The Eruptive Chronology of the Ampato-Sabancaya Volcanic

Home , Ampato, Aucanquilcha, Casiri (Tacna), Chachani, Chimborazo, Colca Canyon

Journal of Volcanology and Geothermal Research 323 (2016) 110–128

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

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

The eruptive chronology of the Ampato–Sabancaya volcanic complex (Southern Peru)

Pablo Samaniego a,⁎,MarcoRiverab, Jersy Mariño b, Hervé Guillou c,CélineLiorzoud, Swann Zerathe e, Rosmery Delgado b, Patricio Valderrama a,b,VincentScaoc a Laboratoire Magmas et Volcans, Université Blaise Pascal - CNRS - IRD, 6 Avenue Blaise Pascal, TSA 60026 - CS 60026, 63178 Aubière, France b Observatorio Vulcanológico del INGEMMET, Dirección de Geología Ambiental y Riesgo Geológico, Urb. Magisterial B-16, Umacollo, Arequipa, Peru c Laboratoire des Sciences du Climat et de l'Environnement, LSCE/IPSL, CEA-CNRS-UVSQ, Université Paris-Saclay, F-91198 Gif-sur-Yvette, France d Laboratoire Domaines Océaniques, Institut Universitaire Européen de la Mer, Université de Bretagne Occidentale, Rue Dumont d'Urville, 29280 Plouzané, France e Institut des Sciences de la Terre, Université Grenoble Alpes – CNRS - IRD, 1381 rue de la piscine, 38400 Saint Martin d'Hères, France article info abstract

Article history: We have reconstructed the eruptive chronology of the Ampato–Sabancaya volcanic complex (Southern Peru) on Received 14 January 2016 the basis of extensive fieldwork, and a large dataset of geochronological (40K–40Ar, 14Cand3He) and geochemical Received in revised form 1 April 2016 (major and trace element) data. This volcanic complex is composed of two successive edifices that have experi- Accepted 29 April 2016 enced discontinuous volcanic activity from Middle Pleistocene to Holocene times. The Ampato compound Available online 07 May 2016 volcano consists of a basal edifice constructed over at least two cone-building stages dated at 450–400 ka and – fi fi Keywords: 230 200 ka. After a period of quiescence, the Ampato Upper edi ce was constructed rstly during an effusive – – Ampato stage (80 70 ka), and then by the formation of three successive peaks: the Northern, Southern (40 20 ka) and Sabancaya Central cones (20–10 ka). The Southern peak, which is the biggest, experienced large explosive phases, resulting Central Andes in deposits such as the Corinta plinian fallout. During the Holocene, eruptive activity migrated to the NE and con- Eruptive chronology structed the mostly effusive Sabancaya edifice. This cone comprised many andesitic and dacitic blocky lava flows Eruptive rates and a young terminal cone, mostly composed of pyroclastic material. Most samples from the Ampato–Sabancaya Volcanic hazards define a broad high-K magmatic trend composed of andesites and dacites with a mineral assemblage of plagioclase, amphibole, biotite, ortho- and clino-pyroxene, and Fe–Ti oxides. A secondary trend also exists, corresponding to rare dacitic explosive eruptions (i.e. Corinta fallout and flow deposits). Both magmatic trends are derived by fractional crystallisation involving an amphibole-rich cumulate with variable amounts of upper crustal assimilation. A marked change in the overall eruptive rate has been identified between Ampato (~0.1 km3/ka) and Sabancaya (0.6–1.7 km3/ka). This abrupt change demonstrates that eruptive rates have not been homogeneous throughout the volcano's history. Based on tephrochronologic studies, the Late Holocene Sabancaya activity is characterised by strong vulcanian events, although its erupted volume remained low and only produced a local impact through ash fallout. We have identified at least 6 eruptions during the last 4–5 ka, including the historical AD 1750–1784 and 1987–1998 events. On the basis of this recurrent low-to-moderate explosive activity, Sabancaya must be considered active and a potentially threatening volcano. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Zone (CVZ) results from the subduction of the oceanic Nazca plate below the South American continental lithosphere. As a result, the Reconstructing the eruptive chronology of active volcanic systems volcanic front includes at least twelve volcanic centres of Pleistocene represents a key step for any hazard assessment initiative. However, age (Fig. 1a) of which seven have experienced historical eruptive the recent eruptions of Chaitén (2008, Major and Lara, 2013)and activity (i.e. since the arrival of the Spanish conquistadors in the 16th Reventador volcanoes (2002, Hall et al., 2004) showed that the eruptive century). These volcanoes include El Misti (Thouret et al., 2001; chronology of many active volcanic complexes remains poorly known. Harpel et al., 2011), which threatens the city of Arequipa, the active vol- In the Andean cordillera, the Peruvian segment of the Central Volcanic canoes of Ubinas (Thouret et al., 2005; Rivera et al., 2014)and Sabancaya (Gerbe and Thouret, 2004), and Huaynaputina volcano (Thouret et al., 1999; Adams et al., 2001), which has had the biggest ⁎ Corresponding author. historical eruption in the Andes. However, little is still known about E-mail address: [email protected] (P. Samaniego). the eruptive chronology of some of these volcanic centres, such as the

74º 72ºHistorically active 70º Plio-Quaternary15º a Potentially active volcanic front SARA SARA SOLIMANA ANDAHUA c AMPATO- JULIACA COROPUNA TITICACA BOLIVIALAKE RIO COLCA SABANCAYA

CHALA PUNO Madrigal 16º CHACHANI 16º Pinchollo MISTI Cabanaconde Lari

UBINAS 000 AREQUIPA HUAYNAPUTINA CAMANA Maca TICSANI

8270 Y TUTUPACA MOLLENDO MOQUEGUA YUCAMANE Achoma

ILO CASIRI

5-6 cm/y HUALCA Qda. Huayuray 18º HUALCA 18ºS TACNA CHILE 0 50 100 km

74º 72º 70º 000 Río Sepina

8260 Mucurca 72º Cabanaconde Chivay Trigal Río Colca lake Maca SABANCAYA Ichupamba Solarpampa Hualca Sepina Colihuiri

Hualca 000 Huambo

8250 AMPATO Sabancaya Cajamarcana

Ampato Collpa Río Parcomayo

Sallalli Normal fault Huanca Japo

Strike slip 000 Corinta Lineaments 8240

0 5 km Baylillas b 72º 16º 180000 190000 200000 210000

Fig. 1. (a) The Peruvian volcanic arc. (b) Structural context of the Ampato–Sabancaya region, including the Colca river valley. Main structures from Mering et al. (1996) and Gerbe and Thouret (2004) Sabancaya and Hualca Hualca complexes and the nearby Colca canyon. 112 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

Sabancaya volcano, and its neighbouring Ampato edifice. Rare historical 3. Methodology accounts mention eruptive activity that occurred in AD 1750 and 1784 (Siebert et al., 2010; Travada y Córdova, 1752; Zamácola y Jaúregui, Fieldwork was carried out during several field campaigns between 1888). More recently, Sabancaya entered a new eruptive phase in 2009 and 2012, which included geological mapping and sampling of 1988, which lasted until at least 1997 (Global Volcanism Program, most volcanic units. At high altitude (above 5000 m asl), fieldwork 1988, 1997). During this period, Sabancaya experienced low to moder- was complicated by the presence of a large icecap as well as voluminous ate explosive eruptions (VEI 1–2) that were characterised by violent glacial deposits. However, the presence of numerous deep glacial vulcanian explosions accompanied by small (up to 5–7 km height) valleys allowed sampling of almost all volcanic units, resulting in a eruption columns with a local ash fallout impact. The most significant broad sample array for petrographic and geochemical studies (Fig. 2). activity was observed between April–May 1990 and April 1991 Major and trace element whole-rock analyses were obtained from (Global Volcanism Program, 1990, 1991). Since March–April 2013, agate-crushed powders of 133 samples spanning the entire volcanic Sabancaya has shown increased fumarolic activity, accompanied by complex, at the Institut Universitaire Européen de la Mer, Université frequent seismic swarms (Global Volcanism Program, 2013; Jay et al., de Bretagne Occidentale (Brest, France), using an Inductive Coupled 2015). Plasma-Atomic Emission Spectrometer (ICP-AES) and following the Following its reactivation in 1988, several studies have been analytical procedure described by Cotten et al. (1995). These data, carried out on Sabancaya. These works include an initial geological together with petrographic descriptions, have been used to characterise reconnaissance, comprising a hazard assessment (Thouretetal., and correlate the different volcanic units. 1994), a regional tephro-chronological survey (Juvigné et al., 1998, We constrained the Pleistocene eruptive chronology via the 2008) and a petrological description of the last eruption products unspiked 40K–40Ar dating method at the Laboratoire des Sciences du (Gerbe and Thouret, 2004). Based on detailed field work and Climat et de l'Environnement (LSCE/IPSL, Gif-sur-Yvette, France). We geochronological and petrological studies, we reconstruct the obtained 10 ages covering the entire history of this volcanic complex structure and the volcanic and magmatic history of the Ampato– (Table 1). The Holocene chronology is based on 14 new radiocarbon Sabancaya volcanic complex from the Pleistocene to the present day. ages mainly obtained from peat and soil samples from several peatbogs around the volcanic complex. Most samples (8) were analysed at the Laboratoire de Mesure du Carbone 14 (LMC14, Gif-sur-Yvette, France) 2. Geological setting and an additional group (6) were analysed at the Centre for Isotope Research (CIO), Groningen University (Netherlands). Table 2 shows The Ampato–Sabancaya Volcanic Complex (ASVC, 15° 49.3′S, 71° the conventional 14Cages(±1σ), the 13δ values, and the calibrated 52.7′W) is located 70–75 km NW of Arequipa (Fig. 1). It is construct- ages are given at 2σ confidence levels. Conversion from conventional ed upon the Western Cordillera of the Peruvian Andes, which is 14C ages to calendar ages was carried out using the Calib 7.1 code composed of Mesozoic and Cenozoic volcanic and sedimentary for- (Stuiver and Reimer, 1993; Stuiver et al., 2005) and the recently mations (Klinck et al., 1993; Sébrier and Soler, 1991). To the north, updated Southern Hemisphere calibration curve (SHcal13, Hogg et al., the ASVC borders the older and highly eroded Hualca Hualca volcano 2013), which is available back to 50 ka cal BP. Finally, additional (6025 m above sea level – m asl), located at the southern margin of constraints on the eruptive chronology of Sabancaya volcano were Colca canyon. An 40Ar–39Ar age of 0.80 ± 0.04 Ma has been deter- obtained by cosmic ray exposure dating of lava flows using the couple mined by Gerbe and Thouret (2004) for an andesitic lava flow from 3He/pyroxene. Three samples (Table 3) were collected from the tops this edifice, which represents the base of the Ampato–Sabancaya of undisturbed lava boulders and were processed at the CRPG noble volcanic complex. Southwards, the ASVC dominates a high plateau gas lab (Centre de Recherches Pétrographiques et Géochimiques, with an overall southward slope, which is composed of Mio- Nancy, France). Additional information concerning the geochemical Pliocene volcanic formations comprising lava flows and large and geochronological methods is included in the Supplementary ignimbritic deposits that overlie the Western Cordillera basement material. (Klinck et al., 1993; Mamani et al., 2010). Thouretetal.(2007) obtained a 40Ar–39Ar age of 2.20 ± 0.15 Ma for a dacitic ignimbrite 4. Morphology and structure deposit on top of the Patapampa plateau, located to the east of the volcanic complex. The Ampato–Sabancaya volcanic complex has a roughly elliptical Regional tectonic investigations have identified three main fault basal outline (16–20 km NE–SW by 12–14 km NW-SE; Figs. 2, 3) and systems in this part of the Andes (Fig. 1c; Sébrier and Soler, 1991; is composed of two main edifices: The older Ampato compound volcano Mering et al., 1996). The first one corresponds to NW-SE striking faults (6280 m asl) and the younger Sabancaya edifice (5967 m asl). The lower that are oblique-slip extensional structures, with a minor left-lateral flanks of the Ampato edifice have gentle slopes (5–10°), and are strong- component. These faults correspond to regional structures such as ly glacially eroded. As a result, the flanks are coated by a thick layer of the Huanca and Ichupampa faults (located to the SW and NE of the moraine deposits, especially on the southern and western sides. At ASVC, respectively). A second system is composed of E-W striking higher altitude, slopes are steeper (10–30°), locally up to 45°. The faults with a southward normal, dip-slip component, notably the upper part of the Ampato edifice is composed of three major cones Trigal and Solarpampa faults. These structures are almost parallel to oriented NE-SW (Figs. 2 and 3), hereafter termed the Southern, Central, the Colca river valley, and located to the NW of the complex. A recent and Northern cones, respectively located 1.2 and 2.5 km from the Mw 5.9 earthquake was associated with these faults (Jay et al., 2015). summit (Southern cone). The Ampato edifice reaches an elevation of The third, NE-SW striking faults represent local extensional ~1800 m on the west and east sides, and slightly higher (~2250 m) on structures, such as those crossing the Patapampa plateau, the most the southern side. The summit area is characterised by these three prominent being the Sepina fault. This latter structure seems to peaks and several unconformities that delimit the different cones. The be the focus of several seismic events (Mw 4.5–5) that have summit zone of the Southern cone displays a 1-km-long scar open to occurred during the last few decades (Antayhua et al., 2001; Jay the east testifying to an explosive phase associated with the upper et al., 2015). The Ampato and Sabancaya vents, as well as several part of Ampato. It also marks the transition from the older Southern glacial valleys in the western part of Ampato, are roughly aligned cone to the younger Central cone. in a NE-SW direction. This observation suggests that the NE-SW The Sabancaya edifice is located 4–5 km to the NE of Ampato's Sepina fault probably controlled the structural development of the Southern peak. It is built on the remnants of Ampato and Hualca Hualca ASVC. (Figs. 2 and 3), and reaches 1300–1500 m in elevation on the west and P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 113

180000 190000 200000 Mucurca Lake 09-17

11-68 11-67 To Chivay 000 000 11-54 11-64 Colihuiri 8250 8250 10-14/16/17 To Huambo 11-71A 10-20 10-13

Qda. Sahuancaya 10-24 11-38 09-30 11-79 10-19 11-37 10-18 Cajamarcana Qda. Huaraya Yanajaja CollpaQda. Vizcachane 11-28 11-03 11-42 11-15 Moldepampa Sallalli Río Parcomayo Jatún Pampa

oyamarcahC.adQ Japo 11-11 000 000 11-20 Corinta 11-19 11-44 8240 8240

Qda. Huaycumayo 10-33 10 km

Qda. Baylillas To Taya 11-33 180000 190000 200000

Fig. 2. Digital elevation model of the Ampato–Sabancaya volcanic complex showing locations of rock samples (open circles) and dated rock samples and/or other studied sections (solid circles). east sides. Its lower part is characterised by gentle slopes (10–15°), free around 11–12 ka. On the basis of these data, we consider that the of glacial erosion. Two peaks compose the upper part of this edifice: A old moraines around the ASVC (M1 in Fig. 4) are associated with dome-like structure characterised by steep slopes (30–45°) and a the LGM period (i.e. 17–25 ka) whereas the younger moraines younger main cone with moderate slopes (20–30°), located to the (M2) are probably associated with a late glacier re-advance just northeast, which comprises an active crater (300–400 m in diameter) after the Holocene–Pleistocene boundary (10–13 ka). Other small with continuous fumarolic activity. moraines (M3) observed at higher altitudes (above 4800 m asl) at The morphology of the Ampato–Sabancaya volcanic complex was Ampato are clearly associated with younger Holocene fluctuations shaped by the late Pleistocene glaciations. With the exception of the (cf. Jomelli et al., 2011). north and north-eastern flanks, covered by younger Sabancaya lavas, medium-sized glacial valleys have been carved all around the edifice. 5. The eruptive chronology of the Ampato volcano Alcalá et al. (2011) identified at least two well-defined groups of moraines radially oriented around the volcano. The oldest and biggest Our geomorphologic, stratigraphic, and geochronological data show moraines, which are morphologically well-defined, bear witness to the that Ampato is a compound volcano comprising (Fig. 4, Table 4): (1) The Last Glacial Maximum (LGM) and extend down to 4250–4450 m asl. Basal edifice, which is an old, highly eroded volcano; (2) the Upper In contrast, the younger moraines reach higher elevations, in the edifice, which started with the Yanajaja stage and continued with the range of 4400–4650 m asl, and are interpreted as resulting from late successive construction of the Northern, Southern, and Central peaks. glacier re-advances. Based on cosmogenic 3He surface-exposure dating of Coropuna 5.1. The Ampato Basal edifice volcano moraine deposits, Bromley et al. (2009) propose that the LGM in this part of the Andean Cordillera was synchronous with 5.1.1. A former andesitic stage that at other latitudes (i.e. 17–25 ka). These authors also observe a The older remnants of the Ampato Basal edifice correspond to a 600- younger glacier re-advance in the range of 10–13 ka. These ages m-thick volcanic pile that crops out on the south-western and western agree with some 36Cl cosmogenic dates on samples from the flanks of Ampato. This sequence is composed of 40–60 m-thick lava Huayuray valley on the northern flank of the Hualca Hualca edifice flows interlayered with proximal scoria-flow deposits and spatter ag- that give ages of around 17–18 ka (Alcalá et al., 2011). These authors glutinates (Figs. 3, 5a). Volcanic units corresponding to the subsequent also propose that a younger re-advance occurred at Hualca Hualca volcanic stages discordantly cover this unit. Another remnant of this 114 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

σ stage consists of a 200–400-m-thick sequence of sub-horizontal and highly eroded lava ﬂows that crop out in the Jatún Pampa plain, at the south-eastern foot of the volcano (Figs. 2, 4). Lavas and tephras from Age (ka) ± 2 both units are dark-grey, aphanitic, olivine-pyroxene andesites (57.2– 40 40 59.8 wt.% SiO2). Two K– Ar ages (Table 1, Fig 3, 4), indicate that σ

12 these two units are almost contemporaneous. A lava sample from the − proximal unit is 440 ± 8 ka old (SA-11-64), and a second sample from the Jatún Pampa sequence is 410 ± 10 ka old (SA-11-03). Thus, the Ar* 10 Weighted mean 40 (mol./g) ± 1 oldest part of the Ampato Basal ediﬁce was possibly built up between

σ 400 and 450 ka. 12 −

5.1.2. Moldepampa stage Ar* 10

40 (mol./g) ± 1 This unit consists of a 200–300-m-thick sequence of lava ﬂows representing the main cone-building stage of the Ampato Basal ediﬁce

Ar* % (Figs. 2, 4, 5). To the south, lavas rest discordantly on the lavas of the 40 previous stage. Based on the radial distribution of these lavas, and its average extension (8–10 km) and slope (7–8°), we infer that the summit of the Ampato Basal ediﬁce was located at roughly the same location as the current summit, at an elevation of 5200–5400 m asl. – Mass molten (g) 0.51325 0.064 0.017 ± 0.017 1.99591 0.665 0.083 ± 0.015 1.05451 1.108 0.163 ± 0.026 1.18860 0.824 0.209 ± 0.008 1.27769 2.024 0.363 ± 0.019 1.56919 5.43 1.263 ± 0.011 1.01345 6.498 1.333 ± 0.010 1.033401.01938 8.182 12.503 1.275 ± 0.012 1.233 ± 0.010 1.03500 8.666 1.490 ± 0.018 0.54099 11.581These 2.425 ± 0.024 lavas are porphyritic dacites (62.9 65.1 wt.% SiO2), with plagioclase, amphibole, biotite, Fe–Ti oxides and minor pyroxenes. Thus, a

σ sharp mineralogical and geochemical contrast (see below) exists between these lavas and those of the previous andesitic stage. Two samples from the south-western side of the cone yielded coherent

Split K (wt.%) ± 2 ages of 217 ± 5 ka and 226 ± 4 ka, which correspond to the upper part of this lava sequence (SA-11-37 and SA-11-38, respectively, Table 1). A third sample, from the southern ﬂank yields a similar age (210 ± 3 ka, SA-11-15, Table 1). On the basis of these dates we (m asl) constrained this cone-building stage to 230–200 ka.

5.1.3. Rhyolitic fallout deposits The eroded plateau located to the west of ASVC is partially covered by thick rhyolitic tephra fallout deposits. At 10 km to the southwest of the summit, in the Chacramayo valley, crops out a 4–5-m-thick sequence of tephra fallout deposits. This sequence consists of four pumice and lithic lapilli horizons of metre-sized thickness, interlayered with altered and indurated ash layers. Pumice clasts reaching a maximum size of 3–4cm(at10–12 km from the Ampato summit) are fibrous ank 189,043 8,244,402 4682 2.806 ± 0.028 0.42555 1.095 0.386 ± 0.021 0.373 ± 0.014 77 ± 4 ank 187,567 8,246,350 4710 3.512 ± 0.035 0.54128 6.336 1.310 ± 0.010 1.327 ± 0.009 217 ± 5 ank 188,471 8,247,137 4854 3.188 ± 0.032 1.00949 5.136 1.241 ± 0.018 1.248 ± 0.007 226 ± 4 ank 196,910 8,255,870 5010 2.457 ± 0.025 0.47434 0.017 0.007 ± 0.016 0.012 ± 0.011 3 ± 5 ank 192,823 8,248,561 5680 2.789 ± 0.028 0.39318 0.556 0.168 ± 0.031 0.165 ± 0.019 34 ± 8 ank 194,252 8,244,656 4542 2.067 ± 0.021 1.03014 2.312 1.446 ± 0.020 1.471 ± 0.013 410 ± 10 fl fl fl –

fl and rhyolitic in composition (74.2 76.9 wt.% SiO ) and contain ank 193,676 8,249,484 5760 2.632 ± 0.026 0.41030 0.323 0.053 ± 0.034 0.078 ± 0.014 17 ± 6 ank 193,200 8,247,108 5155 3.055 ± 0.031 0.51211 1.159 0.214 ± 0.012 0.211 ± 0.007 40 ± 3 ank 192,405 8,243,491 4589 3.512 ± 0.035 1.51905 9.803 1.296 ± 0.010 1.281 ± 0.007 210 ± 3 2 fl fl fl fl fl plagioclase and biotite. We also observed accidental andesitic lithics (≤2 cm). This sequence is covered by 20–25 m of dm-sized layers of ow, NE ow, E ow, SE ow, S ow, SW ow, S ow, SW ow, SW ow, SE fl fl fl fl fl fl fl fl fl andesitic gravel and sand. At a distal location (20–25 km west- southwest of the summit), this sequence is thinner and the pumice smaller (≤2–2.5 cm). Here, the fallout deposits are interlayered with layers of reworked ash. Unfortunately we could not obtain an age for this unit because pumice clasts are altered. However, these distal deposits are probably not related to this volcanic complex, because ce Lava

fi these deposits have no proximal counterparts around Ampato. In addition, the lack of amphibole and the different geochemical signature of these samples do not correlate with the Ampato magma series (see below). Sabancaya volcanic complex. – ce Volcanic stage Unit and location UTM Easting UTM Northing Altitude 5.2. The Ampato Upper edifice fi Edi 5.2.1. The Yanajaja stage This unit consists of a sequence of lava flows reaching 200–300 m in thickness that crops out on the south and south-western flanks of Ampato (Figs. 2, 4). These lavas rest on top of the remnants of Experiment number 8464 8545 8546 8446 8729 8830 8420 8721 8722 8593 8707 the Basal edifice (i.e. Moldepampa stage) defining a marked erosional discontinuity (Fig. 5c). The lava sequence includes porphyritic andesites 28 8720 Upper Ampato Yanajaja Lava – and dacites (61.8–63.5 wt.% SiO2) of similar mineralogy to the Ar ages for rocks from Ampato 40

– Moldepampa stage. A sample from this unit (SA-11-28) was dated at SA-09-17 8448 Sabancaya Basal edi Sample number SA-11-54 8521 Upper Ampato Central Cone Lava SA-11-71A 8522 Upper Ampato South cone Lava SA-09-30 8430 Upper Ampato South cone Lava SA-11 SA-11-15 8801 Basal Ampato Moldepampa Lava SA-11-37 8411 Basal Ampato Moldepampa Lava SA-11-38 8706 Basal Ampato Moldepampa Lava SA-11-03 8561 Basal Ampato Former andesitic stage Lava SA-11-64 8692 Basal Ampato Former andesitic stage 189,617 8,249,025 5392 3.221 ± 0.032 0.51816 4.371 2.506 ± 0.030 2.456 ± 0.019 440 ± 8 K

Table 1 40 77 ± 4 ka. P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 115

Table 2 New 14C data for Ampato–Sabancaya volcanic complex. We also include additional radiocarbon data from the literature.

14 13 ⁎ Lab code Sample no. Locality UTM UTM Type of C age (aBP) δ C Calendar age range Relative area Lab Reference Easting Northing sample (o/oo) (2σ)§ (%)

GrA 57971 SA-11-68C Qda. Huaraya 186,942 8,251,723 Peat 85 ± 35 −26.39 1697–1725 cal AD 12 CIO This study 1807–1870 cal AD 35 ⁎ 1875–1954 cal AD 53 SacA 27951 SA-11-68C (bis) Qda. Huaraya 186,942 8,251,723 Peat 265 ± 30 −25.20 1627–1680 cal AD 52 LMC14 This study 1731–1801 cal AD 47 GrA 50536 SA-10-13D Colihuiri 201,497 8,248,928 Peat 730 ± 35 −25.23 1269–1326 cal AD 58 CIO This study 1341–1390 cal AD 42 SacA 27943 SA-10-13B Colihuiri 201,497 8,248,928 Peat 810 ± 30 −24.40 1218–1286 cal AD 100 LMC14 This study GrA 57885 SA-11-79C Sallalli (III) 202,341 8,246,831 Peat 2925 ± 45 – 2876–3159 cal BP 100 CIO this study SacA 27952 SA-11-79C (bis) Sallalli (III) 202,341 8,246,831 Peat 3080 ± 30 −25.80 3144–3358 cal BP 99 LMC14 This study GrA 50535 SA-10-13F Colihuiri 201,497 8,248,929 Peat 3105 ± 40 −23.80 3156–3380 cal BP 100 CIO This study GrA 56327 SA-10-19C Sallalli (II) 201,980 8,246,726 Peat 3815 ± 35 −26.50 3986–4051 cal BP 13 CIO This study 4062–4258 cal BP 86 GrA 50533 SA-10-19F Sallalli (II) 201,980 8,246,726 Peat 4150 ± 40 −24.53 4450–4464 cal BP 2 CIO This study 4518–4821 cal BP 98 SacA 27946 SA-10-19C Sallalli (II) 201,980 8,246,726 Peat 5050 ± 30 −26.30 5651–5773 cal BP 63 LMC14 This study 5779–5795 cal BP 2 5804–5892 cal BP 35 SacA 27947 SA-10-19D Sallalli (II) 201,980 8,246,726 Peat 5830 ± 35 −22.20 6477–6676 cal BP 100 LMC14 This study SacA 27948 SA-11-20A Japo 194,357 8,240,560 Soil 9480 ± 40 −26.90 10,553–10,786 cal BP 98 LMC14 This study 11,036–11,057 cal BP 1 SacA 27944 SA-10-18C Sallalli (I) 202,016 8,246,444 Soil 9705 ± 35 −26.60 10,792–10,964 cal BP 38 LMC14 This study 11,004–11,024 cal BP 2 11,065–11,201 cal BP 60 SacA 27945 SA-10-18G Sallalli (I) 202,016 8,246,444 Peat 11,165 ± 45 −24.90 12,831–13,089 cal BP 100 LMC14 This study Beta-126965 – Sallalli –– Peat 300 ± 50 – 1485–1678 cal AD 83 Juvigné et al. (2008) 1733–1800 cal AD 17 Lv-2184 – Sallalli (S2) –– Peat 2050 ± 70 – 1813–2154 cal BP 99 Juvigné et al. (2008) GrN-25586 – Sallalli (S2) –– Peat 2370 ± 90 – 2150–2710 cal BP 100 Juvigné et al. (2008) Hv-24660 – Sallalli (S3) –– Peat 1790 ± 110 – 1409–1919 cal BP 100 Juvigné et al. (2008) Hv-24662 – Sallalli (S3) –– Peat 2955 ± 80 – 2849–3258 cal BP 97 Juvigné et al. (2008) 3290–3333 cal BP 3 Hv-24661 – Sallalli (S3) –– Peat 4500 ± 125 – 4823–5469 cal BP 100 Juvigné et al. (2008) –– Qda. Huaraya –– Peat 5440 ± 40 – 6014–6081 cal BP 12 Gerbe and Thouret (2004) 6104–6158 cal BP 14 6171–6292 cal BP 74 Lv-2107 – Sallalli (A) –– Peat 8520 ± 80 – 9281–9597 cal BP 100 Juvigné et al. (1998) Lv-2110 – Sallalli (A) –– Peat 9650 ± 170 – 10,420–11,358 cal BP 100 Juvigné et al. (1998)

§ Calendar ages were obtained using the updated SHcal13 calibration curve (Hogg et al., 2013) except for SA-11-68C (GrA57971) for which we used the former SHcal04 calibration curve (McCormac et al., 2004). ⁎ CIO, Centre for Isotope Research, Groningen University (Netherlands).LMC14, Laboratoire de Mesure du Carbone 14, Gif-sur-Yvette (France).

5.2.2. North cone maximum thickness of 800–1000 m. We identified several effusive, This small cone is constructed on the northern remnants of the Basal dome-forming phases as well as explosive activity, represented by edifice. It has an almost elliptical basal outline (1.5 × 3 km) located at frequent, albeit highly eroded, pyroclastic sequences. The presence 5200–5400 m asl, steep slopes (30–40°), and consists of viscous, of angular unconformities between lava flows, allows to reconstruct andesitic (61–62 wt.% SiO2) lava flows and breccias with plagioclase, the structure of this cone. The western flank of Ampato, around amphibole, pyroxene, biotite, and Fe–Ti oxides. Co-magmatic, cm- 5600–6000 m asl, is marked by an unconformity separating sized aphanitic enclaves of andesitic composition (57–58 wt.% SiO2) the older volcanic sequences of the Basal edifice from a 150–200- are common. Given the degree of erosion of these lavas, we consider m-thick subhorizontal sequence of lavas. This sequence corresponds this cone to be the oldest of the three peaks comprising the Ampato to the infilling of a depression, probably associated with major Upper edifice (Figs. 3, 5). explosive activity of the Basal edifice (Fig. 5).

5.2.3. South cone 5.2.3.1. First cone-building stage. Lavas overlying the previously This ediﬁce is constructed on the remnants of the Yanajaja mentioned subhorizontal lavas form the bulk of the Southern cone. stage and the Basal ediﬁce. It includes several volcanic units with a They originate from the Ampato summit and descend to 5100 m asl

Table 3 Cosmogenic 3He data and exposure ages.

Sample Altitude Latitude Longitude Sampling Sampling thickness Topographic Scaling Mass 4He 3He Li Exposure ⁎ 6 (m) (°S) (°W) thickness (cm) correction shielding factor (mg) (at/g) (10 at/g) (ppm) age (ka)

SA1 4905 15.8222 71.827667 3 0.954 0.997 11.89 43.2 b.d.l 9.37 ± 0.47 9.5 6.30 ± 0.31 SA4 4754 15.823417 71.820067 4 0.94 0.995 11.15 42 b.d.l 9.11 ± 0.44 15.1 6.65 ± 0.32 SA5 4740 15.82505 71.817383 3.5 0.947 0.999 11.05 31.6 b.d.l 17.0 ± 0.76 14.1 12.34 ± 0.55 b.d.l.: below detection limit, indistinguishable with the 4He blank at 1 sigma of uncertainty. ⁎ Time-dependent scaling of Stone (2000) considering a speciﬁc Andes atmosphere model (Farber et al., 2005). Calculated using CosmoCalc (Vermeesch, 2007). 116 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

AMPATO SABANCAYA

Central Southern cone Terminal cone cone Northern & active crater cone

Sallalli peatbog

Fig. 3. Panoramic view of the Ampato–Sabancaya volcanic complex. View from the SE showing the Ampato and Sabancaya edifices. Note the discordances between the Northern, Southern and Central cones. Sallalli peatbog is in the foreground. on the southern flank, whereas they reach down to only 6000 m asl on 5.2.3.2. Block-and-ash flow deposits. A N20-m-thick sequence crops out the western flank. Two lavas from the south and south-east flanks of the 5–6 km from the present summit to the east and west of the volcano cone yielded ages of 40 ± 3 and 34 ± 8 ka (SA-09-30 and SA-11-71A re- (Fig. 4), and lies on lava flows from the Basal edifice. These deposits spectively, Table 1). are massive, indurated, and matrix-supported, with 20–40 vol.% of

000 000 000 000 000 000 184 188 192 196 200 204 Younger moraines (M3) Intermediate moraines (M2) M1 Older moraines (M1)

GLACIAL Undifferentiated fluvio-glacial HH DEPOSITS sediments (FGS) M1 HH M1 vent 000 000 3±5 ka SA-2 crater 8256 8256 caldera M1 M1 SA-4 HH Terminal cone SA-4 SA-2 Lavas and pyroclastics (SA-4)

SA-2 Basal edifice 000 000 SA-1 SA-1 Satellite vent lavas (SA-3) 5440±40 (SA) SA-4 Upper lavas (SA-2) aBP AMII-2 8252 8252

SA-2 SABANCAYA M2 Lower lavas (SA-1) AMI-1 D M1 HH AMII-6 SA-3 Central cone C1 17±6 ka SA-1 6.65±0.32 D Young dome (AMII-9) FGS M2 440±8 ka AMI-1 ka M3 6.30±0.31 M1 C2 12.34±0.55 Central cone lavas (AMII-8) 34±8 ka ka 000 000 M3 ka South cone AMI-1 M2 SA-3 Pyroclastic deposits (AMII-7) M2 M1 226±4 ka 8248 AMII-3 40±3 ka FGS 8248 Thick lavas (AMII-6) M2 M3 2925±45 aBP

217±5 ka (AMII) Block-and-ash sequences (AMII-5) AMII-1 FGS AMII-1 Sub-horizontal lavas (AMII-4) AMI-3 UPPER AMPATO UPPER AMI-2 South cone lavas (AMII-3) FGS M2 77±4 ka M2 000 410±10 ka 000 North cone AMI-3 M2 Northern cone lavas (AMII-2) M1 AMI-3 210±3 ka Yanajaja stage 8244 8244 M1 FGS Lower cone lavas (AMII-1)

AMI-2 AMI-3 Moldepampa stage FGS Lower cone lavas (AMI-3) FGS Former andesitic cone 000 000 Distal lavas (AMI-2) (AMI) Proximal lavas and 8240 8240 pyroclastics (AMI-1) BASAL AMPATO BASAL FGS 10 km Older formations Hualca Hualca lavas (HH) Undifferentiated volcanics 184000 188000 192000 196000 200000 204000

Fig. 4. Geological map of the Ampato–Sabancaya volcanic complex. P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 117

Table 4 Generalized chronostratigraphy showing the main eruptive stage at Ampato–Sabancaya volcanic complex.

Ediﬁce Volcanic stage Units Structure Age Magma composition

Sabancaya Terminal cone Lavas and pyroclastics (SA-4) Explosive activity ~3 ka Amphibole-biotite, silicic andesites and

(SA) Satellite vent Lavas (SA-2) Lava flow emission - Cone building 3–10 ka dacites (60–66 wt. SiO2) Basal edifice Lavas (SA-1,3) - No glacial erosion Upper Ampato Central cone Young dome (AMII-10) and Central Lava flows and dome growth 10–20 ka Amphibole-biotite, silicic andesites and

(AM-II) cone lavas (AMII-9) dacites (61–66 wt. SiO2) South cone Upper lavas (AMII-6, 7, 8) Mainly lavas 20–40 ka Amphibole-biotite, siliceous andesites and

Block-and-ash ﬂow deposits (AMII-5) Dome growth dacites (61–67 wt. SiO2)

Corinta pyroclastic deposits Explosive activity - caldera Amphibole-biotite dacites (65–69 wt. SiO2)

(AMII-4) - Scoriaceous tephra formation (C2) and andesitic scoria (57–59 wt. SiO2)

Lower lavas (AMII-3) Cone-building Amphibole-biotite dacites (63–64 wt. SiO2) North cone Lavas (AMII-2) Cone-building ~50 ka (?) Amphibole-biotite silicic andesites (61–62

wt. SiO2) Yanajaja Thick lava ﬂows (AMII-1) Cone-building 70–80 ka Amphibole-biotite, siliceous andesites and

dacites (62–64 wt. SiO2)

Basal Ampato Terminal Rhyolitic tephra (AMI-4) Explosive activity b200 ka (?) Biotite-bearing rhyolites (74–77 wt. SiO2) (AM-I) explosive activity (?)

Moldepampa Thick, viscous lava ﬂows (AMI-3) Two main cone-building stages 200–230 ka Amphibole-rich dacites (63–65 wt. SiO2) separated by quiescence period

Former andesitic Proximal lavas and pyroclastics 400–450 ka Two-pyroxenes andesites (57–60 wt. SiO2) cone (AMI-1) and distal lavas (AMI-2)

subangular blocks (Fig. 6a-d). The matrix is composed of pink Petrographic and chemical characteristics of the juvenile material ash enriched in lithic material and free crystals. Two main lithologies link these scoria fallout deposits to the scoria flow deposits that crop are present: The dominant group consists of dark-grey porphyritic out at the south and south-western foot of Ampato. For instance, at amphibole-rich andesites, whereas the second group consists of Japo, we found at least four scoria-flow deposits, each up to 3 m in thick- porphyritic amphibole-biotite dacitic blocks. This pyroclastic sequence ness (Figs. 7d, 8). These deposits are matrix-supported (around 30 vol.% testifies to a large dome-forming stage during the Ampato Upper edifice bombs), with 10–40-cm-sized andesitic blocks and bombs (59–60 wt.% development. SiO2) and a reddish-grey consolidated matrix. These andesitic deposits The above sequence is overlain by thick lava flows related to the testify to an intense explosive activity associated with this edifice. upper part of the Southern cone. On the eastern flank, these lavas Because these deposits are highly eroded, we infer that they pre-date consist of a 150–200-m-thick sequence of at least three to four the LGM period roughly dated at 17–25 ka (Bromley et al., 2009; heterogeneous lava flows, with individual thicknesses of 30–50 m, and Alcalá et al., 2011). that are characterised by the ubiquitous presence of two end- members: A more abundant porphyritic, dark grey andesite (62.9– 5.2.3.4. Corinta pumice fallout and flow deposits. Athickpumice-and-

63.6 wt.% SiO2), and a subordinate, highly porphyritic, light grey dacite lithics fallout deposit crops out in the south-western sector of (66.9–67.3 wt.% SiO2). These two contrasting compositions, similar to Ampato, from the Collpa and Quebrada Vizcachane in the north to the those observed in the previous block-and-ash flow deposits, are distrib- Quebrada Baylillas in the south. In the Corinta area, 10–12 km from uted as both dm-sized bands and magmatic inclusions. The amount of the present volcano's summit (Figs. 7b, 8), the tephra deposit is 3–4- these inclusions is variable, but it increases towards the base of the m-thick and is composed of two layers separated by a distinctive lava sequence until it resembles a breccia in the lower part of the lithic-rich coarse ash horizon 3–4 cm in thickness. In this section, the sequence. On the western flank, a conspicuous lava sequence with a lower layer displays a minimum thickness of 2.5 m whereas the upper similar stratigraphic position and degree of erosion is present. There, layer is 0.8–1.0-m-thick. Pumice fragments, reaching a maximum size the andesitic (61–63 wt.% SiO2)lavaflows reach 4800 m asl at 5 km of 7–8 cm, are white, dacitic in composition (65–68 wt.% SiO2)with from the present summit and have a thickness of 40–60 m. These plagioclase, amphibole, biotite, and minor of pyroxene. These deposits lavas spilled into an ancient glacial valley and overlie the older and are covered by a 1–1.5 m ash-rich horizon with disseminated pumice larger moraines, but they also show marked relief inversion that lapilli. Above an erosional contact, we observed a 1–2m-thick,roughly suggests intense glacial erosion. Based on these field constraints, we stratified, matrix-supported deposit, bearing subrounded lithic and infer that this lava sequence was contemporaneous with the LGM (i.e. pumice fragments (b20 cm in diameter) in an indurated ash-rich 17–25 ka). matrix. This deposit is interpreted as being associated with fluvio- glacial activity. This suggests a major explosive eruption prior to or 5.2.3.3. Baylillas scoria fallout and flow deposits. A5–10-m-thick sequence during the LGM period (i.e. 17–25 ka). of tephra fallout deposits crops out in the southern part of Ampato, near Encircling the south and south-western base of Ampato we the Quebrada Baylillas, at 10–12 km from the summit (Figs. 7a, 8). identified the vestiges of a thick, reddish-orange pumice flow deposit. Across this section, at least 6 scoria-lapilli fallout deposits are In the Collpa and Quebrada Vizcachane zone, this deposit displays a interlayered with ash-rich horizons. Their thickness ranges from ~10– roughly metric stratification, with an overall thickness of N10 m. Other 20 cm, although a 50-cm-thick scoria fallout layer marks the middle outcrops occur along the Quebrada Chacramayo (south-western flank) part of the sequence (Figs. 7a, 8). The 80-cm-thick pumice fallout depos- and around the Japo zone (south flank). In this latter location, this unit it at the top is described below. This sequence crops out southwards, rests subhorizontally on top of the pre-Ampato substratum and is along the road towards Taya village. Despite of the high magnitude of covered by a sequence of younger scoria-flow deposits (Fig. 8). Other the eruptions that originated these deposits (up to 50-cm-thick at 10– outcrops of this unit are scarce, probably because they have been eroded 15 km from the vent), outcrops are scarce and it is not possible to by LGM glaciers. It is composed of pumice fragments (up to 50 cm in trace their extension and dispersion axes, because these layers were diameter) in an ash-rich unconsolidated matrix. The deposit also eroded probably during the Pleistocene glaciations. includes lithics and scoria fragments that suggest magma mixing 118 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

Southern Cone

Basal 440±8 ka Edifice 226±4 ka

Southern Central Cone Cone Northern Cone 34±8 ka 17±6 ka

40±3 ka

Southern Cone

Upper Edifice

40±3 ka Basal edifice

77±4 ka Yanajaja 210±3 ka Basal Edifice

Fig. 5. Panoramic views of the Ampato–Sabancaya volcanic complex, showing some of the main structures. (a) View from the W showing the former andesitic cone (Ampato Basal Edifice), the Ampato Upper Edifice (subhorizontal sequence and the upper part of the South cone). (b) View from the SE showing the discordance between the South and Central cones of the Ampato Upper Edifice. (c) View from the S showing the Ampato Basal Edifice with its former andesitic cone and the Upper edifice. Note the thick package of Pleistocene moraines in the foreground. processes. The juvenile pumice fragments are fibrous and dacitic in shows evidence of glacial erosion, although to a lesser extent than the composition (67.1–69.3 wt.% SiO2), with plagioclase, amphibole, and others. A lava sample from this unit was dated at 17 ± 6 ka (SA-11- scarce biotite. These characteristics suggest a genetic relation with the 54, Table 1). We also included a young dacitic lava dome (63–64 wt.%

Corinta plinian fallout deposit. This explosive phase probably developed SiO2) in this eruptive stage, which extruded on top of the Northern a crater-like structure (C2 in Fig. 4) marking the transition between the cone, between 5700 and 6000 m asl. This structure has an oblate Southern and Central cones. morphology and shows no evidence of glacial erosion. This lava dome represents the youngest volcanic structure of the Ampato compound 5.2.4. Central cone edifice. This edifice was discordantly constructed over the remnants of the Northern and Southern cones of the Ampato Upper edifice (Figs. 3, 5). It is composed of a sequence of andesitic and dacitic (61.3– 6. The eruptive chronology of the Sabancaya volcano

66.0 wt.% SiO2)lavaflows and interlayered breccias (Fig. 4), with a maximum thickness of 400–600 m. These lavas fill a depression Geomorphologic, stratigraphic, and geochronological data point between the Southern and Northern peaks, forming a steep (25–30°) toward a two-stage development of Sabancaya edifice (Fig. 4, cone that extends down to 5300 m on the eastern flank. This cone Table 4): (1) A Basal edifice, and (2) a Young terminal cone. P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 119

a b

c d

Fig. 6. (a) Sequence of indurated block-and-ash flow deposits associated with the Ampato Upper Edifice (site SA-10-17). (b) Block of the deposit in (a). (c) Consolidated block-and-ash flow deposit (SA-10-14). (d) Fragments of blocks with mixing/mingling textures (SA-10-16).

6.1. The Basal edifice as the Holocene–Pleistocene boundary. The ages at 6–7 ka are in agreement with a peat sample that underlies a distal lava flow (unit SA-1, Fig. This stage is represented by a sequence of blocky lava flows that 4) located at 7–8 km to the west of Sabancaya's crater that has been spilled out onto the older rocks of the Ampato and Hualca Hualca radiocarbon dated at 5440 ± 40 aBP (Gerbe and Thouret, 2004). In edifices. These flows extend 6–8 km from the vent and have individual addition, a new 40K–40Ar age of a lava sample from the SA-3 unit yielded thicknesses of 40–80 m, forming a 300–400 m thick lava pile (Fig. 4). On a younger age of 3 ± 5 ka (SA-09-17, Table 1). Based on these the basis of satellite images, Bulmer et al. (1999) identified more than constraints and the lack of glacial erosion, we propose that the 42 distinct lava flows at Sabancaya. Here, using geomorphological data Sabancaya Basal edifice has a lower to middle Holocene age (10– and stratigraphic information, we group these lava flows into two 6 Ma) and that units SA-1 to SA-2 are almost synchronous. successive units (SA-1 and SA-2, Fig. 4). All lava flows show an uneven surface morphology, often with flow structures and levees, and no 6.2. The Young cone evidence of glacial erosion. A lava block, corresponding to SA-1 unit, was sampled for 3He surface-exposure dating and yielded an age of This cone is represented by a sequence of lava flows that lie discor- 6.30 ± 0.31 ka. We include a 400-m-high lava dome in this unit, located dantly on the lavas of the previous stage. This unit also includes a in the south-west part of Sabancaya, which is considered to be older young cone, covered by pyroclastic material, with an active summit than most of the lava flows. These lavas are porphyritic andesites and crater (unit SA-4, Fig. 4). These lava flows reach 4–5 km from the vent, dacites (60.6–65.6 wt.% SiO2) with plagioclase, hornblende, biotite, have a thickness of 40–60 m and consist of porphyritic andesites and and minor ortho- and clinopyroxene and Fe–Ti oxides. dacites (61.7–65.7 wt.% SiO2) with a mineral assemblage similar to the Associated with this unit, we identified a thick lava sequence Basal Sabancaya lavas. that crops out in the Sallalli plain (unit SA-2, Fig. 4). Following Bulmer et al. (1999), we correlate this sequence to a satellite vent 6.3. Holocene tephra deposits located 4 km to the southwest of the current active crater (Fig. 4). This sequence comprises at least 3 subhorizontal, andesitic (60.9–62.5 wt.% In order to better constrain Sabancaya's Holocene eruptive activity

SiO2)lavaflows, with their respective lobes overlying the sedimentary as first established by Juvigné et al. (1998, 2008), we carried out three deposits and the Sallalli peatbog. Two lava blocks were sampled for new excavations in the Sallalli peatbog, and further excavations in 3He surface-exposure dating and yielded ages of 6.65 ± 0.32 ka and other peatbogs on the south-eastern (Colihuiri), southern (Japo), and 12.34 ± 0.55 ka. These age determinations are consistent, and suggest western (Quebrada Huaraya) flanks. We also obtained 14 additional that Sabancaya Basal edifice started its construction probably as early 14C ages on peat and paleosoil samples. The oldest ages come from the 120 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

a mixed with the recent peat layer), we found a 4–8-cm-thick coarse layer that corresponds to the AD 1987–1998 eruption.

7. Main petrological characteristics

The Ampato–Sabancaya samples display a high-K magmatic trend,

ranging from andesites to dacites (57 to 69 wt.% SiO2, Fig. 10, Table 5), with rare rhyolitic compositions (74–77 wt.% SiO2). Four different groups were identiﬁed:

(1) The first group, composed mainly of andesites (57–60 wt.% SiO2), corresponds primarily to lavas from the Former andesitic stage (Ampato Basal edifice), as well as to the scoriaceous tephra fallout and pyroclastic flow deposits of Ampato Upper edifice. These samples are porphyritic lavas and tephras with plagioclase, ortho- and clinopyroxene, and Fe–Ti oxides. (2) The second group corresponds to andesitic and dacitic composi-

tions (60–67 wt.% SiO2) from the Moldepamba stage, the Ampato b Upper edifice and the Sabancaya edifice. No significant differences have been observed between these successive volcanoes. All these samples are porphyritic with phenocryst of plagioclase, amphibole, biotite, ortho- and clinopyroxene, and Fe-Ti oxides. (3) The third group corresponds to dacitic compositions (65–69 wt.%

SiO2) associated with the Corinta plinian fallout and pyroclastic ﬂow deposits of the Ampato Upper ediﬁce. These samples display a similar mineral composition to the second group. (4) Lastly, the fourth group displays rhyolitic compositions (74–

77 wt.% SiO2) and corresponds to tephra fallout deposits croppingoutinthewesternplateau.

In general the major oxides (excepting K2O, Fig. 10) show a negative correlation with silica content, although some scattering is observed for

Na2O (not shown). Light Ion Lithophile Elements (LILE, e.g. Rb, Ba, Th), ﬁ Fig. 7. Pyroclastic deposits of the Ampato Upper edi ce. (a) Scoriaceous tephra fallout Light Rare Earth Elements (LREE, e.g. La, Ce), and some High Field deposits at Quebrada Baylillas (SA-10-33). (b) Corinta plinian fallout deposit (SA-11-12). Strength Elements (HFSE, e.g. Nb) show broad positive correlations (Fig. 10). In contrast, Sr and the transition elements (e.g. Cr, Ni, V) are inversely correlated with silica. Lastly, Zr, Y and Heavy Rare Earth Elements (HREE, e.g. Dy, Yb) display contrasting behaviour with silica increase. Samples from groups 1 and 2 have an inverse correlation Sallalli - Section A (Juvigné et al., 1998) and Sallalli I and Japo sections with silica, while group 3 dacites display higher HREE concentrations,

(this work). These sections are characterised by peat and/or soil as well as elevated values of Al2O3,K2O, and LILE (e.g. Rb, Th). In horizons interlayered with reworked gravel and sand layers (Fig. 9). In addition, this latter group has lower concentrations of CaO, MgO, Sr, these sections, we found medium-to-coarse ash layers ranging from 5 and transition elements. Lastly, the rhyolites of group 4 plot as an to 30 cm in thickness. On the basis of granulometry and microscope extension of the trend defined by groups 1 and 2, except for LREE and examination, Juvigné et al. (1998) concluded that some of these MREE. deposits represent pristine ash fallout layers. Thus, we consider that Trace element variations clearly indicate mineral fractionation, for there are at least 4–5 ash fallout layers in the period between 11 and instance a decrease in Ni and Cr with increasing silica suggests olivine 8 ka. Younger ages (b4.5 ka) have been obtained for the Sallalli S2 and/or clinopyroxene fractionation, whereas Sr and Eu decrease and (not shown) and S3 (Juvigné et al., 1998), as well as the Sallalli II and depletions in MREE and HREE indicate plagioclase and amphibole III, Colihuiri, and Quebrada Huaraya I and II sections (Fig. 9). The period crystallisation. Trace elements ratios plotted against a differentiation between 4.5 and 3 ka shows two well-preserved ash layers, whereas index place additional constraints on the differentiation process. In a stratigraphic correlations are difficult for the period younger than Rb versus Rb/Sr diagram (Fig. 11), groups 1 and 2 samples define a 3 ka, and reveal thick layers that represent either sequences of ash single linear trend (the ASVC main trend), probably controlled by fallout deposits or reworking of individual layers. At least 6 ash fallout fractional crystallization processes. In contrast, samples from group 3 deposits were observed in the Sallalli II section. In the upper part of show a different slope, and those of group 4 have a completely different the sequence, we found a 5–10-cm-thick, biotite-bearing, fine yellow trend. These differences suggest a prominent role of assimilation of ash layer that represents a key marker horizon. We interpret this layer the local upper crust (represented by the Charcani gneiss; Rivera, as being associated with the 1600 AD Huaynaputina eruption 2010) in the petrogenesis of these magmas. On the basis of these trace (Thouret et al., 1999). Above this layer, and only at the Quebrada element constraints, we tested some first-order geochemical models Huaraya section, we found a 10–30-cm-thick coarse ash layer, whose based on major and trace element concentrations. Major element underlying peat has been dated at 265 ± 30 aBP (Fig. 9). This age is in mass-balance calculations were carried out between mafic and siliceous agreement with the 300 ± 50 aBP date obtained by Juvigné et al. end-members in order to estimate the modal compositions of the (2008) and confirms the occurrence of explosive eruptions during the fractionating cumulate. These results were compared with the observed 17th or 18th century (Travada y Córdova, 1752; Zamácola y Jaúregui, mineralogy and subsequently used in trace element modelling, 1888). Finally, in the upper part of most sections (and sometimes using fractional crystallisation and assimilation models. As a result, a P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 121

Qda. Baylillas Route to Taya Taya-Sallalli road 0190776 - 8237067 0185941 - 8233726 0195619 - 8240453 [cm] [cm] [m] >200 [m] 50 0 >3

1 0.4 10 40 >1 2 30 0.5 0.3 3 0.6 12 [cm] 40-120 14 4 0 10 SA- 5 3 20 15 11-33A Corinta Collpa 0184069 - 8240343 0184842 - 8245165 40 30 50 [m] [m] 1 SA- 5 >1 60 10-33A 2 13 13 >3-5 2 80 1-2 SA- 39 10-33B 100 0.8-1.5 1.5 SA- SA- 6 0.8 1.3 6 11-39A 11-45 110 30-40 [cm] >2.5 2.7 SA- SA- 12 15 11-39B 8 10-33C 15 5 SA- 10-30 10 10-31 >1.5 SA- 10 11-39C 10 80 40 20 40 SA- 1.7 10-30 10 SA- 15 50 10-33D 10-20 SA- 5 50 >2 11-33B 25 SA- >20 12 10-33E pumice and lithics 11 fine ash 6 lapilli fallout 20 medium ash scoria lapilli fallout

pumice flow 20 coarse ash 15 reworked material scoria flow

Fig. 8. Stratigraphic sections of the tephra fallout deposits associated with the Ampato Upper Edifice. UTM Easting and Northing are included below the section's name. fractional crystallisation model (FC in Fig. 11) can be used to explain the interpretation of the diverse Ampato–Sabancaya magmas. A detailed evolution of the main ASVC trend by fractionation of a cumulate reconstruction of the magmatic processes in operation during the life composed by pl + amph + bio + cpx + mag ± ol. However, this of this magmatic system is beyond the scope of this work, and will be model does not explain the composition of the silica-rich magmas developed elsewhere. (namely the dacites of group 3). In contrast, an assimilation-fractional crystallisation model (AFC1 and 2 in Fig. 11) can account for the 8. Edifice volumes and eruptive rates chemical composition of these silicic magmas. Thus, the differences in magma chemistry of the ASVC reflect Using a 40-m digital elevation model (obtained from 1:50,000 complex crustal process involving fractional crystallisation of an topographic maps), we obtained the volcano's morphometric parame- amphibole-rich cumulate, together with variable amounts of upper ters (volcano basal area, height, and volume), following the methodolo- crustal assimilation. In addition, the frequent banding textures observed gy of Grosse et al. (2014). This consists of computing the volcano basal at Ampato Upper edifice rocks as well as in the 1990–1998 eruptive area (the edifice outline) and then fitting a 3D volcano basal surface products of Sabancaya (Gerbe and Thouret, 2004) clearly indicate that corresponding to the substratum topography. Using this surface, it is magma mixing and mingling must be taken into consideration in the possible to compute the volcano's maximum height and volume. Here, Q. Huaraya II Q. Huaraya I Colihuiri Sallali II Sallalli III Sallali (S3) Sallali I Sallali (Section SA-11-68 SA-11-67 SA-10-13 SA-10-19 SA-11-79 Juvigne et al. (2008) SA-10-18 0186942 - 8251723 0187021 - 8251578 0201497 - 8248928 0201980 - 8246726 0202341 - 8246831 202016 - 8246444 [cm] [cm] [cm] [cm] [cm] [cm] [cm] 2 5 5 5 4-6 12 4-6 2-4 20 Late 20 14 2 13 Holocene 3 24 1-3 6-7 1987- activity 3 5 HP? 4 2 300 3-4 1998 AD 10 3-5 7 2 6 3-4 5 2 [cm] 3 2 5 SA-10-13D 5-6 0 50 25 730 ± 35 15 20 SA-10-18C 10 15 aBP 14 2 9705 ± 35 5 2-3 1-2 2 aBP 20 13 HP? 7 3-4 1-2 3 SA-11-79C 14 20 1790 ± 110 30 13 2925 ± 45 15 aBP 32 18 aBP 10 6 40 8-9 3080 ± 30 4 5 aBP 6 30 4 3 7 3-4 50 1 5-8 6 10 SA-10-19C 8 5 15 2 3815 ± 35 25-26 40 SA-11-68C 5 3 aBP 8 3 265 ± 30 5 15 aBP 6-8 30 2955 ± 80 12 22 aBP > 30 40 5 30 46 5 100 5-6 SA-10-13F 15 8 3105 ± 40 12-13 6 5 aBP 6-7 7 2 1 6 7 5 1 20 6 6 30 25 2 6 4500 ± 125 1-3 7 aBP 2-3 SA-10-19F 20 20 5 4150 ± 40 aBP 50 5 5 SA-10-18G > 8-10 m 25 11165 ± 45 20 aBP 35

fine ash peat >15 medium ash soil 35 coarse ash reworked gravel and ash

diatomite block-and-ash flow

Fig. 9. Stratigraphic sections of the Holocene explosive activity of the Sabancaya ediﬁce. UTM Easting and Northing are included below the section's P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 123

6 20 G1 5 18

3 G3

HK O O 4 2 2 16

3 MK wt.% Al

wt.% K G4 G2 14 2 a AD R b 1 12

3 800

2 Sr ppm

wt.% MgO 400 1

c d 0 0

300 200

200 Zr ppm Rb ppm 100 100

e f 0 0

40 1 La ppm Yb ppm

g h 0 0 55 60 65 70 75 80 55 60 65 70 75 80 wt.% SiO2 wt.% SiO2 SABANCAYA UPPER AMPATO BASAL AMPATO Corinta tephra and related Undifferentiated lavas Rhyolitic tephra scoriaceous deposits Moldepamba stage Yanajaja stage Former andesitic cone North, South and Central cones

Fig. 10. Selected major and trace elements for the Ampato–Sabancaya samples, plotted against silica. (a) SiO2 vs. K2O classification diagram. A andesites, D dacites, R rhyolites, MK medium- potassium, HK high-potassium (after Peccerillo and Taylor, 1976). (b–h) Variation diagrams for Al2O3, MgO, Rb, Sr, Zr, La, and Yb respectively. G1–4 in (b) correspond to the four groups identified (see text for more details). the edifice outline was obtained from the geological map (Fig. 4) interpolation techniques (linear, cubic, inverse distance weighting taking into account that the edifice's base is definedbyaconcave method) to delimit the substratum topography. This procedure yielded break in slope (Grosse et al., 2014). Then, we applied different a basal area of 170–180 km2, a maximum height of 1.6 km, and a volume Table 5 Geochemical analyses representative of main volcanic units of the Ampato–Sabancaya volcanic complex.

Edifice Ampato Basal edifice Ampato Upper edifice

Volcanic Former andesitic Moldepampa Riolitic Yanajaja North South cone Central stage cone sequence cone cone

Sample no. SA-11-64 SA-11-03 SA-11-15 SA-11-38 SA-11-37 SA-11-84A SA-11-28 SA-11-75A SA-09-30 SA-11-71A SA-10-33C SA-11-27A SA-11-12A SA-11-54 UTM-North 8,249,025 8,244,656 8,243,491 8,247,137 8,246,350 8,234,467 8,244,402 8,251,299 8,247,108 8,248,561 8,237,095 8,242,318 8,243,120 8,249,484 UTM-East 189,617 195,706 192,405 188,471 187,567 809,521 189,043 193,121 193,200 192,823 190,769 183,658 184,012 193,676 Nature lava lava lava lava lava tephra lava lava lava lava fallout fallout PF bomb lava tephra tephra

SiO2 (wt.%) 60.28 57.37 64.22 62.78 61.90 70.62 61.06 60.27 61.79 62.82 56.71 63.23 65.82 60.79

TiO2 0.86 1.17 0.80 0.88 0.85 0.16 0.93 0.97 0.83 0.86 1.23 0.69 0.60 0.95

Al2O3 15.76 16.52 15.51 15.49 16.04 12.47 16.11 16.24 15.56 15.96 17.59 15.86 15.01 16.96

Fe2O3 * 5.46 7.16 4.54 4.95 4.83 0.98 5.42 5.61 4.98 5.26 7.44 3.83 3.24 5.22 MnO 0.07 0.09 0.05 0.06 0.06 0.08 0.07 0.07 0.07 0.07 0.10 0.05 0.04 0.07 MgO 2.20 3.59 1.83 1.99 2.00 0.21 2.66 2.60 2.21 2.25 2.83 1.01 0.82 2.49 CaO 4.57 6.08 3.72 3.91 3.84 0.89 4.81 5.08 4.41 4.61 5.87 2.51 2.30 5.14

Na2O 3.61 4.32 4.12 4.19 4.01 2.84 4.18 4.26 4.07 4.35 4.12 3.18 3.64 4.51

K2O 3.74 2.53 3.57 3.44 3.45 5.18 2.93 2.84 3.11 3.23 2.17 4.37 4.75 2.74

P2O5 0.28 0.44 0.28 0.32 0.30 0.03 0.32 0.33 0.26 0.33 0.34 0.17 0.16 0.33 LOI 1.81 0.00 0.64 0.26 0.74 4.71 1.22 0.31 1.23 −0.02 1.96 4.66 2.88 0.37 TOTAL 98.65 99.26 99.30 98.28 98.02 98.16 99.71 98.57 98.50 99.72 100.37 99.58 99.26 99.56 Sc (ppm) 9.8 13.5 7.5 7.5 7.8 2.6 10.6 10.1 9.2 9.1 13.0 6.1 5.3 9.0 V 124.2 169.0 97.9 109.1 103.4 5.6 124.6 128.6 113.2 115.0 181.1 69.6 58.6 117.3 Cr 19.0 104.9 34.3 39.8 39.7 1.7 46.1 49.3 34.2 42.3 7.5 1.3 0.0 37.7 Co 15.0 22.7 11.8 14.1 13.1 0.3 16.0 16.1 14.5 14.2 21.2 9.3 6.2 15.5 Ni 18.0 53.4 20.6 21.8 22.8 1.2 30.9 23.4 24.1 23.7 16.4 5.1 1.8 25.3 Rb 104 69 140 134 117 135 87 89 77 101 50 182 204 84 Sr 679 911 693 704 734 141 771 821 678 817 835 438 410 893 Y 16.5 15.1 13.1 13.1 13.2 12.0 12.8 12.6 12.5 12.3 16.8 18.2 17.3 12.4 Zr 214 238 146 143 152 91 133 129 134 148 207 342 333 120 Nb 9.6 9.2 9.3 9.3 9.4 12.8 7.8 8.2 8.0 9.7 7.2 12.6 12.4 7.4 Ba 1014 1168 1020 983 1065 1174 977 1036 1005 1154 949 1186 1137 1045 La 40.9 42.1 46.2 44.0 44.7 20.0 38.2 38.1 37.0 45.8 33.8 49.1 40.3 37.4 Ce 84.8 85.2 94.6 89.4 89.8 56.4 78.7 80.1 74.0 94.0 71.7 108.6 98.5 73.1 Nd 35.6 38.7 40.1 38.7 41.3 15.2 34.6 34.9 32.0 38.3 35.8 40.6 33.7 35.3 Sm 6.5 6.9 6.1 6.7 6.6 3.1 6.1 6.5 5.4 6.5 6.7 7.1 5.8 6.2 Eu 1.4 1.7 1.3 1.4 1.4 0.6 1.5 1.5 1.3 1.4 1.6 1.2 1.0 1.6 Gd 4.4 4.7 4.2 4.5 4.3 2.3 4.2 4.0 3.7 4.3 5.0 4.9 4.7 4.1 Dy 3.0 3.0 2.5 2.6 2.7 1.8 2.6 2.5 2.3 2.4 3.2 3.3 2.9 2.4 Er 1.7 1.0 0.7 1.2 1.0 0.8 1.0 1.2 1.0 0.9 1.5 1.5 0.9 0.9 Yb 1.4 1.2 1.0 1.0 1.1 1.2 1.0 1.0 1.0 0.9 1.3 1.6 1.6 0.9 Th 9.8 4.3 16.4 14.5 15.7 8.8 9.7 8.5 11.0 9.7 4.8 21.8 19.7 8.1 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 125

0.8 0.8 Rhyolithes: AFC1 Rb: 140-240 ppm AFC2 0.6 Rb/Sr: 0.6-2.9 0.6 UC

0.4 0.4 FC Rb/Sr Rb/Sr

0.2 0.2

ab 0.0 0.0 0 100 200 300 0 100 200 300 Rb Rb 70 70 FC

ASVC 50 50 main trend

Corinta

La/Yb dacites La/Yb AFC1

30 30

AFC2 cdUC 10 10 0 100 200 300 0 100 200 300 Rb Rb

Fig. 11. (a, c) Trace element ratios (Rb/Sr and La/Yb) plotted against Rb. Same symbols as Fig. 10. (b, d) Fields for Ampato–Sabancaya main trend (group 1 and 2, grey ﬁeld) and Corinta dacites (group 3, blue ﬁeld). Note that rhyolites (group 4) display very high Rb/Sr values. FC: fractional crystallisation model from a parental composition (SA-11-62) and a amphibole- bearing cumulate (62% pl + 30% amph +1% cpx + 7% mag). AFC: assimilation-fractional crystallisation models from the same parental composition and fractionating phases. AFC 1, r = 0.5 and AFC2, r = 1.0, r = ratio between assimilated and fractionated mass. The assimilated upper crustal material (UC) is represented by the Charcanigneiss(Rivera, 2010).

Table 6 Magma eruptive rates at Ampato–Sabancaya Volcanic Complex, compared to eruptive rates from other dacitic volcanoes in active continental margins.

Volcano Compositional range Arc Volume Lifespan Average eruptive rate Peak eruptive rates Reference 3 3 ⁎ 3 (km ) (ka) (km /ka) (km /ka)

Mt. Baker Andesitic to rhyolitic Cascades 161 1300 0.12 Hildreth et al. (2003) Mt. Adams Basaltic to dacitic Cascades 230–400 940 0.24–0.42 1.5–2.5 Hildreth and Lanphere (1994) Mt. Mazama Andesitic to dacitic Cascades 176 420 0.42 0.8–2.5 Bacon and Lanphere (2006) Ceboruco Dacitic to rhyolitic Mexico 81 800 0.10 0.6 Frey et al. (2004) Tancitaro Andesitic Mexico 97 556 0.17 Ownby et al. (2007) Whole Pichincha Andesitic to dacitic Northern Andes 250 850 0.30 Robin et al. (2010) Guagua Pichincha Andesitic to dacitic Northern Andes 30–32 50 0.60–0.64 0.7–2.2 Robin et al. (2010) Chimborazo Andesitic to dacitic Northern Andes 63–100 120 0.53–0.84 1.2–1.6 Samaniego et al. (2012) Parinacota Andesitic to rhyolitic Central Andes 46 163 0.25–0.31 0.5–1.2 Hora et al. (2007) Ollague Dacitic Central Andes 85 ~1000 0.09 Feeley and Davidson (1994) Lascar Dacitic Central Andes 35 ~400 0.08 Mathews et al. (1994) Llullaillaco Dacitic Central Andes 50–60 ~1000 0.04–0.06 Richards and Villeneuve (2001) Aucanquilcha Dacitic Central Andes 38 1000 0.04–0.16 Klemetti and Grunder (2008) Ubinas Andesitic to rhyolitic Central Andes 56 235 0.17–0.22 Thouret et al. (2005) El Misti Andesitic to rhyolitic Central Andes 70 112 0.63 2.1 Thouret et al. (2001) Tatara - San Pedro Basaltic to rhyolitic Southern Andes 40 250 0.16 Singer et al. (1997) Puyehue - Cordon Caulle Basaltic to rhyolitic Southern Andes 131 314 0.42 0.8–0.9 Singer et al. (2008)

Ampato–Sabancaya Volcanic complex Sabancaya Andesitic to dacitic Central Andes 6–10 6–10 0.60–1.70 This work Ampato Andesitic to dacitic Central Andes 38–42 440 0.08–0.09 This work Whole complex Andesitic to dacitic Central Andes 44–54 450 0.10–0.12 This work

⁎ Deduced from estimated total volume and entire duration of the volcanic activity. 126 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 of 44–54 km3 for the whole Ampato–Sabancaya volcanic complex. This 80 Mt. Mazama estimate is highly dependent on the substratum topography, and is El Misti responsible for the uncertainty (10–20%) in the bulk volume assess- 70 ~1 km3/ka ment. These bulk volumes take into account moraine deposits, whose Puyehue 3 volume was estimated at 3–5km . Given that the older ages for the Tatara-San Pedro 60 ) 3 volcanic complex are ~450 ka, we estimated an average eruptive rate Parinacota – 3 of ~0.10 0.12 km /ka (Table 6). Using a similar approach, we also Ampato-Sabancaya 50 – 3 estimated the volume for Ampato, which is 38 42 km . Given a duration average from ~450 ka until the end of the Pleistocene, its average eruptive rate min-max 40 was calculated at 0.08–0.09 km3/ka (Table 6). The young Sabancaya range /ka ﬁ – 3 3 edi ce has a smaller volume (6 10 km ). This large uncertainty is relat- 30 ed to imprecise constraints on the substratum topography because ~0.1 km

Sabancaya is constructed on remnants of the older Ampato. Sabancaya's 20 Cumulative volume (km construction lasted from 10 to 6 ka until ~3 ka, at an average eruptive rate of around 0.6–1.7 km3/ka (Table 6), which is an order of magnitude 10 higher than for Ampato. Estimating the volume of the main eruptive stages of the ASVC 0 is challenging given the uncertainties associated with the lateral extent 500 400 300 200 100 0 of the older units (Table 7, Fig. 12). We used a GIS-based procedure Age (ka) to estimate the surface of each volcanic unit and its minimum and maximum thickness and applied these parameters and simple Fig. 12. Cumulative volume versus age diagram for the Ampato–Sabancaya volcanic geometrical forms (cone, frustum, parallelepiped) to constrain the complex. bulk volume for these units. This highlights significant variations between the long-term eruptive rates associated with the whole edifice, and the eruptive rates for individual eruptive stages. For instance, the average eruptive rate for Ampato was estimated at (~0.1 km3/ka), Cordón Caule (Singer et al., 2008), Mt. Adams (Hildreth and Lanphere, whereas we obtained higher eruptive rates for some volcanic 1994), and Mount Mazama (Bacon and Lanphere, 2006). Therefore, stages such as Moldepampa (0.38–0.45 km3/ka) or Yanajaja (0.35– we consider that the eruptive rate estimated for Sabancaya volcano 0.50 km3/ka, Table 7). (0.6–1.7 km3/ka; Fig. 12) represents peak eruptive rates associated The bulk eruptive rates of the ASVC are comparable to those with a main cone-building stage, and corroborates the hypothesis that from other silicic volcanic systems in continental arc settings (0.04– composite volcanoes grow in “spurts” with peak eruptive rates as high 0.2 km3/ka, Table 6), such as the Central Andes (i.e. Aucanquilcha; as 1–2km3/ka (Hildreth and Lanphere, 1994; Davidson and de Silva, Klemetti and Grunder, 2008), the Cascades (Mt Baker; Hildreth et al., 2000). 2003) and the Trans-Mexican volcanic belt (i.e. Ceboruco, Tancítaro; Frey et al., 2004; Ownby et al., 2007). In contrast, ASVC eruptive rates are notably lower than those of other Andean volcanoes, particularly 9. Late Holocene eruptive activity and hazards the neighbouring El Misti volcano (0.63 km3/ka, Thouret et al., 2001) and Parinacota (0.25–0.31 km3/ka, Hora et al., 2007) and others from Our new data confirms previous tephrochronological studies the Northern Andes, such as Guagua Pichincha (0.6 km3/ka, Robin (Juvigné et al., 1998, 2008) and enables better constraints to be placed et al., 2010), Chimborazo (0.5–0.8 km3/ka, Samaniego et al., 2012), on the Holocene explosive activity of Sabancaya. Several ash layers and Puyehue-Cordón Caulle in the Southern Andes (0.42 km3/ka, dated between 11 and 8 ka point to Early Holocene explosive activity Singer et al., 2008). associated with this complex. Given that the younger ages of the We stress that these estimates suffer from an averaging effect, since Ampato edifice fall in the range 20–15 ka and that the younger volcanic the long repose times are also taken into consideration. In fact, large unit of this edifice (the NE dome) lacks glacial erosion, the early eruptive rate variations have been observed at several continental arc Holocene ash layers preserved in the peatbogs of the southern flank volcanoes, such as Chimborazo (Samaniego et al., 2012), Puyehue- could be related to the last eruptive phases of Ampato or the beginning of volcanic activity at Sabancaya. In contrast, the Late Holocene tephrochronology reveals recurrent explosive activity, with at least 6– 8 eruptions during the last 4000–5000 years, including the historic Table 7 eruptive phase of the 17th–18th century and the eruptive episode at Minimum–maximum volumes and eruptive rates for the main eruptive stages of ASVC. the end of the 20th century (AD 1987–1998). Sabancaya Min–max Time Lifespan Min–max An eruption of Sabancaya potentially threatens the Colca valley volume (km3) (ka) (ka) eruptive rate (located 23–26 km to the north), which is an important tourist destina- (km3/ka) tion in Southern Peru. Based on the tephrochronological and historical Terminal cone (SA4) 1.4–2.2 3–0 3 0.5–0.7 – eruptive records, as well as the geological information for both the Satellite vent (SA3) 0.3 0.7 fi Basal edifice (SA1–2) 5.0–8.1 (10–6)-3 3–7 1.5–2.7 Ampato and Sabancaya edi ces, the most probable scenario for an erup- Whole Sabancaya 6–10 (10–6)-0 10–6 0.6–1.7 tion of Sabancaya would be a low-to-moderate magnitude (VEI 1–2) vulcanian or phreatomagmatic activity, accompanied by notable tephra Ampato Upper edifice Central 0.2–0.5 20–10 10 0.02–0.05 emissions that would produce a local impact. In addition, due to the South 1.5–5.2 40–20 20 0.08–0.26 large icecap on Ampato and Sabancaya volcanoes, secondary lahars North 0.1–0.7 might be triggered, as during the AD 1987–1998 eruption (Global Yanajaja 3.5–5.0 70–80 (?) 10 0.35–0.50 Volcanism Program, 1991). Moreover, given the major explosive activi- Ampato Basal edifice ty (subplinian to plinian) of the Ampato volcano, higher magnitude Moldepampa 19–22 250–200 50 0.38–0.45 scenarios (VEI ≥ 3) have also been taken into account for the recently Former andesitic cone 12–16 450–400 50 0.24–0.32 published volcanic hazard map (Mariño et al., 2013; Global Volcanism Whole Ampato 38–42 450–10 440 0.08–0.09 Program, 2013). P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128 127

10. Conclusions References

The Ampato–Sabancaya volcanic complex (ASVC) comprises two Adams, N.K., de Silva, S.L., Self, S., Salas, G., Schubring, S., Permenter, J.L., Arbesman, K., fi 2001. The physical volcanology of the 1600 eruption of Huaynaputina, southern successive edi ces. The oldest one, the Ampato Basal volcano, was Peru. Bull. Volcanol. 62, 493–518. built during at least two cone-building stages: The Former andesitic Alcalá, J., Palacios, D., Zamorano, J.J., Vázquez-Selem, L., 2011. Last Glacial Maximum and cone, and the Moldepampa stage dated at 450–400 ka and 230– deglaciation of Ampato volcanic complex, Southern Peru. Rev. Cuaternario Geomorfol. 25, 121–136. 200 ka, respectively. After a period of quiescence, the Ampato Upper Antayhua, Y., Tavera, H., Bernal, I., 2001. Análisis de la actividad sísmica en la región del edifice developed on the remnants of the Basal edifice about 80 to volcán Sabancaya (Arequipa). Bol. Soc. Geol. Peru 92, 79–88. 70 ka ago with the lava sequence of the Yanajaja stage. This edifice Bacon, C.R., Lanphere, M.A., 2006. Eruptive history and geochronology of Mount Mazama – comprises several cone-building stages that successively shaped the and the Crater Lake region. Or. Geol. Soc. Am. Bull. 118, 1331 1359. Bromley, G.R.M., Schaefer, J.M., Wincler, G., Hall, B.L., Todd, C.E., Rademaker, C.K.M., 2009. current peaks of Ampato: The Northern, Southern, and Central cones. Relative timing of last glacial maximum and late-glacial events in the central tropical The Southern peak, which is the largest, was built over several effusive Andes. Quat. Sci. Rev. 28, 2514–2526. episodes spanning the time period ka. The Baylillas tephra fallout se- Bulmer, M., Johnston, A., Engle, F., 1999. Analysys of Sabancaya Volcano, Southern Peru Using Radarsat and Lansat TM Data. Application Development and Research quence and the Corinta plinian fallout deposit testify to the intense ex- Opportunity (ADRO) (10 pp). plosive activity during this volcanic stage. The last cone-building stage Cotten, J., Le Dez, A., Bau, M., Caroff, M., Maury, R.C., Dulski, P., Fourcade, S., Bohn, M., of Ampato constructed the Central cone from 20 to 10 ka, and probably Brousse, R., 1995. Origin of anomalous rare-earth element and yttrium enrichments in subaerial exposed basalts: evidence from French Polynesia. Chem. Geol. 119, continued until Holocene times. During the Holocene, eruptive activity 115–138. migrated to the NE and built up the mostly effusive Sabancaya edifice, Davidson, J., de Silva, S., 2000. Composite Volcanoes. In: Sigurdsson, H., Houghton, B., between 6 and 3 ka. This cone comprises andesitic and dacitic blocky McNutt, S., Rymer, H., Stix, J. (Eds.), Encyclopedia of Volcanoes. Academic Press, San – fl Diego, pp. 663 681. lava ows which were emplaced during at least two eruptive stages. Farber, D.L., Hancock, G.S., Finkel, R.C., Rodbell, D.T., 2005. The age and extent of tropical During the Late Holocene (4.5–3 ka), Sabancaya's activity turned mostly alpine glaciation in the Cordillera Blanca, Peru. J. Quat. Sci. 20, 759–776. explosive, although the erupted volume remained low. Feeley, T.C., Davidson, J.P., 1994. Petrology of calc-alkaline lavas at Volcán Ollagüe and the origin of compositional diversity at Central Andean Stratovolcanoes. J. Petrol. 35, Samples from the ASVC define a broad high-K magmatic trend 1295–1340. composed of andesites and dacites with a mineral assemblage of Frey, H.M., Lange, R.A., Hall, C.M., Delgado-Granados, H., 2004. Magma eruption rates plagioclase, amphibole, biotite, ortho- and clino-pyroxene, and Fe– constrained by 40Ar/39Ar chronology and GIS for the Ceboruco–San Pedro volcanic fi – Ti oxides. A secondary trend is also identified, and corresponds to eld, western Mexico. Geol. Soc. Am. Bull. 116, 259 276. http://dx.doi.org/10.1130/ B25321.1. rare dacitic explosive eruptions (Corinta fallout deposits). Both Gerbe, M.C., Thouret, J.C., 2004. Role of magma mixing in the petrogenesis of lavas magmatic trends are derived by fractional crystallisation involving erupted through the 1990–98 explosive activity of Nevado Sabancaya in south – an amphibole-rich cumulate with variable amounts of upper crustal Peru. Bull. Volcanol. 66, 541 561. Global Volcanism Program, 1988. Report on Sabancaya (Peru). In: McClelland, L. (Ed.), assimilation. Scientific Event Alert Network Bulletin, 13:6. Smithsonian Institution http://dx.doi. The overall eruptive history of the Ampato–Sabancaya volcanic org/10.5479/si.GVP.SEAN198806-354006. complex reveals major changes in the bulk eruptive rates during the Global Volcanism Program, 1990. Report on Sabancaya (Peru). In: McClelland, L. (Ed.), Bulletin of the Global Volcanism Network, 15:5. Smithsonian Institution http://dx. volcano's development. Throughout Ampato's evolution, peak eruption doi.org/10.5479/si.GVP.BGVN199005-354006. 3 rates reached values of 0.2–0.5 km /ka, although the greatest variation Global Volcanism Program, 1991. Report on Sabancaya (Peru). In: McClelland, L. (Ed.), has been identified between the overall eruptive rate for Ampato Bulletin of the Global Volcanism Network, 16:5. Smithsonian Institution http://dx. 3 – 3 doi.org/10.5479/si.GVP.BGVN199105-354006. (0.09 km /ka) and that for Sabancaya (0.6 1.7 km /ka). This abrupt Global Volcanism Program, 1997. Report on Sabancaya (Peru). In: Wunderman, R. (Ed.), change highlights variable eruptive rates throughout the volcano's Bulletin of the Global Volcanism Network, 22:7. Smithsonian Institution http://dx. history. doi.org/10.5479/si.GVP.BGVN199707-354006. Based on tephrochronologic studies, the Late Holocene activity at Global Volcanism Program, 2013. Report on Sabancaya (Peru). In: Wunderman, R. (Ed.), Bulletin of the Global Volcanism Network, 38:2. Smithsonian Institution http://dx. Sabancaya is characterised by vulcanian events with local ash fallout doi.org/10.5479/si.GVP.BGVN201302-354006. impact. We identified at least 6 eruptions during the last 4–5 ka, includ- Grosse, P., Euillades, P.A., Euillades, L.D., van Wyk de Vries, B., 2014. A global database of ing the historical AD 1750–1784 and 1987–1998 events. On the basis of composite volcanomorphometry. Bull. Volcanol. 76, 784. http://dx.doi.org/10.1007/ s00445-013-0784-4. this recurrent low-to-moderate explosive activity, Sabancaya must be Hall, M.L., Ramón, P., Mothes, P., Le Pennec, J.-L., García, A., Samaniego, P., Yepes, H., 2004. considered an active volcano, as corroborated by the frequent signs of Volcanic eruptions with little warning: the case of Volcán Reventador's surprise No- – unrest since 2013. vember 3, 2002 eruption, Ecuador. Rev. Geol. Chile 31, 349 358. Harpel, C.J., de Silva, S., Salas, G., 2011. The 2 ka eruption of Misti Volcano, southern Peru—the most recent Plinian eruption of Arequipa's iconic volcano. Geol. Soc. Am. Spec. Pap. 484, 1–72. http://dx.doi.org/10.1130/2011.2484. Acknowledgements Hildreth, W., Lanphere, M.A., 1994. Potassium-argon geochronology of a basalt-andesite- dacite arc system: the Mt. Adams volcanic field, Cascade Range of southern This work is part of a Peruvian–French cooperation programme Washington. Geol. Soc. Am. Bull. 106, 1413–1429. Hildreth, W., Fierstein, J., Lanphere, M., 2003. Eruptive history and geochronology of the carried out between the Instituto Geológico, Minero y Metalúrgico Mount Baker volcanic field, Washington. Geol. Soc. Am. Bull. 115, 729–764. (INGEMMET, Peru) and the Institut de Recherche pour le Développement Hogg, A.G., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Guilderson, T.P., Heaton, T.J., (IRD,France).Itwaspartiallyfoundedbya“Jeunes Equipes Associées à Palmer, J.G., Reimer, P.J., Reimer, R.W., Turney, C.S.M., Zimmerman, S.R.H., ” 2013. SHcal13 southern hemisphere calibration, 0–50,000 years cal BP. Radiocar- l'IRD (JEAI) project that is an initiative designed to promote and bon 55, 1889–1903. strengthen new research teams in developing countries. Several radiocar- Hora, J.M., Singer, B.S., Wörner, G., 2007. Volcano evolution and eruptive flux on the thick bon dates were obtained thanks to IRD support to Laboratoire de Mesure crust of the Andean Central Volcanic Zone: 40Ar/39Ar constraints from Volcan Parinacota, – du Carbone 14 (LMC14), UMS 2572 (CEA-CNRS-IRD-IRSN-MCC), Saclay, Chile. Geol. Soc. Am. Bull. 119, 343 362. http://dx.doi.org/10.1130/B25954.1. Jay, J.A., Delgado, F.J., Torres, J.L., Pritchard, M.E., Macedo, O., Aguilar, V., 2015. Deformation France. We thank B. Jicha, C. Siebe and J.L. Le Pennec for their constructive and seismicity near Sabancaya volcano, southern Peru, from 2012 to 2015. Geophys. reviews of a previous version of this manuscript, and F. van Wyk des Vries Res. Lett. http://dx.doi.org/10.1002/2015GL063589. for improving the English. This is Laboratory of Excellence ClerVolc Jomelli, V., Khodri, M., Favier, V., Brunstein, D., Ledru, M.P., Wagnon, P., Blard, P.H., Sicart, J.E., Braucher, R., Grancher, D., Bourlès, D.L., Braconnot, P., Vuille, M., 2011. Irregular contribution no 200. tropical glacier retreat over the Holocene epoch driven by progressive warming. Nature 474, 196–199. http://dx.doi.org/10.1038/nature10150. Juvigné, E., Thouret, J.C., Gilot, E., Leclercq, L., Gourgaud, A., 1998. L'activité du Appendix A. Supplementary data volcan Nevado Sabancaya (Pérou) au cours de l'Holocene. Quaternary 9, 45–51. Juvigné, E., Thouret, J.-C., Loutsch, I., Lamadon, S., Frechen, M., Fontugne, M., Rivera, M., Supplementary data to this article can be found online at http://dx. Marino, J., 2008. Téphrostratigraphie et chronologie dans la region des Nevados doi.org/10.1016/j.jvolgeores.2016.04.038. Ampato et Sabancaya, sud du Pérou. Quaternaire 19, 161–178. 128 P. Samaniego et al. / Journal of Volcanology and Geothermal Research 323 (2016) 110–128

Klemetti, E.W., Grunder, A.L., 2008. Volcanic evolution of Volcán Aucanquilcha: a long- Sébrier, M., Soler, P., 1991. Tectonics and Magmatism in the Peruvian Andes from Late Ol- lived dacite volcano in the Central Andes of northern Chile. Bull. Volcanol. 70, igocene Time to Present. In: Harmon, R.S., Rapela, C.W. (Eds.), Andean Magmatism 633–650. and its Tectonic Setting. Geol. Soc. Am. Spec. Paper 265, pp. 259–278. Klinck, B., Ellison, R., Hawkins, M., Palacios, O., de la Cruz, J., de la Cruz, N., 1993. Geología Siebert, L., Simkin, T., Kimberly, P., 2010. Volcanoes of the World. third ed. Smithsonian de la Cordillera Occidental y Altiplano al Oeste del lago Titicaca- Sur del Perú. Boletín Institution and University of California press (551 pp). INGEMMET N° 42, Serie a (257 p). Singer, B.S., Thompson, R.A., Dungan, M.A., Feeley, T.C., Nelson, S.T., Pickens, J.C., Brown, Major, J.J., Lara, L.E., 2013. Overview of Chaitén volcano, Chile, and its 2008–2009 L.L., Wulff, A.W., Davidson, J.P., Metzger, J., 1997. Volcanism and erosion during the eruption. Andean Geol. 40, 196–215. http://dx.doi.org/10.5027/andgeoV40n2-a01. past 930 ky at the Tatara–San Pedro complex, Chilean Andes. Geol. Soc. Am. Bull. Mamani, M., Worner, G., Sempere, T., 2010. Geochemical variation in igneous rocks of the 109, 127–142. Central Andean orocline (13 °S to 18 °S): tracing crustal thickening and magmas Singer, B.S., Jicha, B.R., Harper, M.A., Naranjo, J.A., Lara, L.E., Moreno-Roa, H., 2008. Eruptive generation through time and space. Geol. Soc. Am. Bull. 97, 241–254. history, geochronology, and magmatic evolution of the Puyehue-Cordón Caulle volca- Mariño, J., Samaniego, P., Rivera, M., Bellot, N., Manrique, N., Macedo, L., Delgado, R., 2013. nic complex, Chile. Geol. Soc. Am. Bull. 120, 599–618. http://dx.doi.org/10.1130/ Mapa de peligros del Complejo Volcánico Ampato-Sabancaya. Esc. 1:50.000. Edit. B26276.1. INGEMMET-IRD. Stone, J.O., 2000. Air pressure and cosmogenic isotope production. J. Geophys. Res. 105, Mathews, S.J., Jones, A.P., Gardeweg, M.C., 1994. Lascar volcano, Northern Chile: evidence 23753–23759. for steady-state disequilibrium. J. Petrol. 35, 401–432. Stuiver, M., Reimer, P.J., 1993. Extended 14C database and revised CALIB radiocarbon McCormac, F.G., Hogg, A.G., Blackwell, P.G., Buck, C.E., Higham, T.F.G., Reimer, P.J., 2004. calibration program. Radiocarbon 35, 215–230. SHCal04 Southern Hemisphere Calibration 0-11.0 cal Kyr BP. Radiocarbon 46, Stuiver, M., Reimer, P.J., Reimer, R.W., 2005. CALIB 5.0. [WWW Program and 1087–1092. Documentation]. Mering, C., Huaman, D., Chorowicz, J., Deffontaines, B., Guillande, R., 1996. New data on Thouret, J.C., Guillande, R., Huaman, D., Gourgaud, A., Salas, G., Chorowicz, J., 1994. the geodynamics of southern Peru from computerized analysis of SPOT and SAR L'activité actuelle du Nevado Sabancaya (Sud Pérou): reconnaissance géologique et ERS-1 images. Tectonophysics 259, 153–169. satellitaire, evaluation et cartographie des menaces volcaniques. Bull. Soc. Geol. Fr. Ownby, S., Delgado Granados, H., Lange, R.A., Hall, C.M., 2007. Volcán Tancítaro, Michoacán, 165, 49–63. Mexico, 40Ar/39Ar constraints on its history of sector collapse. J. Volcanol. Geotherm. Res. Thouret, J.C., Dávila, J., Eissen, J.P., 1999. Largest historic explosive eruption in the Andes at 161, 1–14. Huaynaputina volcano, South Peru. Geology 27, 435–438. Peccerillo, P., Taylor, S.R., 1976. Geochemistry of Eocene calc-alkaline volcanic rocks from Thouret, J.C., Finizola, A., Fornari, M., Suni, J., Legeley-Padovani, A., Frechen, M., 2001. the Kastamonu area, northern Turkey. Contrib. Mineral. Petrol. 58, 63–81. Geology of El Misti volcano nearby the city of Arequipa, Peru. Geol. Soc. Am. Bull. Richards, J.P., Villeneuve, M., 2001. The Llullaillaco volcano, Northwest Argentina: 113, 1593–1610. construction by Pleistocene volcanism and destruction by sector collapse. Thouret, J.C., Rivera, M., Wörner, G., Gerbe, M.C., Finizola, A., Fornari, M., Gonzales, K., J. Volcanol. Geotherm. Res. 105, 77–105. 2005. Ubinas: the evolution of the historically most active volcano in southern Rivera, M., 2010. Genèse et évolution des magmas andésitiques à rhyodacitiques récents Peru. Bull. Volcanol. 67, 557–589. des volcans Misti et Ubinas (Sud du Pérou). Unpublished PhD thesis, Université Blaise Thouret, J.C., Wörner, G., Gunnell, Y., Singer, B., Zhang, X., Souriot, T., 2007. Geochronolog- Pascal, Clermont-Ferrand II, France, 407 pp. ic and stratigraphic constraints on canyon incision and Miocene uplift of the Central Rivera, M., Thouret, J.C., Samaniego, P., Le Pennec, J.L., 2014. The 2006–2009 activity of Andes in Peru. Earth Planet. Sci. Lett. 263, 151–166. http://dx.doi.org/10.1016/j.epsl. Ubinas volcano (Peru): petrology of the 2006 eruptive products and insights into 2007.07.023. genesis of andesite magmas, magma recharge and plumbing system. J. Volcanol. Travada y Córdova, V., 1752. El Suelo de Arequipa Convertido en Cielo. In: de Odriozola, Geotherm. Res. 270, 122–141. http://dx.doi.org/10.1016/j.jvolgeores.2013.11.010. M. 1877. Documentos Literarios del Perú. Tomo X. Imprenta del Estado, Lima, p. 415. Robin, C., Samaniego, P., Le Pennec, J.L., Fornari, M., Mothes, P., van der Plicht, J., 2010. Vermeesch, P., 2007. CosmoCalc: an excel add-in for cosmogenic nuclide calculations. New radiometric and petrological constraints on the evolution of the Pichincha Geochem. Geophys. Geosyst. 8, 1525–2027. volcanic complex (Ecuador). Bull. Volcanol. 72, 1109–1129. http://dx.doi.org/10. Zamácola y Jaúregui, J.D., 1888. Apuntes Para la Historia de Arequipa, Año de 1804. 1007/s00445-010-0389-0. Imprenta de la Bolsa, Arequipa, p. 96. Samaniego, P., Barba, D., Robin, C., Fornari, M., Bernard, B., 2012. Eruptive history of Chimborazo volcano (Ecuador): a large, ice-capped and hazardous compound volcano of the Northern Andes. J. Volcanol. Geotherm. Res. 221–222, 33–51. http:// dx.doi.org/10.1016/j.jvolgeores.2012.01.014.