This is a re-post that has been re-posted before. But we attract new readers and sometimes it is worth republishing something. Especially since this is about Icelandic (and other) calderas, something that has been raised in the comments recently. This post was originally written during the Holuhraun eruption and associated Bardarbunga collapse, and you will find references to this, in part dated. Over to Henrik.

Volcano? Erm... I don't see a volcano? Toba caldera wall seen from Samosir Island, a resurgent dome near the middle of the caldera. (Photo, Sebastian Hubarat, Tobaexplorer)

Volcano? Erm… I don’t see any volcano? Toba caldera wall seen from Samosir Island, a resurgent dome near the middle of the caldera. (Photo, Sebastian Hubarat, Tobaexplorer)

It does not take long for a newcomer to volcano-watching, if we are to call our hobby that, to come across the term “caldera”, meaning cauldron. The term is very loosely used to describe large volcanic depressions ranging from volcanic craters a few hundred metres in diameter up to the huge depressions left behind by the very largest volcanic eruptions that may approach 100 km in their greatest extent. Consciously or subconsciously, there is always a sensationalist undertone evoked in the reader when the term caldera is used. So, what is a caldera?

The archetypal caldera is Yellowstone. For many decades, scientists had realised that Yellowstone was a large volcanic feature with its geyser fields, tuffs and lava flows, but where was the volcano? There was no sign of it! It wasn’t until the advent of space flight and high resolution orbital photography that the truth was finally realised: Yellowstone and its major explosive eruptions were so huge that the volcanic scars left were too large to leave a recognisable volcano behind. Almost wherever one is in Yellowstone National Park, one is standing on top of the volcano. Let us just add immediately that the vast majority of its volcanic or volcano-related eruptions are nowhere near huge, that there is no periodicity to the largest, caldera-forming eruptions and that Yellowstone is in no way “overdue”, whatever Michio Kaku or Steve Quayle may say. Any eruption from Yellowstone is currently (this millennium) exceedingly unlikely.

The trail of "supervolcanic" eruptions across the state of Idaho and through the Rockies known as the Snake River Plain (ISU)

The trail of “supervolcanic” eruptions across the state of Idaho and through the Rockies known as the Snake River Plain (ISU)

With this epiphany, the scars left behind by several more such very large eruptions were identified. West of Yellowstone running through the state of Idaho and cutting a swathe through the Rockies lies the Snake River Plain, a trail of at least seven calderas left behind by several ultra-colossal (VEI 8), nowadays sometimes referred to as apocalyptic eruptions, and mega-colossal (VEI 7) eruptions dating as far back as 16 million years as the North American plate travelled over the Yellowstone hotspot. It can be traced back to the Columbia River Flood Basalts in Oregon, a large igneous province formed over a period of 10 – 15 million years and most vigorous 17 – 14 million years ago that deposited some 174,300 km3 over an area of 163,700 km2 with a maximum thickness of over 1.8 km. This is not the only area in the USA where such eruptions have occurred. Long Valley in California was home to a mega-colossal VEI 7 eruption about 760,000 BP. The Valles Caldera, New Mexico, was formed by two VEI 7 eruptions some 1.15 million years ago. What has long been regarded as the largest eruption ever identified is the massive VEI 8 eruption of 27.8 million years ago that formed the La Garita Caldera and Fish Canyon Tuff in Colorado.

Interesting though the American caldera-forming eruptions are, they are not unique, nor are they the most interesting. The Atana Ignimbrite eruption (VEI 8) that formed the Pacana Caldera in Northern Chile some 4.0 million years ago was as large as the largest of Yellowstone’s eruptions. Lake Toba in Sumatra Indonesia was even larger, occurred as recently as 74,000 BP, and was somewhat controversially linked to a bottle-neck in the human genome. The Toba Catastrophe Theory claimed that as many as 60% of all humans on Earth at the time could have perished. The evidence for this has evaporated over time. The 74,000 BP Toba Eruption was preceded by at least one other VEI 8 and two VEI 7 eruptions dating as far back as c 1.2 million years. Subsequent to the most recent eruption, a resurgent dome, Samosir Island, has appeared.

Beautiful Lake Taupo, New Zealand. The only hint of its volcanic origin are the distant volcanoes on the horizon, built at the edges of the ring fault system. (Provenance unknown)

Beautiful Lake Taupo, New Zealand. The only hint of its volcanic origin are the distant volcanoes on the horizon, built at the edges of the ring fault system. (Provenance unknown)

No disrespect to the USA, but the undisputed World Champion of ultra-colossal and mega-colossal eruptions is the Taupo Volcanic Zone located on New Zealand’s North Island. Over a period from 280,000 BP to as late as 1,800 years ago, the TVZ has seen no less than two VEI 8 and five VEI 7 eruptions. The latest of those two are the Oruanui Eruption (VEI 8) 26,500 BP and the Hatepe Eruption (VEI 7) 1,800 BP of Lake Taupo.

Other recent mega-colossal eruptions that formed large calderas include the 3,600 BP civilisation-destroying eruption of Thera (Santorini) from which we have derived words such as “terrible”, “terrific” and the Tera of terawatt/terabyte as well as the legend of Atlantis. It may well be mentioned in the Biblical description of the Plagues of Egypt. Two further recent such eruptions are the Kikai Caldera, Ryukyu Islands, Japan (6300BP)and Macauley Island in the Kermendec Islands, New Zealand(6,300BP) eruptions.

But there have also been two recent VEI 7 eruptions that did not lead to a caldera collapse even if the huge summit craters, partially the result of collapse, are often referred to as “calderas”, by scientist – Mount Baekdu/Changbaishan (945 AD) North Korea-China border and Tambora, Sumbawa Island, Indonesia (1815 AD) that caused “The Year Without a Summer”, 1816.

Mount Baekdu or Changbaishan on the Chinese - North Korean border, source of the 969 AD VEI 7 eruption (Wiki)

Mount Baekdu or Changbaishan on the Chinese – North Korean border, source of the 945 AD VEI 7 eruption (Wiki)

These very largest of volcanic eruptions begin with an initial eruption that removes a sufficient amount of magma that the ~5 ~ 10 km thick roof over the magma reservoir begins to buckle and drop. This forms a series of concentric ring fractures which eventually begin to erupt after which the now unsupported roof collapses into the magma reservoir leading to the major and usually tuff-forming ignimbrite eruption which partially fills the caldera. Later eruptions and erosion further fills in the caldera and intrusions thousands of years later may cause a resurgent dome to emerge near the middle. Note that neither Baekdu nor Tambora fit this pattern for caldera formation very well as the progenitor volcanoes survived!

In comparison to these giant scars, often tens and several tens of kilometres in extent, it seems ridiculous to refer to the 1½ km wide, shallow summit crater of Iceland’s Eyjafjallajökull, a volcano that gently spews out no more than some 0.1 cubic kilometres or so over a period of months with its eruptions, as a “caldera”, but even scientists sometimes do in serious academic papers.

The 3 x 4½ km summit caldera of Mount Katmai, Alaska, that formed after the 1912 VEI 6 Novarupta eruption (Wiki)

The 3 x 4½ km summit caldera of Mount Katmai, Alaska, that formed after the 1912 VEI 6 Novarupta eruption (Wiki)

But the very large, ultra- and mega-colossal, eruptions are not the only eruptions to form collapse calderas. In addition to this, there are relatively speaking smaller eruptions that form them, mostly from the colossal VEI 6 eruptions. The most recent such example is the 1912 VEI 6 Novarupta eruption when the magma reservoir under the Katmai volcano, Alaska, emptied through a new vent well to the side of the volcano to form the “Valley of Ten Thousand Smokes”. Over the next few years, much of the edifice of the Katmai volcano collapsed into the emptied magma chamber to form a 3 x 4½ km summit caldera.

A better known and researched caldera-forming eruption is that of Mount Mazama (7,700 BP), Oregon. This followed the pattern of a ring fault forming followed by an eruption at the ring fault and the collapse of the volcanic edifice into the relatively shallowly located and emptied magma reservoir. The caldera has been filled with water to form the 8.0 x 9.7 km diameter Crater Lake and, in contrast to the non-eruptive resurgent domes of very large calderas, subsequent eruptive activity has built a central cone, Wizard Island.

Topographical map of the Crater Lake caldera showing the features of subsequent intra-caldera eruptions. (USGS)

Topographical map of the Crater Lake caldera showing the features of subsequent intra-caldera eruptions. (USGS)

Calderas such as these are plentiful. Europe alone, perhaps not noted for great volcanic eruptions, has several; Laacher See in Germany, Vulsini, Colli Albani, Roccamonfina in Italy, the Rodalquilar Caldera complex in Spain to name but a few. In addition there are several ancient ones such as the caldera complex upon which Oslo, the capital of Norway, is built, Mount Snowdon, the Lake District, UK, etc. etc.

But there is one type of caldera that defies the traditional description and which is impossible to tie down to a single, specific event and that is best exemplified by the large Icelandic calderas of subglacial origin, i.e. Katla and Askja. While it is thought that the calderas of both Torfájökull and Tindfjallajökull are the result of single large, possibly VEI 6 eruptions, some 50-70,000 years ago, those of Katla, Askja and the volcanoes below both Hofsjökull and Langjökull are much harder to explain in those terms. The calderas of both Askja and Katla are far larger than can be explained by eruptions known to have occurred from analysis of ash strata. Both volcanoes seem to be limited to no more than medium-sized VEI 5 eruptions whereas in both cases, at least a low-end VEI 7 – which we know never has happened – would be required to produce their large calderas. It is here that what we are now seeing at Bardarbunga might have a bearing.

Aerial photograph of the sub-glacially formed Askja volcano with its large caldera containing the nested caldera of Lake Öskjuvötn, which resulted from post-eruption collapse following the volcano's 1875 VEI 5 eruption (RUV)

Aerial photograph of the sub-glacially formed Askja volcano with its large caldera containing the nested caldera of Lake Öskjuvötn, which resulted from post-eruption collapse following the volcano’s 1875 VEI 5 eruption (RUV)

That Bardarbunga is now undergoing a collapse event is beyond question. Furthermore, to judge by previous eruptions in the area emanating from or involving this central volcano, this is not the first such time in its history. Until this eruption started, the caldera as shown on the IMO maps was significantly smaller but was enlarged to its current extent by human reinterpretation. The earthquake data strongly suggests that a collapse event is happening at the N to NNW margin of the newly defined caldera where the data indicates that the deep magma reservoir extends beyond the limits of the “rescheduled” caldera. If this continues, the caldera will eventually be enlarged in this direction.

If this is indeed what is currently happening and has happened before, we may have obtained a vital clue as to how and why the Icelandic subglacial calderas defy the correlation between the size of the caldera and the eruptions that formed them.


Further reading

34 thoughts on “Calderas

    • I love calderas too. Also highly active subduction zones.

      Sinabung went boom again yesterday.

  1. Re; Crater lake in my area some 600 km from Crater lake, we have 3ft of Crater lake ash in the soil horizon.added to the clay soil of the Grande Ronde valley the blue clay that is the nemesis of any farmer and well driller
    in this area. I won’t go into the Columbia River Basalt..
    Got Boulders?..

  2. Sinabung went boom, plenty of small videos on Twitter. Probably not caldera forming tho.

    • The local people are presumably continually breathing in low levels of ash (ie local dust) between eruptions. I would expect this to produce chronic effects, do we know if these happen?

  3. Volcanodiscovery he say:
    Today at 10:16 local time Volcanic Ash Advisory Center (VAAC) Darwin recorded a powerful vulcanic eruption showering dense dark ash plume which reached approx. 30,000 ft (9,100 m) altitude. Volcanic ash is dispersed towards the west and causes large ash fall onto the volcano’s slopes and resident area.
    The warning bulletin states that ballistic impacts of volcanic bombs and pyroclastic flows could affect an area of about 3 km distance from the main crater and 5 km on the SE flank and 4 km on the NE flank.
    I’m sure there should be a report from malasian vulcanology dept but I can’t see one.

  4. Question: is it possible that the perpetual Pahala earthquake swarm is caused by the inflation of Mauna Loa? The increased weight of the mountain pushes the sides out, and this can have unlocked the fault at the base of the mountain. Just an idea. There is obviously magma involved, but it still needs a pathway before it can move.


      on 3/45 is a map of the yellowstone tectonic parabola. I have seen a better version of it. Just guessing that the Hawaii hotspot has something similar. My metal vision of these hotspots is a bulldozer slowly moving forward with a perpetual supply of gravel with the base in front fracturing from the weight as it moves forward. The sides and back of the blade are constrained by material that is already deposited and the easiest motion is forward. Basically creating a perpetual motion machine. Uhm lol

    • To me, that swarm is a wonder. In addition to the magnitude and frequency of quakes, it has weekly or so 15 minute or so tremors. I think that if it were mostly fault related the quakes would have a more linear distribution. Could it be the result of a blockage in the plumbing of Kilauea? If so, wouldn’t one expect inflation at Pahala? Do the beach balls tell anything?

      • No inflation at Pahala is indicated by the GPS that are accessible by us. Inflation is strongest near the summit of Mauna Loa, and southward motion is strongest halfway on the southern flank of Mauna Loa.

    • Yes, very impressive. Not explosive though – you cannot get explosions at that depth because water can’t turn into a gas under that pressure. The claim for remote explosive deposits cannot be attributed to this caldera. But then, many huge calderas are trapdoors, not explosions.

      • You could, but it would have to be very gassy with a lot of hydrogen and possibly carbon monoxide maybe methane (do you ever get methane?) such that these bubble in substantial volume the whole width and height of the water column. dropping its density enough to allow gaseous steam to form when you would get a runaway decompression that would inevitably be very short, but could be extraordinarily intense, Of course positing a plausible mechanism doesn’t mean it has ever happened..
        Actually I think you could do the same thing with a superheated column of water flashing to steam as it does in conventional volcanoes.

        • Yes, this applies to water. Here is the phase diagram.
          If the pressure is above that of the critical point, the difference between gas and liquid disappears. This happens at 220 atm, or depth. Below that depth, heat water and it turns into a supercritical substance, about ten times less dense than water. It will rise up fast but not explosively. This is exactly what happens in a black smoker. A normal volcano, if you turn water to steam the volume increases by a factor of 1000. That is an explosion. A factor of ten just doesn’t compare. Deep-water eruptions are effusive.

          • OK take this scenario. Large heat output generates a wide column of water at 500C, this is lighter than seawater and so rises rather fast (its 1/10th the density) until it hits the 2km depth at which point it flashes to (pressurised) steam and rises even faster (and reduces the column density basal pressure still more) to explode from the top. You then have your submarine eruption basically by blowing a brief hole in the ocean.
            Would be quite dramatic and would require probably vastly more heat than can be supplied.

          • That would be dramatic. In the ‘best’ case the explosion would happen a kilometer above this particular crater, and so it couldn’t explain that. Shockwaves transmit very well under water, and such an explosion would do widespread damage under water. But how sudden is the transition? If the water remains at 500 C, there is no sudden phase transition. I expect it will be smooth, not explosive. Thing could get really interesting if you hit the critical point exactly. But either way, I would not want to be in a boat above the event. You’d be blown out of the water by the steam, and sink from the lack of buoyancy. Simultaneously.

        • You have just described how a shut in liquid rich hydrocarbon well in equilibrium works when the liquid column becomes depressured above the liquid column.

          • Sorry this last was for Farmeroz in his original post in this thread and the reply ended here out of place.

            Albert I was taught the figure is 1528/1 at atm during my boilerman years. I think it is by far the biggest driver. One thing that should be note is that it cools as it expands. So my thinking is that ash is formed by the expansion and cooling in the eruptive column, probably wrong. lol

  5. Gisli Olafsson confirms on Twitter that the jökulhlaup from Grimsvötn has begun.

    “According to @Vedurstofan (= IMO) there is likely a glacial outburst starting at #Grimsvotn – GPS measurements are indicating a significant rise in elevation. In the past they have indicated that an #eruption is potential following an outburst. Stay tuned for more information. ”

    “Scientists have indicated that it takes some time for the glacial outburst to exit the glacier, but once it does it is likely that nearby roads will be closed as a precaution.”

    “Confirmation of whether the outburst has truly started may not be until tomorrow when glacial water with higher conductivity would reach sensors that are outside the glacier.”

    • The GRVC GPS went funny around mid-day yesterday, fluctuated, but is now increasing again.

    • It is not yet confirmed. There is a small subsidence according to GPS data, which could indicate the onset of a hlaup, but there is still no increase of water in Gígjukvísl and there are no signs of an ongoing hlaup in the seismometer data.

      It takes a couple of days for the water to travel under the ice, so it’s still too early to tell if it has started or not.

Comments are closed.