Ubinas: the evolution of the historically most active volcano in southern Peru

Abstract

Ubinas volcano has had 23 degassing and ashfall episodes since A.D. 1550, making it the historically most active volcano in southern Peru. Based on fieldwork, on interpretation of aerial photographs and satellite images, and on radiometric ages, the eruptive history of Ubinas is divided into two major periods. Ubinas I (Middle Pleistocene >376 ka) is characterized by lava flow activity that formed the lower part of the edifice. This edifice collapsed and resulted in a debris-avalanche deposit distributed as far as 12 km downstream the Rio Ubinas. Non-welded ignimbrites were erupted subsequently and ponded to a thickness of 150 m as far as 7 km south of the summit. These eruptions probably left a small collapse caldera on the summit of Ubinas I. A 100-m-thick sequence of ash-and-pumice flow deposits followed, filling paleo-valleys 6 km from the summit. Ubinas II, 376 ky to present comprises several stages. The summit cone was built by andesite and dacite flows between 376 and 142 ky. A series of domes grew on the southern flank and the largest one was dated at 250 ky; block-and-ash flow deposits from these domes filled the upper Rio Ubinas valley 10 km to the south. The summit caldera was formed between 25 and 9.7 ky. Ash-flow deposits and two Plinian deposits reflect explosive eruptions of more differentiated magmas. A debris-avalanche deposit (about 1.2 km³) formed hummocks at the base of the 1,000-m-high, fractured and unstable south flank before 3.6 ka. Countless explosive events took place inside the summit caldera during the last 9.7 ky. The last Plinian eruption, dated A.D.1000–1160, produced an andesitic pumice-fall deposit, which achieved a thickness of 25 cm 40 km SE of the summit. Minor eruptions since then show phreatomagmatic characteristics and a wide range in composition (mafic to rhyolitic): the events reported since A.D. 1550 include many degassing episodes, four moderate (VEI 2–3) eruptions, and one VEI 3 eruption in A.D. 1667.

Ubinas erupted high-K, calc-alkaline magmas (SiO₂=56 to 71%). Magmatic processes include fractional crystallization and mixing of deeply derived mafic andesites in a shallow magma chamber. Parent magmas have been relatively homogeneous through time but reflect variable conditions of deep-crustal assimilation, as shown in the large variations in Sr/Y and LREE/HREE. Depleted HREE and Y values in some lavas, mostly late mafic rocks, suggest contamination of magmas near the base of the >60-km-thick continental crust. The most recently erupted products (mostly scoria) show a wide range in composition and a trend towards more mafic magmas.

Recent eruptions indicate that Ubinas poses a severe threat to at least 5,000 people living in the valley of the Rio Ubinas, and within a 15-km radius of the summit. The threat includes thick tephra falls, phreatomagmatic ejecta, failure of the unstable south flank with subsequent debris avalanches, rain-triggered lahars, and pyroclastic flows. Should Plinian eruptions of the size of the Holocene events recur at Ubinas, tephra fall would affect about one million people living in the Arequipa area 60 km west of the summit.

List of data repository

Data Repository 1

Sketch map of the Ubinas volcano and surroundings showing the location of the analysed lavas and tephra. In addition, dated ⁴⁰Ar/³⁹ Ar lavas (bold numbers) and tephra-fall deposits (¹⁴C) are shown, as well as the sites of the drilled peat cores (see Fig. 2).

Data Repository 2: sample preparation and ⁴⁰Ar/³⁹Ar dating method

Four biotite (samples Ubi 46, Ubi 71, Ubi 120, and Ubi 127) and two amphibole (samples Ubi 122 and Ubi 123) concentrates and a whole rock (Ubi 122) were used (see Fig. 2 for location). In order to separate the mineral concentrates, the rocks were crushed and sieved through the 800–500 and 500–300 μm. The minerals were concentrated using bromoform and eventually selected under binocular microscope. The mineral concentrates were wrapped in copper-foil packets (furnace experiment) and in aluminum-foil (laser experiment) and irradiated in the 5C position at the nuclear reactor of McMaster University (Hamilton, Canada). Irradiation lasted for one hour with cadmium shielding except for the biotites of the sample Ubi 127, which were irradiated two hours. The samples were associated with the Fish Canyon sanidine (FCS) as a neutron flux monitor of the reactor (J-value determination) assuming a FCS age of 28.02 Ma (Renne et al. 1998). A whole-rock thin slab of 205 mg was also cut from the sample Ubi 122 and irradiated for two hours.

Age determination was performed in the Laboratory of Geochronology of UMR Géosciences Azur in the university of Nice, France. Step heating of individual grains of biotite were carried out using a 50-W Synrad CO₂ laser. Each step lasted 5 min, including 1 min for heating and 4 min for cleanup of the released gas, before introducing the gas in the spectrometer. Isotopes were measured statically with a VG3600 mass spectrometer working with a Daly detector system. The gas extractions for bulk samples (amphiboles and biotites) and whole rock were performed in double-vacuum, high frequency heated furnace, connected to a stainless steel purification line and analyzed with a mass spectrometer composed of a 120°M.A.S.S.E tube, a Bauer-Signer GS98 source, and a Balzers electron multiplier.

Ages were calculated from measured isotope ratios corrected for mass discrimination, system blanks, and interfering isotopes produced during irradiation. For the laser experiment, blanks routinely measured every three steps, were in the range 40–90, 2–10, 2–6×10⁻¹⁴ ccSTP for the masses 40, 39, and 36, respectively. The argon isotopes were on the order of up to 2000, 5–100, 2–100 times the blank levels, respectively, in part related to the weight of the analyzed mineral.

During the furnace experiments, heating lasted 20 min for each temperature step followed by 5 min for cleanup of the released gas, before introducing the gas in the spectrometer. Argon isotopes were of the order of 100–2,000, 100–1,000 and 2–200 times the blank levels for the masses 40, 39, 36, respectively. In some high temperature steps, measured ³⁶Ar was near the blank level. Mass discrimination was monitored by regularly analyzing air pipette volume and averaged 1.00721±0.19% over a 2-year period for the laser system and was ranging from 1.00740 to 1.00606 (±0.15%) for the HF furnace system.

The criteria used for a plateau age are: (1) the plateau should contain at least 70% of ³⁹Ar, (2) the plateau should include at least three following steps of temperature, and (3) the integrated age of the plateau should concur within 2 σ with each apparent age of the plateau.

For each biotite sample, single grains were fused with laser to check the possibility of heterogeneity of the sample. Each single grain of biotite was first moderately heated (one or two steps) before fusion in order to release part of the atmospheric contamination. The aim of this procedure was to reduce the atmospheric contamination before the fusion step. However, this procedure was not very efficient and in most of the sample, all the steps include high level of atmospheric contamination. Generally, we retained the age of the fusion step, as it includes an important part of the total ³⁹Ar released but when the sample released ³⁹Ar in the lower temperature steps, we calculated a plateau age. Integrated (total gas) ages appear in the data table but were not used for the discussion.

Data Repository 3: ⁴⁰Ar/³⁹Ar analytical data

Laser experiments: high-temperature (fusion) step ages are generally preferred but some plateau age are also used (see text for explanation).

Table 4

³⁹Ar (%) = fraction of ³⁹Ar released for each step; ³⁷ArCa/³⁹ArK = Ar isotopes produced by Ca and K neutron interferences; ⁴⁰Ar* = radiogenic ⁴⁰Ar. The error bar is at the 1s level and does not include the error of the J irradiation parameter (estimated at 0.2%). Corrections factor for the interfering isotopes produced by nuclear reactions on potassium and calcium in the McMaster reactor were $\left( {^{39} \text{Ar}/^{37} \text{Ar}} \right)\text{Ca} = 7.06 \times 10^{ - 4}$, (±4%), $\left( {^{36} \text{Ar}/^{37} \text{Ar}} \right)\text{Ca} = 2.79 \times 10^{ - 4}$ (±3%) and $\left( {^{40} \text{Ar}/^{39} \text{Ar}} \right)\text{K} = 1 \times 10^{ - 3}$ (±4%). Isotopic ratio were corrected for blank and mass discrimination (1.00721±0.19% for the laser system and 1.00740 to 1.00606 (±0.15%) for the HF furnace system. Ages were calculated using the decay constants proposed by Steiger and Jaeger (1977).

Representative electron microprobe analyses of phenocrysts in the Ubinas rocks. Fe₂O₃* calculated after Papike et al. (1974) and Tindle and Webb (1994); End-member components of “quadrilateral” pyroxenes (mol%) normalized to atomic $\text{Ca} + \text{Mg} + \Sigma \text{Fe} = 100$ with $\Sigma \text{Fe} = \text{Fe}^{2 + } + \text{Fe}^{3 + } + \text{Mn}$ (Morimoto et al. 1988); Mg-no. = 100 Mg/(Mg + Fe), atomic; mght magnesio-hastingsite; c phenocryst core; r phenocryst rim.

Table 5

Table 6

Table 7

Table 8

Table 9

Table 10

Table 11