Lake Toba is a beautiful place. The largest and deepest lake in Southeast Asia, it contains an island (Samosir) almost the size of Singapore – the fifth largest lake island in the world, and the largest island-within-an-island. Samosir contains two smaller lakes, Lake Sidihoni and Lake Aek Natonang – the former even contains its own island. The best views of Lake Toba and Samosir, in the highlands of Northern Sumatra, are from Tele. Most tourists end up on the TukTuk peninsula. The tourist blurb calls it a place ‘to sit back, relax, and absorb’, and ‘a beautiful place to do nothing at all’. But this tranquility hides a violent secret. Lake Toba is the place of the largest volcanic eruption on Earth of the past 2 million years.
The lake is 100 by 30 km in size, has a surface area of 1100 km2, and a water volume of 240 km3. It is up to 450 m deep. The volume makes it the 24th largest lake in the world (admittedly this list excludes the Caspian Sea), and bigger than the Dead Sea. The full extent traces a caldera with a staggering size. Toba, together with Yellowstone, Long Valley, Valles Caldera, Taupo Volcano, and Aira Caldera, is one of the ‘secret six’: the current supervolcanoes. (But Aira seems too small for this list!) Of these six, Toba is the most densely populated. Samosir, the centre of the massive Toba caldera, is home to more than 300,000 people. In contrast, the town of Taupo only has a population of 25,000. The term ‘supervolcano’ was first introduced to describe Toba. It is now defined as a volcano capable of ejecting over 1000 km3. Since Toba, only the Taupo Oruanui eruption in New Zealand (26,500 years ago, 530 km3 DRE) has come close.
The region of lake Toba has suffered four major eruptions, starting 1.2 million years ago from a large strato-volcano at the northwest end of Lake Toba (35 km3 DRE), than 840,000 years ago towards the south of Lake Toba (500 km3 DRE, possibly more), 500,000 years ago at the northern end of the lake (80 km3 DRE), and the fourth and final eruption occurred 74,200 +- 900 years ago from multiple locations, perhaps the entire surrounding ring fracture (2800 km3 DRE). The eruptions are named after the tuff layers they produced: the Haranggoal dacite tuff (HDT), the oldest (OTT), the middle (MTT), and the youngest Toba tuff (YTT). (The naming lost some of its logic when a layer older than the oldest was found.) The first three all created calderas, but the final eruption was much larger and its caldera encompasses the earlier three. Nowadays the region is considered inactive, but the volcanic roots are still apparent in the hot (or even very hot) springs, on the west of the island. (Hot springs seem a fairly common occurrence in supervolcanoes.) There is on-going activity in the wider area: Mount Sinabung is only 40km from the edge of the lake.
Lava (or rather, tuff) from the YTT eruption covers an area of more than 20,000 km2 with a typical thickness of 50 meter (but up to 400 meter in some places and more than 600 meter inside the caldera); the flows reached the ocean on both sides of Sumatra. The lack of ash at the bottom of the lava indicates that there was no plinian phase at the start of the eruption: there was fountaining probably to great height but but no large explosions, and the velocities were on the low side. The distant flows came from silica-rich part of the magma chamber, while the deeper silica-poor magma mainly filled the caldera and was probably erupted while the caldera collapse was in progress.
The distant ash blanket from Toba covers an area estimated at 7 million square kilometre (close to the area of the US) and has been found as far away as the Arabian Sea and Lake Malawi. A lot ended up in the Indian ocean. The average thickness of the ash layer is 10 cm, but in Malaysia, 350 km away, it averages 90 cm and is up to 3 meter in places, and in India, 3100 km away, a thickness of 50cm – 1m is common and in one place a thickness up to 6 meter has been reported. The thickest layers were probably concentrated by rivers. The ash shows two layers: the lower one is coarse-grained and the upper one more fine-grained unit. The lower layer is not seen in the more distant locations. This suggests that the ash was produced in two phases. The total volume of the ash corresponds to 800 km3 DRE.
The dispersion of the ash shows that the eruption took place during the monsoon season, in the northern summer. The winds at that time blow towards India. A winter eruption would have deposited the ash further south. The duration of the eruption is only known approximately, from the structure of the ash layers in the ocean at different depths. This has shown that the ash producing phase lasted between 9 and 14 days. For comparison, the main eruptions of Tambora and Krakatoa both lasted 24 hours, although Krakatoa also had a lesser eruption for the preceding 4 months.
Toba erupted a volume estimated at 2800 km3 DRE, including about 1000 km3 of lava. Imagine this: Bardarbunga took 6 months to erupt 1 km3. Laki, while suffocating Iceland, erupted 1 km3 every two weeks. Toba erupted this much every 10 minutes! Its eruption rate was ten times faster even than that of Tambora.
Putting this together, we can get a rough idea how the eruption proceeded. The huge magma chamber was located some 10 km below the surface. It had been present for 150,000 years, and during this time crystallization had occurred. The crystals sank to deeper levels, leaving the upper part of reservoir silica-rich and the bottom silica-poor and crystal-rich. Eventually, the heat began to weaken and melt the roof of the magma chamber, and buoyancy of the magma pushed up the roof, until eventually it started to crack. Once the melt fraction reaches 50%, the overpressure of a huge magma chamber can spontaneously crack 10km of rock: unlike normal eruptions, supervolcanoes do not need a new influx of magma to initiate an eruption. The cracking may have been preceded by major inflation (perhaps hundreds of meters, albeit over a long time), and finally caused big earthquakes. In Krakatoa, a big earthquake happened three years before the eruption.
The first crack provided a small outlet for the first lava. As magma escaped, the changing pressure below caused more cracks and the eruption rapidly intensified. Lava began to come from other places along the ring fracture, and finally, sometime in July or August, along the entire ring. It is a guess how long this initial phase would have lasted; for Krakatoa, it took several months. There was probably no central eruption: all came from the caldera ring. As has been pointed out in this blog, very large eruptions can have multiple exit points simultaneously. The lava now fountained out at incredible rates, dwarfing anything the world has seen since. Fountain collapse, mostly during the early and middle part of the eruption, gave rise to pyroclastic flows traveling hundreds of kilometers. The lava and tuff eventually covered the width of Sumatra to a depth of 50 meters or more. The cliffs around the lake, 300 meters high, were also created from these. There may have been one or more very large explosions at this time, with an eruption column tens of kilometer high and depositing ash across many parts of the Indian ocean, but perhaps not reaching India. However, this is disputed: some people see little or no evidence for such high eruption columns and think all the lava was erupted at low velocity.
The heat of the pyroclastic flows propelled smaller ash particles into the air, up to 10 kilometer high (these are called secondary or co-ignimbrite plumes, and are known from Mt St Helens). This ash traveled far and wide, following the trade winds, and blanketed nearby Malaysia but also India. Most of the ash came from this.
Eventually the magma reservoir had lost so much pressure that the ridge began a rapid collapse, within days to weeks forming a hole 2 kilometer deep. As the caldera collapsed deeper the eruption began to wane and the ejecta now stayed mostly within the caldera. By this time, the magma came from the bottom of the original reservoir which had grown silica-poor. The eruption finally ended with a whimper.
The volcano now remained quiet for a very long time. Perhaps 40,000 years after the eruption, the magma reservoir began to refill and the centre of the caldera began to inflate. Very small eruptions occurred, mainly along the ring fault, mostly from old magma remnants but in one or two cases new magma also (just) made it to the surface and formed some cones. The rise became a resurgent dome, formed an island, and is now called Samosir. It has risen by some 1100 meters. The large magma chamber underneath very closely follows the contours of Samosir. There is another, smaller magma chamber to the north, partly underneath the lake and partly west of the caldera. Samosir is no longer rising.
Why did Toba become a supervolcano?
Why did Toba become so large? This is not really known: something must have allowed a very large magma reservoir to build up, relatively close to the surface. Once you get a very large, single reservoir, it can force an eruption through slow pressure. Small reservoirs, often divided into honeycomb chambers do not do this. It has been argued that volcanoes can slowly develop these huge chambers as they grow old, so that all it takes is a constant, slow magma supply over very long periods. But at the same time, the new magma should enlarge the reservoir rather than erupting at the surface. In the case of Toba, we can make a guess what may have caused this.
All Indonesian volcanoes, including Toba, are fed by subduction, and not by a hotspot. Unusually, the Sumatran volcanoes are also located very close to a fault. The Sumatra fault runs along the entire length of the island, including the western edge of the Toba caldera. This fault runs parallel to the off-shore Sunda megathrust fault (the location of the 2004 earthquake), where the Indian-Australian plate subducts underneath the South-East Asian plate. In Sumatra, this subduction occurs not perpendicular to the faults, but at an angle: while the plate slides underneath, it also moves to the side, northward. Some of this slip motion occurs along the Sumatra fault. This can in fact be seen in the YTT ejecta: on either side of the fault, they are displaced by 2 km. The fault must therefore be moving at 2.7 cm per year. The slip is faster to the northwest of Toba than it is in the south. This causes a problem, as over time a gap is created. This currently happens just south of Lake Toba. The close association of a fault and a volcanic arc in Sumatra is unusual and may even be a coincidence. But this thinning of the crust could be one reason for the very large magma reservoir underneath Toba: it continuously creates space (or a region of reducing pressure) for the magma to move into.
The effects of the eruption are often discussed but not that well known. Certainly, much of Sumatra would have been covered by lava, poisoned by gases, and sterilized by pyroclastic heat – a hell worthy of Dante. The ash would have wiped out most life to the west and northwest, luckily largely oceanic regions but including Malaysia and part of India. Other places were spared, such as Indonesia further to the east and the Philippines, due to the luck of the prevailing winds. The effect on humanity is not clear. If air travel had been discovered by than, it would have had to be abandoned for perhaps years. But that would have been the least of their worries, as it is argued that most of the emerging Homo Sapiens were killed off, leaving only a few thousand people to populate the world. There is some genetic evidence for such a bottle neck. However, it is not known whether this bottle neck coincided with Toba. In India, a few stone tools below the Toba ash are similar to the (many more) tools above it, suggesting that the culture survived – even if the local population did not. The Earth’s climate and the monsoon would have been shot for years to decades, and famine inevitable, but the amount of global cooling depends on the amount of sulphur which was ejected which is very uncertain. Cooling of between 1 and 6C has been estimated but different authors.
Were the climatic consequences even worse? The eruption happened almost exactly at the onset of a deep glacial period lasting 1000 years. Toba could not have caused this by itself: remember that volcanoes cause cool summers but not cold winters. But Toba could have caused a change in the ocean currents. This may have pushed the Earth over the brink into a new but short-lived ice age. This is pretty speculative but not impossible; however the ‘Toba catastrophe theory’ contains a fair amount of hyperbole. But if this event had happened in the modern era, it would indeed have been catastrophic. The casualties would not have been counted in the millions, but in hundreds of millions. Our society is not resilient to a Toba-sized event.
Luckily, eruptions the size of Toba are expected only about once every million years. A low VEI8 (as in Taupo) may occur every 100,000 years. A low VEI7 may happen once a century or so. From the point of risk management, we should be prepared for one-in-ten-thousand-year events. That corresponds to roughly a medium VEI7, of 100 km3 DRE. Toba was completely off this scale.
The risk of another large eruption from Toba is currently extremely small. The repose times have been fairly consistent around 400,000 years, and the next one is therefore not due until around AD 300,000 (but the sizes of the eruption are highly variable and unpredictable, and do not always reach the defined limit for supervolcano size). AD 300,000 is a very long time away! Yellowstone has even longer repose times than Toba (600,000 years on average), and Valles Caldera is similar to Toba at 335,000 years. Supervolcanoes do it more than once, but not frequent.
Another way to look at this is by using average eruption rates. Averaged over millennia, oceanic arc volcanoes erupt typically 0.002-0.005 km3 per year. Toba’s average rate is 0.005 km3 per year. It became a supervolcano not because it produces magma faster than normal volcanoes, but because its magma chamber has a much larger storage capacity which can cope with the long times between eruptions. The road to destruction is traveled slowly.
Lake Toba is a tranquil place, far from the hustle of modern life. But what a past.