A burning topic for Bronze Age archaeologists, such as the delightful  Bettany Hughes – biographer of Helen of Troy, is the explosive collapse of the volcano Thera (modern Santorini) whose distant effects (ash and tsunamis)wiped out the Minoan civilisation of Crete around 1600 BCE, giving rise to Plato’s legend of Atlantis. It was a big one alright, hurling of the order of 60 km³ of pulverised magma skywards, though not the largest historic eruption: that involved 160 km³ from the Tambora volcano on Indonesia’s island of Sumbawa in 1815. The inhabitants of Santorini simply disappeared, after evacuating their homes during precursor earthquakes and small eruptions, which were then buried beneath many metres of tephra when Thera literally ‘blew its top’. Little ash fell on Crete, yet its northern coast shows clear signs of a major tsunami. The reason for such an engulfing wave is revealed by the nature of Thera’s eruption: after evacuating magma, the edifice collapsed to form a caldera clearly revealed by the elliptical bay around which the remnants stand as the various islands of Santorini.  Caldera formation would have displaced vast amounts of sea water.

Santorini has been well studied by volcanologists, still being an astonishingly awesome spectacle as well as preserving the full record of the eruption and the archaeology that it buried (http://santorini-eruption.org.uk/). Empirical research reveals four distinct eruptive phases probably over a period of a few months. The explosive force of the final catastrophe probably resulted from seawater reaching the sub-volcanic magma chamber: not a difficult feat of imagination. What has not been known is how the magma evolved over times leading up to the cataclysm, and that is a knotty issue for all volcanoes that pose a major threat because of evidence for repeated and perhaps cyclic activity. A new technique is now capable of lifting the veil on such purely magmatic evolution, and is based on the changes that took place in minerals that crystallised over lengthy periods while the magma cooled slowly at depth but was periodically added to (Druitt, T.H et al. 2012. Decadal to monthly timescale of magma transfer and reservoir growth at a caldera volcano. Nature, v. 482, p. 77-80).

Such phenocrysts are commonly found in fragments of pumice that make up Theran tephra, and they are commonly zoned in a concentric fashion, especially those of the mineral feldspar, each zone marking a phase of growth that occasionally traps samples of magma in the form of now glassy inclusions. The zones mark chemical changes in the magma as new pulses are added in the sub-volcanic chamber, and sometimes temperature changes and loss of gas. Although the zone boundaries a are expected to be sharp in terms of chemical differences, in practice they are blurred as a result of element diffusion at high temperatures. Diffusion is a predictable process and so the degree of blurring indicates the time at which a new zone formed relative to that of eruption and cooling, when diffusion would have stopped abruptly. Rates of high-temperature diffusion depend on the element concerned. So using a suite of trace elements in feldspar zones gives a variety of chronometers. A fast-diffusing element such as Mg can chart changes of the order of decades to months, while a more sluggish trace element – for instance titanium – can examine changes on longer timescales.

The results obtained by the authors present a surprise: although Thera had last erupted catastrophically 18 ka previously, additional magma recharged the volcano only in the last few decades before it extinguished life on Santorini and set the Minoan civilisation on a downward spiral. Indeed, magma continued to be added even in the last few months. Calderas, such as that at Yellowstone in the western US, to which are linked ancient ash layers covering areas hundreds and thousands of kilometres away, pose threats as large and even bigger than Thera. If Thera is anything to go by, they lie in repose long after an eruption and signs of recharge may herald eruption in the near future. The Yellowstone caldera, that has lain dormant for 640 ka is indeed showing signs of magmatic ‘stoking’, as the Earth’s surface there is slowly bulging. It produced ‘supereruptions’ that dwarfed Thera at 2.1 Ma (2500 km³), 1.3 Ma (280 km³) and 0.6 ka (1000 km³). For each of these and several other calderas there are abundant tuffs that carry phenocrysts, whose zonation is yet to be checked for signs of past behaviour by their local magma chambers.

Related articles

• Caldera Eruption “Early Warning System”? Not so Fast. (wired.com)
• Ami Iida: Decadal to monthly timescales of magma transfer and reservoir growth at a caldera volcano http://www.nature.com/nature/journal/v482/n7383/full/nature10706.html (friendfeed.com)
• Santorini comes under close observation after Crete rattled by second 5.2 earthquake (theextinctionprotocol.wordpress.com)