 2002 2002 2002 2002 2004 2004 2005 2005 Year 2002 2002 2002 2002 2002 2002 2002 2002 2004
Unit PFs PFs PFs PFs LF1 LF1 LF2 LF2 LF3 LF3 LF6 LF6 Unit PFs PFs PFs PFs LF2 LF2 LF2 LF2 LF3
Sample REV16A REV16A REV16A REV16A REV20 REV20 REV31B REV31B REV42A REV42A REV62 REV62 Sample REV15E REV15E REV16F REV16F REV31B REV31B REV31B REV31B REV42A
C R C R C R C R C R C R C R C R C R C R C

SiO2 47.89 56.31 55.83 51.94 51.72 56.2 55.25 58.28 50.96 57.92 48.17 52.83 SiO2 49.38 52.53 51.68 51.22 51.43 51.1 51.64 48.67 51.39
TiO2 0.02 0.03 0,00 0.03 0,00 0.043 0.03 0,00 0.06 0.05 0.04 0.07 TiO2 1.14 0.45 0.46 0.55 0.76 0.88 0.52 1.37 0.51
Al2O3 31.92 26.31 27.31 30.1 30.23 27.05 27.98 26.27 30.81 26.84 31.68 28.33 Al2O3 4.64 1.79 2.84 2.15 3.72 3.21 2.32 5.11 3.48
FeO 0.55 0.62 0.38 0.44 0.551 0.6 0.26 0.32 0.4 0.55 0.54 0.63 FeO 8.85 8.14 6.24 9.01 6.59 8,00 8.37 8.39 5.98
MnO – – – 0.04 0.02 0.03 0.02 – – – – – MnO 0.3 0.34 0.25 0.33 0.1 0.19 0.26 0.25 –
MgO 0.01 0.04 0.03 0.02 0.05 0.041 0.01 0.02 0.01 0.09 0.08 0.06 MgO 14.48 16.5 15.95 15.3 14.95 14.82 15.28 13.75 15.42

P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
CaO 16.08 9.53 9.93 13.43 13.98 9.639 10.52 8.74 14,00 8.96 15.66 11.77 CaO 20.96 20.67 21.94 20.95 22.78 21.71 20.98 21.91 21.56
Na2O 2.5 5.83 5.82 3.97 3.66 5.96 5.61 6.9 3.72 6.35 2.56 4.69 Na2O 0.45 0.36 0.34 0.4 0.37 0.39 0.45 0.47 0.29
K2O 0.14 0.45 0.31 0.214 0.19 0.44 0.35 0.44 0.21 0.51 0.11 0.29 K2O 0,00 0,00 0,00 0,00 0,00 0.01 0,00 0.01 0,00
Total 99.13 99.13 99.67 100.19 100.39 100.04 100.11 100.99 100.18 101.27 98.85 98.66 Total 100.23 100.82 100.11 99.93 101.1 100.31 99.8 99.95 98.65
An% 77.43 46.23 47.67 64.38 67.16 46.01 49.89 40.21 66.73 42.53 76.67 57.16 Mg# 74.49 78.29 81.95 75.16 80.16 76.79 76.5 74.51 82.13
En% 41.77 45.68 45.09 42.97 42.6 42.31 43.41 40.02 45,00

Clinopyroxene Orthopyroxene Amphibole

Year 2004 2005 2005 2005 2005 Year 2002 2002 2002 2002 2004 2004 2005 2005 Year 2002 2002 2002 2002 2002 2002 2002
Unit LF3 LF4 LF4 LF6 LF6 Unit PFs PFs LF1 LF1 LF3 LF3 LF4 LF4 Unit PFs PFs PFs PFs LF1 LF1 LF2
Sample REV42A REV56 REV56 REV62 REV62 Sample REV15B2 REV15B2 REV19 REV19 REV42A REV42A REV49 REV49 Sample REV15B2 REV15B2 REV16A REV16A REV19 REV19 REV31B
R C R C R C R C R C R C R C R C R C R C

SiO2 51.00 51.75 52.15 52.48 50.89 SiO2 54.81 54.33 54.36 53.91 53.99 54.21 53.88 53.58 SiO2 42.96 42.98 41.22 41.95 41.94 41.44 41.9
TiO2 0.74 0.44 0.54 0.44 0.99 TiO2 0.24 0.26 0.17 0.13 0.32 0.22 0.26 0.28 TiO2 3.85 3.64 2.86 2.58 2.39 2.03 2.75
Al2O3 2.67 2.12 2.16 1.61 3.05 Al2O3 0.63 1.02 0.74 0.4 1.56 0.86 1.44 1.67 Al2O3 9.96 9.96 13.45 13.25 13.01 12.74 13.36
FeO 8.11 8.26 7.97 8.55 8.86 FeO 16.64 17.41 16.81 17.72 15.33 14.75 16.17 16.86 FeO 12.7 12.51 9.63 10.04 10.43 12.64 8.98
MnO – – – – – MnO 0.68 0.59 0.65 0.71 – – – – MnO 0.18 0.22 0.05 0.1 0.17 0.24 0.08
MgO 14.68 15.06 16.02 16.44 15.75 MgO 26.67 26.57 26.45 22.91 26.4 26.47 26.4 26,00 MgO 13.8 13.87 14.97 14.78 15.24 13.39 15.52
CaO 20.9 21.2 20.45 19.42 19.69 CaO 1.35 1.29 1.38 2.27 1.38 1.51 1.32 1.34 CaO 11.19 11.01 11.99 11.85 11.43 11.33 11.91
Na2O 0.45 0.41 0.28 0.38 0.42 Na2O 0.01 0.04 0,00 0.01 0.04 0.06 0.03 0.02 Na2O 2.56 2.83 2.8 2.78 2.68 2.46 2.84
K2O 0,00 0.01 0,00 0.01 0.01 K2O 0,00 0.04 0,00 0,00 0,00 0.03 0.01 0,00 K2O 0.82 0.83 0.6 0.53 0.72 0.63 0.46
Total 98.56 99.24 99.56 99.32 99.65 Total 101.04 101.57 100.59 98.06 99.06 98.12 99.52 99.76 Total 98.07 97.91 97.68 98.15 98.15 96.94 97.94
Mg# 76.3 76.49 78.16 77.41 75.96 Mg# 74.09 73.13 73.72 69.75 75.45 76.17 74.43 73.33 Mg# 65.96 66.41 73.48 72.4 72.26 65.37 75.5
En% 42.85 43.14 45.52 46.71 45.14 En% 71.39 70.64 71.03 65.69 73.36 73.88 72.5 71.4

Amphibole Olivine Magnetite

Year 2002 2004 2005 2005 year 2002 2002 2002 2002 2002 2004 2005 2005 2005 2005 Year 2002 2002 2002 2004 2005 2005
Unit LF2 LF3 LF4 LF6 Unit PFs PFs LF1 LF1 LF2 LF3 LF4 LF4 LF6 LF6 Unit PFs PFs LF1 LF3 LF4 LF6
Sample REV31B REV42A REV56 REV62 Sample REV16A REV16A REV20 REV20 REV31B REV42A REV49 REV49 REV62 REV62 Sample REV15B2 REV15B2 REV19 REV50 REV49 REV62
R C C C C C C R C C C R C R C C C C C C

SiO2 41.27 44.68 42.88 42.82 SiO2 37.83 39.12 41.06 37.56 38.31 39.04 39.28 38.54 39.25 37.48 SiO2 0.08 0.05 0.13 0.09 0.07 0.09
TiO2 2.89 2.49 2.44 4.09 TiO2 0,00 0,00 0,00 0,00 0,00 0.03 0,00 0.01 0.01 0.02 TiO2 8.4 13.63 12.24 9.74 11.05 14.17
Al2O3 13.5 10.87 13.35 10.63 Al2O3 0.01 0.02 0,00 0.01 0.01 0,00 0.03 0,00 0.04 0,00 Al2O3 3.22 2.63 3.92 3.37 3.25 2.45
FeO 9.02 12.2 10.56 10.99 FeO 26.58 17.72 12.89 27.79 25.61 18.29 18.29 22.35 18.19 26.45 FeO 79.51 76.52 76.3 77.32 77.29 73.74
MnO 0.11 – 0.13 0.19 MnO 0.65 0.19 0.22 0.67 0.52 – – – – – MnO 0.4 0.52 0.43 – – –
MgO 15.53 13.45 12.67 14.6 MgO 35.94 43.35 47.45 34.5 37.18 41.08 42.38 38.08 42.42 35.11 MgO 3.06 2.71 2.88 3.39 3.17 2.95
CaO 12.02 11.41 10.87 11.27 CaO 0.1 0.14 0.17 0.12 0.21 0.1 0.11 0.12 0.12 0.13 CaO 0.05 0,00 0.03 0.03 0,00 0.12
Na2O 2.93 2.41 2.3 2.53 Na2O 0.04 0,00 0,00 0.02 0.02 0.01 0,00 0.01 0,00 0.03 Na2O 0,00 0.05 0.02 0.01 0.01 0.02
K2O 0.59 0.68 0.82 0.66 K2O 0,00 0.03 0.01 0.01 0,00 0.01 0,00 0.01 0.02 0,00 K2O 0.02 0,00 0,00 0.01 0,00 0.01
Total 98.32 98.24 96.04 97.82 Total 101.2 100.76 102.12 100.72 101.91 98.63 100.09 99.11 100.05 99.21 Total 94.84 96.27 96.18 93.94 94.85 93.55
Mg# 75.42 66.27 68.14 70.32 Fo 70.7 81.3 86.8 68.9 72.1 80 80.5 75.2 80.6 70.3
 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93 89

Table 3
Major element (wt.%) and volatile contents (ppm) with standard deviation of Matrix Glasses (MG) and Glass Inclusions (GI) from the 2002 El Reventador eruption. Analyses have been
carried out using a CAMECA SX 100 microprobe (Université Blaise Pascal, Clermont–Ferrand, France). The operating conditions were 5 um beam size, 15 kV accelerating voltage, 8 nA
beam current and 10 s counting time for Si, Ti, Al, Fe, Mg, Ca, Na and K; and 15 kV accelerating voltage, 50 nA beam current, and between 50 and 100 s counting time for Mn, P, Cr, Ni, F,
Cl and S. Volatile limit of detection are 400 ppm for F, and 20 ppm for S and Cl. ⁎ number of analyses for major/volatile elements

MG of PFs MG in enclaves MG of LF1 GI in Pl GI in Cpx GI in Opx GI in Ol GI in Pl2

54/33⁎ ± 14/11 ± 10/6 ± 33 ± 32 ± 14 ± 5 ± 7 ±

SiO2 64.82 0.88 72.04 1.2 68,00 3.02 66.23 2.02 66.23 1.01 65.35 0.96 62.99 1.41 57.77 2.26
TiO2 0.75 0.11 0.56 0.22 0.61 0.22 0.74 0.25 0.65 0.24 0.75 0.14 0.51 0.03 1.55 0.35
Al2O3 16.56 0.49 14.41 0.56 17.01 2.2 15.31 1.42 16.59 0.79 16.9 0.67 17.86 0.52 13.73 1.97
FeO 4.16 0.58 2.22 0.06 2.09 0.62 3.49 0.9 3.39 0.53 3.11 0.65 3.4 1.12 9.14 2.8
MnO 0.1 0.03 0.06 0.02 0.05 0.03 0.1 0.03 0.09 0.02 0.09 0.02 0.08 0.02 0.17 0.04
MgO 1.42 0.36 0.45 0.18 0.23 0.31 1.09 0.5 0.55 0.2 0.47 0.24 1.9 1.15 3.32 1.05
CaO 3.46 0.43 1.52 0.37 2.63 1.31 2.15 0.81 1.94 0.4 2.64 0.4 3.77 0.22 1.81 0.48
Na2O 4.69 0.5 4.05 0.33 5.29 1.22 3.13 0.71 4.36 0.76 4.25 0.49 4.07 0.75 2.52 1.55
K2O 3.47 0.3 4.37 0.17 3.88 1.04 3.24 0.62 4.03 0.24 3.66 0.26 3.63 0.3 4.16 1.03
P2O5 0.32 0.05 0.12 0.06 0.11 0.05 0.37 0.25 0.25 0.1 0.23 0.07 0.45 0.1 0.9 0.33
Total 99.75 99.78 99.85 95.86 98.1 1.15 97.46 1.37 98.66 95.06
S 116 66 40 20 26 13 181 70 206 86 293 106 601 104 565 221
Cl 1522 395 1448 201 1517 492 2862 1095 2765 545 2962 1182 3149 149 3147 1652
F 449 153 549 170 398 279 427 223 461 204 814 315 549 72 1615 722

allowing us to assume that measured S is mobilized mostly as SO 2. We
estimate the amount of the volatiles (SO2 and HCl) released to the
atmosphere during the paroxysmal phase of the eruption following
the methodology proposed by Self et al. (2004):
E ¼ m Mvð1 — WXLSÞðCINC — CMTXÞ=100
these features and the change in the eruptive style point to the in-
jection of mafic magma in 2004–05. For comparison, lavas erupted in
1912, 1944, 1972, 1974 and 1976 have similar, although somewhat
narrower, silica ranges (55–58 wt.% SiO2) and their mineralogy is
similar to that of the 2004–05 lavas (Pl+ Cpx + Opx+ Ol+ Mag±Amph).

where E is the SO2 or HCl emission in kg, MV corresponds to the mass

of erupted magma in kg, WXLS represents the mass fraction of crystals
in the magma, and CINC − CMTX is the difference (in wt.%) between the
average volatile concentrations of the glass inclusions (the pre-
eruptive magma) and the matrix (the degassed magma). The factor m
accounts for the difference between the molecular weights of SO 2 or
HCl, and S or Cl. We obtain ~ 80 kT SO 2 and ~ 280 kT HCl for the
paroxysmal phase of the 2002 eruption. This estimate is based upon
the total erupted mass of tephra and lavas in 2002 (MV =3.1 × 1011 kg),
the mass fraction of crystals (W XLS = 0.35, assuming a modal abun-
dance of ~ 30 vol.% and appropriate magma and crystal densities), the
average abundance of volatiles in glass inclusions (0.0227 wt.% S and
0.2863 wt.% Cl, Table 3), and the matrix glasses of lava flows
(0.0026 wt.% S and 0.1517 wt.% Cl). Considering S and Cl measured
in the volatile-rich glass inclusions (0.0583 wt.% S and 0.3148 wt.% Cl,
Table 3), the bulk mass of volatiles released into the atmosphere
during the entire 2002 eruption amounts to 225 kT SO2 and 340 kT
HCl. Our SO2 release estimates are consistent with those derived from
Total Ozone Mapping Spectrometer satellite data (TOMS), which
yielded values of 100 kT SO2 for the paroxysmal phase and 300 kT SO2
for the entire 2002 episode (Carn et al., 2008-this issue). The notable
HCl release relative to that of SO 2 is probably related to the Cl-rich
signature of El Reventador magmas.

4.2. Magmatic episodes

Although the silica contents in the 2002 eruptive products

decrease from 62–58 wt.% in pyroclastic deposits to 58–56 wt.% in
LF1 and 54–53 wt.% in LF2, the mineral assemblage Pl + Cpx + Opx+
Amph+ Mag± Ol remains unchanged. As in the 2002 lavas, the 2004–
05 samples show decreasing silica contents from 56 to 54 wt.%. For
similar silica content, the 2004–05 lavas are slightly enriched in Sr,
and depleted in HREE and Y compared to the 2002 eruptive products
(Fig. 3). Higher trace element ratios (e.g. La/Yb, Dy/Yb, Sr/Y, Fig. 5) are
noted in the 2004–05 lava compositions. This compositional differ-
ence coincides with a shift to the Pl + Cpx+ Ol+ Mag± Amph± Opx
mineral assemblage. In these lavas, amphibole was not in equilibrium Fig. 5. Trace elements ratios versus a differentiation index. (a) La/Yb, (b) Dy/Yb, and
with the liquid, while plagioclase, clinopyroxene and olivine crystals (c) Sr/Y versus Rb. The schematic insets in La/Yb and Dy/Yb versus Rb diagrams show
are euhedral and normally zoned, suggesting equilibrium. Altogether, expected fractionation effects for an amphibole-bearing cumulate and a gabbroic
cumulate (composed of plagioclase, pyroxene and olivine).
90 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93

Fig. 6. (a) Pseudo-ternary projection from PL into the plane OL-CPX-QZ (Sisson and Grove, 1993). (b) Pseudo-ternary projection from OL into the plane CPX-PL-QZ. Continuous line
represents liquids saturated with olivine, clinopyroxene and plagioclase under 0.2 GPa and water-saturated conditions. Dashed line represents a reaction boundary along which
olivine and liquid react to form amphibole.

4.3. Phase equilibrium constraints on pre-eruptive P–T conditions plagioclase, sanidine, amphibole, biotite, quartz, sphene, and magne-
tite/ilmenite. However, if quartz is absent in the equilibrium assem-
Experimental work on hydrous andesitic magmas have shown that blage, the geobarometer yields anomalously high pressures (Johnson
a stable assemblage of Pl+ Cpx + Amph+ Mag is attained under limited
P–T–H2O conditions (100–300 MPa, ~ 1000 °C, and N 5 wt.% H2O;
Moore and Carmichael, 1998; Pichavant et al., 2002; Grove et al.,
2003). We investigated the phase relations by comparing bulk-rock
compositions with the experimentally-determined phase stability in
the system Ol–Cpx–Pl–Qz at 200 MPa and water-saturated conditions
(Sisson and Grove, 1993). A plot of 2002 and 2004–05 whole-rock data
in the pseudo-ternary diagrams reveals a trend that crosses the mul-
tiple saturation boundary, where olivine, clinopyroxene and plagio-
clase coexist (Fig. 6a). Instead, the El Reventador trend plots on the
boundary along which olivine and liquids react to form amphibole
(Fig. 6b). The observations of amphibole rims around olivine and/or
clinopyroxene crystals (cf. Grove et al., 2003) and the high water
contents of glass inclusions (up to 6 wt.%) provide further support to
this process. Taking into account these observations, thermo-
barometers were applied in order to constrain the P–T conditions.
The composition of co-existing ortho- and clinopyroxene has been
used to estimate equilibrium temperatures using the QUILF program,
which assesses the oxides and silica equilibria in the reaction Quartz+
Ulvöspinel= ILmenite+ Fayalite (Andersen et al., 1993). Conditions for
application of this method include euhedral phenocrysts, similar Mg#,
homogeneous composition, and parallel tie-lines in the pyroxene
tetrahedron. Calculated temperatures range from 1000 to 1040 °C ±
27–43 °C for 2002 pyroclastic bombs and 1015 to 1025 °C ± 34–58 for
lavas. Since the QUILF formulation is not sensitive to pressure change
(only 5 °C difference for a change of 100 MPa), we fixed the pressure at
300 MPa (see below). The lack of ilmenite phenocrysts precludes the
use of coexisting Fe–Ti oxides thermometer. However, the existence of
some ilmenite exsolution bands in magnetite phenocrysts permits the
estimation of temperature and fO2 using the QUILF program, which
yields temperature values in the range 776–846 °C ± 28–46 and fO2
from − 11.3 to − 13.9 ± 0.5. Fig. 7. Fractional crystallization models for El Reventador magmas illustrated by (a) La/
The Al-in-hornblende geobarometer (Johnson and Rutherford, Yb and (b) Sr/Y versus Ni diagrams. Mineralogical composition of cumulate is estimated
1989) is widely used to estimate the total pressure in the pre-eruptive by mass-balance. M1: Fractional crystallisation model with a cumulate composed of
55%Pl + 22%Cpx + 17%Ol + 6%Mag. M2: Fractional crystallisation model with a cumulate
magma storage region. This geobarometer assumes that the alumi-
composed of 51%Pl + 27%Amph + 11%Cpx + 5%Ol + 6%Mag. Partition coefficients used for
nium content of the amphibole is proportional to the total pressure in these models are given in Martin (1987). Marks in model represent 10% of
particular conditions specified by a mineral assemblage comprised of crystallization. B: Primitive basaltic magma.
92 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
and Rutherford, 1989). On the other hand, for magmas crystallizing at
temperatures above those of the experimental calibration (740–
780 °C), the formulation needs adjustments (Anderson and Smith,
1995). Assuming that amphibole rims crystallized shortly before the
end of the crystallization sequence, we set the temperature at 850 °C.
The geobarometer yields an average pressure of 250 ± 80 MPa for the
2002 and 2004–05 eruptive products. Given that the mineral assem-
blage and geochemical compositions of our samples do not match the
original formulation of the geobarometer, these estimates correspond
to maximum values. The results of Al-in-hornblende geobarometry
are in agreement with water saturation pressure in glass inclusions
estimated using VolatileCalc 1.1 software (Newman and Lowenstern,
2002). For a rhyolitic melt at 850 °C and 5 wt.% H 2O, the saturation
pressure varies from 150 to 230 MPa (depending on the CO2 content of
the magma).
In summary, phase equilibrium constraints for hydrous andesitic
magmas and thermo-barometric estimates for the El Reventador
mineral assemblage point to pre-eruptive pressure-temperature con-
ditions ranging from 170 to 330 MPa, ~ 1050 to 850 °C, and a high
water content of the magmas (~ 5 wt.%). Assuming a bulk rock density
for crustal rocks of 2600 kg/m 3, the last crystallization phases (i.e.
amphibole crystallization) occurred at a depth of between 6.5 and
12.5 km.
4.4. Petrogenesis of the 2002 and 2004–05 magmatic suite
4.4.1. Magma mixing
Several disequilibrium features suggest that a mixing process
followed the injection of a mafic magma. The evidence is based upon
wide plagioclase composition ranges, reverse and oscillatory zonings
Fig. 8. Sketch model for the magmatic system below El Reventador volcano, depicting (a) recovery interval following the 1976 eruption ; (b) time period prior to the 2002 eruption;
(c) period immediately prior to the 2002 eruption; and (d) the 2004–05 eruptive episode. Vertical dimension and reservoir size are not drawn to scale. The colour gradation in the
reservoir reflect their chemical zonation. Dotted line represents the avalanche scar.
in plagioclase and clinopyroxene, resorption surfaces, and phenocrysts
not in equilibrium with the matrix. Additionally, some minerals, such
as Ca-rich plagioclase, Al-rich clinopyroxene, and olivine have not
been reproduced in experimental studies in andesitic magmas (cf.
Moore and Carmichael, 1998; Pichavant et al., 2002). These composi-
tions may be attributed to phenocrysts crystallized in a different
magma.
Two different mixing mechanisms are generally invoked to explain
these disequilibrium features including: (1) a physical mixing of
magmas of contrasted composition, temperature and physical proper-
ties, and (2) the recharge of a magmatic reservoir by mafic magma,
producing temperature increase and thermal convection (Couch et al.,
2001). This latter mechanism is known as self-mixing or thermal
mixing. The lack of textural banding and mafic enclaves in El Rev-
entador rocks does not favour the physical mixing scenario involving
magmas of contrasting compositions. But, for magmas with similar
compositions, the petrological identification of magma mixing will be
difficult and this process cannot be completely ruled out. However, the
disequilibrium features observed in minerals point to a mafic recharge
into an evolved magma reservoir, favouring the self-mixing
mechanism.
4.4.2. Fractional crystallization
The whole-rock geochemical trends of the 2002 and 2004–05
rocks (Figs. 2 and 3), the good correlations in the variation diagrams
(Fig. 3), and the occurrence of normally-zoned minerals support a
petrogenesis dominantly controlled by fractional crystallization
processes. In particular, trace element trends point to olivine and/or
clinopyroxene fractionations (Ni and Cr decrease with differentiation),
and to plagioclase fractionation (Sr decrease). Low MREE and HREE
 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
concentrations point to amphibole fractionation. Because amphibole
preferentially incorporates MREE over HREE, the amphibole role is
illustrated by a La/Yb increase, and a decrease in Dy/Yb during differ-
entiation (Fig. 5). In contrast, fractionation of a gabbroic assemblage
(Pl+ Cpx+ Ol) has no influence on REE patterns.
The above geochemical and mineralogical constraints allow us to
test a fractional crystallization model based on major element mass-
balance calculations. The calculated modal compositions of the cumu-
late and the degree of fractionation are subsequently used in trace
element modelling (Rayleigh-type fractional crystallization). In Fig. 7,
La/Yb and Sr/Y have been plotted against Ni to illustrate this model.
From hypothetical hydrous basalts (B in Fig. 7) to basic andesite
compositions, the differentiation is dominantly controlled by fractio-
nation of Pl + Cpx+ Ol+ Mag (gabbroic assemblage). The fractionation
of this cumulate (M1 in Fig. 7) may account for the compositional
diversity of LF2 lavas as well as many other historical lava flows. The
transition from the 2002 andesites to basic andesites is consistent
with a model in which an amphibole-bearing cumulate fractionates.
Importantly, a crystal fractionation model fails to explain the
petrological relation between the 2002 eruptive products and 2004–
05 lavas. These lava compositions may have resulted from early
fractionation of amphibole that probably occurred in a deep reservoir
or during the ascent of hydrous mafic magmas.
5. Concluding reTharks: characteristics of the ThagTha
pluThbing systeTh
Field observations and petrological studies of the 2002 and 2004–
05 eruptions of El Reventador highlight the existence of two volcanic
and magmatic episodes. The 2002 episode started with an explosive
phase removing an andesitic magma reservoir, which was then fol-
lowed by effusive activity releasing andesites and basaltic andesites
lavas. In contrast, during the mostly effusive 2004–05 episode, only
basaltic andesitic lava flows were emplaced.
Mass balance calculations based upon microprobe analyses of
matrix glasses from pyroclastic bombs and glass inclusions trapped
inside phenocrysts suggest that ~ 80 kT of SO2 and 280 kT of HCl were
degassed during the paroxysmal phase of the eruption. The highly
explosive character of this eruption is explained by elevated water
(and other volatile species) content of the pre-eruptive magma. A
closed-system degassing regime is also consistent with the rapid
increase in eruption intensity and corresponding magnitude of the
initial explosive onset (VEI 4). Phreatomagmatism probably played an
additional role in enhancing the explosivity, as suggested by phe-
nomenological and geological observations.
Both the 2002 and 2004–05 magmatic episodes were largely con-
trolled by a two-step fractional crystallization process. The magmatic
evolution, from basalts to basic andesites was mainly controlled by
fractionation of Pl + Cpx + Ol+ Mag. This was followed by Ol and/or Cpx
reaction with the magma to produce amphibole, whose subsequent
fractionation produced more evolved magmas.
We synthesize a dynamic model of the El Reventador magmatic
system based upon petrological data. Prior to the 2002 eruption, an
andesitic magma body existed below the volcano at a depth of
about 10 km (Fig. 8a). It was disturbed by the recharge of volatile-
rich basic magma ascending from a deeper reservoir (Fig. 8b). The
arrival of this hydrous basic magma induced amphibole crystal-
lization as well as strong convection in the shallow reservoir
(Fig. 8c). A second, incipiently-crystallized magma batch ascended
later from a deep reservoir to produce the 2004–05 lavas (Fig. 8d).
The magmatic recharges related to the 2002 and 2004–05 erup-
tions, and the fact that historical eruptions of El Reventador have
occurred in clusters of ~ 4 years, similar to the 2002 and 2004–05
activity, suggest that this mechanism is common at this volcano
and leads to the high frequency of its eruptions. On the other hand,
the crystallization of amphibole at an early stage of differentiation
91
represents a distinctive and unique aspect of El Reventador vol-
cano, compared to the other volcanoes of Ecuador, where amphi-
bole only occurs in more evolved siliceous andesites and dacites,
such as in Cotopaxi (Hall and Mothes, 2007), Tungurahua (Hall
et al., 1999) or Sangay (Monzier et al., 1999). We propose that the
early amphibole crystallization is related to the volatile-rich sig-
nature of primitive El Reventador magmas.
AcknowledgThents
This work was financially supported by the Instituto Geofísico of
the Escuela Politécnica Nacional (IG-EPN) of Quito (Ecuador) and the
French Institut de Recherche pour le Développement (IRD) in the
framework of a cooperative agreement.
We thank J.-L. Devidal for help with the acquisition of microprobe
data and D. Andrade, D. Barba and J. Johnson for providing samples
from the 2005 eruptions. D. Andrade also provide a previous version
of the Fig. 1. Constructive reviews of two anonymous reviewers helped
to improve the manuscript.
References
Aguilera, E., Almeida, E., Balseca, W., Barberi, F., Innocenti, F., Coltelli, M., Pasquare, G.,
1988. El Reventador: an active volcano in the sub-andean zone of Ecuador.
Rendiconti della Societa Italiana di Mineralogia e Petrologia 43, 853–875.
Andersen, D.J., Lindsley, D.H., Davidson, P.M., 1993. QUILF: a Pascal program to assess
equilibria among Fe–Mg–Mn–Ti oxides, pyroxenes, olivine, and quartz. Comput.
Geosci. 19, 1333–1350.
Anderson, J.L., Smith, D.R., 1995. The effects of temperature and fO2 on the Al-in-
hornblende barometer. Am. Mineral. 80, 549–559.
Burgisser, A., Scaillet, B., 2007. Redox evolution of a degassing magma rising to the
surface. Nature 445, 194–197.
Carn, S.A., Krueger, A.J., Arellano, S., Krotkov, N.A., Yang, K., 2008. Daily monitoring of
Ecuadorian volcanic degassing from space. J. Volcanol. Geotherm. Res. 176, 141–150
(this issue). doi:10.1016/j.jvolgeores.2008.01.029.
Couch, S., Sparks, R.S.J., Carroll, M.R., 2001. Mineral disequilibrium in lavas explained by
convective self-mixing in open magma chambers. Nature 411, 1037–1039.
Cotten, J., Le Dez, A., Bau, M., Caroff, M., Maury, R.C., Dulski, P., Fourcade, S., Bohn, M.,
Brousse, R., 1995. Origin of anomalous rare-earth element and Yttrium enrichments
in subaerial exposed basalts: evidence from French Polynesia. Chem. Geol. 119, 115–
138.
Devine, J.D., Sigurdsson, H., Davis, A.N., Self, S., 1984. Estimates of sulphur and chlorine
yield to the atmosphere from volcanic eruptions and potential climate change.
J. Geophys. Res. 89, 6309–6325.
Gill, J.G., 1981. Orogenic Andesites and Plate Tectonics. Springer-Verlag, New York.
389 pp.
Grove, T.L., Elkins-Tanton, L.T., Parman, S.W., Chatterjee, N., Müntener, O., Gaetani, G.,
2003. Fractional crystallization and mantle-melting controls on calc-alkaline
differentiation trends. Contrib. Mineral. Petrol. 145, 515–533.
Gutscher, M.-A., Malavieille, J., Lallemand, S., Collot, J.-Y., 1999. Tectonic segmentation of
the North Andean margin: impact of the Carnegie Ridge collision. Earth Planet. Sci.
Lett. 168, 255–270.
Hall, M.L., 1980. El Reventador, Ecuador, un volcán activo de los Andes Septentrionales.
Revista Politécnica, Monografía de Geología 5, 123–136.
Hall, M.L., Robin, C., Beate, B., Mothes, P., Monzier, M., 1999. Tungurahua Volcano, Ecuador:
structure, eruptive history and hazards. J. Volcanol. Geotherm. Res. 91, 1–21.
Hall, M.L., Ramón, P., Mothes, P., Le Pennec, J.-L., García, A., Samaniego, P., Yepes, H.,
2004. Volcanic Eruptions with little warning: the case of Volcán Reventador's
Surprise November 3, 2002 Eruption, Ecuador. Rev. Geol. Chile 31, 349–358.
Hall, M., Mothes, P., 2007. The rhyolitic–andesitic eruptive history of Cotopaxi Volcano,
Ecuador. Bull. Volcanol. doi:10.1007/s00445-007-0161-2.
Johnson, J., Ramón, P., Andrade, D., Hall, M.L., 2006. Reventador volcano: 2002 to
present, explosive and effusive activity. Field Guide A5, Cities on Volcanoes 4
Conference, Quito. 14 pp.
Johnson, M.C., Rutherford, M.J., 1989. Experimental calibration of the aluminium-in-
hornblende geobarometer with application to Long Valley caldera (California)
volcanic rocks. Geology 17, 837–841.
Leake, B.E., Woolley, A.R., Arps, C.E.S., Birch, W.D., Gilbert, M.C., Grice, J.D., Hawthorne,
F.C., Kato, A., Kisch, H.J., Krivovichev, V.G., Linthout, K., Laird, J., Mandarino, J.A.,
Maresch, W.V., Nickel, E.H., Rock, N.M.S., Schumacher, J.C., Smith, D.C., Stephenson,
C.N., Ungaretti, L., Whittaker, E.J.W., Youzhi, G., 1997. Nomenclature of amphiboles:
report of the Subcommittee on Amphiboles of the International Mineralogical
Association, commission on new minerals and minerals' names. Am. Mineral. 82,
1019–1037.
Le Pennec, J.-L., Hidalgo, S., Samaniego, P., Ramón, P., Yepes, H., Eissen, J.-P., 2003.
Magnitud de la erupción del 3 de noviembre del 2002 del volcán El Reventador.
Memorias, Quintas Jornadas en Ciencias de la Tierra, 8–11 abril 2003. Escuela
Politécnica Nacional, Quito, pp. 97–99.
Martin, H., 1987. Petrogenesis of Archaean trondhjemites, tonalites, and granodiorites
from eastern Finland: major and trace element geochemistry. J. Petrol. 28, 921–953.
92 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 176 (2008) 82–93
Monzier, M., Robin, C., Samaniego, P., Hall, M.L., Cotten, J., Mothes, P., Arnaud, N., 1999.
Sangay volcano, Ecuador: structural development, present activity and petrology.
J. Volcanol. Geotherm. Res. 90, 49–79.
Moore, G., Carmichael, I.S.E., 1998. The hydrous phase equilibria (to 3 kbar) of an
andesite and basaltic andesite from western Mexico: constraints on water content
and conditions of phenocryst growth. Contrib. Mineral. Petrol. 130, 304–319.
Newman, S., Lowenstern, J.B., 2002. VolatileCalc: a silicate melt-H2O–CO2 solution
model written in Visual Basic for EXCEL. Comput. Geosci. 28, 597–604.
Oppenheimer, C., 2003. Volcanic Degassing, pp. 123–166. In The Crust (ed. R.L. Rudnick)
Vol. 3. Treatise on Geochemistry (eds. H.D. Holland and K.K. Turekian), Elsevier-
Pergamon, Oxford.
Pichavant, M., Martel, C., Bourdier, J.L., Scaillet, B., 2002. Physical conditions, structure,
and dynamics of a zoned magma chamber: Mount Pelée (Martinique, Lesser
Antilles Arc). J. Geophys. Res. 107, 2093. doi:10.2029/2001JB000315.
Ramon, P., Andrade, D., Hall, M.L., Troncoso, L., Stevens, N., Samaniego, P., Rivero, D.,
Yepes, H., Boker, F., 2006. Reventador volcano 2004 renewed activity. Monitoring
activities and early warning alert. Abstract, Cities on Volcanoes 4 Conference, Quito,
p. 154.
Saito, G., Uto, K., Kazahaya, K., Shinohara, H., Kawanabe, Y., Satoh, H., 2005. Petrological
characteristics and volatile content of magma from the 2000 eruption of
Miyakejima Volcano, Japan. Bull. Volcanol. 67, 268–280.
Samaniego, P., Eissen, J.P., Hall, M.L., Mothes, P., Le Pennec, J.L., Ramón, P., Cotten, L.,
Chavrit, D., 2006. Petrological monitoring of the 2002–2005 eruption of El
Reventador volcano (Ecuador). Abstract, Cities on Volcanoes 4 Conference, Quito,
p. 72.
Scarpa, R., Tilling, R.I. (Eds.), 1996. Monitoring and Mitigation of Volcano Hazards.
Springer, Berlin. 841 pp.
Self, S., Gertisser, R., Thordarson, T., Rampino, M.R., Wolff, J.A., 2004. Magma volume,
volatile emissions, and stratospheric aerosols from the 1815 eruption of Tambora.
Geophys. Res. Lett. 31, L20608. doi:10.1029/2004GL020925.
Simkin, T., Siebert, L., 1994. Volcanoes of the World, 2nd edition. Geoscience Press,
Phoenix, AZ. 349 pp.
Sisson, T.W., Grove, T.L., 1993. Experimental investigations of the role of H2O in calc-
alkaline differentiation and subduction zone magmatism. Contrib. Mineral. Petrol.
113, 143–166.
Troncoso, L., Johnson, J.B., Ruiz, G., Andrade, D., Barba, D., Ramón, P., Rivero, D., Hall, M.L.,
Lema, V., Viracucha, G., 2006. Evolution of the seismic activity of El Reventador
volcano during its current eruptive phase (2004–2005). Abstract, Cities on
Volcanoes 4 Conference, Quito, p. 112.
Wallace, P.J., 2001. Volcanic SO2 emissions and the abundance and distribution of
exsolved gas in magma bodies. J. Volcanol. Geotherm. Res. 108, 85–106.
Wallace, P.J., 2005. Volatiles in subduction zone magmas: concentrations and fluxes
based on melt inclusion and volcanic gas data. J. Volcanol. Geotherm. Res. 140,
217–240.