Some nations seem decidedly unvolcanic. They lack the geology and the temperament for eruptions. Take Norway – a no-nonsense nation which prefers a bit of solitude and does not go for attention grabbing. A search for Norwegian volcanoes does come up with some good examples, but these are distant islands which Norway acquired accidentally. Of course this is also a nation with a Viking history where marauding Norwegians (hard to imagine) pillaged and conquered half the known world just because their own harvests hadn’t done well. The past, of course, is different. In the end all nations have a volcanic origin. Volcanoes have made every continent’s base rock. But nations cannot be blamed for their ancient roots. Still, even a nation as inoffensive as Norway is not entirely devoid of active volcanoes. It has Haakon Mosby, a continuously active mud volcano all of ten meters high. And in typical Norwegian fashion, they keep it hidden at the bottom of the Barents Sea, 200 km north of Tromso. Excitement is not for them.
South Korea is another such quiescent country. The official description says that it is “known for its green, hilly countryside dotted with cherry trees and centuries-old Buddhist temples, plus its coastal fishing villages, sub-tropical islands and high-tech cities”. A search for volcanic activity comes up blank, perhaps not surprising as this is solidly continental, even cratonic. But like Norway, there is more here than meets the eye.
The four musketeers
The Korean peninsula of carries two nations. North Korea has a rather more eruptive reputation then is southern brother. Its self-defense knows no borders. It has two volcanoes, both borderline. One is very well known: Baekdu (also known as Tianchi), the mountain with the Heaven Lake, origin of the devastating 946 AD eruption, a high VEI-6 (not VEI-7, according to recent research). It is exactly on the border between North Korea and China and is claimed by both nations. It lies on a large volcanic plateau with many different cones. One of those cones self-destructed in 946. It has erupted several times since then, although obviously not with the same intensity.
The second North Korean volcano is almost entirely unknown: it is Chugaryong. It straddles the border between North Korea and South Korea, and is inconveniently located exactly in the demilitarised zone. Good luck investigating that one. It may have erupted in the holocene and should not be considered extinct, but finding even maps of the area is near impossible. Chugaryong lies at the end of the Chugaryeong rift structure (a variable western spelling may be noticed). This fault runs south-southwest (15 degrees east of north, to be precise) from Wonsan to the border at Jeongyeon-ri, where a basaltic plateau has formed. From there it continues to Seoul. The volcano produced large lava flows about half a million years ago and again 150,000 years ago. The lava flows reached tens of kilometers into South Korea, just west of the Hantan river. The eruption site itself is unstudied because of its location just north of the border
South Korea also has two potentially active volcanoes. One is a 10-km wide island in the Sea of Japan, called Ulleungdo. It is the tip of a submarine volcano which last erupted 5000 years ago. Perhaps optimistically, some 10,000 people live on the island. It is reported to be one of the most beautiful islands in the world but beauty can come from danger. Ulleungdo has a 3 km wide caldera which formed 10,000 years ago in a Pinatubo-sized eruption. One wonders whether it is wise to live on such a volcano with no means of escape.
The second South Korean volcano is far more obvious and much larger. A look at the map of Korea shows a large island off the south coast. This is Jeju island. The satellite image (show above) immediately draws attention to the dark area in the centre. Indeed, Jeju island is a volcano. A large one.
Jeju island measures 80 by 40 km, oriented almost east-west. It is a perfect holiday destination, easily reachable and with beaches all around it. Jeju island is especially popular for honeymooning. The island is located on the continental shelf, surrounded about 100 meter deep water. This already makes it very different from Ulleungdo.
The island forms a neat ellipse . The interior is mountainous, rising towards the volcano at the centre. The central mountain is called Hallasan, or Yeongjusan. But the volcano is more than the mountain. It covers the entire island.
Hallasan reaches 1950 meters in height. If there was any doubt about its nature, there is a nice crater next to the summit, 400 meters wide and 100 meters deep. It even has a crater lake which can be small or large depending on the season. The name of the crater is Baengnokdam; it formed by collapse, possibly during a rift eruption
There have been many eruptions on Jeju, often not at Hallasan. The island is mostly a basaltic lava plateau (as are other volcanic areas in Korea). Hallasan grew with the same elongated shape as the island. A wide, lower ridge extends it further to the east. The central mountain looks eroded: it has clearly not erupted much for quite a long time. But there are many other peaks on the island, including some 360 cinder cones. The cones are typically between 150 and 200 meters high. A few lava domes line the coast, but almost all eruptions were on land.
The main eruptions started around 1.9 million years ago. They continued into the historical period. Four eruptions have been reported in old documents, in 1002 and 1007AD. This volcano is not extinct.
Jeju island build up over four stages. In the first two stages, a lava bed formed. This happened while the region was submerged. The low-viscosity lava was erupted on a granite basement. This is basement is no longer exposed anywhere, but was included in some pyroclastic ejecta. The initial eruptions were centred first in the northern coastal part of the area and later moved southeast. A 90 meter deep lava bed formed. Mudstones show that the island suffered frequent collapses. In the second phase the activity shifted to the centre and increased. A thick lava plateau formed which grew to around the current size of Jeju island. The lava was basaltic at first, but towards the end of this phase became andesitic. The lava plateau retained a gentle slope, indicative of lower viscosity.
During the third phase Hallasan build up. This happened again in stages, and it appears the centre of the eruptions moved around a bit. Some of the other peaks may have formed at this time, before the eruptions focussed on Hallasan. The eruptions initially remained basaltic. But above 500 meters Hallasan becomes steeper: as it grew, the viscosity of the lava was increasing. A volcanic field formed around the summit, which at this time grew to 1700 meters. At the end of this phase, a lava dome 300 meters tall formed at Hallasan’s summit. An eruption on the eastern flank of this dome formed the current crater. It has a double ring structure (now eroded) and may therefore have formed in phases. A thick layer of scoria on the southeast side is a remnant of this phase. In contrast, the southwestern side of the crater is trachyte lava.
In the fourth and current phase, Hallasan ceased and from now on eruptions formed the many individual cones across the island. No dikes have been seen, which suggests that during this stage there was no longer a central magma reservoir. This phase has clearly lasted a long time, as shown by the sheer number of domes. The domes have an average length of 1-2.5 km and a height of 50-250 m. They contain scoria, volcanic bombs, and what the Koreans rather charmingly call ‘volcanic rice cakes’. (At least, that is what giggle translate makes of it.) One large cinder cone erupted basaltic to andesitic lava, a characteristic of a stratovolcano, but it did not manage to grow into one.
There are brief historical descriptions of volcanic event on Hallasan in 1002 AD and off-shore to the west of the Jeju island in 1007 AD.
The ‘History of Koryo’ reported that in 1002 AD, four holes opened in the mountain and after five days, turned water into tiled stones. This requires some interpretation! It is read as meaning that four of the parasitic cones erupted and produced rope-like lava. The eruption occurred in July and/or August and lasted less than a month.
In November 1007 AD, a mountain in the sea erupted amidst dense air. It is thought to have been at Feiyang island (Biyangdo), just northwest of Jeju Island. This has two craters, and one of these may have formed in the 1007 eruption. The description suggests that the island already existed prior to the eruption: this fits with the double crater.
We know a bit about older eruptions. Carbon dating of charcoal found below a lava layer shows that Byeongak Oreum erupted around 5000 years ago. It is a parasitic cone, 10 km west southwest of the summit of Hallasan and some 500 meters tall. It is 1 km north of the Camellia Hill botanical gardens. Before that, Mount Songak erupted about 7000 years ago. Songak is at the southwestern tip of Jeju island, and has several craters. The central cone appears to be the most recent and may have been the one that erupted. The two sites are not far apart, and are on the same western side of Jeju as the 1007 eruption.
Jeju island is famous for its lava tunnels. They occur in the gentle-sloping lava beds of the second phase. There are over 50 of these tunnels. The largest is the Manjanggul tube which is almost 9 kilometers long and in places is 30 meters tall. It is a double-story tunnel, and in one place there is a frozen lava column, 7 meters tall, where the lava flowed from the upper tunnel into the lower one. The tunnel shows the lava-equivalent of features of normal caves: wikipedia mentions lava stalactites and lava stalagmites, lava columns, lava flowstone, lava helictites and lava blister, cave corals, benches, lava raft, lava bridges, lava shelves and striations.
Jeju Island is a funny place for such a major volcano. It is far from the Pacific subduction arcs, sitting on a stable continental plate and without any indication of a hot spot. Something is clearly generating magma and something is letting it reach the surface, but what that something is is a mystery. It is a volcano without a cause, one that shouldn’t be there.
The same problem exists for the other volcanoes in Korea. They too are far from anywhere volcanic and have no reason to erupt. There are some other similar volcanic fields further north, in China. And these are not incidental volcanoes. Korea has had two VEI-6 eruptions in the past 10,000 years – not bad for a peninsula with only four volcanoes! All four have had voluminous eruptions in the past, building up large basalt plateaus. And all four date from around 2 million years ago. Whatever cause they don’t have, it is the same one for all four. And they are largely immobile. The four volcanoes have been erupting in the same location (roughly) for 2 million years. They are not being affected by moving plates. This means that the volcanic activity is attached to the plate, not to the deep mantle. The only change one can see is that on Jeju island, the volcanic activity may have moved a bit further west over time, at something like 1 cm per year.
So what is going on here?
Let’s have a look at the compositions. All four volcanoes have mostly had basaltic eruptions. Such low viscosity lavas would normally build a shield volcano, but because the eruption locations were moving around a bit, rather than single volcanoes they formed plateaus. Only Ulleungdo, as a deep marine volcano, has managed to grow a proper strato-seamount.
Basalt would suggest a mantle origin, but the evidence mentioned above suggests otherwise. The other source of basalt can be oceanic crust, as this is largely made of (enriched) mantle material.
The figure belows shows the compositions of the four volcanoes and a few related volcanoes in China. Light grey are the volcanoes of the subduction arc. The other circles are the intraplate volcanoes discussed here and in adjacent China. These volcanoes have the same range of SiO fraction as the arc volcanoes, but are enriched in potassium and sodium – the two northernmost in sodium, the others in potassium. Location is clearly important. The SiO range shows that the magma evolves somewhat over time, increasing the SiO, but this is happening in arc volcanoes as well.
Looking at the trace elements, a very clear picture arrives. There are two main sources of basalt: mid-oceanic ridges and oceanic islands. Arc volcanoes show compositions that are similar to the mid-oceanic ridges. This is not unexpected, since they are fed by the melting of subducting oceanic plates and these plates are formed at the mid-oceanic ridges – from the basalt that is there. Their basalt comes from the upper mantle. Oceanic islands get their magma from a deeper part, the lower mantle. All the Korean and Chinese volcanoes discussed here have very similar trace element distribution, and this is similar to that of oceanic islands. They do not get their magma from subduction melt. But is it from the lower mantle? Or from a different region of the upper mantle?
Could we have mantle plumes? This is always the tempting magic solution to any volcanic problem. If a volcano needs explaining, there will be someone who has an unused mantle plume in the cupboard and offers it as a solution. But in this case, this is not the case. Plumes give increased 3He/4He ratios and cause broad swells on the surface. Neither is seen here.
In this area, the upper mantle is very confused. We have come across that in the post about Fuji. In the east is the Pacific plate, old and cold and quick to subduct. In the south is the Philippine plate, also oceanic but younger, warmer, and more resilient against subduction. Both subduct underneath the Eurasian plate but in different directions and at different depths.
The Eurasian plate is actually fragmented here. Korea lives on the Amur plate. However, that does not seem to be an important distinction as far as our mysterious volcanoes are concerned.
Here we can see what is happening in the upper mantle. The colours are velocity measurements of earthquake waves, where lower velocities (red) indicate warmer and the higher velocities (blue) colder material. The top right panel shows where the cross sections were taken – all of course intercept one of the mystery volcanoes. Dark blue tends to show the subducting Pacific plate. A slightly lighter blue will be the Philippine plate. In places they are stacked on top of each other! (You may remember the layer cake of Mount Fuji.) The subduction slows down at a depth of 410 km and again at 660 km: these are two transition regions in the mantle which the plates have difficulty transcending. Below these lies the lower mantle.
The intraplate volcanoes are located above places where the subducting slabs are at depths between 410 and 660 km. This is too deep for subduction melt, and this explains why these volcanoes are different from subduction arc volcanoes. They cannot get their magma from those plates. All but one of the volcanoes are close to places where the subducting plate ends or has a gap. Is that the secret?
One suggestion is that at such an end or gap, a convection flow starts. The slab depresses the mantle below, and it escapes by flowing around the edge, upward. As it does so the pressure decreases (the heavy oceanic slab no longer pushes on it) and decompression melt sets in. The melt now continues its upward journey because it has a lower density than the solid mantle rock around it. This model seems to work quite well for most of the volcanoes, but not Ullengdo which is not sitting above such a gap. Other suggestions have been made for that volcano.
The model explains why the volcanoes remain fixed in place on the surface: they stay at a constant(ish) distance to the subduction front. So they move with the plates rather than with the mantle. It puts the source of the magma at the top of the lower mantle, which fits with the lava composition. As an interesting aside, it relates the volcanoes to the edge of the subducted Pacific plate, but not Jeju which would be driven by the Philippine plate.
It is not the full answer, though. Why does the volcanism only happen in a few places, rather than all along the subduction edge? And why are all four Korean volcanoes 2 million years old? Subduction has been going on for much longer than that.
The answer to the first question may perhaps be found in the mantle. There may be some old recycled crust in the mantle, from an older subduction event. The Pacific ocean has existed since a long time while one plate was replaced by another. Oceanic plates that existed before the current Pacific plate became dominant have subducted all around its edges. Perhaps we are seeing some variability in the mantle itself, underneath Korea.
Did anything happen here 2 million years ago? There are two suspects. The first is to the south. As recent as 3 million years ago, the Philippine plate was moving to the north. Then it changed direction, and now it is moving west to northwest. The direction of subduction changed. Plates don’t decide to do this themselves. They are being pulled around, and it is the location of the subduction zones that determines which way the plate moves. There was a subduction change.
The second suspect is to the west: the Sea of Japan. Japan rifted from Eurasia because of the pull from the subduction of the Pacific plate. The Sea of Japan formed behind, with some oceanic crust forming in the process. The connection between Korea and Japan, the Korea Strait, opened 2.6 million years ago. This is when Japan became an island.
The two events are of course related. Subduction of the Philippine plate ceased for a while, then restarted 5 million years ago. As the subduction extended, the plate changed direction. This change caused the opening of the Korea Strait.
It is clear how this could have affected Jeju Island which depends on the Philippine plate for whatever causes its magma. But the other volcanoes live off the Pacific plate subduction: why would this have been affected? One can only speculate. Underneath Baekdu, the Pacific plate shows a gap. Is this gap caused by the changes along the Philippine plate? Plates affect each other, and the changing stress may have caused breaks in the Pacific plate.
We don’t know. The Sea of Japan has a complicated history, being at the mercy of this double subduction of plates of different ages. Somewhere in that history lies the answer to the mystery of Korean volcanoes.
Korea should not be underestimated. It may not have had eruptions over the last few hundred years, but that is the exception, not the rule. When it goes, it can go big. 2 VEI-6 eruptions in 10,000 years should be enough warning. And when so many people live on sleeping volcanoes, even small eruptions can cause havoc.
For all its civilized quiescence, Korea is not Norway. It has volcanic potential. One day, some people will get a honeymoon to remember.
Albert, November 2022
Jack Ward et al, Geophysical and geochemical constraints on the origin of Holocene intraplate volcanism in East Asia. Earth-Science Reviews 218 (2021) 103624
Moon Won Lee. Geology of Jeju Volcanic Island , Korea. Journal of the Japanese Society of Rock and Mineral Deposits 77, 55-64, 1982
Marco Brenna, et al. Spatio-temporal evolution of a dispersed magmatic system and its implications for volcano growth, Jeju Island Volcanic Field, Korea, Lithos, Volume 148, 2012, Pages 337
Van Horn et al. Evolution of the Sea of Japan back-arc and some unsolved issues. Tectonophysics. Volumes 710–711, 25 July 2017, Pages 6-20