<<

b-

Largest in historical times in the Andes at A.D. , 1600, southern

Jean-Claude ThOuret* IRD and Instituto Geofisicodel PerÚ, Calle Calatrava 21 6. Urbanización Camino Real, La Molina, Lima 12, Perú Jasmine Davila Instituto Geofísico del Perú, Lima, Perú Jean-Philippe Eissen IRD Centre de Brest BP 70, 29280 Plouzane Cedex, France

ABSTRACT lent [DREI volume. km': Table The The largest esplosive eruption (volcanic explosivity index of 6) in historical times in the 4.4-5.6 I). tephra is composed of -SO% with rhyolitic Andes took place in 1600 at Huaynaputina volcano in southern Peru. According to chroni- A.D. glass, crystals (mostly amphibole and bio- cles, the eruption began on February 19 with a Plinian phase and lasted until March 6. Repeated 15% tite), and 5% xenoliths including hydrothermally tephra falls, pyroclastic flows, and surges devastated an area 70 x 40 kni? west of the vent and altered fragments from the volcanic and sedimen- affected all of southern Peru, and shook the city of Arequipa 75 km away. Eight tnry bedrock, and a small amount of accessory lwa deposits, totaling 10.2-13.1 km3 in bulk volume, are attributed to this eruption: (1) a widespread, fragments. The erupted pumice is a white, , -S.1 kn3 pumice-fall deposit; (2) channeled (1.6-2 km3) with (3) ground-surge and medium-potassicdacite (average SiO, and ash-cloud-surge deposits; (4) widespread co- ash layers; (5) base-surge deposits; (6) 63.6% 1.85% hlgO) of the potassic calc-alkalic suite. The unconfined ash-flow deposits; (7) crystal-rich deposits; and (S) late ash-fall and surge deposits. phenocryst assemblage comprises plagioclase, bio- Disruption of a hydrothermal system and hydromagmatic interactions are thought to have tite, amphibole, magnetite, quartz, ilmenite, and fueled the large-volume explosive eruption. Although the event triggired no collapse, apatite. The white color, phenocryst assemblage, ring fractures that cut the vent area point to the onset of a funnel-type caldera collapse. and stratigraphic position distingirish the A.D. 1600 tephra from ash from adjacent volcanoes. INTRODUCTIOS LARGE VOLUME AND WIDE 2. Three types of nonwelded ignimbrites, The largest explosive eruption in historical SPECTRUM OF TEPHRA 1.6-2 km' in volume, are preserved in radial tinies in the Andes took place from 19 February Eight types of enipted deposits encompass an valleys and on the plateau. Lowest in strati- to 6 March 1600 at Huaynaputina. a small vol- estimated bulk volume of 10.2-13.1 km? (Figs. 2 graphic order, an ash-flow deposit is interbedded canic center 75 kin east of Arequipa in southern and 3. Table I). in the pumice-fall deposit in proximal sections Peni (Fig. 1: Thouret et al., 1997). Huaynaputina I. The Plinian pu,niice-fall deposit was dis- (Fig. 3A). Second, pyroclastic-flow deposits is not a typical ; instead, four vents persed in a widespread, I-cni-thick lobe 11 15000 5-15 m thick above the Plinian fallout are chan- are nested on the floor (4200 m) of a horseshoe- kni? to the west and west-northwest toward the neled in all catchments and as far as 45 kni away shaped 2.5 x 1.5 km' caldera. open on the east Pacific coast (Fig. 2A). According to a log-nomial from the vent in the Rio Tambo canyon: they are edge of a high toward the plot of thickness versus the square root of area composed of inversely graded pumice and canyon of the Rio Tambo. This caldera breached (Fig. 2B) and to the 360000 hn2 area of reported normally graded vitreous clasts in ash. ivith the summit (4SOO ni) of the volcano before the ash fall (Fig. 2C), the computed bulk volume of a few breadcrust bombs (Fig. 3A). Third, thick A.D. 1600 eruption. The plateau has been built up the Plinian fallout is -8.1 km'(dense-rock equiva- ash- and pumice-rich pyroclastic-flow deposits, of no more than 500-m-thick lava flows and ignimbrites of late hliocene to Quatemary age. tvhich overlie sedimentary and intrusive rocks of hlesozoic age. Repeated tephra falls. pyroclastic flows, and surges buried seven villages, killed -1500 people (Navarro. 1994). and devastated the landscape ivithin an elliptical area about 70 x 40 km west of the volcano. Pyroclastic flows choked the Rio Tambo canyon and reportedly dammed two teni- porary lakes (Barriga. 195 1). The catastrophic breaching of the lakes released large-scale debris Figure 1. Schematic map flows that swept down the 120-km-longvalley to of area of Huaynaputina the Pacific . volcano and Rio Tambo Understanding the voltrminous A.D. 1600 area. Inset: Location of Plinian eruption at Huaynaputina is important Huaynaputina and central because of the correlation of a wide spectrum of Andean volcanic zone (CAVZ) in Peru. erupted deposits with reported events, the severe afterniuth on the once-populated area around a poorly known volcano, and the probable global climatic impact. This paper shows the large mag- nittide of the eruption. the variety of the erupted deposits across an area of high relief, and the sig- nificance of the event.

v. 1 : Mny 1999: 27: no. 5: p. 4354&; 4 fi~ures:I table. 435 I I Fonds Documentaire ORSTOM Figure 2. A: Map showing ex- tent and isopachs (thickness A.D. in centimeters) of 1600 Plinian tephra fall from Huay- naputina (HP). 8: Plot of thickness vs. square root of area (after Pyle, 1989) for computing bulk volume of Plinian tephra-fall deposit: 6.91 km3 within 1 cm isopach and 8.12 km3 within limit of reported ash fall, shown in sketch map C.

preserved on the plateau, are underlain by cres- (Fig. 3A). The sandwave facies and cross-bedded ash and lapilli. Both thickness and grain size of cent-shaped dunes as far as 1.5 km from the cal- beds as well as the pinch-and-swell and dunelike the ash layers increase toward the valley-con- dera rim (Fig. l).The dunes consist'of , a features point to ground-surge deposits. Other fined ignimbrites. mixture of lava clasts, pumiceous and porphyritic thin, fine-ash layers from ash-cloud surges over- 5. Cross-bedded ash and pumice deposits, lumps, and xenolithic blocks. lie the pyroclastic-flow units (deposit 3, Fig. 3A). 1-3 m thick and showing sandwave facies, accre- 3. One-decimeter- to several-decimeters-thick 4. Several massive ash layers totaling 5-20 tionary lapilli, and soft deformation, are interbed- ash and pumice deposits, fine grained and with cm in thickness mantle the pumice-fall deposit ded in the middle part of the ignimbrite sequence few'lithic fragments, underlie most of the pyro- as far as 75 km away from the vent (Fig. 3B). as far as 15 km from the vent in the Rio Tambo clastic-flow units in proximal valley sections They consist of fine ash alternating with coarse valley and tributaries (Fig. 3A). The base-surge

Figure 3. Composite strati- graphic sections of tephra from A.D. olluviurn; ash-fall deposit 1600 eruption. Deposits fine vitric numbered 1 to 8 are described sh-cloudsurge, ash in text. rystal-rich surge deposit sh-cloud surge. fine vitic ash rystal-rich lapilli and ash

Pre-A.D.1600soil Massive with increasing Unconfined medial to distal sections I3 %15km from the vent area L c Pre-A.D.1600soil in ash

436 f 1999 GEOLOGY, May e k-.-#f LI F

deposits of phreatomagmatic origin represent an TABLE 1. VOLCANOLOGICAL DATA FOR A.D. 1600 PLINIAN ERUPTION estimated volume of 0.05 km'. Extent of the Plinianfallout (within 1cm isopach) 5 115 O00 km2 6. Unconfined ash-flow deposits2-3 m thick in Extent of reported ash and dust fallout (Fig.' 2C). -360 O00 kmz proximal areas (Fig. 3A) correlate with a massive Volume of Plinian (within 1cm isopach and -6.9 and 8.1 km3 vitric-ash layer 3-4 cm thick as far as 70 km west- from Fig. 2C) northwest of the vent. The flows that emplaced Minimum volume percent beyond 1cm isopach 17.5% (80% 4.4-5.6 the ash-flow deposits must have surmounted the Minimum volume DRE pumice, average km3 density 1.5 g/cm3) 2400-m-high left wall of the Rio Tambo canyon Minimum volume of lithic fragments 1.O5 km3 15 km away from source toward the southeast and Minimum volume of 3.754.25 km3 a series of 1000- to 1400-m-high topographic Duration of the Plinian phase 13-19 h 9 barriers at least as far as 60 km to the west. Mass eruption rate (1 h average, DRE magma density 1.3-1.5.105 kgls 2.4 g/cm3) 7. A gray, crystal-rich deposit, 15-35 cm thick io4 0.5-0.6 Volumetric eruption rate (same characteristicsas above) 7.~7.7 m3/s in proximal to medial sections and km3 Peak mass eruption rate* 1-2 10' kg/s in volume, overlaps a large part of the area man- Column height' 33-37 km tled by the Plinian fallout as far as 75 km west- Peak wind velocity' 10-20 mls northwest of the vent. The crystal-rich deposit, Weight percent ejecta

EMPLACEMENT OF DEPOSITS AND size in the middle part of the medial sections and A hydrothermal system was probably de- ERUPTION MECHANISMS by the coarser pumice-fall deposit that directly pressurized and fragmented during the Plinian According to chronicles (Barriga, 195 1; rests on the pre-A.D. 1600 soil in upwind sections. phase, evidenced by (1) the abundance of hydro- Vasquez de Espinosa, 1942; Mateos, 1944), the The high column remained sustained almost thermally altered lithic fragments from the base eruption began on 19 February 1600 following throughout the Plinian phase, for the mean size of to the middle part of the pumice-fall sections four days of intense seismic activity and lasted the pumice decreases only in the uppermost and (2) fumarolic stains coating lithic frag- until 6 March. Correlating historical accounts to 10 cm of the tephra-fall deposit. The intra-Plinian ments and pervading pumice. Lithic-rich units measured tephra and inferred eruptiveprocesses, flow deposit in the middle of the sections in the with oxidized boulders in the second half of the we distinguish five eruptive phases. Rio Tambo valley (Fig. 3A) shows that the erup- near-vent Plinian fallout indicate that the con- tion column was unstable toward the second half duit walls widened in the hydrothermally Plinian Phase and Dynamics of the Eruption of the Plinian phase. The flow deposit correlates altered host rocks during the second part of the Column stratigraphically with the increase in hydrother- Plinian phase. From the drastic decrease in the The Plinian phase lasted 13 to 19 h (between mally altered lithic fragments in the pumice-fall amount of the hydrothermal component up- 5 P.M. on February 19 and 5 A.M. to midday on deposit, indicating either vent erosion, higher ward in the Plinian fallout, we infer that the February 20) (Barriga, 1951) and delivered the mass eruption rate, or instabilitydue to local over- flaring process stopped just before the end of widespread pumice fallout at the onset of the loading of the column or increased water input. the Plinian phase. eruption. By using characteristics of the Plinian fallout (Fig. 4) and Carey and Sparks's method k c L (1986), we computed a 33-37-km-high and a high discharge rate (Table 1). The wide dispersal of tephra across a distance of -600 Figure 4. Areas (in square kilo- meters) enclosed by isopachs km toward the west and west-northwest (Fig. 2A) and isopleths of maximum pum- is due to the pattern of wind directions prevailing ice (MP) and maximum lithic from December to March in southem Peru. fragments (ML) sizes (in millime- The eruption column rapidly became high and ters) of A.D. 1600 Plinian tephra- sustained, for coarse grain size and reverse grad- fall deposit, for comparison with six well-characterized Plinian ing appear as low as in the first decimeter above eruptions, all referenced in Hil- the base of the deposit (Fig: 3). An umbrella dreth and Drake (1992, their plot and/or a shift in wind direction developed only on p.111). -during the second half of the Plinian phase, as demonstrated by the abrupt increase in pumice 1 10 100 1000 lW00 10 1W 10 100 ... GEOLOGY, May 1999 -? 437 k

.4 Ignimbritic Phase that the crystal enrichment through density sege- Second, no caldera formed despite the total Following a lull, the explosive activity in- gation and elutriation of vitric-fine ash occurred 6-8.8 km3 DRE volume of erupted tephra creased during February 24-27, when earth- during the flow transport over the rugged topog- (Table 1). The lack of caldera collapse may be quakes shook down the cathedral of Arequipa. raphy. The crystal-rich deposit implies that an due to the lithologic behavior of the bedrock The increase may correspond to the emplace- originally crystal-rich magma was tapped toward relative to the small, eroded Huaynaputina ment of the 1.6-2 km3 channelized ignimbrites, the end of the eruption, a fact that is reflected by stratovolcano.But ring fractures that cut the vent underlain by and interspersed with ground- and the increase in density and crystal content of the area point to a funnel-shaped structure beneath ash-cloud surges (deposits 2 and 3, Fig. 3A). pumice throughout the erupted sequence. the crater complex and to the onset of a funnel- The ignimbrite-formingeruptive phase tapped type collapse (Fig. I). another magma batch, as shown by the more Phase of Emplacement of Ash Flows and These clues suggest that the A.D. 1600 eruption evolved composition of the porphyritic, crystal- Surges may have been a single event in a longer process. rich pumice of the ignimbrites: they are more Intermittentash falls that reportedly continued Should an eruption of this magnitude occur again porphyritic,richer in silica (average 65.2% SiO,), in Arequipa from February 28 to at least March 6 in southern Peru, fallout, pyroclastic flows, and and poorer in magnesium (average 1.72% MgÕ) may have derived from ash flows and surges that surges would wreak havoc in an area where than the pumice from the Plinian fallout. Post- left deposits 6 and 8 at top of sections (Fig. 3). -30000 people are now directly at risk and would Plinian flows, first lithic rich and then pumice Unconfined ash-flow deposits traveled more than affect the -700000 inhabitants of Arequipa. and ash rich, deposited a significant volume of 60 km over a relief of six ridges 1000 to 1400 m ignimbrites (Table l), whose lack of either weld- high above valley floors and normal to flow ACKNOWLEDGMENTS v This research was supported by the Institut de ing or gas pipes suggests that the magma was direction. The equation = @, where g =force Recherche pour le Développemen\ (IRD) and the Insti- neither very hot nor gas rich. The lithic-rich of gravity, h = height of the topographic barriers, tuto Geofísico del Pení (IGP). We \hank J.-L. Bourdier ignimbrite, channelized in the Rio Tambo valley and v = velocity, produces a minimum ash-flow and S. Young for improving an eqrly manuscript and R. Waitt for a constructive review. and tributaries, also deposited ground velocity of 50 to 1 10 ds. The rugged topography that formed crescent-shaped dunes on near-vent and high relief controlled the flow transport and REFERENCES CITED slopes 1.5 km from the caldera rim. deposition: massive ash-flow deposits on ridges Barriga, V. M., 1951, Los terrerpotos en Arequipa The massive ash layers (deposit 4) in distal change laterally to bedded surge deposits. (1582-1868): Arequipa, La Çolmena, 426 p. S., S. areas >45 km away from source (Fig. 3B) are Carey, and Sparks, R. J., 1986, Quantitative interpreted as Co-ignimbrite ash, although some Posteruption Emplacement of Mass-Flow models of the fallout and disnersal of tephra from volcanic eruption columns: Bulletin of Volcanol- of the ash layers may have been emplaced by ash Deposits ogy, V. 48, p. 109-126. falls from a dwindling eruption column during According to Barriga (195 l), the sky remained Hildreth, W., and Drake, R. E., 1992, Volcan Quizapu, the events recorded on February 20-22. The wide hazy until April 2 when the air eventually Chilean Andes: Bulletin of , v. 54, 1 extent (Fig. 2) and the 1.2-2.4 km3 volume of co- cleared, but Mateos (1944) and Ocaña (1969) p. 93-125. ignimbrite ash and co-Plinian ash are consistent stated that dust remained in the air for nine Juvigné, E., Thouret, J.-C., Gilot, E., Gourgaud, A., Legros, E, Uribe, M., and Graf, K.,1997, Etude with contemporary reports that ash fell repeat- months, a fact that we attribute to the emplace- téphrostratigraphique et bioclimatique du Tardi- edly on Arequipa during the two weeks after the ment of voluminous, posteruption mass-flow glaciaire et de I'Holochne de la Laguna Salinas, Plinian phase. deposits. Reworked ignimbrite and debris-flow Pérou méridional: Géographie Physique et Qua- deposits 5-20 m thick (Fig. 3A) mantle the edges ternaire, v. 51,p. 219-231. Mateos, E, 1944, Historia general de la Companía de Hydromagmatic Phase of the plateau and are channeled as far as 15 km Jesus en la provincia del Pení: Madrid, Consejo Four additionalevents occurred from February away in the radial valleys. Superior de Investigaciones Científicas, 240 p. 28 to March 2. Of these, one or two may cor- Navarro, R., 1994, Antologia del valle de Omate: respond to the hydromagmaticphase, which em- SIGNIFICANCE OF THE ERUPTION Arequipa, Universidad Nacional San Agustin, placed base surges intërbedded in the ignimbrite Such a voluminous explosive eruption (Table 1) 76 p. Ocaña, D. de, 1969, Un viaje fascinante por la America sequence. Evidence for repeated hydromagmatic occurred at this site because of the disruption of a Hispana del Siglo XVI: Madrid, Edición A. explosions includes accretionary lapilli and soft hydrothermal system during the Plinian phase and Alvarez, Historia 16,256 p. deformation in base-surge deposits and breadcrust because of hydromagmatic interactions inter- Pyle. D. M., 1989, The thickness, volume, and grain- bombs in ignimbrites and ground breccia. This spersed in the eruptive sequence. Does this event size of tephra fall deposits: Bulletin of Volcanol- ogy, V. 51, p. 1-15. ,evidence suggests that magma-discharge decreased represent the onset or the end of a long-lasting Sparks, R. S.J., 1986,The dimensions and dynamics of because the vent area widened owing to cata- process? Two preliminary clues are available. volcanic eruption columns: Bulletin of Volcanol- strophic erosion concomitant with the eruption of First, the horseshoe-shaped caldera had ogy, V. 48, p. 3-15. lithic-rich ignimbrites. The hydromagmatic com- already breached the stratovolcano and had been Thouret, J.-C., Davila, J., Rivera, M., Eissen, J.-Ph., Gourgaud, A.. Le Pennec, J.-L., and Juvigné, E., ponent and the base-surge deposits point to an ex- modified by earlier eruptions. The A.D. 1600 1997, L'éruption explosive de 1600 au Huayna- plosive episode that likely opened the three funnel- eruption blew away domes that had previously putina (Pérou), la plus volumineuse de l'histoire shaped vents now in the pre-A.D. 1600 caldera. grown in the area, led to the formation of the dans les Andes centrales: Comptes-Rendus crater complex (Fig. l), but induced no debris Académie des Sciences Paris, Elsevier, v. 325, Phase of Emplacement of Crystal-Rich Flow : the Plinian fallout rests on a thin soil p. 931-938. Vasquez de Espinosa, A., 1942, Compendium and and Surge Deposits above the weathered top of older debris- description of the West Indies (translated by C. U. Of the four events of February 28 to March 2, at avalanche deposits and pyroclastic deposits that Clark): Smithsonian Miscellaneous Collection, least two violent explosions likely released the two fill the Rio Tambo valley (Fig. 1). A pumice-fall v. 102,862 p. crystal-rich layers of deposit 7 (Fig. 3A). Each deposit interbedded in the pyroclastic deposits crystal-rich flow and surge deposit and its over- correlateswith a similar pumice-fall deposit from Manuscriptreceived September 25, 1998 lying vitric ash form a genetically related pair: the- Huaynaputina that yielded a radiocarbon age of Revised manuscript received January 14,1999 Manuscript accepted February I, 1999 percentage of crystals in the crystal-rich deposit 9700 f 190 yr B.P. (Juvigné et al., 1997). and the thickness of each vitric-ash layer both in- creasêaway from source. The related pair suggests

438 * Printed in U.S.A. ' GEOLOGY, May 1999