A shifty Toba magma reservoir: improved eruption chronology and petrochronological evidence for lateral growth of a giant magma body

Abstract

Polycyclic caldera complexes hold clues to understanding why some magmatic systems develop into supersized magma bodies and how they can recover to produce several caldera-forming eruptions. However, the geologic records of the transitions between successive caldera events are very often inaccessible due to limited preservation of eruptive products of inter-caldera activity, prompting the search for alternative archives of magma evolution such as accessory minerals. Here we applied multiple geochemical tools to study one of the most active caldera centres of the Quaternary, the Toba caldera complex in Sumatra (Indonesia), which produced at least four caldera-forming eruptions in the last 1.6 My, including the iconic Youngest Toba Tuff at 74 ka. We combined feldspar 40Ar/39Ar and zircon U–Pb geochronology of proximal pyroclastic deposits with glass and mineral chemistry of both the tuffs and distal marine tephra to revise the eruption chronology of Toba, obtaining new eruption ages of 1417 -31/+14 ka (zircon) or 1339 ± 39/39 ka (plagioclase, internal/full external 2σ uncertainty) for the Haranggaol Dacite Tuff, 783.81 ± 0.85/1.32 ka (sanidine) for Oldest Toba Tuff, and 503.61 ± 1.36/1.50 ka (sanidine) for Middle Toba Tuff. Isotope dilution thermal ionisation mass spectrometry (ID-TIMS) U–Pb crystallisation ages, trace element contents and Hf isotopic ratios of zircons illuminate changes in the shallow magma reservoir which saw near-continuous zircon crystallisation over 1.6 My. Prolonged build-ups to each eruption with highly scattered zircon trace element compositions reflect a complex, heterogeneous character of the shallow reservoir, without a clear temporal trend or indications of the eruption trigger. In contrast, hafnium isotopes in zircon display a pronounced shift towards unradiogenic values immediately after the OTT caldera collapse, followed by a gradual recovery to a baseline value of εHf = -7 at the time of YTT eruption, interpreted as a reflection of the shift in magma reservoir position corresponding to change in the character of assimilated crust. We can show in unprecedented detail how a large caldera collapse affects magma geochemistry; however, identification of patterns in the behaviour of the Toba system and making geochemistry-based predictions about its future development remain a challenge.