Bawean Island

And then there was a volcano no one has ever heard off. It is on an island that is far off-grid and which takes a 9-hour ferry to get there. (There is a small plane, if you are happy to travel without luggage.) The volcano has long gone extinct. All that remains is this island, 15 km across with a rugged, hard-to-reach interior. The main road on the island is narrow, staying near the coast. Tracks that go into the interior are for four-wheel drive vehicles only. The interior is hilly and bumpy, never more than about 650 meters tall but with many different peaks. The dense forest makes it difficult to move around without a machete. The Java sea stretches on all sides, with not another island in sight. The sea is not deep: below it lies another world that once was land. But there are no other volcanic islands here, and it is far from the volcanoes of Java. Bawean Island is a bit of a mystery. There is a hidden history here.

The people

Bawean island covers around 192 km2 and has some 75,000 people across 30 villages. Rice is grown but much of the island is unused. Part of the island is a protected national forest, with the main aim to protect the deer. Eco-tourism is on the agenda, with good diving and snorkeling on the coral reefs: 30% of the reefs are pristine. But the island is hard to reach and tourism remains limited. This is probably a good thing: small islands and mass tourism are a poor combination and lead to rapid degradation of the environment. The local port is too small to accommodate cruise ships: a blessing in disguise.

For those interested in military history, Bawean island was the location of one the less successful battles of the second world war, when a convoy of three ships, the USS Pope, HMS Exeter and HMS Encounter were sunk by Japanese cruisers. This happened on 1 March 1941.

The people on Bawean have a diverse origin. They are a mixture of settlers from all across the region, especially Java and Borneo. Many came from Madura (an island just off the north coast of Java), and the local language is still a variation on Madurese though with significant differences in pronunciation. The name for the island is supposed to mean ‘there is light’, referring to the sunlit peaks which could be seen from sea at sunrise. Nowadays there are strong links with Singapore, where many men go to find work. It is a tradition here that men work away from the island, without their families. So many do this that Bawean is also known as the ‘island of women’: some 77% of the population is female.

The life

Isolated islands often have a limited but unique fauna. That is true here too. There is an endemic species of deer on the island which is highly protected but in severe danger of extinction. One wonders how it got there: presumably it has been here since the ice age when the Java sea was dry land. There are only about 200 Bawean deer left. The other endemic mammal is the Bawean warty pig, a dwarf species of which some 250 are left. In spite of being the world’s rarest pig, it is not specifically protected, perhaps because pigs are too often seen as pests! There is also a rare subspecies of serpent eagle, now down to 30 pairs. The top predator on Bawean Island is a snake: the Malayan krait, highly venomous and nocturnal. It is the only venomous snake on the island but more than makes up for that! The trees on the hills are covered with hanging mosses, sometimes so dense nothing else is visible – a sign of danger when there are venomous snakes hiding in the moist forest! The name of one of the peaks, Mount Lumut, means ‘mossy mountain’. (There is more than one ‘Mount Lumut’ in Indonesia!) Part of the island is a protected reserve, aimed to protect the deer. It benefits the pig too. Illegal logging is a problem even in the protected forest. The most recent survey found remarkably few land birds on Bawean island, attributed to the Asian cage bird trade. Nowadays, even the most isolated islands suffer from civilization.

Bawean deer. Source

The moss forest is an eery place. The scarcity of land birds means there is little bird song, and the moss dampens the sounds further. The silence is deafening. The most forest is found in elfin forest, which is typical for higher tropical mountains. Elfin forest consists of mixed trees some 10 meters tall creating a single canopy layer above the ground cover, with small leaves. Moss grows on the ground and lower stems but can also hang thickly from the branches. It requires a lot of moisture. Moss forest is therefore found in foggier places. This can be depressions of the crest of the mountains, as they may contain more standing water and are therefore subject to fog. The altitude of the moss forest varies with location: moss forest can form lower on small ocean islands than on the large land masses. On Bawean Island it is found as low as 550 meters, but on Java it can occur as high as 2 km. In effect, the moss follows the typical height of the cloud layer. A long dry season stops moss forest from growing.

There is a volcanic connection here. Mount Pangrango on Java has moss forest, but its twin Mount Gede does not. Gede is volcanically active whilst Pangrango has long been extinct. The sulphur emission may be part of the problem, but young volcanoes also are more free draining than old, eroded volcanoes. The moisture levels may differ. In any case, to find moss forest, avoid the most lively volcanoes. Bawean Island is extinct. It is perfect.

The land

The island is a raised area above the bottom of sea. Under water the island extends about twice as far as above. There is a steep edge to 50 meters depth, perhaps from a time of lower sea levels. The Java sea here is about 60 meters deep.

The interior of the island has many peaks and domes, with some argument (even now!) about which peak is the highest. They reach about 650 meters. Mount Besar is considered to (just about) be the highest. Three other peaks are over 600 meters. There is also a 1-km wide lake called Kastoba which is surrounded by an elevated rim.

Lake Katoba

Landslides are common around the lake. The topsoil is a 5-meter thick layer of leave litter, sitting on a layer of clay. The topsoil can, and often does, slide. Underneath the clay is lava with a basaltic andesitic composition. Clearly, if you look below the surface everything on Bawean island is volcanic in origin: the land, the peaks and the lake.

A typical landslide in Bawean Island

The volcano

The volcanic origin is surprising in view of the lack of visible volcanoes within the Java sea. We are far from the subduction-driven volcanic arc for which Java is famous. The thick leaf mould shows the activity is not recent. The island is extinct. But the ruggedness still reminds of the old activity. The lake is the remnant of a volcanic crater. The various peaks are plugs and domes – there are a staggering 44 of them! This volcano is very different from the stratovolcanoes further south. There were many eruptions here but they build a bumpy plateau, not a mountain.

The composition is also different. There are two different lava types on Bawean Island. The lava flows deep below the clay are andesitic basalt. To be precise, these flows are basanitic and tephritic, lowish in silica, and are enhanced in sodium and potassium. The lava flows contain leucite, where the quartz is replaced by potassium silicates. The domes and plugs are nepheline-bearing phonolites, like leucite but with more sodium. Detailed analysis shows that the Bawean lavas are high in potassium. The K2O fraction of the rocks are around 5% and the NaO fraction is up to 10%.

Dating has found that the phonolites have ages between 380,000 and 840,00 years, while the other lava flows have ages between 250,000 and 670,000 years. Often the plugs rise up through the lava flows and so are younger, but the opposite also occurs. Older lava may be present, of course. Dating shows what is at or near the surface, not what lies underneath.

The measured fractions make the island ultra-potassium-rich. These types of magma are rare in subduction environments. There are a few other potassium-rich lavas around Java among current volcanoes. The volcanic arc on Java does not have them. Away from this main line of volcanoes, Java’s ultra-potassic rocks occur in three groups of volcanoes. They are located at 20-30 km (Ringgit: East Java), 100km (Muria: central Java) and 250km (Bawean Island) from the volcanic arc. Bawean is the only one of these which is not on Java itself. The subducted plate here is deep, 250 km at Ringgit and 650 km at Bawean island. If this plate is the source of the magma, as it is in the volcanic arc, then the magma here comes from a very large depth.

Location of the potassium-rich volcanic centres on and near Java (coloured dots). From Handini et al 2022, IOP Conf. Ser.: Earth Environ. Sci. 1071 012013

The images (all from Handini et al.) show the locations of the ultra-potassic volcanoes and their compositions. The colours indicate the volcano, and the size of the dots reflects the depth of the subductin plate underneath. These are not unique. Similar magmas to Bawean island (although not identical) are found, amongst others, in the volcanic area near Rome in Italy and in the intraplate volcanics of Australia. What these places have in common is that continental crust is involved in the melt.


Source: Wikipedia

So is there continental crust involved? Indeed, there is. Indonesia is a continent hiding behind an island arc. The large western islands of Indonesia (Sumatra and Java) lie on the rim of a submerged continental plate known as Sundaland. It was dry land during the ice age. That 9-hour ferry to Bawean Island sails above a drowned land which people once called home.

Sundaland connected much (but not all) of Indonesia to much of Southeast Asia. There were major river valleys here, huge drainage systems flowing toward the South China Sea. The continent was used for migrations and probably for living. There are scattered populations of people in Malaysia and Indonesia with ancestry that predates the current populations. They are closer related to the indigenous populations of Papua New Guinea and Australia. Flores, a more eastern island, had a human group that was even older, which later became extinct. All must have lived on Sundaland. But as the far-away ice melted, the land became submerged. It must have been terrifying for the people who lived here. This is the ultimate Atlantis, a true continent now gone, with only scattered islands remaining. One wonders how much archaeology lies hidden below the Java sea.

Before the flood, Bawean Island was a 700-meter tall, bumpy plateau. Now it has become an island, far from anywhere, best known for catching the early morning sun while the sea is still in darkness. It is the sea that is the intruder here.

It is not the only continent here. The volcanic arc of Java is caused by subduction of the Australian plate, migrating into Indonesia at headlong speed. This plate carries continental shelf with it. This continental crust has not reached Indonesia yet. When it does, expect the mountains to rise! Already, the subduction is pushing up the edge of Sundaland. This push-up is the origin of the large islands of Sumatra and Java. They are the raised rim of the sunken land, pushed up by the approaching continent.


Sundaland is an accumulation of pieces. It grew to its final size by accretion of fragments.

200 million years ago, this was part of Pangea, the supercontinent which was split east to west by the Tethys ocean. To the south, on the far side of the Tethys, was Gondwana. The northern edge of Gondwana was a complicated place. It rifted in several places; fragments split off and were pulled in by subduction zones to the north. We have come these fragments before when discussing the fossils of Mount Everest.

One fragment came from what is now western Australia, crossed the Tethys, and merged with Eurasia. This is now known as the Billiton lineament. It finished the crossing some 100 million years ago. The southwestern part of Borneo and western Java came from this accretion. Some 90 million years ago it was followed by a second fragment; this added southeast Borneo, east Java and part of Sulawesi. This fragment originally came from northwestern Australia and Bird’s head peninsula on New Guinea. It added a different type of crust.

From Pacey et al. Earth and Planetary Science Letters, Volumes 369–370, May 2013, Pages 24-33

On the image above, the dashed lines show the various additions to Sundaland. ‘1’ shows the edge of the first accretion. ‘2’ is the limit of the second one. Later, other, more oceanic parts were added.

For the next 40 million years or so not much happened. The subduction which had pulled in the fragments had ceased. Around 45 million years ago it restarted: a new volcanic arc formed south of Sundaland and Australia began to move north. Debris from these new volcanoes covered part of Sundaland. But volcanism ceased again. The island arc moved northwards and merged with Sundaland between 5 and 10 million years ago. The merged arc formed the Southern Mountains on Java, and the impact raised the region to form the rest of the island of Java. The modern volcanoes of Java are build in the past 2 million years or so on this raised basement. The volcanoes further to the north are build on the Sundaland continental plate. It is these volcanoes that produce the potassium-rich rocks. Bawean Island is the furthest of those volcanoes. In age it belongs to the modern volcanic era – in composition it belongs to an older era.

There is something funny here. The potassium-rich volcanoes on Java are on the part that was added in the second accretion: they are absent from the area added in the earlier accretion event. Where is Bawean island? It is also on this part of the crust, close to the old suture where these two fragments joined. The secret to these volcanoes is not in the deep subduction zones. It is the crust that lies above it.

The fault and the ridge

We have come across this suture before. In an earlier post, Merapi was found to be located near the Progo Muri fault, the line along which Java was stitched together. The fault ends at Muri which, of course, is the potassium-rich volcano on the north coast of Java. This is an ancient fault, but on Java it still capable of generating substantial and damaging earthquakes. Why? The fault should longe have died, but it has been reactivated by pressure from the approaching Australian plate. There is now a bit of rotation and extension and it makes use of this ancient weakness.

The Progo-Muri fault (it has several different names) ends at Muri volcano. But in the Java sea it continues as the Muria trough, a deeper section of the Java sea, and along this trough lies the Muria fault, running towards the northeast. It seems inactive at the current time. The trough seems a very shallow one but in the rock underneath it can be traced to a depth of 4 km. It has almost completely filled with sediment. The Muria fault borders it on the east side. The Muria fault passes some 30km west and northwest of Bawean Island. On the east side of the fault is the Bawean arch, a volcanic ridge which runs from Muria volcano to Bawean island. Muria volcano and Bawean Island are somehow connected: 100 km apart, but on the same volcanic axis and active at the same time.

Source: Lunt, P. 2019, Marine and Petroleum Geology Volume 105, Pages 17-31

So the suture where an ocean disappeared is still there. It separates the stable Sundaland continental crust from the squashed and broken crust (‘melange’) to the east where the later block joined. The volcanic ridge follows this line, and Muria and Bawean Island are the two end points of the this arch. Everything is connected and the dots line up.

The making of an island

Bawean Island is a funny place. It is a reluctant volcano where none should exist, producing a lava type that shouldn’t be there. It all stems from a collision 90 million years ago when a lost fragment of northwestern Australia traveled across an ocean that no longer exists to find safety in the north. The potassium-rich Javanese volcanoes are all on this fragment. The safe home has lasted all this time, while Gondwana split and Australia separated into isolation.

But now all of Australia is following the route of this ancient fragment. Land is rising in anticipation and Java has started to rotate under the pressure. Exactly what happened is unclear, but the trough suggests there may have been a bit of extension in the Java sea. The old suture began to leak. Deep underneath, the subducting plate is producing a bit of melt. Somehow this found its way up through the leaky fault, in the process melting and mixing with the local crust. The melt fraction is probably quite small. For some time, perhaps a million years, lava was produced and build up the arch, stretching from Muria to Bawean Island. The latter formed at the northeastern end of the arc, not as a stratovolcano but as a series of close vents and domes. Muria formed at the southwestern end, also originally as an island that only very recently joined Java. Like Bawean Island, Muria has lava domes although it did build a stratovolcano. Muria began to form around 1 million years ago, probably at a similar time as Bawean island. They also went extinct at similar times.

The time scale makes it unlikely that any hot spot was involved. The arch is a bit more than 100 km long. The subducting plate is moving at 6-7 cm per year. (In fact rather slower than that because here it is sharply bending down, so mainly moving vertically into the abyss.) To move 100 km at that speed takes 1.5 million years. The ages of Muri and Bawean Island are much closer than that – they are similar. Therefore, it is perhaps more likely that the cause is in the fault. Under pressure from Australia, Java started to slowly rotate: we have seen this in the Merapi area. This rotation caused a bit of extension in the old suture, and that this allowed the two volcanoes to form, as well as the connecting arch.

This is speculation. But it is nice the think that this isolated island in the middle of the Java sea exists because of that isolated continent to the south. Bawean Island is part of Australia’s history.

70 years ago nations fought battles near Bawean island. They did so at a place where people from many different places settled and mixed, and which formed from mixing continents. It is an echo of the ever-changing kaleidoscope that became the world we know today. There is history everywhere.

Albert, November 2022

Postscript. There is one other thing which I was unsure whether it is related or not. There is a line of earthquakes that runs across the Java Sea. The image below is taken from and shows earthquake larger than 4.5 over the last 100 years. The largest events are M6.6. The line has two clusters, which are near the two largest islands in the Java Sea. These are extraordinarily deep, 540 to 640 km. This is where the subducting plate hits the mantle barrier and sinks through. Is this line related to the Bawean island? With a bit of roll-back, in the past it may have been directly underneath the island. I do not know.

Leterrier et al., Journal of Southeast Asian Earth Sciences, Volume 4, 1990, Pages 171-187. Potassic volcanism in Central Java and South Sulawesi, Indonesia

Handini et al 2022, IOP Conf. Ser.: Earth Environ. Sci. 1071, 012013. Geochemistry of arc alkaline magmatism of Java Island, Sunda Arc: a statistical review

Pacey et al. Earth and Planetary Science Letters, Volumes 369–370, 2013, Pages 24-33. Linear volcanic segments in the central Sunda Arc, Indonesia

Susilohadi and Soeprapto, Berita Sedimento, 2015. Plio-Pleistocene Seismic Stratigraphy of the Java Sea between Bawean Island and East Java

Lunt, P. 2019, Marine and Petroleum Geology Volume 105, Pages 17-31. The origin of the East Java Sea basins deduced from sequence stratigraphy

Karimah et al., Jurnal Penelitian Pendidikan IPA 8(3):1495-1502 (2022). 3D Modelling of Geoelectrical Resistivity Data to Determine the Direction of Landslides in Kastoba Lake, Bawean Island, Indonesia

Albab et al. 2017, Bulletin of Marie Geology, 32. Seismic Facies of Pleistocene–Holocene Channel-fill Deposits in Bawean Island and Adjacent Waters, Southeast Java Sea

117 thoughts on “Bawean Island

  1. Interesting. Might this someday result in a smorgasbord of terranes like south-central Alaska and most of California?

    Why wait? Start the hydrocarbon assays while the getting’s good!

  2. Interesting, I did not know there was a volcano there. Also, it is young geologically speaking.

    Most volcanoes in odd places are highly potassic and erupt lavas such as alkali basalts, basanites, or phonolites. Most intraplate volcanoes, both oceanic and continental, are potassic, including every volcano in Africa, most Polynesian volcanoes (Marquesas, Tahiti, Samoa, Pitcairn-Gambier, Easter…), or European volcanic fields such as Olot, Chaine des Puys, or Eifel.

    • Yes! Funny volcanoes produce funny magmas – that is normal. I had not realized that Indonesia had these different types of volcanoes, until I started wondering why that island was there in the middle of the Java Sea. There are volcanoes even in Borneo (otherwise very non-volcanic) but these are much older: 5-10 million years and are not so related to the current events. I think that Tambora may also be potassic, although not as much as Bawean

      • Batu Tara is similarly potassic to Bawean if memory serves right. It is a back-arc volcano too, but an stratovolcano. Tambora is almost as potassic as Bawean and Batu Tara, although it is in the proper volcanic arc.

    • Wow as potassic as Nyiragongo in some of the drill core samples, its just more evolved in Sio2 .. ultrapotassic Phonolites then. Indicates a deep melting source and small ammounts of melting of the mantle source materials

    • I’ve taken a capture from a map I have in Google Earth with the potassium levels of magma. It goes red, orange, yellow, green, blue, dark blue, purple, pink, in order of increasing alkalinity/potassium content. Indonesia has a couple of ultrapotassic volcanoes, labelled in black, most of them are behind the arc. Part of the Sunda Arc is also very potassic too, in particular the area from eastern Java to Tambora/Sangeang Api is more potassic than usual for a continental volcanic arc.

        • The pink data points seem to have the coordinates wrong, there is no volcano there as far as I can tell. It is possibly misplaced data from nearby Ringgit Beser volcanic complex that was featured in Albert’s article.

  3. Since its an article in the tropics

    Here is a very Impressive video of a tropical Vietnam Thunderstorm, with a very high ligthning intensity. Its in Real Speed too.

    I Did not knew that tropical thunderstorms can have the same electric intensity as midwest Supercells, But this one appears to have. This is likley a Severe Summer Moonsoon CB.

    In the tropics these kind of flash intensity is rare ? Most tropical thunderstorms are the garden type vaierty, just everyday in an unchaging tropical climate

    But You also have very high temperatures and extremely high humidity in Vietnam in summer so perhaps explosive results..

    Now back to volcanoes

  4. Away from the tropics, the seismic moment curve at Grimsvotn has taken a turn up. It is now much steeper than before, with a similar rate of increase as seen in the few hundred days before the previous two eruptions. The difference is that we have larger individual quakes which did not happen in previous run-ups. I am ignoring these for comparing the rates. The current build-up is not identical to the previous two. But I still would expect an eruption here in 2023.

    Of course volcanoes do not listen to expectations!

  5. Hello Albert,
    can you explain why it takes 56 days to find a person (deceased, RIP) two to three miles from La Silla Obs.?

    having little time I can’t comment, but will be back.

    • Tom was a friend. He disappeared in an inaccessible area, very rocky, steep, and no roads, no water, no people or wildlife. Finding someone in those circumstances over 100 square kilometers is unbelievably difficult. It took one of the rescue groups 6 hours to walk that distance. It is rare for someone lost in the Atacama to ever be found.

      • Now this (Parque Nacional Bosque de Fray Jorge) is 100 km north of La Silla (crow’s fly). The picture might explain some. The fog is called Camanchaga. It is the nearly only source of humidity in the Atacama Desert.

  6. One of the best bits of information about Hunga Tonga Hunga Ha’apai I have seen.

    Excellent visuals too.

    To summarize, the eruption was twice as voluminous as Pinatubo, 9.5 km3 is the volume difference from the pre-eruption bathymetric data that was made in 2017.
    Ignimbrite flows also went 80 km away from the volcano, possibly as far as 100 km, and are what destroyed the internet connection to Tonga after flows topped a ridge and buried the cables. I was actually not aware ignimbrites could flow underwater at all, but evidently the ocean presents little if any obstruction. If anything, it seems the flows went further underwater…

    There is no data on the bulk volume, but the 9.5 km3 is the caldera volume, reflecting the DRE, so the bulk volume is presumably much larger. In mafic-intermediate magmas the bulk tephra is as much as 3x the DRE, so this might approach 30 km3, and is easily bigger than the 15 km3 of Novarupta. This is one of the biggest eruptions we have ever seen in the historic record.

    After reading this I am now far from surprised it is being linked to the weather this year.

      • In that bbc link there’s an interesting recorded interview with the project leader Kevin Mackay. One of the things they found out during their investigations, that he talks about in the interview, is that there is a new cone inside the caldera. That means the eruptive activity didn’t end with the big bang, but the eruption has continued afterwards.

    • Fascinating. By the way, I read the article as that the 9.5km3 is the tephra volume, not DRE and not caldera volume. You could not fit 10km3 of rock into the new caldera. The hole is (I am kind of happy to see) similar in size to our own first estimates where we had the tephra estimated at 6 to 8 km3. That also fitted with the amount of ash seen at the main islands. The new number is more accurate but not wildly different. We are talking about a top-end VEI 5, perhaps just about VEI 6. It is the largest eruption since Hudson, I think.

      As for the flows, ignimbrite has a density of about twice that of water. It is dense enough not to mix with the water, but low enough to benefit from the water. Still, much of the power must have come from the 2 kilometer elevation difference.

      • “They found the Hunga Tonga-Hunga Ha’apai spewed out at least 9.5 cubic kilometres of material – a third more than the team’s initial estimates.

        By comparison, the 1991 eruption at Mt Pinatubo in the Philippines discharged 5.5 cubic kilometres of material and 4 cubic kilometres was erupted over Naples from Mt Vesuvius in 79AD.”

        I think they actually are talking about the DRE Albert, 5.3 km3 is the DRE for Pinatubo, and I am sure they would have known the difference before using that number as a comparison.

        Maybe assuming it is bigger than Novarupta is jumping too far but in any case this eruption was way bigger than either of the 1991 eruptions. The new caldera of HTHH is also actually about as big as the caldera on Katmai, slight smaller in diameter but deeper. It is a lot bigger than the caldera of Pinatubo, 4x as deep and 10% wider in diameter.

        I imagine the flows underwater must have been a bit more liek lava flows than they would have looked on land, semi-solid partly molten passes of rock, probably kept lubricated by a steam cavity and gliding over the seafloor on the steam too. The magma was probably not that viscous, crystal poor basaltic andesite, so maybe in these situations the line between an ignimbrite and a lava flow dissolves and the flow is basically the same either way mechanically speaking. On land though there would be a more obvious difference. Still 80 km is an enormous distance, the great depth of the ocean here is probably the only thing that saved the residents of Tongatapu from a fiery fate like those on shores facing Krakatau…

        • Interesting article. I expected the cables had go down that that way. Mount Pele cut the underwater telegraph cables in the 1902 explosion that destroyed Saint Pierre. Submarine pyroclastic flows were though to be the likely culprit. Submarine ignimbrites are very common. Most calderas in Tonga are surrounded by large fields of concentric dunes that were formed in earlier ignimbrite eruptions which advanced tens of kilometres in all directions from the volcanoes.

          I have tried to estimate the size of the caldera by using the volume of cylinders that could more or less fit inside or slightly exceed the caldera. I think the DRE is more than 2.5 km3 but less than 7.5 km3. Somewhere around 5 km3 DRE might be the true size. I don’t think it is possible to fit 9.5 km3 in the caldera as Albert says. Also take into account that Katmai and Pinatubo had tall mountains where their caldera formed, Hunga Tonga collapsed a flat area, so Hunga Tonga will be deeper for the same size.

          • Hunga Tonga is about the same size as Pinatubo, Quizapú and Santa María, and given that the magma is mafic, more dense. Then the erupted mass could surpass Pinatubo. The bulk is likely to be smaller than Pinatubo I think, because mafic magmas tend to make scoria in ignimbrite eruptions, which is denser than the pumice of silicic magmas. So depending upon considerations it might be the largest eruption since 1991 Pinatubo or since 1932 Katmai.

          • A suspension of mud in water will flow downslope, and sometimes even upslope if it is pushed away by the weight of the mud nearby, making something that looks a lot like a pyroclastic flow in appearance, I’ve seen that happening in patches of water. It is the same process as turbidity currents:


            “The driving force behind a turbidity current is gravity acting on the high density of the sediments temporarily suspended within a fluid.”

            Hunga Tonga would have generated powerful turbidity currents due to a large volume heavy scoria and ash mixing with water.

          • Another interesting excerpt from the Wikipedia entry:

            “Taiwan is a hot spot for submarine turbidity currents as there are large amounts of sediment suspended in rivers, and it is seismically active, thus large accumulation of seafloor sediments and earthquake triggering.[51] During the 2006 Pingtung earthquake off SW Taiwan, eleven submarine cables across the Kaoping canyon and Manila Trench were broken in sequence from 1500 to 4000 m deep, as a consequence of the associated turbidity currents”

          • Part of my other comment was to address the enormous intensity of HTHH too though, i forgot the number but it was all said and done in an hour, nothing in the 20th century was even close. The volume is one thing but the actual eruption itself was off the charts.

            Turbidity currents were my second guess, but I also wanted to play with the leidenfrost effect helping the flow move an idea.

    • I double checked the numbers.

      Santa Maria 1902, 8.5 km3 DRE/20 km3 bulk.
      Bimodal dacite and basalt plinian eruption. 20 hrs.

      Novarupta/Katmai 1912, 14 km3 DRE/ 28 km3 bulk.
      Rhyolitic to andesitic ignimbrite plinian eruption and caldera formation. 60 hrs.

      Cerro Azul/Quizapu 1932, unclear DRE/ 9.5 km3 bulk.
      Mostly rhyodacitic plinian eruption. 11 days, but majority in 18 hours.

      Pinatubo 1991, 5.3 km3 DRE/ 11 km3 bulk.
      Dacitic plinian eruption and caldera formation.

      Hunga Tonga Hunga Ha’apai 2022, 9.5 km3 DRE, bulk unclear
      Basaltic andesite ultraplinian ignimbrite eruption and caldera formation.

      Cerro Hudson 1991, was much smaller, still a sizable VEI 5 at 2.7 km3 DRE/ 4.3 km3 bulk, but it is nowhere near a 6. There are a number of other large VEI 5s in the 20th century, like St Helens and El Chichon, smaller but nuch more powerful than Hudson.

      But as can be seen HTHH was enormous. By volume DRE it is only behind Novarupta. It could be even bigger if the magma was highly fragmented. Either way you look this was a massive evebt.

      • Chad, do you have a publication for Santa Maria? I thought the old cone was basaltic andesite and since 1902 it’s been nothing but rhyolite/dacite.


          There is probably a lot more, I just wanted volume numbers. But it does seem that the basalt was not a big part and probably a trigger for the dacite. Eventually I think Santiaguito will erupt basalt, the dacite is magma that didnt erupt in 1902. There is a lava flow active there now though which while still dacite looks much less viscous than earlier lavas, formed a thin flow only a few meters thick initially.

  7. Albert, thank you for this article. So where did the Tethys close, exactly?

    These deep earthquakes are very interesting to me, as they occur in a liquidish environment.

    Also, looks like a nice place to visit.

    • There was more than one Tethys ocean. Every time a part of continent crossed the ocean, it obliterated the ocean in front and created a new one behind. That happened in Sundaland, the Himalayas and Europe and in a way is still happening in Italy. For the oldest ocean, you can trace the black see, the caspian sea, the Aral sea, as depressions left by the Tethys. But other remnants are further south

      • Is the large Basin-like desert in Xinjiang (Taklimakan or Tarim basin) related at all? There is a string of basins between Europe and China.

        Also, what is driving the movement of terrains? Is it the subduction or the mantle plume?
        Why would a chunk of Australia break off and move faster than the rest of Australia in the same direction?

        Subduction zones make sense to me. Plates or bits moving away from each other makes sense. A chunk breaking off and saying “I’ll race you to Indonesia” to the rest of the continent, less so.

        • It has happened several times in Africa, producing Madagascar and India. You need a spreading rift inside the continent. After that the subduction zone on the far side of the ocean pulls in the bit that split off. Somalia is going through this but hasn’t gone anywhere for lack of a subduction pull.

        • Probably. China was consisting of two parts in Tethys’ times, a peninsula and an island. They are supposed as sort of hanging down resp. sitting in the east of Tethys. On the other side of that early China we have Panthálassa, later the Pacific Ocean. That one was always the larger ocean it seems.

          A part of Thethys was in the middle of the Austrian and Italian Alps (rocks and fossils for prove).

        • Concerning your question about China you can also trace Tethys with oil. In the North and South of the Earth other seas have left traces. Think Venezuela, Libya, Iran, China, and there must be oil under the Himalayas.

          It is the remnant of tiny marine creatures surrounded with a shield of chalk. So when they dig in today’s oceans they first find chalk, and when they find chalk they dig on. When they died they fell to the ground of their sea.

          The following map – if it loads – will answer your question about Xinjiang:

          • I have always wondered if gas, and at least some oil, derives from methyl hydrates buried below sediments. Given how easily these form in cold anaerobic situations, and how big current seafloor deposits are, it seems more plausible than thermal degradation of decomposing microbes. I would expect more nitrogenous compounds rather than the relatively pure methane (and other hydrocarbons) and hydrogen sulphide.that we find, if derived directly from rotten biomass.

      • One witness of the closing (late) is The Valley of the Whales, Wadi al-Hitan in Egypt, others are Tethys rocks on the coast of Oman or also Spain and other places.

        The first closing probably happened between India and the Asian continent where Tethys was subducted first. The Ocean behind though is not called Tethys, but Indian Ocean.

        The situation in Middle America is unclear so far as it is not precisely known whether the Caribbean Plateau came in from the West.

        • India was actually quite late. There is a mountain range just north of the Himalayas which came from an earlier collision. The Sundaland collisions are also much older. The closing of the Tethys was a very drawn-out process.

          • Technically it is ongoing, the Mediterranean is Tethys crust south of Greece, where subduction still occurs.

          • Yes, the closure of the Persian Gulf and the easter Mediterranean can be seen as the last gasps of the Tethys – or at least of its descendents

    • Rough idea of Tethys between Eurasia and Middle America. First closure India around 59 plus/minus 1 Ma, last closure in the area of the Mediterranian and Red Sea around 10 Ma or still going on.

        • Looks to me more like tethys nearly gone, pacific shrinking and Atlantic increasing. Three intraplate oceans in various stages of the cycle. In a few hundreds of millions of years maybe it will all recycle one step and the old tethys (then landlocked) will open as the pacific is a small suture and the atlantic starts to shrink.
          The landshapes will, of course, be wildly different.

  8. There are warnings that Shiveluch volcano in Kamchatka is in danger of erupting.

    • Disputed. This is the Christmas island chain of seamounts. The mounts are old (50-100 million years) but they re-activated more recently (10-20 million years). One suggestion is that the reactivation is from bulging behind the subduction. As the plate subducts, it forces some mantle material backward and this can push the plate there upward. The old plate cracks and magma comes up through the cracks. Others put several hot spots underneath the chain (one won’t do as the chain goes in the wrong direction). Pick your choice

      • This sounds sensible:

        The Cocos-Nazca spreading center (CNS) has a complicated plate tectonic history. New magnetic data allow a detailed reconstruction. After the break-up of the Farallon plate at 22.7 Ma the newly formed CNS had a SE-NW strike direction and a high spreading rate. At 19.5 Ma, the spreading axis jumped south. The jump was coincident with a change of the strike direction to almost E-W and a significant decrease of the spreading rate. Another jump of the spreading axis to the south combined with a slight change in the strike direction occurred at 14.5 Ma. In younger times, multiple small scale jumps to the south are being observed. We interpret this pattern of jumps as the influence of the Galapagos hotspot: The plate motion vectors of the Cocos and Nazca plates result in a constant movement of the CNS to the north as long as spreading is symmetric. The repeated jumps to the south keep the spreading axis in the vicinity of the Galapagos hotspot. Similar behaviour of spreading axes has been observed around hotspots in the South Atlantic and Indian oceans.

        No full text though

        • I had assumed the question was about the Cocos island northwest of Australia. The same name also occurs near the Galapagos, which is what this paper is about.

          • Yes. That’s funny. Didn’t occur to me that there is another Cocos Island Down Under. Keeling should have made me suspicious.

            I think the tho articles BBC and ABC abiut HTHH are impressive. I’ve read that Tofua (Tonga Islands) has also been debated for one of the mystery eruptions in the 15th century (Károly Németh). Looking at its crater and then seeing the articles I could also imagine it.


          • Another set of Cocos Islands to the NW of Mauritius, also known as the Cargados Carajas Islands depending on who’s chart you use.

    • At least 9.5 km^3! So at least a very, very high-end VEI 5! Sounds like the article is confusing DRE and bulk volume with its comparison to Pinatubo.

  9. Was this Island a forested hill in an ocean of Savannah during the last Ice Age ?

    The LGM was both cooler and much much drier than todays equator and rainforests only existed along otographic rainfall areas


    Finally, someone from HVO has gone down and walked on the new lava at Kilauea 🙂
    Lots of pictures of the hornitos all over the lake surface, most of them still glow and act as fumaroles so are probably points that will reactivate again.

    The summit quakes also have mostly stopped at shallow depths in the past 2 days. There is also decreased SO2, and the lake has been falling for several days. Despite this it is still as active as it has been since the recent intrusion in September, the eruption isnt stopping but it is showing all the signs of doing so… I expect, if there wasnt an open conduit the eruption would have ended.

    The shallow quakes stopping is interesting though, they are indicating pressure, which would make sense if the lake was not rising or was getting too high above the vent. But there has been no larger eruption or shallow intrusion, so the idea of it going into a deep intrusion might be a likely case and taking out all of the pressure from Halemaumau. The GPS all around the summit are showing uplift and southwards movement, maybe there is a large deep sill intruding under the summit, possibly a very significant volume if barely a trickle of magma has been reaching the lake for 2 months and the whole south flank has been pushed away.

    The quakes of the last month do show a lot more of a stack underneath the summit area of Kilauea than is typical to see there.

    • Maybe magma is accumulating in the rift system : ) and with an over 0,1 km3 yearly supply, something will happen soon.

    • The vent Wont die.. its circulating and the magma system remains open. But its true that gas emissions are very low compared to the Overlook Lava Lake that had 3300 tons a day of sulfur. But this vent lake is still connected to the magma system below. Since the summit eruption became steady since 2021 the sulfur emissions have not been as high as it was during the Puu Oo years. Perhaps the channel in the lake is not as wide as the Overlook Shaft was in 2017, to explain the lower So2, But With Kilaueas large supply .. something will always happen.. soon

      • Overlook was excessive, the lava erupted by Pu’u O’o was more degassed in 2018 than in 2008, so the degassing was greater than the supply. But the SO2 rate from Pu’u O’o before 2008 was still about 3x the rate now, so clearly not all the magma is getting out. I would expect most of the Halemaumau magma chamber is degassed, but deeper down it is fresh and in the other magma chambers the magma probably isnt degassed either, so it is probably a deeper connection problem than a narrow conduit to the lake.

      • Leilani was a sulfur rich dragon of a beast .. and perhaps deeper materials dragged up then with lots of decompression of sulfur gas

        And Ionian basalt lavas seems even more crazy becomming sulfur yellow as soon as they cool below 300 C

        • I have an expectation that the 2018 eruption did involve a lot more of the volcano than the eruptions before it. Pu’u O’o was fed by an intrusive complex at 2-3 km depth, with maybe some minimal feed from the deeper magma system but largely from the caldera. the initial intrusion into the LERZ in 2018 came from this too, becoming a dike east of the JOKA station.

          But after the earthquake, I think this changed completely. The magma flow rate was low until then, and the summit was not really deflating. But after the quake the summit drained out rapidly and the frequency of eruptiosn and quakes relating to the intrusion increased. After the quake of 1975 a lot of magma went into deep storage and it may have been this that encouraged a long lived ERZ eruption, as Pu’u O’o erupted from the section of rift that was directly adjacent to the part of the flank that slipped that year (and in 2018).

          Lava erupted between the 16th and 27th of May 2018 came from Pu’u O’o, it was fluid but had no olivine or very little. The early lava was probably some evolved stuff that was in a pocket somewhere inbetween Pu’u O’o and the highway, maybe similar to the stuff erupted in 1977. Fissure 17 is thought to be from the dacite magma that got drilled into in 2006 under PGV, from that magma mixing with the basalt, as such it is probably a lot older than any historical eruption.

          The lava erupted from fissure 8 though was very magnesian, with abundant olivine. Some of the crystals were formed from magma that was very close to ultramafic composition in the melt phase.

          “In addition, the erupted crystal cargo from F8 contained some of the most forsteritic olivines (Fo88-89) erupted at Kı̄lauea since 1974, which must have grown in melts with 13–14 wt% MgO (Gansecki et al., 2019). Some of these crystals also contain prominent kink bands (Gansecki et al., 2019), indicating that their crystal lattices have been deformed (Wieser, Edmonds et al., 2020). Previous work has suggested that highly forsteritic, deformed olivines are derived from the deeper, SC reservoir at 3–5-km depth (Helz et al., 2015, 2014; Wieser et al., 2019; Wieser, Edmonds et al., 2020), or Kı̄lauea’s deep rift zones at 6–9-km depth (Clague & Denlinger, 1994; Vinet & Higgins, 2010). Alternatively, Lynn et al. (2017) suggest that highly forsteritic olivines from the Keanakāko‘i Tephra may originate from deeper crustal storage reservoirs, perhaps located near the base of the volcanic pile at ∼8–10-km depth.”

          So it seems at least a very good case that the 2018 eruption began as a normal dike which is what was traced and followed, but later incorporated a much larger intrusion that involved basically the entire volcano. An intrusion like what Hector proposed for 1868 is one way it might have looked.

        • What was the temperatures of the 2018 olivine high magneisan stuff? Well it have formed at over 1300 C and the source is 1600 C

          But was the eruption temperatures for that 1140 C as well ?

          • There isnt anything I have seen that gives a number more than 1140 C. The furthest east that the flank moved was to about where Heiheiahulu is, at JOKA station, and most of the movement was closer to Pu’u O’o and durectly upslope of and at Kalapana. The dike also began at this location, right next to the quarry west of the highway, which is at one of the 1955 vents. Magma got within 100 meters of the surface here but the dike advanced downrift before it erupted above the highway. It seems this dike formed when magma stopped moving in the deep rift and got shallower.

            Lava in 1955 was found to be erupting with high fluidity at 1020-1070 C, the measurement done with an optical pyrometer. The real numbers might have been as much as 100 C more, which is about the same as the 2018 numbers. The vents of that eruption that were furthest east had a lower temperature though. 1960 was up to 1100 C, it isnt clear if it was measured the same way as above, but it seems the temperature of eruptions here is generally fairly consistent. Maybe the fact these eruptions need to move through a dike so far from the summit gives a much higher cooling value than for a dike that forms up at the summit.

            I do somewhat suspect the 2018 temperatures could be underestimations though at least for fissure 8. The samples were from the channel, not the vent for obvious reasons. The upper part of the channel was very turbulent compared to the rest of the flow so probably lost an enormous amount of heat, possibly the major reason why there was such intense storms above. The lava at the vent itself was probably at least similar to Pu’u O’o temperature, although so far from the summit it probably isnt going to be as hot as the lava from Halemaumau or Kilauea Iki can get. This could also be true of the 1960 eruption, pyrometers get minimum values as they are based on surface temperature. And it is known at least some of the 1959 magma was very primitive and extremely hot, maybe some of that went east too.

  11. Just in case I interprete this in the right way, the whole area might haev been partly connected in the Pleistocene:

    The Indonesian archipelago nearly reached its present form in the Pleistocene period. For some periods, the Sundaland was still linked with Asian mainland, creating the landmass extension of Southeast Asia that enabled the migrations of some Asian animals and hominid species. Geologically the New Guinea island and the shallow seas of Arafura is the northern part of Australia tectonic plate and once connected as a land bridge identified as Sahulland.

    Also interesting: Same pigs in the Philippines and Sulawesi:

    In the course of this survey, we also discovered tusks of Celebochoerus, an extinct pig first thought to be restricted to the island of Sulawesi in Indonesia South of the Philippines. This fossil, which we described and named, is important since it hints on the possible Philippine origin of at least some of the well-described Sulawesi fauna.

    So, Bawean might have been part of a larger subareal structure.

  12. Some additional interesting papers about Potassium-rich lavas in collision and subduction zones around former Tethys Ocean.

    Constraints on the Genesis of Potassium-rich Italian Volcanic Rocks from U/Th Disequilibrium

    Subduction, ophiolite genesis and collision history of Tethys adjacent to the Eurasian continental margin: new evidence from the Eastern Pontides, Turkey

    Identifying Tethys oceanic fingerprint in post-collisional potassium-rich lavas in Tibet using thallium isotopes

    • To be precise the middle article from Turkey doesn’t mention Potasium, but describes very neatly the setting of the suture line in Turkey.

      In the first article Vesuvius is described as ultra-potassic (under results).

  13. Controversial:

    In one of the articles about HTHH there was an urgent appeal to build more equipment between Samoa and NZ.
    If I consider this roughly as VEI 6 (higher VEI 5) and add the other two VEI 6, Kuwae or Tofua (mystery eruption 1452/53) and Taupo (230) I wonder whether this makes sense at all. VEI 6 are not that frequent.

    In the 20th century there were three VEI 6, Santa Maria, Novarupta and Pinatubo, so all elsewhere. There were five VEI 5, in Chile, Russia, Japan and Indonesia.

    So how much sense does it make to build an equipment which will not be used in foreseeable time in a remote area? Statistically seen, of course.

    In the Philippines and Indonesia much more damage is caused by volcanoes with a VEI 3 or 4, say Taal or Semuru last year because of dense population. The deadly Mount Pélé Eruption in 1902 was a VEI 4. And Guatemala speaks its own language and so on.

    I believe this is cheap screaming for money. It’s over. Probably for half a millenium or so. Statistically seen.

    As VC pointed out many times there’s more to a volcano than VEI, there’s above all population density and equipment and the nature of the beast. Pyroclastic flows on land are much more dangerous than in the South West Pacific submarine.

    This one day demands an article of one of the big guys, Carl or Albert, imho.

    To be reminded: Six deaths Jan 2022, about 30.000 deaths Japan 2011 because of an earthquake resp. seaquake, more than 3000 deaths Indian Ocean because of a similar quake near Sumatra.

    • As of now we do not know how to forecast caldera collapses. More data could help understand calderas. Had Hunga Tonga Hunga Haapai been monitored before its eruption we may have gained insight into how a caldera collapses.

      Tonga-Kermadec calderas are not particularly dangerous to people given their remote location, but that doesn’t mean the equipment will be useless, at contrary. The equipment at major continental calderas like Yellowstone or Taupo might end up being less used because there might not be too much going on there for centuries or thousands of years. The Tonga-Kermadec arc has probably the largest concentration of caldera volcanoes in the planet, and many of them are highly active.

      I am aware of at least two volcanoes that produce cyclic CVLD earthquakes in Tonga-Kermadec (earthquakes of caldera ring faults), an unnamed volcano right next to Tongatapu, and Curtis Island which produces the largest CLVD earthquakes in the planet. I also suspect Volcano F threw a CLVD right before its 2019 eruption, and that Hunga Tonga Hunga Haapai may have done a few of them in the decades before this year’s collapse. There are only 6 or so other volcanoes on Earth that have produced cyclic CLVDs related to caldera inflation in recent decades, only 3 of which are in accessible locations (Rabaul, Sierra Negra, and Bardarbunga). Several Tongan calderas have produced large eruptions in recent decades. For example, Havre erupted 1.5-2 km3 of rhyolite in 2012. We do not know how often caldera collapses happen in Tonga -Kermadec, these smaller calderas have faster collapse cycles than larger calderas. Hunga Tonga previously collapsed around 1000 AD, which makes for a 1000 year interval between its two last collapses. Most calderas of this arc are not studied, and the timing of their last collapses are not known.

      I do agree that volcanoes that are endangering populated areas are more important to monitor. But there could be great insight to be gained from monitoring some of the more active Tongan calderas. Knowledge that may help understand the more sleepy calderas elsewhere in the world.

    • I would have to whole-heartedly disagree with that statement, Denailwatch. if we had 3 VEI 6s in the last century, then that means there is a 1 in 33.3 chance of VEI 6 taking place at any given year. Which are not low odds by any means, and we have to remember that ice-core and tree ring data only give information concerning eruptions with global impacts. As Tonga has shown, these large eruptions can take place and produce negligible SO2 along with negligible global impacts that won’t be readily recognized after a few decades. Which means that these eruption could be more common than previously assumed.
      Significant climate altering eruptions( with climate impacts surpassing Pinatubo) happen around once every 150-181 years with some events happening a few years apart.
      This is peachy bias, there is no evidence to suggest that we’re safe from volcanoes for the next 500 years and in fact, the evidence points to the contrary, Before Tonga, it was generally thought that eruptions couldn’t produce long distance tsunamis like earthquakes and now we see how moronic that assumption was. As the HTHH eruption just barely made it to VEI 6 which is nowhere near close to a large eruption, even by historically standards and still produced a significant tsunami that killed people on an entirely different continent. No one can say when we’ll get the next big one, it could next year or in the next 100 years
      It’s not over because nothing has started.

      • 1. I wouldn’t have suspected, Tallis, that it is you who objects as you created the alternative scale always concerning population.

        2. You are making a mistake in your arguing when you figure out the chances for an eruption of a scale of VEI 5 or 6 as you don’t include that all the VEI 5 or 6 in the 20th century happened in Asia or Middle America, and that Campi Flegrei or Vesuvius in Europe are also candidates for big disastres.

        So I must wonder whether you read the whole of my comment. My comment is a plea for good monitoring in populated areas and again I post a reminder that one of the most deadly eruptions was the 1902 VEI 4 of Mount Pélé that annihilated a whole town, well described in Clive Oppenheimer’s “Eruptions that shook the World”.

        3. Science, mentioned by Héctor, is another thing. And it is very interesting that one of the nations that started and financed the science around HTHH is Japan. Japan is much more endangered by such catastrophes than the South Pacific:
        RIP, Henrik.

        It’s worth reading the list again, it is a good list. I don’t think that anybody who sticks to the criteria for the list would add the South Pacific. And before the South Pacific, long before, there must be Iceland, of course. But those guys do their research and stoically continue living on the MAR and possibly (Carl would say safely) a hotspot instead of screaming. Europe would get a volcanic winter with a larger eruption in Iceland and should certainly save its money for that area plus Italy.

        4. Then there are also things like the earthquake in Managua or the other legendary earthquakes in populated areas. It should be population density that decides about these finances not dead fish or some UNESCO site and touristic hotspot. In three years nature will start to recover. Nature is stronger than people living on volcanoes. Taal or Fuego are definitely more important than a remote area, no matter how high the column was. Pitturesque, good for ah’s and oh’s. Btw, we are already having sort of a winter here.

        • To be added, Nyarongo caused (only) 32 deaths (26 more than HTHH), but much more displacement and homeles people and is an ongoing threat.
          So, Nyaragongo doesn’t only need equipment, but also money for teams who supervise it as otherwise it is stolen and sold on markets.

        • Eruptions with large population densities should be prioritized but more isolated volcanoes should not be neglected, Volcanic eruptions are capable of producing distal impacts that the could cause significant damage through tsunamis and volcanic winter. a relatively modest sized eruption produced a 20 m high tsunami for the country of Tonga and killed 3 people. Just because an eruption doesn’t destroy some densely populated area doesn’t mean it isn’t important. eruptions in the south pacific can still threaten over a 100,000 people. Certainly causing more of an impact than “some dead fish.” The Tonga eruption was more an exciting research opportunity, it was a tragedy for many people and there are more volcanoes in the region that could cause similar damage. The world has more than enough money to monitor remote active volcanic regions, just take away the budget for the maintenance of nuclear weapons.

          I don’t see why where the eruptions took place matters when it comes to calculating the odds. The eruptions happened and that all that matters, no one should care what continent it takes place unless it’s on Antarctica when assessing the risk for volcanic disasters. Middle America and Asia have highly population dense regions and some of the most active regions on the planet so it’s natural that they would have more frequent large eruptions. No matter how you look at it VEI 5-6 eruptions are pretty common throughout history. Ice core data has show that there has been at least 25 eruptions that produced more sulfate than Pinatubo over the 2500 years which again, doesn’t account for sulfur poor eruptions.
          Whatever winter we’re having right now is nothing compared to what a volcanic winter is capable of.

          • “Volcanic eruptions are capable of producing distal impacts that the could cause significant damage through tsunamis and volcanic winter.”

            That is right, Tallis, but you cannot avoid an eruption, you can only evacuate people. And the more people are concerned the more important this seems to me. Think of some pics Carl posted in an article about Fuego.

            So, however you monitor you cannot avoid a volcanic winter or a tsunami. So it seems that highly populated areas should be closely monitored, and the money should go there in the first place.

            It would be not too small a catastrophe if the money went to the South Pacific, and then Nyaragongo or Nyramura erupted and took 1000 this time. Everybody would shake their heads.

          • The world has more than enough money to monitor every active volcanic region in the world USGS has 1.7 billion dollars in funding and global military spending over 1000 times that so just 10% percent of the military budget in addition to current monitoring resources would be enough. You don’t need to monitor every volcano but you can monitor every region that poses an appreciable risk. Knowing when a large eruption would take place would aid in the preparation. THere is no need to monitor one region and abandon another due to abundant resources that could be pooled from other areas along with international cooperation.

    • Distributed acoustic sensing, a technique that uses optical fiber cables to record seismicity, could be suitable to study Tonga-Kermadec calderas. The best target to monitor would probably be the two large unnamed submarine calderas that are 40 km to the southwest of Tongatapu, closer to the main island of Tonga than Hunga Tonga. These two calderas are dacitic and have produced enormous ignimbrites that are visible in the seafloor, in eruptions that were probably much bigger than Hunga Tonga. Thus they could be hazardous. One of the calderas produces cyclic CLVD earthquakes, so we know it is actively inflating, the other has an spectacular resurgent dome that overtops the caldera rim in places, and makes for a rare structure.

      Niuatoputapu island in the northern end of Tonga is surrounded by major calderas. Curacoa and Niuatahi to the north. Five unnamed calderas in a row to the south which are flanked by massive ignimbrite deposits that overlap with each other and make the largest field of submarine ignimbrites I know of. Niuatoputapu itself may have been a caldera at some point. It would be a good location to study, even if those calderas are not dangerous due to their remoteness.

      • Could you specify which two calderas southwest of Tongatapu. I believe I may have read a paper on one of them. It could include valuable info.

        • These two calderas:

          They are much larger than Hunga Tonga with sizes of 7×4 km and 7×5 km, Hunga Tonga in comparison has a diameter of 3 km. The southern caldera is think should be studied given that it is active, as evidenced by the large ring fault earthquakes, is known to have erupted in 1916 and 1932, and is also close enough to Tongatapu that a major eruption could be dangerous.

        • The northern caldera is probably hiding a large magma chamber under that resurgent block. So that also makes it worthwhile studying this volcanic complex.

          • Those links are interesting. The northern caldera (Volcano 1) has spectacular pyroclastic density current deposits from caldera-forming events. And according to the second article you’ve linked these deposits make up the entire 300 meter tall caldera scarp. Likely this caldera has collapsed again and again, several times, in large VEI 6 magnitude eruptions. Unfortunately the authors of those articles do not seem to have recognized the enormous Suswa-like resurgent structure that occupies the centre of the caldera, which shows the caldera floor has been pushed up by several hundred meters and is probably ready to caldera collapse again.

            The northern caldera (Volcano 1) seems younger to me than the southern. Volcano 1 pyroclastic density current deposits are un-modified so they probably overlie the caldera-forming events from the southern caldera. This is why I’m surprised that the CLVD earthquakes have been located in the southern one, it would be more likely to me the CLVD originated from the northern one. The location of earthquakes in these remote areas is very poor, so maybe Volcano 1 is actually the one making the ring fault earthquakes as it pushes the floor up into the resurgent block.

            The northernmost 800 km of the Tonga-Kermadec arc has an spectacular collection of calderas surrounded by seemingly young pyroclastic density current deposits. There are 12 calderas here surrounded by dramatic pyroclastic deposits like those of Volcano 1. There are as many older calderas, whose PDC deposits seem to have been buried under younger material or eroded. Two or three volcanoes lacking calderas but with large scale PDC deposits in their flanks. There are also Tofua and Hunga Tonga two young caldera systems that still lack clearly defined submarine aprons of PDC deposits. And the 9 km Niuatahi caldera lies in the back-arc, which is probably a deep water effusive caldera with a most impressive submarine ring of lava domes that is probably unique in the world.

          • I managed to find some a paper on volcano 2. It’s the one with the clvd earthquakes. There is also info on other nearby volcanoes and calderas.
            It’s 100% an active region. I hope the hthh eruption will bring more attention to these. Maybe one of the mystery eruptions are around there. According to the link below, volcano 1 had a very violent eruption in the last 200 years. 1808?


        • I should also note that multiple calderas collapsing in close succession in the same area in known to have happened multiple times. Hunga Tonga itself previously collapsed at around the same time neighbouring Tofua collapsed, around 1000 AD.

          Other cases known are Kikai and Mashu in Japan, whose pyroclastic deposits overlie each other with no intervening paleosoil.

          Or a wave of three major caldera-forming events in the New Britain arc in the 7th century:

          While it is probably unlikely that the next caldera will form in Tonga, it is a possibility that cannot be neglected either

          • Wide chunk of mountain. When you let the water go it appears before the MAR and is very thick. As it is full of volcanoes it might show us how the Andes might have started under water. Back then (starting point around 20 Ma) the Andes raised their head out of the water. Before they might have looked like the Tonga-Kermadec-Trench and the trenches further north. Subduction would have been steeper then, possibly.
            So, this might be an orogeny in the making that has already popped out of the water in New Zealand and other places.

            The Pacific Ocean is sort of a bath tub with a rough bottom though, once you let the water go. Very different from other oceans.

    • New Zealand has in the past been hit by massive tsunamis, with run-up height on the north coast of tens of meters. The causes are not well known, I believe. Kuwae is one of the suspects. Hunga Tonga has now confirmed that an eruption can cause on ocean-wide tsunami. And this eruption wasn’t huge. (Large, yes. Huge, no.) Kuwae is suspected (but not proven) as causing the collapse of intra-ocean trace in the Pacific. There are many volcanoes around the Tonga region. There are also many large earthquakes. The latter can trigger the former, for instance with a flank collapse, one of the possibilities for what happened at Kuwae. In my opinion, a warning system is warranted. It is a trade-off between the frequency of events and their impacts. The Netherlands tries to protect its polders against 1 in 10,000 year events. The UK doesn’t even do 1 in 100 years. Should Basel worry about the next earthquake? Koeln about the next eruption? The world about the extermination of insects? One article I saw argues that by 2050 people will start dying of whatever has decimated the insect population across the world. Choices have to be made, but it is the task of the scientist to point out what the choices and consequences are. They may not have the final decision, but that doesn’t mean they should be silenced. Although some do try.

      • Great statement. Yes, it is all about choices.

        I believe the Netherlands are doing the right thing being densely populated as the Netherlands are small and a big chunk of land has been won from the water and can easily be flooded. The country is basically densely settled everywhere and flat like a flounder.

        The Netherlands have offered – I read a few years ago after a large storm in GB – to give the UK advise, but the offer wasn’t accepted. In earlier times there might have been more wisdom. The only dry place in Tewkesbury in storm is the Abbey.

        Managua has made the wrong decision – everything has been built up – as far as I know – on the same terrain that had been destroyed. Monitoring is not everything. Monitoring is good when the lessons are learnt. Some people (comfy politicians) seem to falsely think that monitoring is a protection. It is only a system for early alert.
        So, as Henrik rightly wrote, there might be an alert without an consequence.

      • We have a warning system, and the heart of it is our DART buoy network.

        And in your list of potential tsunami sources, you missed the one that is perhaps the ‘quietest’ and least known, but may be among the most dangerous; the New Hebrides trench.

        • That is near Kuwae.Aside from that, first thing that pops up when googling is a 1939 earthquake. No damage. This is better compared with San Francisco, Lisbon, Managua or Japan.

          • I would say that the New Hebrides trench does not appear to have released much tension recently, just a bunch of large-ish 7.6/7.8 earthquakes (of which one was a doublet), the largest yet being 8.1-8.2. The trench is subducting an ocean ridge (d’Entrecasteaux ridge) in the centre which slows the subduction but at the northern end it is 170mm/yr and at the southern end 120mm/yr. It is a megathrust fault but not how much of it can rupture given the ridge-subduction is seemingly stuck.

  14. There’s a reasonable moan that crops up on here from time to time, and that is that there’s too much attention paid to Iceland. And yes, I plead guilty to having a great love for that place, its people, its landscape and its geology. But: Katla has been seeking our attention again.

    22.11.2022 20:11:12 63.655 -19.097 0.0 km 3.5 99.0 7.7 km ENE of Goðabunga
    22.11.2022 20:10:42 63.657 -19.078 0.0 km 2.5 99.0 8.6 km N of Hábunga
    22.11.2022 20:09:14 63.662 -19.072 0.0 km 2.5 99.0 9.1 km ENE of Goðabunga
    22.11.2022 20:09:14 63.568 -19.159 0.2 km 2.1 90.01 3.2 km WSW of Hábunga
    22.11.2022 20:07:57 63.663 -19.112 0.0 km 3.8 99.0 7.3 km ENE of Goðabunga
    22.11.2022 19:56:06 63.668 -19.102 0.1 km 2.9 99.0 7.9 km ENE of Goðabunga
    22.11.2022 19:55:15 63.659 -19.145 0.1 km 3.9 99.0 5.6 km ENE of Goðabunga

    We had another swarm like this (but lower magnitude) in recent weeks, but in the last few years these shallow swarms occurred in late summer when, it was explained, snow melt decreased pressure on the subglacial geothermal areas. I wonder whether the pattern is changing.

  15. Past year of quakes at mauna Loa. The rate today is not as hig has it was when the alert was brought up but the background level is now a lot higher than it was before then. 300 quakes a week.

    I dont know what percent of those are of a size that would have been detected in 1975 or 1984 but HVO has said in recent public meetings that the rate seen now is very comparable to before those eruptions, it isnt just a sensitive instrument bias…

    1975 took about a year of high quakes and inflation before it erupted. 1984 was much more rapid, but could have been partly opened from the other eruption still. It is still far from certain but after inflating for 20 years and holding high levels of shallow quakes like this for 2 months I think this is now in the final stretch. Personally giving this a greater than 50% chance to erupt before the end of 2023.
    As to what sort of eruption, probably a fissure like 1975 going across Mokuaweoweo and staying at high elevation. But 1975 only had a year of inflation, where there has been 20 years of inflation to date up to today… so I think the chance the next eruption might be a lot more serious is probably higher than many would like it to be…

  16. Quakes before 1984. 300 quakes per week is about 42 daily. Actual rates seem slightly less, closer to 30 a day. But as can be seen, the rate in 1984 didnt get above this until only a couple months before erupting…

    The GPS is also going up, and the caldera is still extending rapidly

  17. Another quick Ingenuity update.

    Yesterday Ingenuity completed Flight 34, which was a back to flight school first solo. New 3D aware navigation and landing hazard avoidance software has been uploaded and this was a simple hover test to make sure everything worked as expected. Ingenuity is preparing to resume scouting ahead of Perseverance with its upgraded capabilities.

    Flight 34 Was Short But Significant

    Flight 34 may not seem like much, but it was Ingenuity’s first with this software update. The team will use results from this simple flight to start testing these new capabilities, ensuring that everything works as expected on the surface of Mars. The update brings out new functionality in Ingenuity, making it a far more capable vehicle and effective scout for Perseverance. We’re all excited to see where this update will allow us to take Ingenuity’s journey next!

    that will allow it to safely fly up the steep terrain of the Jezero river delta, scouting ahead of the rover Perseverance as it searches for signs of past life on Mars.

  18. CRYOVOLCANIC ERUPTION ON A COMET: The British Astronomical Association (BAA) is reporting a new outburst of cryovolcanic comet 29P/Schwassmann-Wachmann. On Nov. 22nd, amateur astronomer Patrick Wiggins watched 29P increase in brightness by more than 4 magnitudes, promping a worldwide call for observations. On Nov. 23rd, André Debackère used the Faulkes Telescope North in Hawaii to photograph an expanding shell of debris:

    Image at the link.

    • And this on the heels of only the 6th asteroid that has ever been detected prior to impact…in this case the meteor/fireball fell over Canada with hundreds of ground reports.
      The asteroid, though only a meter or so in size, was detected by an amatuer ~ 3 hours ahead of time who then relayed the info to some observatories for a look.

      • 7th .. We know that Phobos will eventually impact Mars!

        And I could add Shoemaker-Levy although that was officially a comet. That would add more than 20 impacts to your list

  19. Definitely a bit of a stack of quakes under Kilauea this month, all the way down to the same depth as the Pahala quakes. There also look to be more quakes that are at 5-10 km depth than in most other weeks on the weekly plot although far fewer shallow quakes, steadily increasing since a complete stop about a week ago. It looks like another intrusion happened again on that day, but this one was deeper than 3 km so was not in the zone of brittle rock and so didnt result in an earthquake swarm on one of the rift zones. There are some moderate isolated quakes that are out in strange areas of the volcano, like the quake to the east of Kilauea Iki north of the ERZ, or west of the caldera on the lower flank of Mauna Loa. The latter at this particular depth seem to be a Kilauea thing more than a Mauna Loa thing despite the surface location, more concentrated swarms in this area have often preceded intrusions at Kilauea although that is only in regards to shallow activity.

    It is very hard to figure out what this really is, the volcano was clearly pressurizing until recently and had low rate of eruption to show magma was accumulating somewhere else. All of the stations on the Hilina pali and over to Kalapana are showing accelerated movement to the south, my only conclusion is that this has allowed magma to intrude deeper than the shallow systems. This coudl well be a continued response to the 2018 activity, where a deeper intrusion probably happened and may have either been caused by or was the trigger for the big quake of that year. South flank quakes have never stopped since then, I guess it was only a matter of tiem before magma began flowing back there.

    I wouldnt be surprised if this results in a few years of low activity, or even if the eruption stops completely. But at the same time, the last time this probably happened was after 1975, and after the volcano recovered from that it created Pu’u O’o… So high ERZ activity could be triggered by these deep intrusions, either by direct contribution or just by keeping the area open, likely a bit of both. Consequently things might well get very interesting in the later part of the 2020s 🙂


    Some information on Katla. Seems it might actually be doing something for real now, although 14 cm of uplift is not necessarily a big number for a caldera like this. Does Katla have an active ring fault?

    Definitely not the most dangerous volcano in the country though, not when places like Grindavik or Hafnarfjordur have the possibility to see eruptions within their borders in the near future, or Hengill wiping out the primary power station to Reykjavik. But those would all be VEI 0 so not scary enough…

  21. Interesting article, I had never heard of this island/volcano before.

    I’m just a bit uncomfortable with the following sentence: “If this plate is the source of the magma, as it is in the volcanic arc, then the magma here comes from a very large depth”. I’m always uncomfortable with such sentences, that could be interpreted (I’m not saying that’s what you meant) as melting of the subducted plate itself—unless we are talking adakites, of course. The plate is not the source of the magma, the mantle wedge is.

    Sorry, I’m a bit fussy about this… Otherwise great read, thanks!

    This comment was held back for approval by the system – something that happens to al new commenters (and is sadly necessary). Happily approved and future comments shouls appear without delay

Leave a Reply

Your email address will not be published. Required fields are marked *