lable at ScienceDirect

Journal of South American Earth Sciences 48 (2013) 193e208

Contents lists avai

Journal of South American Earth Sciences

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

Crystallization conditions and petrogenesis of the lava dome from thew900 years BP eruption of Cerro Machín Volcano, Colombia

Kathrin Laeger a, Ralf Halama a,b,*, Thor Hansteen c, Ivan P. Savov d, Hugo F. Murcia e,Gloria P. Cortés f, Dieter Garbe-Schönberg a

a Institut für Geowissenschaften and SFB 574, Universität Kiel, Ludewig-Meyn-Str. 10, 24118 Kiel, Germanyb Institute of Earth and Environmental Science, University of Potsdam, Karl-Liebknecht-Str. 24-25, 14476 Potsdam, GermanycGEOMAR Helmholtz-Zentrum für Ozeanforschung Kiel, Wischhofstrasse 1-3, 24148 Kiel, Germanyd School of Earth and Environment, University of Leeds, Leeds LS2 9JT, United Kingdome School of Environment, The University of Auckland, Private Bag 92019, Auckland 1142, New ZealandfObservatorio Vulcanológico y Sismológico de Manizales, Servicio Geológico Colombiano, Avenida 12 de Octubre No. 15-47, Manizales, Colombia

a r t i c l e i n f o

Article history:Received 1 March 2013Accepted 26 September 2013

Keywords:Colombian AndesCerro Machín VolcanoMagma mixingAmphibole geothermobarometryTrace element geochemistry

* Corresponding author. Institute of Earth and Envirof Potsdam, Karl-Liebknecht-Str. 24-25, 14476 Potsdam5783; fax: þ49 331 977 5700.

E-mail address: rhalama@geo.uni-potsdam.de (R.

0895-9811/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jsames.2013.09.009

a b s t r a c t

The last known eruption at Cerro Machín Volcano (CMV) in the Central Cordillera of Colombia occurredw900 years BP and ended with the formation of a dacitic lava dome. The dome rocks contain bothnormally and reversely zoned plagioclase (An24e54), unzoned and reversely zoned amphiboles ofdominantly tschermakite and pargasite/magnesio-hastingsite composition and olivine xenocrysts(Fo¼ 85e88) with amphibole/clinopyroxene overgrowth, all suggesting interaction with mafic magma atdepth. Plagioclase additionally exhibits complex oscillatory zoning patterns reflecting repeated replen-ishment, fractionation and changes in intrinsic conditions in the magma reservoir. Unzoned amphibolesand cores of the reversely zoned amphiboles give identical crystallization conditions of 910 � 30 �C and360 � 70 MPa, corresponding to a depth of about 13 � 2 km, at moderately oxidized conditions(fO2

¼ þ0.5 � 0.2 DNNO). The water content in the melt, calculated based on amphibole chemistry, is7.1 � 0.4 wt.%. Rims of the reversely zoned amphiboles are relatively enriched in MgO and yield highercrystallization temperatures (T ¼ 970 � 25 �C), slightly lower melt H2O contents (6.1 � 0.7 wt.%) andoverlapping pressures (410 � 100 MPa). We suggest that these rims crystallized following an influx ofmafic melt into a resident magma reservoir at mid-crustal depths, further supported by the occurrence ofxenocrystic olivine. Crystallization of biotite, albite-rich plagioclase and quartz occurred at comparativelylow temperatures (probably <800 �C) during early stages of ascent or storage at shallower levels. Basedon amphibole mineral chemistry, the felsic resident melt had a rhyolitic composition (71 � 2 wt.% SiO2),whereas the hybrid magma, from which the amphibole rims crystallized, was dacitic (64 � 3 wt.% SiO2).The bulk rock chemistry of the CMV lava dome dacites is homogenous. They have elevated (La/Nb)Nratios of 3.8e4.5, typical for convergent margin magmas, and display several geochemical characteristicsof adakites. Both Sr and Nd isotope compositions (87Sr/86Sr w0.70497, 143Nd/144Nd w0.51267) are amongthe most radiogenic observed for the Northern Volcanic Zone of the Andes. They are distinct fromoceanic crust that has been subducted in the region, pointing to a continental crustal control on theisotope composition and hence the adakitic signature, possibly in a crustal “hot zone”.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Cerro Machín Volcano (CMV), a composite volcano located inthe Colombian Central Cordillera (Fig.1A), is considered to be one of

onmental Science, University, Germany. Tel.: þ49 331 977

Halama).

All rights reserved.

the most dangerous active volcanoes in Colombia due to its provenability to produce large explosive eruptions and its location in astrategic region for the country (Cortés, 2001; Murcia et al., 2008,2010). In the last 5000 years, CMV has produced at least six ma-jor dacitic eruptions, four plinian e sub-plinian and two vulcanian,with volcanic activity generating pyroclastic flows, pyroclasticsurges, pyroclastic falls and lahars (Cortés, 2001; Rueda, 2005;Murcia et al., 2008, 2010). Eruptions were dated at w5000,w4600, w3600, w2600, w1200 and w900 years BP based on av-erages of several individual 14C radiometric ages (Méndez et al.,

Fig. 1. Map and photographs of Cerro Machín Volcano. A) Map of the western part of Colombia with three groups of active volcanoes. CMV is situated at the southern end of thenorthern group, about 20 km west of Ibagué. B) Crater and lava dome of CMV in viewing direction NE with sample locations GEO-D1, -D2A and -D3B. Nevado del Tolima Volcano isseen in the distance. C) View onto CMV towards the SW with sample location GEO-D4 and Cajamarca in w6 km distance.

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208194

2002; Rueda, 2005). The last eruption of CMV, w900 years BP,produced pyroclastic flows associated with a vulcanian eruptionthat terminated with the emplacement of an intra-crater daciticdome (Thouret et al., 1995; Rueda, 2005; Murcia et al., 2010). Five ofthe major eruptions have generated lahars, the biggest of whichreached distances of more than 100 km from the volcano edifice(Cepeda et al., 1995, 1999; Cortés, 2001; Cortés et al., 2006; Murciaet al., 2008). Today, similar eruptions would affect nearly 1 millioninhabitants (Méndez et al., 2002; Murcia et al., 2008) in an area ofw2000 km2, including the town of Ibagué in the West, where py-roclastic deposits of previous eruptions have been found. From2000 to 2010, seismic activity has increased at CMV, reaching up to9000 volcano-tectonic (VT) earthquakes per year in 2010 (Londoño,2011). 6500 and about 2000 VT earthquakes were recorded in 2011and 2012, respectively. Two VT earthquakes on the 7th of October2012 reached magnitudes of 4.7 and 4.1 (Londoño, 2011). Hypo-center locations clearly show threemain seismic zones; one locatedbeneath the central dome at 3e5 km depth; the second one, locatedto the SE of the central dome (5 km away) at 5e8 km depth; and thelast, located 8e10 km to the SE of the central dome at 12e18 kmdepth (Londoño, 2011). It seems that these seismic sources canrepresent a fault/dyke system related to the main path of magmaascent at CMV (Londoño, 2011). At the time of writing (February2013), swarms of volcano-tectonic events continue to occur underthe volcano. Moreover, ground deformation, changes in composi-tion and temperature of hot springs and fumarolic activity withradon outputs were observed, suggesting that the volcano has beengradually increasing its activity (Londoño et al., 2007), and thisconstitutes a potentially increasing threat to the local population(Murcia et al., 2008). A better understanding of the CMV magmatic

system is therefore particularly urgent, given the recent observa-tion that ascent of hydrous felsic magmas can occur rapidly andlead to unexpected explosive eruptions (e.g., Castro and Dingwell,2009).

For an assessment of volcanic hazards in the case of futureeruptions, it is essential to understand past volcanic eruptions andthe dynamics of a given magmatic system (Sparks, 2003). Criticalinformation about volcanic eruptions includes knowledge aboutmagma storage conditions, manner and rate of magma ascent,volatile concentrations and eruption trigger processes (e.g.,Bachmann et al., 2002; Browne and Gardner, 2006; Pallister et al.,1992; Rutherford and Devine, 2003; Prouteau and Scaillet, 2003;Savov et al., 2008). Geothermobarometers can be used to deter-mine PeT-conditions during magma storage and ascent, and theyprovide information about water contents and oxygen fugacities ofthe erupted magmas (Putirka, 2008). Amphibole is a particularlyimportant phenocryst phase in many convergent margin magmasand its composition can provide information about magma storageand ascent (Thornber et al., 2008; Browne and Gardner, 2006;Bachmann and Dungan, 2002; Sato et al., 1999; Savov et al.,2008). Combined with seismic observations, petrologic con-straints on the magmatic system of a volcano are a powerful tool tounderstand magma dynamics and eruption triggering processes inmore detail. Magma mixing, where fresh magma batches interactwith residual magma and cause an increase in heat flux and volatilecontents, is of particular importance for explosive eruptions(Sparks et al., 1977; Eichelberger, 1980, 1995). Magma mixing ap-pears to be a ubiquitous process in many ancient and recenteruptions (e.g., Bachmann et al., 2002; Browne et al., 2006; Di Muroet al., 2008; Gourgaud and Thouret, 1990; Swanson et al., 1994),

Table 1Petrography of the lava dome dacites from Cerro Machín Volcano.

Texture Porphyritic, partially glomerophyric

Vesicles �0.5 mm in diameter, 3e5 vol.%

Mineral Size (mm) vol. % Grain shape Inclusions Characteristics Occurrence

Plagioclase 0.08e6.4 25e35 � Prismatic� Euhedral� Subhedral

� Fluids� Ilmenite� Amphibole� Apatite� Zircon

� Polysynthetic twinning� Twins� Dissolution rims� Oscillatory zoning

� Fragments� Matrix-forming aggregates

with amph þ bt

Quartz 0.5e4 5e10 � Prismatic� Anhedral� Rounded

Amphibole 0.06e1.6 5e7 � Acicular� Euhedral� subhedral

� Biotite� Apatite� Ilmenite

� Fragments� Rim around ol� Aggregates with plag þ bt

Biotite 0.1e2.0 3e5 � Tabular� Euhedral� Subhedral

� Ilmenite� Apatite� Zircon

� Fragments

FeeTi-oxides 0.1e0.2 <2 � EuhedralOlivine 0.1e0.7 <2 � Anhedral � Reaction rims to amphibole

and pyroxeneApatite <0.05 <2 � Euhedral

� AcicularPyroxene <0.15 <<1 � Subhedral � Rim around olivine

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 195

including the 1985 eruption of Colombia’s Nevado del Ruiz volcano(Gourgaud and Thouret, 1990; Sigurdsson et al., 1990).

In this study, we investigate dacites from the lava dome of theyoungest eruption of CMV, dated at w900 years BP (Thouret et al.,1995; Rueda, 2005; Murcia et al., 2010), to determine melt watercontents, crystallization temperatures and magma chamber depthsusing geothermobarometric methods. Chemical characteristics ofmagmas in equilibrium with amphibole phenocrysts are recon-structed using the methodology of Ridolfi and Renzulli (2012).Moreover, we provide geochemical data of the CMV dacites in orderto constrain their petrogenesis and compare them to other vol-canoes of the Andean Northern Volcanic Zone (NVZ), many ofwhich are characterized by an adakitic signature.

2. Geological setting

Cerro Machín Volcano is one of the several active volcanoes inthe northern Central Cordillera of Colombia (Fig. 1A). It is located150 km southwest of the capital Bogotá and 17 km northwest ofIbagué at the intersection of the Cajamarca and Machín faults(Rueda, 2005). CMV is part of the Northern Volcanic Zone (NVZ) ofthe Andes, which stretches from Ecuador to central Colombia andresults from subduction of the Nazca Plate beneath the SouthAmerican Plate in eastward direction (Hall and Wood, 1985; Ramosand Alemán, 2000). The continental crust has a thickness of 40e50 km (Taboada et al., 2000) and the basement of CMV is formed byvariably deformed, metasedimentary rocks of the CajamarcaComplex (Maya and González, 1995; Toro et al., 2012). The Caja-marca Complex comprises para- and orthogneisses, phyllites,quartzites, greenschists, graphitic schists and locally marbles(Vargas et al., 2005; Villagómez and Spikings, 2013). These rockswere deposited during Triassic rifting between South and NorthAmerica and subsequently underwent anatexis. The maximumdepositional age is constrained to w220e240 Ma based on UePbdetrital zircon ages (Villagómez et al., 2011). The Cajamarca Com-plex represents a part of the continental crust of the CentralCordillera that defines the pre-Cretaceous continental margin.Based on geophysical data, the Cajamarca Complex appears todefine a broad syncline, reaching a maximum thickness ofw10 km,underlain by amphibolites (Vargas et al., 2005). Subsequent to

Triassic rifting, subduction-related magmatism along the Colom-bian margin occurred from w180 to w147 Ma. In the CajamarcaComplex, this phase is represented by the intrusion and contactmetamorphism by Jurassic, calc-alkaline I-type granitoids of theIbagué Batholith (Villagómez et al., 2011). The intersection of twomajor faults, the Cajamarca and Machín faults, coincides with thelocation of CMV (Rueda, 2005; Murcia et al., 2008). CMV belongs tothe Cerro Bravo-Cerro Machín volcanic chain (Fig. 1A) and, withonly 2750 m altitude, has the most inconspicuous relief comparedto its neighbors. However, based on evidence from past eruptions,CMV is one of the most explosive volcanoes in Colombia and it canbe considered as the most dangerous active volcano of the countrybecause of the inhabitants and the infrastructure that might beaffected (Murcia et al., 2008). The volcanic edifice consists of a ringof pyroclastic material with a diameter of 2.4 km and comprises adacitic intra-crater lava dome (Arango and Castañeda, 2012,Fig. 1B,C).

3. Petrography

Four samples from the central dome of CMV were investigated(Fig. 1B,C), one from the Southwest (D1), two from the West (D2Aand D3B) and one from the North (D4). Coordinates of the samplesare given in the electronic Supplementary Material. All samplesshow a fine-grained, porphyritic texture with plagioclase, amphi-bole, quartz, and biotite phenocrysts and accessory FeeTi oxides,olivine and apatite (Table 1). Themicrocrystalline groundmass (45%on average), dominated by plagioclase microphenocrysts, containsvesicles (3e5%) up to 0.5 mm in diameter and is characterized bylight and dark gray streaks. Plagioclase phenocrysts (25e35%) areeuhedral to subhedral and occasionally broken (Fig. 2AeC). Theycan be grouped as three distinct petrographic types. Type A hasresorbed cores, overgrown by rims with strong oscillatory zoningforming euhedral crystals (Fig. 2A). Type B shows a continuousoscillatory zoning from core to rim (Fig. 2B). In Type C, compara-tively homogenous cores that preserve the euhedral shape ofplagioclase are surrounded by extremely resorbed outer rims,leading to rounded anhedral phenocrysts (Fig. 2C). Amphibolephenocrysts (5e7%) are also euhedral to subhedral and two typescan be distinguished petrographically (Fig. 2DeF). Type I comprises

Fig. 2. Photomicrographs showing characteristic textures of plagioclase (AeC; crossed polarizers) and amphibole (DeF; plain polarized light) in Cerro Machín dacites;plag ¼ plagioclase, amph ¼ amphibole, bt ¼ biotite, ol ¼ olivine. A) Plagioclase phenocryst with corroded core and oscillatory zoned rim (Type A). B) Oscillatory zoning in plagioclasecrystals of Type B. C) Plagioclase with resorbed rim and rounded shape (Type C). D) Euhedral amphibole with limited core-to-rim chemical variation (Type I). E) Amphibole crystal ofType II with dark brown core and light brown, more Mg-rich rim. F) Olivine xenocryst overgrown by amphibole. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208196

optically nearly homogenous crystals with olive-green to browncolors (Fig. 2D). These amphiboles typically form aggregates andalso occur as inclusions in plagioclase. Some Type I crystals lackreaction rims, whereas others have reaction rims of up to 50 mmthickness. Type II amphibole phenocrysts are characterized by adark brown core with a yellowish rim and a small (w5e25 mm)reaction rim (Fig. 2E). Quartz (5e10%) occurs as strongly roundedphenocrysts. Biotite (3e5%) is euhedral to subhedral, and somebiotites have coarse reaction coronas of variable thickness (w25e50 mm; Fig. 2F). The accessory phases ilmenite and apatite occur asinclusions in plagioclase, amphibole and biotite and also in thematrix. Rare olivine xenocrysts are anhedral (0.1e0.7 mm) and areovergrown by amphibole and clinopyroxene (Fig. 2F).

4. Analytical methods

Major and trace elements of whole rock samples were analyzedat the Institute of Geosciences at Kiel University by X-Ray Fluo-rescence (Philips PW 1480 X-ray spectrometer) and InductivelyCoupled Plasma Mass Spectrometry (Agilent 7500cs ICP-MS),respectively. Mineral chemical compositions were determined us-ing a JEOL JXA 8200 electron microprobe at GEOMAR, Kiel. Stron-tium (Sr) and neodymium (Nd) isotope analyses were performedon a Thermo Finnigan Triton multicollector mass spectrometer atthe School of Earth and Environment, University of Leeds. Detailsabout the analytical methods are given in the electronicSupplementary Material.

5. Results

5.1. Whole rock geochemistry

Whole rock SiO2 (65.3e66.2 wt.%) and total alkali (6.7e6.8 wt.%)contents show little variation (Table 2, Fig. 3A). All dome samplescan be classified as subalkaline dacites, plotting in the field of themedium-K series from Le Maitre et al. (1989; not shown). Thecompositional variation in incompatible trace element

concentrations is also very limited. Chondrite-normalized (CN) REEpatterns are fractionated with (La/Yb)CN ratios between 14 and 17and (La/Sm)CN between 3.4 and 3.7, similar to dacites from otherColombian NVZ volcanoes (Fig. 3B). The samples display markeddepletions in Nb and Ta relative to elements with a similar degreeof incompatibility (Fig. 3C), typical of calc-alkaline suites atconvergent plate margins, as exemplified by (La/Nb)N ratios(N ¼ normalized to primitive mantle values) of 3.8e4.5 (Table 2).Among the fluid-mobile trace elements, Ba, Pb, Sb and Li showstrong relative enrichments. The individual SreNd isotopic ana-lyses overlap within analytical uncertainty, and average 87Sr/86Srand 143Nd/144Nd ratios are w0.70497 and w0.51267, respectively(Fig. 3D).

Based on their major and trace element composition, the CMVdacites can be classified as adakites. We use “adakite” as adescriptive term, based on the distinctive geochemical parameters,without any petrogenetic implication. In particular, “adakite” is notused as an equivalent to “slab melt” since the origin of the adakiticsignature is debated (see Discussion). The CMV dacites displaychemical characteristics typical for high-silica adakites such asrelatively low K/Rb ratios (Fig. 4A), high Cr/Ni ratios at moderateTiO2 contents (Fig. 4B), and elevated SiO2 at moderate K2O/Na2O(Fig. 4C). High SiO2 (>56 wt.%) and elevated Al2O3 contents(>15 wt.%), as well as Na2O contents between 3.5 and 7.5 wt.%, arealso consistent with an adakitic geochemical signature (seeSamaniego et al., 2005; for a list of parameters used to discriminateadakites from other volcanic arc rocks). Moreover, Sr is >400e600 ppm, Y is <18 ppm and Yb <1.9 ppm, resulting in La/Yb ratios>20 and Sr/Y > 40 (Table 2, Fig. 4D). On diagrams that were con-structed to discriminate between high- and low-silica adakites(Martin et al., 2005;Moyen, 2009), the CMV dacites consistently fallinto the field of high-silica adakites, and the REE patterns are alsovery similar to the average high-silica adakite from Martin et al.(2005). The Sr isotopic composition of the CMV dacites is justbelow or slightly higher than the compositional boundarycommonly used for adakites, depending on whether <0.705(Martin, 1999) or <0.7045 (Samaniego et al., 2005) is used.

Table 2Whole-rock geochemical data of Cerro Machín dacites.

Rock type Dacite Dacite Dacite Dacite

Sample# Comment D-1 D-2A D-3B D4

Major elements (wt.%) by XRFSiO2 66.18 65.47 66.12 65.25TiO2 0.53 0.54 0.54 0.55Al2O3 16.15 16.05 16.17 16.26Fe2O3 3.33 3.41 3.44 3.57MnO 0.06 0.06 0.06 0.06MgO 1.99 2.03 2.04 2.21CaO 3.98 4.11 4.12 4.25Na2O 4.58 4.59 4.57 4.54K2O 2.18 2.16 2.17 2.13P2O5 0.18 0.19 0.19 0.19LOI 0.28 0.28 0.31 0.37Total 99.44 98.89 99.73 99.38Mg# 62.4 62.3 62.2 63.2

Trace elements (ppm) by ICP-MSLi 20.7 21.6 19.3 20.9Sc 8.05 8.41 8.10 8.77V 68.7 73.0 69.9 75.9Cr 47.8 50.6 49.5 59.3Co 6.38 7.06 6.55 7.03Ni 10.8 11.8 11.5 13.2Cu 6.49 5.76 6.19 3.75Zn 80.4 82.2 78.6 78.7Ga 20.5 21.0 20.4 20.6Rb 65.2 64.7 63.0 62.9Sr 756 778 739 760Y 11.5 11.9 11.3 11.6Zra 161 164 161 162Nb 5.65 5.81 5.76 5.76Mo 0.351 0.244 0.304 0.386Sn 1.35 1.42 1.48 1.31Sb 0.232 0.159 0.237 0.297Cs 2.14 1.72 2.17 2.22Ba 1521 1532 1481 1507La 24.8 22.4 24.0 21.5Ce 47.5 43.8 46.4 41.8Pr 5.92 5.49 5.69 5.25Nd 22.7 21.4 21.9 20.5Sm 4.23 4.16 4.09 4.00Eu 1.11 1.12 1.09 1.10Gd 3.50 3.50 3.41 3.38Tb 0.461 0.466 0.449 0.452Dy 2.40 2.44 2.35 2.39Ho 0.429 0.444 0.429 0.435Er 1.13 1.16 1.12 1.14Tm 0.156 0.160 0.156 0.160Yb 1.00 1.03 1.01 1.03Lu 0.142 0.146 0.143 0.145Ta 0.339 0.344 0.342 0.348W 0.282 0.211 0.261 0.279Tl 0.533 0.533 0.680 0.368Pb 16.7 16.8 16.6 16.4Th 7.00 6.31 6.68 5.91U 2.35 2.13 2.24 1.92La/Yb 24.8 21.6 23.8 20.9Sr/Y 65.7 65.3 65.3 65.6(La/Nb)N 4.45 3.91 4.22 3.79(La/Yb)CN 16.7 14.6 16.1 14.1

Isotopic ratios by TIMS87Sr/86Sr 0.704958 (7) 0.704990 (8) 0.704967 (13)143Nd/144Nd 0.512648 (5) 0.512684 (43)

N ¼ Normalized to primitive mantle values (McDonough and Sun, 1995);CN ¼ Chondrite-normalized (Boynton, 1984).Mg# is calculated assuming Fe2O3/FeO ¼ 0.4. Numbers in parentheses give the in-ternal error of the isotope ratio measurement referring to the last significant digit(s).

a Measured by XRF.

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 197

5.2. Mineral chemistry

Plagioclase phenocrysts (Table 4, Fig. 2AeC) have predomi-nantly andesine composition and both normal and reverse zoning

occurs. The resorbed cores of Type A phenocrysts show relativelylittle chemical variation (XAn ¼ 31e36), but some Type A crystalshave significantly more Ca-rich cores (up to XAn ¼ 47). A high-Ca(XAn ¼ 37e49) innermost growth zone of w20e50 mm width sur-rounds the resorbed cores of Type A crystals (Fig. 2A). The sur-rounding, oscillatory zoned mantles vary compositionally fromXAn ¼ 40 to XAn ¼ 28. The continuous oscillatory zones in Type Bphenocrysts have a compositional range of XAn ¼ 24e42. Type Cphenocrysts have variable core compositions (XAn ¼ 29e41) thatare similar to Type A cores. In contrast to Type A crystals, the outerzones are chemically quite homogenous and distinctly more Ca-rich (XAn ¼ 49e54) than the cores. Very few homogenous Type Ccrystals lack the outer high-Ca zone. In the more calcic plagioclasesof the CMV dacites, there is a subtle indication of higher Fe con-tents. At XAn w0.3, FeO is <0.1 wt.%, whereas at XAn w0.5, FeOincreases to w0.25 wt.% (Table 4). Overall, plagioclase shows anegative correlation of XOr with XAn, forming a continuouscompositional trend (Fig. 5). This trend shows an inflection atwXAn ¼ 32e36, where the shallow trend at higher XAn becomessteeper and more scattered towards lower XAn, following theexperimentally determined trends of decreasing melt H2Ocontents.

All amphiboles (Table 3, Fig. 2DeF) are calcic amphiboles withrelatively constant CaO contents of 11.4 � 0.2 wt.%. Based on Leakeet al. (1997), they can be classified as magnesio-hornblendes,tschermakites, pargasites and magnesio-hastingsites (Fig. 6A).Type I amphiboles are mainly tschermakites (Si < 6.5 andA(Na þ K) < 0.5 p.f.u.) with some pargasites (Si < 6.5 andA(Na þ K) > 0.5 p.f.u.) and magnesio-hornblendes (Si > 6.5 p.f.u.).Type II cores are of tschermakite and pargasite (VIAl � Fe3þ) withsubordinate magnesio-hastingsite (VIAl < Fe3þ) composition, whileType II rims are mainly magnesio-hastingsites. Type I amphiboles(MgO ¼ 10.6e13.2 wt.%, Al2O3 ¼ 9.6e13.9 wt.%) and the dark coresof Type II amphiboles (MgO ¼ 10.6e12.2 wt.%, Al2O3 ¼ 10.7e13.2 wt.%) are chemically overlapping. The petrographicallydistinct, light-colored rims of Type II amphiboles (MgO ¼ 12.4e17.7 wt.%, Al2O3 ¼ 10.6e14.2 wt.%) also show differences in chem-istry. They are more variable in composition and have higher Mg#,IVAl, CTi, and slightly elevated A(Naþ K) relative to Type I and Type IIcores (Fig. 6BeD).

Olivines are Mg-rich with Fo85e88 and have amphibole and cli-nopyroxene (14e22 wt.% CaO, 17e23 wt.% MgO) overgrowths. FeeTi oxides are ilmenite with XIlm (calculated after Carmichael, 1967)ranging from 0.67 to 0.82. Biotite (XFe ¼ 0.39e0.46) contains 3.5e4.2 wt.% TiO2.

5.3. Crystallization conditions (P, T, fO2, aH2O)

Chemical differences between or within amphibole crystals canbe related to variations in temperature, pressure and/or SiO2-saturation of the melt (e.g., Johnson and Rutherford, 1989; Satoet al., 1999; Scaillet and Evans, 1999; Rutherford and Devine,2003). Ridolfi et al. (2010) and Ridolfi and Renzulli (2012) derivedempirical thermobarometric formulations based on experimentalstudies on amphibole compositions that are applicable toamphibole-bearing, calc-alkaline volcanic rocks at a wide range ofPeT conditions (130e2200MPa and 800e1130 �C). We focus on theamphibole thermobarometric formulations of Ridolfi et al. (2010)for the CMV dacites because the CMV lava dome samples fall intothe compositional range for which these formulations are appli-cable and because amphibole is ubiquitous and characterized bycompositional and textural variations. Moreover, the amphibole-based formulations are applicable to hybrid and re-homogenizedmagmas, whereas the application of geothermobarometers cali-brated on the equilibrium of two or more phases is not

50 55 60 65 70 750

2

4

6

8

10

SiO2 (wt.%)

Na 2

O +

K2O

(wt.%

)

Cerro MachínNevado del RuizNVZ volcanoes, Ecuador

basalticandesite

andesite dacite

trachyandesite trachyte

rhyolite

A B

C D

0.7025 0.703 0.7035 0.704 0.7045 0.705 0.70550.5124

0.5126

0.5128

0.5130

0.5132

143 N

d/14

4 Nd

87Sr/86Sr

Cerro MachínEcuadorian volcanoesLower crustal xenolithsAccreted oceanic rocksGalapagos IslandGalapagos Ridge

1

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

c sam

ple/c

chon

drite

Colombian dacites

D-1D-2AD-3BD-4High-silica adakite

CsRb

BaTh

UNb

TaK

LaCe

PbSr

NdZr

SmEu

SbTi

GdTb

DyLi

YEr

YbLu

1

10

100

c sam

ple/c

PM

Colombiandacites

D-1D-2AHigh-silica adakite

D-3BD-4

Fig. 3. Geochemical characterization of the lava dome dacites from Cerro Machín Volcano. Data for Nevado del Ruiz are pumice and scoria from the 1985 eruption (Sigurdsson et al.,1990; Gourgaud and Thouret, 1990), data for Ecuadorian volcanoes from the Northern Volcanic Zone are from Schiano et al. (2010). A) TAS diagram with alkalkine-subalkalinesubdivision from Irvine and Baragar (1971). B) K2O vs. SiO2 diagram with boundaries from Le Maitre et al. (1989). C) Chondrite-normalized (values from Boynton, 1984) REE di-agram. For comparison, dacites from the Colombian part of the NVZ e one airfall pumice from Nevado del Ruiz (Melson et al., 1990) and five lavas from Doña Juana and Chiles (Drouxand Delaloye, 1996) e are shown as light gray field. The high-silica adakite is the average from Martin et al. (2005). D) SreNd isotope diagramwhich includes data for the GalapagosRidge (Geist et al., 2008), Galapagos Islands (White et al., 1993), accreted oceanic plateaux and island arcs fromwestern Ecuador (Reynaud et al., 1999), lower crustal xenoliths fromSW Colombia (Weber et al., 2002), and Ecuadorian volcanoes (Samaniego et al., 2005, 2010; Barragan et al., 1998; Bourdon et al., 2003).

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208198

straightforward in hybrid rocks due to the presence of xenocrystsderived from distinct magma batches (Ridolfi et al., 2010) (Table 5).

The results of the thermobarometric calculations are summa-rized in Table 6. For Type I amphibole phenocrysts, temperaturesand pressures range from 855 to 955 �C and 200e570 MPa,respectively (Fig. 7). Assuming a crustal density of 2700 kg/m3

(Ridolfi et al., 2010), this corresponds to an average depth ofw13 � 2 km. The average oxygen fugacity (fO2

) is w0.5 log unitsabove the NieNiO (NNO) buffer, and the average melt H2O contentis w7.1 wt.%, assuming water saturation. Calculations for Type IIcores give a temperature range of 883e947 �C and pressures of270e500 MPa, corresponding to an average depth of 14 � 2 km.Both fO2

(þ0.4 � 0.2 DNNO) and melt H2O contents (7.0 � 0.2 wt.%)are indistinguishable from Type I amphiboles (Fig. 7). The Type IIamphibole rims yield temperatures of 926e1004 �C and pressuresof 240e610MPa, corresponding to an average depth ofw16� 4 km(Fig. 7). These temperatures overlap with those calculated for Type Icrystals and Type II cores. However, there is a difference in T of 46e

70 �C between corresponding rims and cores. Type II rims crystal-lized from a moderately oxidized melt (fO2

¼þ1.2 DNNO) with 4.1e7.3 wt.% H2O (average 6.1 � 0.7 wt.%). The range in melt H2O con-tents overlaps with those determined from Type I amphiboles.

6. Discussion

6.1. Amphibole geothermobarometry

Given the similarity in all crystallization parameters, Type Iamphiboles and Type II cores reflect the same main crystallizationenvironment at temperatures of 910 � 30 �C, pressures of360 � 70 MPa, oxygen fugacities of þ0.5 � 0.2 DNNO and melt H2Ocontents of 7.1 � 0.4 wt.% (Table 6). The inferred main crystalliza-tion environment (w12e16 km depth) is located at greater depthsrelative to those where the majority of recent earthquake hypo-centers underneath the central dome have been located, but itoverlaps with depths of one of the seismic zones southeast of the

K/R

b

1200

1000

800

600

400

200

020 40 60 80 1000

SiO2/MgO

Low-silica

adakites

High-silicaadakites

Cerro MachinAverage high-silica adakiteAverage low-silica adakiteColombian NVZEcuadorian NVZ

A

0 0.5 1 1.5 2

TiO2 (wt.%)

0

1

2

3

4

5

6

Cr/N

i

Low-silica adakites

High-silicaadakites

B

50 55 60 65 70 75

SiO2 (wt.%)

0

K 2O

/Na 2

O

0.2

0.4

0.6

0.8

1.0

1.2

Arc

High-silicaadakites

C

0 20 40 60 80 100 120 140 160

Sr/Y

0

10

20

30

40

50

La/Y

b

calc-alkaline

adakitic

D

Fig. 4. Geochemical diagrams used for identifying adakitic geochemical signatures. Note that Cerro Machín dacites fall into the adakite fields in all four diagrams. For comparison,data for NVZ volcanoes from Colombia (Droux and Delaloye, 1996; Melson et al., 1990) and Ecuador (Schiano et al., 2010) are plotted. (A) K/Rb vs. SiO2/MgO diagram with fields forhigh-silica and low-silica adakite from Martin et al. (2005). B) Cr/Ni vs. TiO2 diagram with fields from Martin et al. (2005). C) K2O/Na2O vs. SiO2 diagram with fields of high-silicaadakites from Moyen (2009). D) La/Yb vs. Sr/Y diagram with cut-off values from Samaniego et al. (2005) based on Defant and Drummond (1990).

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 199

volcano (Londoño, 2011). Taking into account the temperatureuncertainty of the method, �22 �C (Ridolfi et al., 2010) and thesignificant difference in the average temperature compared to TypeII rims, it appears that Type II rims reflect crystallization from ameltwith higher temperatures than the resident silicic magma. Despitethe overlap in T, the clear textural separation and the chemicaldifferences between Type II cores and rims supports the idea thatchanges in the crystallization environment occurred.

Whether an increase in T during crystallization of the Type IIrims occurred can be evaluated by crystal chemical trends in theamphiboles. The effect of increasing temperature in inducinghigher A(Naþ K), CTi and Al2O3 in the rims of Type II amphiboles aresupported by positive correlations of A(Na þ K) and CTi with IVAl(Fig. 6). Most of the variation in IVAl (w0.6 atoms p.f.u.) is accom-modated by variation in A(Na þ K) (w0.34 atoms p.f.u.) and CTi(0.14 atoms p.f.u.). These relations demonstrate that thetemperature-sensitive edenite (A, þ TSi 5 A(Na þ K) þ IVAl) andTi-Tschermak (BMn þ TSi 5 CTi þ IVAl) exchange mechanisms(Thornber et al., 2008; Anderson and Smith,1995; Shane and Smith,2013) were decisively influencing the amphibole chemistry. In

contrast, the poor correlation between IVAl and VIAl (Fig. 6D) showsthat the pressure-sensitive Al-Tschermak exchange(C(Mg,Fe) þ TSi 5 IVAl þ VIAl) was not significant. Although thecalculated pressures for Type II rims are, on average, slightly higherthan those of Type II cores, this apparent difference is not likely tobe real because of the lack of the pressure-sensitive Al-Tschermaksubstitution. Moreover, the statistical uncertainty of P is significant(typically �11.5% and up to 24% for crystal-rich and lower-Tmagmas; Ridolfi et al., 2010) and there is a large overlap betweenType II rims and Type II cores. This result is identical to several otherstudies, which found that temperature exchange mechanisms ac-count for most atomic substitutions in amphibole (Bachmann andDungan, 2002; Rutherford and Devine, 2003; Shane and Smith,2013). We also concur with Shane and Smith (2013) who empha-sized that the results of Al-in-amphibole barometry should beconsidered with care if atomic exchange mechanisms are notassessed.

The melt that replenished the magma reservoir caused theformation of a hybrid magma (DNNO ¼ þ1.2 � 0.3) that attainedequilibrium with the Type II amphibole rims. The distinctly

0.2 0.3 0.4 0.5 0.6 0.70.01

0.02

0.03

0.04

0.05

XAn

X Or

daciteP basaltP

Experimental plagioclase 780° C 834-866° C 899° C

Decreasingmelt H2Ocontent

Fig. 5. Plagioclase compositions of CMV dacites compared to plagioclase trends fromMt. Pinatubo dacites (daciteP) and basalts (basaltP) (Hattori and Sato, 1996). Trends ofexperimentally produced plagioclase show increasing XOr with decreasing melt H2Ocontents (Scaillet and Evans, 1999).

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208200

different average melt H2O contents suggest that the resident meltfromwhich Type I crystals and Type II cores crystallized was slightlyenriched in H2O compared to the melt in equilibrium with Type IIrim compositions, assuming that all amphiboles crystallized underwater-saturated conditions, as implicit in the calculations. Note,however, that the results reflect crystallization conditions of indi-vidual crystals and not necessarily the intrinsic parameters of aspecific magma reservoir.

Assessments of the accuracy of amphibole geobarometry withother petrologic geothermobarometers generally yield a good

Table 3Representative amphibole analyses from Cerro Machín dacites.

Sample D2A D2A D2A D2A D2A D2A

Type I I I I II II

Xtal No. Aa Aa Aa Aa Ab Ab

Location Core Core Rim Rim Core Cor

SiO2 44.32 43.29 44.39 45.37 42.44 43.0TiO2 1.31 1.42 1.30 1.22 1.53 1.4Al2O3 10.96 11.91 11.22 10.38 12.62 12.0Cr2O3 0.02 0.04 0 0 0 0FeO 15.73 16.27 15.94 15.47 16.37 16.3MnO 0.40 0.34 0.44 0.34 0.41 0.3MgO 11.81 11.3 11.72 12.44 10.58 11.3CaO 11.40 11.40 11.43 11.40 11.38 11.4Na2O 1.72 1.96 1.92 1.71 2.20 1.9K2O 0.74 0.85 0.74 0.60 0.89 0.9F 0 0 0 0 0 0.0Cl 0.04 0.04 0.03 0.02 0.04 0.0

Total 98.45 98.82 99.13 98.95 98.46 99.0

Formula on the basis of 13 cations (eCNK)TSi 6.45 6.31 6.43 6.53 6.25 6.2TAlIV 1.55 1.69 1.57 1.47 1.75 1.7CAlVI 0.33 0.36 0.34 0.29 0.44 0.3CTi 0.14 0.16 0.14 0.13 0.17 0.1CCr 0.00 0.00 0.00 0.00 0.00 0.0CFe3þ 0.76 0.74 0.72 0.81 0.58 0.7CMg 2.56 2.46 2.53 2.67 2.32 2.4CFe2þ 1.16 1.24 1.21 1.06 1.43 1.2CMn 0.05 0.04 0.05 0.04 0.05 0.0BCa 1.78 1.78 1.77 1.76 1.80 1.7BNa 0.22 0.22 0.23 0.24 0.20 0.2ANa 0.26 0.34 0.31 0.24 0.42 0.3AK 0.14 0.16 0.14 0.11 0.17 0.1

degree of consistency for several volcanoes (Ridolfi et al., 2010;Shane and Smith, 2013). Calculations for the CMV samples usingthe amphibole-plagioclase geothermometer (Holland and Blundy,1994) yield temperatures that are typically 50e100 �C lower thanthose using the formulations by Ridolfi et al. (2010), with generallysmaller differences if the most calcic plagioclase compositions areused. Whereas the calculations are relatively insensitive to changesin pressure with a change of 50 MPa resulting in temperaturechanges of 3e4 �C only, temperatures calculated using the edenite-tremolite exchange result in slightly better agreement than thosebased on the edenite-richterite exchange. In a study of amphibolecrystals from Okataina volcano (New Zealand), Shane and Smith(2013) also found that temperature estimates based onamphibole-plagioclase thermometry give up to 90 �C lower tem-peratures than the single amphibole geothermometry. The mostlikely explanation for this discrepancy is the significant absolutetemperature deviation of, on average, 61 �C between temperaturescalculated from amphibole-plagioclase equilibria and temperaturesat which the corresponding experiments were run (Blundy andCashman, 2008). The calculated temperatures are typically lowerthan those of the experiments, and the deviation increases withincreasing Mg# of amphibole. Therefore, temperature de-terminations based on amphibole-plagioclase equilibria formagmas with high Mg#, such as the CMV dacites, are likely to betoo low (Blundy and Cashman, 2008). Disequilibrium betweenamphibole and plagioclase may also affect the results.

6.2. Comparison with experimental data on crystallization of silicicmagmas

The ubiquitous presence of amphibole and the absence oforthopyroxene in the CMV rocks require elevated H2O contents.

D2A D2A D3B D3B D4 D4

II II II II II II

Ab Ab Ag Ag Ab Ab

e Rim Rim Core Rim Core Rim

3 42.65 42.21 43.38 43.37 43.28 43.491 2.18 2.11 1.35 1.82 1.38 2.059 12.74 13.16 11.95 12.44 12.06 12.35

0.11 0.03 0.00 0.00 0.02 0.004 11.14 11.91 15.77 10.96 16.08 9.489 0.17 0.18 0.40 0.16 0.36 0.166 14.49 14.07 11.30 15.24 11.31 16.266 11.37 11.49 11.36 11.18 11.26 11.009 2.30 2.51 1.98 2.34 1.99 2.433 0.56 0.58 0.91 0.68 0.84 0.661 0 0 0 0 0 03 0.01 0.02 0.04 0.01 0 0

4 97.72 98.27 98.44 98.20 98.58 97.88

7 6.13 6.07 6.35 6.17 6.31 6.153 1.87 1.93 1.65 1.83 1.69 1.854 0.29 0.30 0.41 0.25 0.38 0.216 0.24 0.23 0.15 0.20 0.15 0.220 0.01 0.00 0.00 0.00 0.00 0.007 0.85 0.83 0.66 1.02 0.77 1.087 3.11 3.02 2.47 3.23 2.46 3.432 0.49 0.61 1.27 0.29 1.19 0.045 0.02 0.02 0.05 0.02 0.05 0.029 1.75 1.77 1.78 1.70 1.76 1.671 0.25 0.23 0.22 0.30 0.24 0.335 0.39 0.47 0.34 0.35 0.32 0.337 0.10 0.11 0.17 0.12 0.16 0.12

Table 4Representative analyses of plagioclase from Cerro Machín dacites.

Sample D2A D2A D2A D2A D2A D2A D2A D2A D2A D3B D3B D3B

Xtal No. Ple Ple Ple Plh Plh Plh Pln Pln Pln Plc Plc Plc

Xtal type A A A B B B C C C C C C

Location Core Core Rim Core Core Rim Core Core Rim Core Rim Rim

SiO2 61.16 56.94 60.80 59.15 59.77 58.46 60.45 59.67 54.83 61.57 60.88 56.49TiO2 0 0.01 0 0.01 0 0 0 0 0 0 0 0.01Al2O3 25.48 27.75 25.33 26.32 25.61 26.70 25.77 25.95 28.91 25.01 25.37 28.07FeO 0.08 0.17 0.11 0.14 0.09 0.25 0.09 0.09 0.26 0.15 0.15 0.25MgO 0 0 0 0 0 0.01 0 0 0.01 0 0 0.02CaO 6.16 9.39 6.30 7.35 6.97 8.29 6.73 7.00 10.79 5.82 6.12 9.68Na2O 7.42 5.88 7.30 6.96 7.17 6.35 7.24 7.09 4.99 7.73 7.55 5.68K2O 0.45 0.32 0.43 0.37 0.39 0.38 0.41 0.40 0.32 0.49 0.46 0.27

Total 100.75 100.45 100.27 100.30 100.00 100.44 100.69 100.19 100.11 100.77 100.53 100.46

Cations per formula unit based on 8 oxygensSi 2.69 2.69 2.69 2.63 2.66 2.60 2.67 2.65 2.47 2.71 2.69 2.53Al 1.33 1.32 1.32 1.38 1.34 1.40 1.34 1.36 1.53 1.30 1.32 1.48Ti e e e e e e e e e e e e

Mg e e e e e e e e e e e e

Fe 0.01 0 0 0.01 0 0.01 0 0 0.01 0.01 0.01 0.01Mn 0 0 0 0 0 0 0 0 0 0 0 0Ca 0.30 0.30 0.30 0.35 0.33 0.40 0.32 0.33 0.52 0.27 0.29 0.46Na 0.63 0.63 0.63 0.60 0.62 0.55 0.62 0.61 0.44 0.66 0.65 0.49K 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.02

Sum 4.97 4.99 4.97 4.99 4.99 4.98 4.98 4.98 4.99 4.98 4.98 4.99

XAn 0.31 0.47 0.32 0.37 0.35 0.42 0.34 0.35 0.54 0.29 0.31 0.49

XAn ¼ Ca/(Ca þ Na).

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 201

Experimental studies show that >6 wt.% H2O and >4e5 wt.% H2Oare required to stabilize amphibole in Montagne Pelée andesite(Martel et al., 1999) and in Mt. Pinatubo dacite (Scaillet and Evans,1999), respectively. The maximum temperature at which amphi-bole is stable in silicic magmas increases with increasing pressure(Sato et al., 1999; Ridolfi et al., 2010). If pressures are sufficientlyhigh, amphibole can be a liquidus phase, as determined by thephase diagram of the Mount St. Helens 1980 dacite at P > 250 MPa(Blundy and Cashman, 2001). For the CMV dacites, average crys-tallization pressures for Type I crystals and Type II cores of>340 MPa, in combination with crystallization temperatures>900 �C and the absence of other phases that crystallize at thesetemperatures suggest that amphibole is a liquidus phase, whichcrystallized at relatively deep levels (13 � 2 km) in the magmaticsystem.

Experimental work on silicic magmas shows that the presenceof Ab-rich plagioclase (XAb> 0.6), biotite and quartz occurs at muchlower temperatures than the amphibole crystallization tempera-tures calculated for CMV dacites (e.g. Holtz et al., 2005; Rutherfordand Devine, 2003). The oscillatory zoning in plagioclase requiresthat the magma repeatedly experienced a range of crystallizationconditions. Factors that potentially cause oscillatory zoning includemelt compositional changes, which directly influence the Ca/Naratio in plagioclase, and variations in T and aH2O. Systematicallydecreasing XAn in plagioclase is observed with decreasing T andfalling aH2O (Hattori and Sato, 1996; Rutherford and Devine, 2003;Scaillet and Evans, 1999). At a given T, a decrease in aH2O movesthe plagioclase composition to higher or content (Scaillet andEvans, 1999). With increasing T, more Kþ can be incorporated intoplagioclase (Hattori and Sato, 1996), which leads to a steeper slopein the XOr versus XAn diagram (Fig. 5). The composition of plagio-clase in the CMV dacites indicates crystallization close to or at H2O-saturation. The continuous trend with little scatter (Fig. 5) showsthat only few plagioclase phenocrysts crystallized from a meltcomposition significantly different from the silicic resident magma

because the linear relationship for XAn ¼ 0.5e0.35 would be diffi-cult to maintain if there was a significant injection of mafic magma(Prouteau and Scaillet, 2003). Hence, the similar XOr contents at agiven XAn suggest that most plagioclase phenocrysts formed in asingle magma reservoir of felsic composition. Most of the compo-sitional variation is likely to be related to changes in T and aH2O andthe steep increase in XOr at very low XAn (<0.3) points to decreasingaH2O due to degassing. However, the concomitant increase in Fewith Ca contents in more calcic plagioclase of the CMV dacitespoints to injections of mafic magma (Hattori and Sato, 1996).

Experimentally produced biotite occurs at T< 800 �C in rhyolitic(Castro and Dingwell, 2009) and T < 780 �C in andesitic melts(Rutherford and Devine, 2003). Similarly, biotite is only stable attemperatures <760 �C and melt H2O contents above w5.6 wt.% inthe Mt. Pinatubo 1991 dacite (Scaillet and Evans, 1999). In thisdacite, plagioclase with XAn ¼ 30 would crystallize at conditionsadequate for biotite crystallization (7 wt.% H2O in the melt, 760 �C).Unrimmed biotites in the rhyodacitic Unzen 1991 magma demon-strate that these crystals have been protected from mixing andheating as the maximum stability of biotite is w800 �C (Venezkyand Rutherford, 1999). The coarse reaction coronas, which arepresent around some biotites in the CMV dacites point to temper-atures beyond the upper stability limit of biotite, which can berelated to the influence of a hotter, more mafic magma. This in-fluence is also evident in resorption of quartz, although decom-pression crystallization may cause similar features in quartz(Blundy and Cashman, 2001). In equilibrium crystallization exper-iments, quartz appears first at 750 �C under H2O-saturated condi-tions in dacite (Scaillet and Evans, 1999) and between 790 and825 �C at pressures of 200e130 MPa in andesite (Rutherford andDevine, 2003). In summary, temperatures <800 �C are indicatedfor the final crystallization of the CMV dacites based on the mineralparagenesis biotite, Ab-rich plagioclase and quartz. In contrast,amphibole crystallized early in the magmatic evolution. The vari-able thickness of reaction rims around amphibole indicates variable

1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1IVAl

0.1

0.15

0.2

0.25

0.3

CTi

Type II core rimD2AD3BD4

Type I

1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1IVAl

0.1

0.2

0.3

0.4

0.5VI

Al

Type I

Si p.f.u.

66.57Si p.f.u.

0

0.5

1A (

Na+

K) p

.f.u.

etisagraPetinedEMg-hastingsite

TschermakiteMg-Hornblende

Type I

Type II core rimD2AD3BD4

A B

C D

1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.10.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

Mg#

= M

g/(M

g+Fe

2++F

e3+)

Type I

IVAl

Fig. 6. Amphibole chemistry of dacites from Cerro Machín Volcano. Type 1 amphiboles from all four samples are shown by the gray field. Note that Mg# was calculated as Mg/(Mg þ Fe2þþFe3þ).

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208202

degrees of re-equilibration at lower temperatures during magmaascent, possibly because of fluctuating kinetic retardation and/orvariations in the availability of chemical components.

6.3. Original melt compositions

To obtain information about the major element compositions ofmelts with which the phenocrysts/xenocrysts were in equilibriumduring crystallization, we calculated melt compositional parame-ters based on olivine and amphibole compositions. Constraints onthe contents of several major element oxides in the equilibriummelts can be obtained from amphibole chemometry (Ridolfi andRenzulli, 2012), which is based only on amphibole compositions.The corresponding melts calculated from amphibole Type I andType II cores have similar compositions (Table 3, Fig. 8) and bothoriginate from an evolved magma (melt A). Compositional pa-rameters of melt A are 68e74 wt.% SiO2, 2.0e4.2 wt.% CaO, 3.6e5.6 wt.% K2O and Mg# ¼ 0.09e0.37. Melt A belongs to the high-Kseries and is dacitic to rhyolitic in composition (Fig. 8A,B). Calcu-lated melt compositions from amphibole Type II rims are moreprimitive than melt A, yielding lower SiO2 (60e70 wt.%) and K2O

(1.8e4.1 wt.%) contents, higher CaO (2.3e6.8 wt.%) contents andelevated Mg# (0.37e0.47). Accordingly, Type II rims crystallizedfrom a melt that was more mafic than melt A.

For ferromagnesian silicates, equilibrium relationships of the Fe/Mg ratio between melt and mineral can be used to constraincrystallization conditions (Streck et al., 2005; Halama et al., 2006).Here, we use the olivine compositions to calculate XMg of the po-tential parental liquid based on equilibrium relationships betweenolivine and natural basaltic liquid (Roeder and Emslie, 1970). Usinga value for KFe�Mg

ol�liquid (FeeMg distribution coefficient betweenolivine and liquid) of 0.3 (Roeder and Emslie, 1970; Putirka, 2008),the Mg numbers of the melt (Mg#melt) were calculated. For themeasured olivine compositions of Fo85e88, Mg#melt ranges from0.63 to 0.69. This value is significantly higher than the Mg# of thebulk dacites and indicates a composition similar to primitive vol-canic rocks of continental arcs (Fig. 8C). It demonstrates that olivineis a xenocrystic phase in the dacite and that a more silicic meltcomponent is required to form the dacites. To our knowledge, thereis no evidence of olivine-bearing basement rocks at CMV, so thatthe olivine most likely originates from the CMV magmatic system.Such a scenario is in accordance with the common occurrence of

0

200

400

600

800

1000

550 650 750 850 950 1050 1150

P (M

Pa)

T (°C)

maximum thermalstability

upper limit ofconsistent

amphiboles

SiO2 52%

SiO2 63%SiO2 70%

SiO2 76%

Type II core rimD2AD3BD4

Type I

A

-14

-13

-12

-11

-10

-9

-8

700 750 800 850 900 950 1000 1050 1100

log

f O2

T (°C)

NNO

NNO +2

Type I

B

550

650

750

850

950

1050

1150

2 4 6 8 10 12

T (°

C)

H2Omelt (wt.%)

maximumthermalstability

lower limit ofconsistent

amphiboles

Type I

C

Fig. 7. Crystallization parameters for Cerro Machín dacites based on amphibole geo-thermobarometry after Ridolfi et al. (2010). A) PeT diagram. Error bars represent theexpected uncertainty in T (�22 �C) and representative errors for P indicating thevariation in accuracy. Isopleths show the anhydrous SiO2 (wt.%) content of the melt.Consistent amphiboles are defined according to Ridolfi et al. (2010) as experimentallyproduced amphiboles that were synthesized in equilibriumwith melts overlapping themain Al2O3 vs. SiO2 pattern. B) log fO2

� T diagram with curves of the nickelenickeloxide (NNO) buffer and NNOþ2. The error bar represents the expected uncertainty in Tand the maximum log fO2

error (�0.4 log units). C) T-H2Omelt diagram. Indicated un-certainty for H2Omelt is �0.4 wt.%.

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 203

forsteritic olivines, mafic recharge and magma mixing in manystratovolcanoes above subduction zones (Ruprecht and Plank,2013). Olivine xenocrysts in intermediate to felsic lavas have alsobeen reported from Nevado del Ruiz (Gourgaud and Thouret, 1990)and Galeras in Colombia (Calvache and Williams, 1997) and fromthe 1991 dacite eruption of Mt. Pinatubo (Di Muro et al., 2008).Even if there is no mafic magma composition among the eruptiveproducts, the presence of a mafic magma at depth can be inferredbased on the presence of magnesian olivine (Gourgaud andThouret, 1990). Several studies have shown that disaggregation ofxenolithic inclusions into individual crystals can be an efficient

magma mixing process, and mineral disequilibrium may be theonly remaining evidence for replenishment and mixing (Clynne,1999; Martin et al., 2006). Even if the magnesian olivine xen-ocrysts in the CMV dacites were derived from older CM rocks, theyclearly preserve a record from a primitive melt of mantle origin andare best interpreted as a liquidus phase from the early silicatecrystallization at depth (Ruprecht and Plank, 2013). Melt in equi-librium with olivine Fo85-88 thus represents a mafic melt B.Althoughmelt Mg# can be calculated from the olivine composition,abundances of SiO2, MgO and FeO are not constrained. In principle,olivine-bearing dacitic melts can also be formed by partial meltingof quartz-normative pyroxenite in the mantle (Straub et al., 2011).However, the amphibole overgrowth around olivine demonstratedisequilibrium conditions that are not consistent with a mantleorigin of the dacites. Experimental work by Coombs and Gardner(2004) demonstrated that amphibole reaction rims grow duringreaction between silicic host melt and olivine xenocrysts, andsimilar features were also described in natural dacite from Mt.Pinatubo (Di Muro et al., 2008). Moreover, broadly decreasing Nicontents with decreasing Fo at Fo < 88 in the CMV olivine xen-ocrysts point to magma differentiation in the crust (Straub et al.,2011). The melt Mg# calculated based on rims of Type II amphi-bole is 0.42 � 0.05 and does not overlap with the more Mg-richmelt B compositions (Mg#melt ¼ 0.63e0.69) based on olivinecompositions. Therefore, we interpret Type II rims as the result ofcrystallization from a hybrid melt that formed by mixing of adominant, silicic component (melt A) and a subordinate, maficcomponent (melt B).

6.4. Magma mixing model

The diversity of phenocryst compositions and textures in theCMV dacites is evidence that the erupted dome magmas representa mixture of crystals with different crystallization histories (e.g.,Murphy et al., 1998; Halama et al., 2006; Streck, 2008). This evi-dence includes (i) reversely zoned plagioclase crystals in additionto the normally zoned ones, with occasionally elevated Fe contentsin the calcic zones (ii) resorbed plagioclase and quartz, suggestingepisodes of heating and resorption (Tsuchiyama, 1985); (iii)amphibole reaction rims around olivine (Fig. 2F; Coombs andGardner, 2004); (iv) co-existence of variably zoned amphiboles(Fig. 2D,E); and (v) forsteritic olivine that is not in equilibriumwitha dacitic bulk composition. These disequilibrium features on themicrometer to millimeter scale are consistent with magmamingling and mechanical mixing and suggest that the dacitic domesamples formed from at least two different precursor magmas.Another indicator for the contribution of basaltic magmas in themixing processes is elevated Cr content in bulk rocks (Pallister et al.,1992; Streck et al., 2007). Compared to eruption products fromMount St. Helens’ 1980 eruption, thought to be the product ofbasalt-dacite mixing, CMV dacites have significantly higher values(34e50 ppm Cr), which emphasize the role of a primitive maficmagma end-member. The relatively high Mg# of the bulk dacites(Mg# ¼ 62e63) also reflects the influence of a mafic melt. Bothfeatures, the high Cr contents and the high Mg#, are unlikely to besolely due to the presence of olivine xenocrysts because thesexenocrysts are extremely scarce in the CMV dacites.

We propose a model where dacitic-rhyolitic magma in a deep-seated magma chamber (9e18 km) was intruded by a maficmagma. The resident silicic magma was water-rich and crystallizedamphibole as a liquidus phase. The mafic magma contained Fo-richolivines and mixing and exchange of crystals between the twomagmas occurred, leading to hybridization. The formation of ahybrid, dacitic magma resulted in the crystallization of Mg-richamphibole rims and amphibole-dominated reaction rims around

45 50 55 60 65 70 75 80 85SiO2 (wt.%)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Mg#

CAB CAA

BONMg# of melt in equilibrium

with xenocrystic olivine

45 50 55 60 65 70 75 80 85SiO2 (wt.%)

0

2

4

6

8

10

12

CaO

(wt.%

)

CAB

CAA

BON

CMV whole rocks NDR glassMelt compositions calculated from type I amphibole type II amphibole cores type II amphibole rims

45 50 55 60 65 70 75 80 85SiO2 (wt.%)

0

1

2

3

4

5

6

K 2O

(wt.%

)

CABCAA

BONlow-K

medium-Khigh-K

B

C

A

Fig. 8. Melt chemistry calculated based on the chemometric equations from Ridolfiand Renzulli (2012). CAB, CAA and BON are average compositions of continental arc

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208204

olivine. Because of the high melt H2O contents and elevated pres-sures for the CMV dacites, nucleation of plagioclase was retarded(Hattori and Sato, 1996). Biotite, albite-rich plagioclase and quartzformed at comparatively low temperatures (probably <800 �C)from the silicic melt during early stages of ascent or storage atshallower levels, as decompression and temperature reductionwould propel plagioclase crystallization (e.g. Holtz et al., 2005). Theoccurrence of resorbed plagioclase and quartz and reversely zonedplagioclase give textural evidence for the occasional influence of ahotter, more mafic melt also at this stage. Due to complex dynamicconvection, phenocrysts of different origin mingled. The investi-gated dacites thus derive from hybrid melts (melt AB), whichcontain zonedminerals that were at some point in equilibriumwitheither mafic, dacitic or rhyolitic compositions. The silicic domeapparently represents the latest magma supply from the deeper,mid-crustal level after the main eruption phase, which was domi-nated by vulcanian eruptions and related pyroclastic flows(Méndez et al., 2002; Rueda, 2005).

6.5. Significance of the adakite geochemical signature

The presence of the adakite signature in the CMV dacites (Fig. 4)is of interest as it extends the area in the NVZ of the Andes from theEcuadorian volcanoes with adakite signature (Bourdon et al., 2002;Hidalgo et al., 2007; Chiaradia et al., 2009; Samaniego et al., 2005;Schiano et al., 2010) to the northernmost group of volcanoes inColombia. Other Colombian volcanoes also appear to have adakite-like geochemical signatures (Droux and Delaloye, 1996; Melsonet al., 1990; Pulgarín et al., 2010; Monsalve et al., 2011). However,the interpretation of these geochemical characteristics remainsdisputed. One possible explanation, based on the original inter-pretation of volcanic rocks from the Aleutian arc (Defant andDrummond, 1990) invokes slab melting (Bourdon et al., 2002;Samaniego et al., 2002, 2005; Hidalgo et al., 2007). Alternatively,normal arc magmatic processes that include mantle wedgemelting, fractional crystallization and assimilation in the lowercrust as well as magma mixing may also account for thegeochemistry of the NVZ lavas (Garrison and Davidson, 2003;Garrison et al., 2006; Chiaradia et al., 2009; Zellmer et al., 2012).The limited compositional variability of the CMV dacites, whichconsistently fall into the fields of high-silica adakites (Fig. 4) asdefined by Martin et al. (2005) and Moyen (2009), precludes anyinterpretation based on geochemical trends. However, the gener-ally low HREE contents and high (La/Yb)CN ratios clearly point to aninfluential role of garnet in the petrogenesis of the CMV dacites. Inthis respect, the high melt H2O contents derived from amphibolethermobarometry are noteworthy because igneous garnet has anincreased stability in silicate liquids with high dissolved H2O con-tents (Alonso-Perez et al., 2009). In such melts, garnet is likely toaffect the trace element composition of the derivative arc volcanicrocks by fractionation (Alonso-Perez et al., 2009). The SreNd iso-topic compositions provide additional clues as to the origin of theadakitic signature. Compared to the Ecuadorian volcanoes of theNVZ, CMV dacites have more radiogenic compositions that overlapwith lower crustal xenoliths (Weber et al., 2002). In contrast, maficrocks that may potentially represent the composition of igneousoceanic crust in the slab (MORB, Galapagos Ridge, uplifted igneousoceanic basement of western Ecuador with arc affinities) aredistinct from the CMV dacites (Fig. 4). In particular, none of thesepotential slab source components has Nd isotopic compositions as

basalt, continental arc andesite and boninite, respectively (data from Kelemen et al.,2003). Glass compositions from eruptive products of Nevado del Ruiz (NDR glass)are shown for comparison.

Table 5Representative analyses of olivine, biotite and ilmenite from Cerro Machín dacites.

Sample D2A D2A D4 D4 D4 D4 D3B D3B D3B D3B D3B D4

Mineral Olivine Olivine Olivine Olivine Olivine Olivine Biotite Biotite Biotite Ilm Ilm Ilm

Xtal No. UOc2 UOc3 UOd1 UOd2 UOd3 UOd4 Btl2 Btl3 Btl6 Tic5 Tik10 Tig5

SiO2 40.64 40.87 41.11 40.92 40.87 40.40 36.19 36.31 36.40 0.12 0.06 0.01TiO2 3.75 3.78 3.80 42.99 36.83 38.84Al2O3 0.01 0.02 0.01 0.02 0.02 0.02 16.16 16.13 16.02 0.22 0.39 0.21Cr2O3 0.07 0.06 0.03 0.06 0.04 0.04 0.04 0.07 0.06FeO 12.18 11.52 12.05 11.52 11.44 11.61 17.16 17.56 17.15 51.43 57.37 56.16MnO 0.18 0.17 0.22 0.17 0.15 0.19 0.15 0.10 0.19 1.16 0.83 0.56MgO 46.9 47.62 47.23 47.94 47.91 47.91 13.38 13.02 13.40 1.70 1.21 1.49CaO 0.26 0.18 0.10 0.09 0.10 0.10 0.02 0.02 0.00NiO 0.30 0.35 0.19 0.25 0.23 0.22 0.16 0.05 0.02Na2O 0.55 0.53 0.57K2O 9.04 9.00 9.09P2O5 0.00 0.00 0.00F 0.00 0.00 0.00Cl 0.04 0.04 0.04

Total 100.53 100.78 100.93 100.97 100.76 100.49 96.45 96.49 96.66 97.82 96.81 97.34

Calculation of Fe2O3 and FeO (in wt.%) for ilmeniteFe2O3 18.89 30.25 27.19FeO 34.44 30.15 31.69Ferric total 99.71 99.84 100.07

Cations per formula unitSi 1.00 1.00 1.01 1.00 1.00 0.99 2.70 2.71 2.71 0.00 0.00 0.00Ti 0.21 0.21 0.21 0.81 0.70 0.74Al 0.00 0.00 0.00 0.00 0.00 0.00 1.42 1.42 1.41 0.01 0.01 0.01Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Fe3þ 0.36 0.58 0.52Fe2þ 0.25 0.24 0.25 0.24 0.23 0.24 1.07 1.10 1.07 0.73 0.64 0.67Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.02 0.02 0.01Mg 1.72 1.74 1.73 1.75 1.75 1.76 1.49 1.45 1.49 0.06 0.05 0.06Ca 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Na 0.08 0.08 0.08K 0.86 0.86 0.86Ni 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Total 2.99 2.99 2.99 3.00 3.00 3.00 7.85 7.83 7.85 2.00 2.00 2.00

Mg# 0.87 0.88 0.87 0.88 0.88 0.88XFe 0.42 0.43 0.42XIlm 0.817 0.705 0.738

Ilm ¼ ilmenite; XIlm calculated after Carmichael (1967) as outlined in Lepage (2003).Cations per formula unit calculated based on 4 and 11 oxygens for olivine and biotite, respectively. For ilmenite, 2 cations and 3 oxygens were used for the calculations.

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 205

low as the CMV dacites. Hence, we interpret the SreNd isotoperatios of the CMV dacites as a signature that has been acquired in acrustal hot zone, where evolved silicic melts are generated both bypartial crystallization of the parental basalts and by partial meltingof surrounding crustal rocks (Annen et al., 2006). The precursormafic magma had crystallized garnet to create the adakiticgeochemical signature. This scenario for the CMV dacites is in linewith a significant crustal influence on the composition of

Table 6Summary of crystallization conditions of amphiboles in Cerro Machín dacites, determine

Amphibole Sample N xtals N points T (�C) T (�C) T (�C) P

Avg � sd Min Max A

Type I D1 9 93 907 � 17 867 955 3Type I D2A 5 56 906 � 16 865 934 3Type I D3B 3 16 909 � 9 897 924 3Type I D4 5 32 904 � 15 855 924 3Type II core D2A 3 12 918 � 17 887 947 3Type II core D3B 2 7 914 � 15 887 934 3Type II core D4 4 13 913 � 12 883 932 3Type II rim D2A 3 10 964 � 18 926 985 3Type II rim D3B 2 9 965 � 17 944 987 4Type II rim D4 4 34 983 � 11 962 1004 4

Ecuadorian NVZ lavas. There, the generation of the adakite signa-ture was attributed to lower-to-mid-crustal processing of mantle-derived melts (Chiaradia et al., 2009; Hidalgo et al., 2012). Sincethe SreNd isotopic composition of regional upper crustal rocks, asrepresented by granites and orthogneisses (87Sr/86Sr ¼ 0.713e0.761, 143Nd/144Nd¼ 0.5121e0.5125; Vinasco et al., 2006) is distinctfrom CMV dacites, a significant geochemical contribution by uppercrustal lithologies can be excluded. In summary, the isotope data

d by amphibole geothermobarometry (Ridolfi et al., 2010).

(MPa) P (MPa) P (MPa) Depth (km) DNNO H2Omelt (wt.%)

vg � sd Min Max Avg � sd Avg � sd Avg � sd

52 � 62 234 568 13.3 � 2.4 0.5 � 0.2 7.1 � 0.442 � 49 227 427 12.9 � 1.9 0.4 � 0.2 7.1 � 0.339 � 21 309 339 12.8 � 0.8 0.5 � 0.1 7.0 � 0.145 � 56 198 432 13.0 � 2.1 0.5 � 0.2 7.1 � 0.375 � 60 271 496 14.2 � 2.3 0.3 � 0.2 7.1 � 0.363 � 57 278 435 13.7 � 2.2 0.4 � 0.2 7.0 � 0.265 � 40 275 426 13.8 � 1.5 0.4 � 0.1 7.0 � 0.281 � 69 240 475 14.4 � 2.6 1.3 � 0.5 5.7 � 1.002 � 68 320 505 15.2 � 2.6 1.2 � 0.3 6.0 � 0.765 � 58 361 606 17.6 � 2.2 1.2 � 0.3 6.2 � 0.5

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208206

suggest that the isotopic signature has been acquired by addition oflower/middle crustal material, whereas the textural diversityformed during magma ascent and crystallization at shallowerlevels.

7. Conclusions

The last eruption of Cerro Machín Volcano in the ColombianCentral Cordillera (Northern Volcanic Zone of the Andes) atw900 years BP terminated with the formation of a dacitic lavadome, which was investigated to unravel the conditions of pre-eruptive magma storage and evolution. We focused on amphibolethermobarometry and chemometry based on Ridolfi et al. (2010)and Ridolfi and Renzulli (2012) because amphibole is a ubiquitousphase in the dome lavas and the two types of amphibole occurringin the rocks provide distinct petrologic information. At mid-crustaldepths, crystallization of amphibole occurred at T ¼ 910 � 30 �Cand P ¼ 360 � 70 MPa from a rhyolitic magma which wasmoderately oxidized and water-rich. This resident magma wasintruded by a mafic, mantle-derived magma that led to growth ofcompositionally distinct amphibole rims and the incorporation ofolivine xenocrysts into the rhyolitic magma. The hybrid magmareached an intermediate, dacitic composition and records slightlyhigher temperatures (T ¼ 970 � 25 �C) than the rhyolitic melt. Thedepth of the magma reservoir in which amphibole crystallizedprior to the dome-forming eruptive phase of the w900 years BPeruption is in a similar depth to one of the three regions whereearthquake hypocenters are currently observed. Shallow-levelcrystallization occurred at lower temperatures (<800 �C) andinvolved fluctuations in the intrinsic conditions. The chemicallyhomogenous lava dome dacites display an adakitic geochemicalsignature similar to many other volcanoes of the Andean NVZ. Therelatively radiogenic Sr- and Nd isotopic compositions point to acrustal compositional control, suggesting that differentiation andassimilation processes rather than slab melting were the mainfactors to produce the adakite signature.

Acknowledgments

This work was initiated during KL’s visit at the ObservatorioVulcanológico y Sismológico de Manizales and is based on hersubsequent BSc thesis. We thank Milton Ordóñez and John M.Londoño for guidance in the field and for providing informationabout the seismic activity at Cerro Machín. Mario Thöner helpedwith electron microprobe measurements at GEOMAR and LindaForbes helped with Sr and Nd column separation in Leeds. ArminFreundt and the Leeds Volcanic Studies Group are thanked forconstructive comments on an earlier version of the manuscript. Wealso thank James Kellogg for editorial handling and two anonymousreviewers for their detailed and constructive comments, whichhelped to improve the manuscript. This publication is contributionno. 260 of the Sonderforschungsbereich (SFB) 574 “Volatiles andFluids in Subduction Zones” at Kiel University.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jsames.2013.09.009.

References

Alonso-Perez, R., Müntener, O., Ulmer, P., 2009. Igneous garnet and amphibolefractionation in the roots of island arcs: experimental constraints on andesiticliquids. Contrib. Mineral. Petrol. 157, 541e558.

Annen, C., Blundy, J.D., Sparks, R.S.J., 2006. The genesis of intermediate and silicicmagmas in deep crustal hot zones. J. Petrol. 47, 505e539.

Anderson, J.L., Smith, D.R., 1995. The effects of temperature and fO2on the Al-in-

hornblende barometer. Am. Mineral. 80, 549e559.Arango, E., Castañeda, D.M., 2012. Morfología, petrografía y geoquímica del cráter

actual del volcán Cerro Machín, Colombia. Tesis de grado. Universidad de Cal-das, Manizales, p. 60.

Bachmann, O., Dungan, M.A., 2002. Temperature-induced Al-zoning in hornblendesof the Fish Canyon magma, Colorado. Am. Mineral. 87, 1062e1076.

Bachmann, O., Dungan, M.A., Lipman, P.W., 2002. The Fish Canyon magma body, SanJuan Volcanic Field, Colorado: rejuvenation and eruption of an upper-crustalbatholith. J. Petrol. 43, 1469e1503.

Barragan, R., Geist, D., Hall, M., Larson, P., Kurz, M., 1998. Subduction controls on thecompositions of lavas from the Ecuadorian Andes. Earth Planet. Sci. Lett. 154,153e166.

Blundy, J., Cashman, K., 2001. Ascent-driven crystallization of dacite magmas atMount St Helens, 1980e1986. Contrib. Mineral. Petrol. 140, 631e650.

Blundy, J., Cashman, K., 2008. Petrologic reconstruction of magmatic system vari-ables and processes. Rev. Mineral. Geochem. 69, 179e239.

Bourdon, E., Eissen, J.P., Gutscher, M.A., Monzier, M., Hall, M.L., Cotten, J., 2003.Magmatic response to early aseismic ridge subduction: the Ecuadorian margincase (South America). Earth Planet. Sci. Lett. 205, 123e138.

Bourdon, E., Eissen, J.-P., Monzier, M., Robin, C., Martin, H., Cotton, J., Hall, M., 2002.Adakite-like lavas from Antisana volcano (Ecuador): evidence for slab meltmetasomatism beneath the Andean Northern Volcanic Zone. J. Petrol. 43, 199e217.

Boynton, W.V., 1984. Geochemistry of the rare earth elements: meteorite studies.In: Henderson, P. (Ed.), Rare Earth Element Geochemistry. Elsevier, Amsterdam,pp. 63e114.

Browne, B.L., Gardner, J.E., 2006. The influence of magma ascent path on texture,mineralogy, and formation of hornblende reaction rims. Earth Planet. Sci. Lett.246, 161e176.

Browne, B.L., Eichelberger, J.C., Patino, L.C., Vogel, T.A., Uto, K., Hoshizumi, H., 2006.Magma mingling as indicated by texture and Sr/Ba ratios of plagioclase pheno-crysts from Unzen volcano, SW Japan. J. Volcanol. Geotherm. Res. 154, 103e116.

Calvache, V.M.L., Williams, S.N., 1997. Geochemistry and petrology of the Galerasvolcanic complex, Colombia. J. Volcanol. Geotherm. Res. 77, 21e38.

Carmichael, I.S.E., 1967. The iron-titanium oxides of salic volcanic rocks and theirassociated ferromagnesian silicates. Contrib. Mineral. Petrol. 14, 36e64.

Castro, J.M., Dingwell, D.B., 2009. Rapid ascent of rhyolitic magma at Chaitén vol-cano, Chile. Nature 461, 780e783.

Cepeda, H., Murcia, L.A., Monsalve, M.L., Méndez, R.A., Núñez, A., 1995. Volcán CerroMachín, Departamento del Tolima, Colombia: Pasado, presente y futuro. Inter-nal report. INGEOMINAS, Popayán, Colombia, p. 48.

Cepeda, H., Cortés, G.P., Macías, J.L., Méndez, R.A., Monsalve, M.L., Rueda, H.,Siebe, C., 1999. El Volcán Cerro Machín, un volcán Holocénico explosivo en elcentro de Colombia. In: Cuarta Reunión Técnica de la Comisión de Geofísica.Departamento de Geodesia y TopografíaeeUNT, Tucumán, Argentina.

Chiaradia, M., Müntener, O., Beate, B., Fontignie, D., 2009. Adakite-like volcanism ofEcuador: lower crust magmatic evolution and recycling. Contrib. Mineral.Petrol. 158, 563e588.

Clynne, M.A., 1999. A complex magma mixing origin for rocks erupted in 1915,Lassen Peak, California. J. Petrol. 40, 105e132.

Coombs, M.L., Gardner, J.E., 2004. Reaction rim growth on olivine in silicic melts:implications for magma mixing. Am. Mineral. 89, 748e759.

Cortés, G.P., 2001. Estudio geológico de los depósitos de lahar asociados a la acti-vidad eruptiva del Volcán Cerro Machín. Internal Report. INGEOMINAS, Man-izales, Colombia, p. 96.

Cortés, G.P., Murcia, H., Hurtado, B., Cepeda, H., Nuñez, A., 2006. Comparison of thelahar deposits of the eruption of Nevado del Ruiz volcano on 13th of November,1985 and the pre-historic eruptions of Cerro Machín Volcano in the central zoneof Colombia. In: Abstract, Cities on Volcanoes 4 Conference, Quito, Ecuador.

Defant, M.J., Drummond, M.S., 1990. Derivation of some modern arc magmas bymelting of young subducted lithosphere. Nature 347, 662e665.

Di Muro, A., Pallister, J., Villemant, B., Newhall, C., Semet, M., Martinez, M.,Mariet, C., 2008. Pre-1991 sulfur transfer between mafic injections and dacitemagma in the Mt. Pinatubo reservoir. J. Volcanol. Geotherm. Res. 175, 517e540.

Droux, A., Delaloye, M., 1996. Petrography and geochemistry of Plio-Quaternarycalc-alkaline volcanoes of southwestern Colombia. J. S. Am. Earth Sci. 9, 27e41.

Eichelberger, J.C., 1980. Vesiculation of mafic magma during replenishment of silicicmagma reservoirs. Nature 288, 446e450.

Eichelberger, J.C., 1995. Silicic volcanism: ascent of Viscous magmas from crustalreservoirs. Annu. Rev. Earth Planet. Sci. 23, 41e63.

Garrison, J.M., Davidson, J.P., 2003. Dubious case for slab melting in the Northernvolcanic zone of the Andes. Geology 31, 565e568.

Garrison, J., Davidson, J., Reid, M., Turner, S., 2006. Source versus differentiationcontrols on U-series disequilibria. Insights from Cotopaxi volcano, Ecuador.Earth Planet. Sci. Lett. 244, 548e565.

Geist, D., Diefenbach, B.A., Fornari, D.J., Kurz, M.D., Harpp, K., Blusztajn, J., 2008.Construction of the Galápagos platform by large submarine volcanic terraces.Geochem. Geophy. Geosys. 9, Q03015. http://dx.doi.org/10.1029/2007GC001795.

Gourgaud, A., Thouret, J.-C., 1990. Magma mixing and petrogenesis of the 13November 1985 eruptive products at Nevado del Ruiz (Colombia). J. Volcanol.Geotherm. Res. 41, 79e96.

Halama, R., Boudon, G., Villemant, B., Joron, J.-L., Le Friant, A., Komorowski, J.-C.,2006. Pre-eruptive crystallization conditions of mafic and silicic magmas at the

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208 207

Plat Pays volcanic complex, Dominica (Lesser Antilles). J. Volcanol. Geotherm.Res. 153, 200e220.

Hall, M.L., Wood, C.A., 1985. Volcano-tectonic segmentation of the northern Andes.Geology 13, 203e207.

Hattori, K., Sato, H., 1996. Magma evolution recorded in plagioclase zoning in 1991Pinatubo eruption products. Am. Mineral. 81, 982e994.

Hidalgo, S., Monzier, M., Martin, H., Chazot, G., Eissen, J.-P., Cotten, J., 2007. Adakiticmagmas in the Ecuadorian volcanic front: petrogenesis of the Iliniza volcaniccomplex (Ecuador). J. Volcanol. Geotherm. Res. 159, 366e392.

Hidalgo, S., Gerbe, M.C., Martin, H., Samaniego, P., Bourdon, E., 2012. Role of crustaland slab components in the Northern Volcanic Zone of the Andes (Ecuador)constrained by SreNdeO isotopes. Lithos 132e133, 180e192.

Holland, T., Blundy, J., 1994. Non-ideal interactions in calcic amphiboles and theirbearing on amphibole-plagioclase thermometry. Contrib. Mineral. Petrol. 116,433e447.

Holtz, F., Sato, H., Lewis, J., Behrens, H., Nakada, S., 2005. Experimental petrology ofthe 1991e1995 Unzen dacite, Japan. Part I: phase relations, phase compositionand pre-eruptive conditions. J. Petrol. 46, 319e337.

Irvine, T.N., Baragar, W.R.A., 1971. A guide to the chemical classification of thecommon volcanic rocks. Can. J. Earth. Sci. 8, 523e548.

Johnson, M.C., Rutherford, M.J., 1989. Experimental calibration of the aluminum-in-hornblende geobarometer with application to Long Valley caldera (California)volcanic rocks. Geology 17, 837e841.

Kelemen, P.B., Hanghøj, K., Greene, A.R., 2003. One view of the geochemistry ofsubduction-related magmatic arcs, with an emphasis on primitive andesite andlower crust. In: Rudnick, R.L. (Ed.), Treatise on Geochemistry, The Crust, vol. 3.Elsevier-Pergamon, Oxford, pp. 593e659.

Le Maitre, R.W., Bateman, P., Dudek, A., Keller, J., Lameyre, J., Le Bas, M.J., Sabine, P.A.,Schmid, R., Sorensen, H., Streckeisen, A., Woolley, A.R., Zanettin, B., 1989.A Classification of Igneous Rocks and Glossary of Terms: Recommendations ofthe International Union of Geological Sciences Subcommission on the Sys-tematics of Igneous Rocks. Blackwell Scientific, Oxford.

Leake, B.E., Woolley, A.R., Birch, W.D., Burke, E.A.J., Ferraris, G., Grice, J.D.,Hawthorne, F.C., Kisch, H.J., Krivovichev, V.G., Schumacher, J.C.,Stephenson, N.C.N., Whittaker, E.J.W., 1997. Nomenclature of amphiboles:report of the subcommittee on amphiboles of the international mineralogicalassociation, commission on new minerals and mineral names. Can. Mineral. 35,219e246.

Lepage, L.D., 2003. ILMAT: an excel worksheet for ilmenite-magnetite geo-thermometry and geobarometry. Comp. Geosci. 29, 673e678.

Londoño, J.M., Cortés, G.P., Méndez, R.A., 2007. Recent volcanic activity of CerroMachín Volcano, Colombia, 1988e2007, and implications of volcanic hazardassessment. In: Abstract, Cities on Volcanoes 5 Conference, Shimabara, Japan.

Londoño, J.M., 2011. 3D seismic tomography of Cerro Machín Volcano, Colombia. In:XIV Latinamerican Congress of Geology. Medellín, Colombia.

McDonough, W.F., Sun, S.-S., 1995. The composition of the Earth. Chem. Geol. 120,223e253.

Martel, C., Pichavant, M., Holtz, F., Scaillet, B., Bourdier, J.-L., Traineau, H., 1999. Ef-fects of fO2

and H2O on andesite phase relations between 2 and 4 kbar.J. Geophys. Res. 104 (B12), 29453e29470.

Martin, H., 1999. Adakitic magmas: modern analogues of Archaean granitoids.Lithos 46, 411e429.

Martin, H., Smithies, R.H., Rapp, R.P., Moyen, J.-F., Champion, D.C., 2005. An over-view of adakite, tonaliteetrondhjemiteegranodiorite (TTG) and sanukitoid:relationships and some implications for crustal evolution. Lithos 79, 1e24.

Martin, V.M., Pyle, D.M., Holness, M.B., 2006. The role of crystal frameworks in thepreservation of enclaves during magma mixing. Earth Planet. Sci. Lett. 248,787e799.

Maya, M., González, H., 1995. Unidades litodémicas en la Cordillera Central deColombia. Bolet. Geológ. Ingeom. 35, 43e57.

Melson, W.G., Allan, J.F., Reid Jerez, D., Nelen, J., Calvache, M.L., Williams, S.N.,Fournelle, J., Perfit, M., 1990. Water contents, temperatures and diversity of themagmas of the catastrophic eruption of Nevado del Ruiz, Colombia, November13, 1985. J. Volcanol. Geotherm. Res. 41, 97e126.

Méndez, R.A., Cortés, G.P., Cepeda, H., 2002. Evaluación de la amenaza volcánicapotencial del Cerro Machín (Departamento del Tolima, Colombia). InternalReport. INGEOMINAS, Manizales, p. 67.

Monsalve, M.L., Correa, A.M., Pulgarín, B., Martínez, L., Navarro, S., Murcia, H., 2011.Geoquímica del Complejo Volcánico Sotará, Colombia. In: Memories XIV Con-greso Latinoamericano de Geología e XII Congreso Colombiano de Geología.Medellin, Colombia.

Moyen, J.-F., 2009. High Sr/Y and La/Yb ratios: the meaning of the “adakitic signa-ture”. Lithos 112, 556e574.

Murcia, H.F., Hurtado, B.O., Cortés, G.P., Macías, J.L., Cepeda, H., 2008. Thew2500 yr B.P. Chicoral non-cohesive debris flow from Cerro Machín Volcano,Colombia. J. Volcanol. Geotherm. Res. 171, 201e214.

Murcia, H.F., Sheridan, M.F., Macias, J.L., Cortés, G.P., 2010. TITAN2D simulations ofpyroclastic flow at Cerro Machín Volcano, Colombia: Hazard implications. J. S.Am. Earth Sci. 29, 161e170.

Murphy, M.D., Sparks, R.S.J., Barclay, J., Carroll, M.R., Lejeune, A.-M., Brewer, T.S.,Macdonald, R., Black, S., Young, S., 1998. The role of magma mixing in triggeringthe current eruption at the Soufrière Hills volcano, Montserrat, West Indies.Geophys. Res. Lett. 25, 3433e3436.

Pallister, J.S., Hoblitt, R.P., Reyes, A.G., 1992. A basalt trigger for the 1991 eruptions ofPinatubo volcano? Nature 356, 426e428.

Prouteau, G., Scaillet, B., 2003. Experimental constraints on the origin of the 1991Pinatubo dacite. J. Petrol. 44, 2203e2241.

Pulgarín, B., Navarro, S., Monsalve, M.L., Murcia, H., Téllez, L., Correa, A.M.,Martínez, L., Martínez, L., 2010. Geología e historia eruptiva del ComplejoVolcánico Sotará. Internal Report. INGEOMINAS, Manizales/Popayán, p. 401.

Putirka, K.D., 2008. Thermometers and barometers for volcanic systems. Rev.Mineral. Geochem. 69, 61e120.

Ramos, V.A., Alemán, A., 2000. Tectonic evolution of the Andes. In: Cordani, U.G.,Milani, E.J., Thomaz Filho, A., Campos Neto, D.A. (Eds.), Tectonic Evolution ofSouth America. FINEP, Rio de Janeiro, pp. 635e685.

Reynaud, C., Jaillard, E., Lapierre, H., Mamberti, M., Mascle, G.H., 1999. Oceanicplateau and island arcs of southwestern Ecuador: their place in the geodynamicevolution of northwestern South America. Tectonophysics 307, 235e254.

Ridolfi, F., Renzulli, A., Puerini, M., 2010. Stability and chemical equilibrium ofamphibole in calc-alkaline magmas: an overview, new thermobarometric for-mulations and application to subduction related volcanoes. Contrib. Mineral.Petrol. 160, 45e66.

Ridolfi, F., Renzulli, A., 2012. Calcic amphiboles in calc-alkaline and alkalinemagmas: thermobarometric and chemometric empirical equations valid up to1130 �C and 2.2 GPa. Contrib. Mineral. Petrol. 163, 877e895.

Roeder, P.L., Emslie, R.F., 1970. Olivine-liquid equilibrium. Contrib. Mineral. Petrol.29, 275e289.

Rueda, H., 2005. Erupciones plinianas del Holoceno en el Volcán Cerro Machín,Colombia. Estratigrafía, petrografía y dinámica eruptiva. MSc Thesis. Uni-versidad Nacional Autónoma de México, México, p. 110.

Ruprecht, P., Plank, T., 2013. Feeding andesitic eruptions with a high-speedconnection from the mantle. Nature 500, 68e72.

Rutherford, M.J., Devine, J.D., 2003. Magmatic conditions and magma ascent asindicated by hornblende phase equilibria and reactions in the 1995e2002Soufrière Hills magma. J. Petrol. 44, 1433e1454.

Samaniego, P., Martin, H., Monzier, M., Robin, C., Fornari, M., Eissen, J.-P., Cotton, J.,2005. Temporal evolution of magmatism in the Northern Volcanic Zone of theAndes: the geology and petrology of Cayambe Volcanic Complex (Ecuador).J. Petrol. 46, 2225e2252.

Samaniego, P., Martin, H., Robin, C., Monzier, M., 2002. Transition from calc-alkalicto adakitic magmatism at Cayambe volcano, Ecuador: insights into slab meltsand mantle wedge interactions. Geology 30, 967e970.

Samaniego, P., Robin, C., Chazot, G., Bourdon, E., Cotten, J., 2010. Evolving meta-somatic agent in the Northern Andean subduction zone, deduced from magmacomposition of the long-lived Pichincha volcanic complex (Ecuador). Contrib.Mineral. Petrol. 160, 239e260.

Savov, I.P., Luhr, J., Navarro-Ochoa, C., 2008. Petrology and geochemistry of lava andash erupted from Volcán Colima, Mexico, during 1998e2005. J. Volcanol.Geotherm. Res. 174, 241e256.

Sato, H., Nakada, S., Fujii, T., Nakamura, M., Suzuki-Kamata, K., 1999. Groundmasspargasite in the 1991e1995 dacite of Unzen volcano: phase stability exper-iments and volcanological implications. J. Volcanol. Geotherm. Res. 89, 197e212.

Scaillet, B., Evans, B.W., 1999. The 15 June 1991 eruption of Mount Pinatubo, I, Phaseequilibria and pre-eruption PeTefO2

eH2O conditions of the dacite magma.J. Petrol. 40, 381e411.

Schiano, P., Monzier, M., Eissen, J.-P., Martin, H., Koga, K.T., 2010. Simple mixing asthe major control of the evolution of volcanic suites in the Ecuadorian Andes.Contrib. Mineral. Petrol. 160, 297e312.

Shane, P., Smith, V.C., 2013. Using amphibole crystals to reconstruct magma storagetemperatures and pressures for the post-caldera collapse volcanism at Okatainavolcano. Lithos 156-159, 159e170.

Sigurdsson, H., Carey, S., Palais, J.M., Devine, J., 1990. Pre-eruption compositionalgradients and mixing of andesite and dacite magma erupted from Nevado delRuiz Volcano, Colombia in 1985. J. Volcanol. Geotherm. Res. 41, 127e151.

Sparks, S.R.J., 2003. Forecasting volcanic eruptions. Earth Planet. Sci. Lett. 210, 1e15.Sparks, S.R.J., Sigurdsson, H., Wilson, L., 1977. Magma mixing: a mechanism for

triggering acid explosive eruptions. Nature 267, 315e318.Straub, S.M., Gomez-Tuena, A., Stuart, F.M., Zellmer, G.F., Espinasa-Perena, R., Cai, Y.,

Iizuka, Y., 2011. Formation of hybrid arc andesites beneath thick continentalcrust. Earth Planet. Sci. Lett. 303, 337e347.

Streck, M.J., Dungan, M.A., Bussy, F., Malavassi, E., 2005. Mineral inventory ofcontinuously erupting basaltic andesites at Arenal volcano, Costa Rica: impli-cations for interpreting monotonous, crystal-rich, mafic arc stratigraphies.J. Volcanol. Geotherm. Res. 140, 133e155.

Streck, M.J., Leeman, W.P., Chesley, J., 2007. High-magnesian andesite from MountShasta: a product of magma mixing and contamination, not a primitive mantlemelt. Geology 35, 351e354.

Streck, M.J., 2008. Mineral textures and zoning as evidence for open system pro-cesses. Rev. Mineral. Geochem. 69, 595e622.

Swanson, S.E., Nye, C.J., Miller, T.P., Avery, V.F., 1994. Geochemistry of the 1989e1990eruption of Redoubt Volcano: part II. Evidence from mineral and glasschemistry. J. Volcanol. Geotherm. Res. 62, 453e468.

Taboada, A., Rivera, L.A., Fuenzalida, A., Cisternas, A., Philip, H., Bijwaard, H., Olaya, J.,Rivera, C., 2000. Geodynamics of the northern Andes: subductions and intra-continental deformation (Colombia). Tectonics 19, 787e813.

Thouret, J.C., Cantagrel, J.-M., Robin, C., Murcia, A., Salinas, R., Cepeda, H., 1995.Quaternary eruptive history and hazard-zone model at Nevado del Tolimaand Cerro Machín Volcanoes, Colombia. J. Volcanol. Geotherm. Res. 66, 397e426.

K. Laeger et al. / Journal of South American Earth Sciences 48 (2013) 193e208208

Thornber, C.R., Pallister, J.S., Lowers, H.A., Rowe, M.C., Mandeville, C.W., Meeker, G.P.,2008. Chemistry, mineralogy, and petrology of amphibole in Mount St. Helens2004e2006 dacite. In: Sherrod, D.R., Scott, W.E., Stauffer, P.H. (Eds.), A VolcanoRekindled: the Renewed Eruption of Mount St. Helens, 2004e2006, pp. 727e754. U.S. Geological Survey Professional Paper, 1750.

Toro, L.M., Moreno, M., Blanco-Quintero, I., Ruiz, E.C., Gómez, A., Vinasco, C., 2012.Composición geoquímica y origen de las metabasitas del Complejo Cajamarca,Cordillera Central de Colombia. In: Cordilleran Section e GSA 108th AnnualMeeting, Querétaro, México.

Tsuchiyama, A., 1985. Dissolution kinetics of plagioclase in the melt of the systemdiopside-albite-anorthite, and origin of dusty plagioclase in andesites. Contrib.Mineral. Petrol. 89, 1e16.

Vinasco, C.J., Cordani, U.G., González, H., Weber, M., Pelaez, C., 2006. Geochrono-logical, isotopic, and geochemical data from Permo-Triassic granitic gneissesand granitoids of the Colombian Central Andes. J. S. Am. Earth Sci. 21, 355e371.

Vargas, C.A., Kammer, A., Valdes, M., Rodríguez, C.E., Caneva, A., Sánchez, J.J.,Arias, E., Cortes, C.A., Mora, H., 2005. New geological and geophysical

contributions in the section Ibagué e Armenia, Central Cordillera e Colombia.Earth Sci. Res. J. 9, 99e109.

Venezky, D.Y., Rutherford, M.J., 1999. Petrology and FeeTi oxide reequilibration ofthe 1991 Mount Unzen mixed magma. J. Volcanol. Geotherm. Res. 89, 213e230.

Villagómez, D., Spikings, R., 2013. Thermochronology and tectonics of the Centraland Western Cordilleras of Colombia: Early Cretaceous e Tertiary evolution ofthe Northern Andes. Lithos 160e161, 228e249.

Villagómez, D., Spikings, R., Magna, T., Kammer, A., Winkler, W., Beltrán, A., 2011.Geochronology, geochemistry and tectonic evolution of the Western and Cen-tral cordilleras of Colombia. Lithos 125, 875e896.

Weber, M.B.I., Tarney, J., Kempton, P.D., Kent, R.W., 2002. Crustal make-up of thenorthern Andes: evidence based on deep crustal xenolith suites, Mercaderes,SW Colombia. Tectonophysics 345, 49e82.

White, W.M., McBirney, A.R., Duncan, R.A., 1993. Petrology and geochemistry of theGalapagos Islands: portrait of a pathological mantle plume. J. Geophys. Res. 98(B11), 19533e19563.

Zellmer, G.F., Iizuka, Y., Miyoshi, M., Tamura, Y., Tatsumi, Y., 2012. Lower crustal H2Ocontrols on the formation of adakitic melts. Geology 40, 487e490.