The large late-glacial Ho eruption of the Hudson volcano, southern Chile

Abstract

Lakes formed in the Aysen region of southern Chile after the retreat of mountain glaciers, established by ~17,900 calendar years before present (cal years BP) or earlier, contain numerous late-glacial and Holocene tephra layers derived from >70 eruptions of the volcanoes in the region, including Hudson, the southernmost in the Andean Southern Volcanic Zone (SVZ). Sediment cores from seven of these lakes contain an unusually thick late-glacial age tephra layer, which based on its distribution and bulk trace-element composition was derived from a large explosive eruption of Hudson volcano between 17,300 and 17,440 cal years BP and is termed Ho. In 13 cores from six of these lakes, each located ~100 km generally northeast of Hudson, the Ho tephra layer ranges between 50 and 88 cm in thickness, and contains pumice grains up to 2 cm in maximum diameter. Comparison with three previously documented large explosive Holocene Hudson eruptions (H1 at 7,750 cal years BP, H2 at 3,920 cal years BP, and H3 in 1991 AD) suggests that Ho was larger, with an estimated tephra volume of >20 km3, the largest post-glacial eruption documented for any volcano in the southern Andes and most likely responsible for the formation of the Hudson caldera. In total, Hudson has erupted ≥45 km3 of pyroclastic material in the last ~17,500 years, making it the most productive volcano in the southern Andes in terms of the total volume erupted since the beginning of deglaciation in the region. Chemical stratification is not seen in the waterlain Ho tephra, but these deposits are bi-modal, consisting of a much greater proportion of dark glassy basaltic-trachyandesite dense fragments and pumice, with glasses which range between 55 and 59 wt.% SiO2, along with volumetrically less-significant lighter-colored trachydacite pumice, with glass of 66 wt.% SiO2. In contrast, H1 products are trachyandesitic in composition, H2 ones are more felsic than H1, being composed essentially of trachydacite, and although H3 1991 AD again produced tephra of bi-modal compositions, it erupted a much smaller proportion of mafic compared to felsic material than did Ho. Thus, the repetitive large explosive eruptions of Hudson volcano have evolved to progressively less-mafic overall compositions from late-glacial to historic times, and their volumes have decreased. Sr-isotopic composition of bulk samples of the most mafic dense glass and most felsic pumice components of the Ho tephra, as well as samples from other Hudson eruptions, which overall range from 51 to 66 wt.% SiO2, with 525 to 227 ppm Sr, are all similar (0.70444 ± 0.00007), indicating that crystal-liquid fractionation rather than crustal assimilation was the main process responsible for these chemical variations.