As more information comes in, it is time for a brief update, and for a theory on why the eruption of Hunga Tonga was so destructive for what was, after all, a fairly small volcano.
The death toll of the tsunami at the moment stands at five, three at Tonga and two in Peru. There is extensive damage along the western shores of the Tonga archipelago. Damage at the capital is limited, but other islands have not yet been contacted, and it appears that they have been affected much worse than initial reports suggested.
There is damage further away, boats in harbours in New Zealand and Japan, and some minor coastal damage in the US. There is an oil spill in Peru, from a tanker that was pumping out oil as the tsunami rolled in. The coast is polluted over 1 or 2 kilometers. But these pale into insignificance compared to the local damage.
Hunga Tonga has been reduced to two small slivers of land, as shown above. (Officially it is called Hunga Ha’apai, and Hunga Tonga is the eastern ridge, but this distinction has quickly been lost.) It is interesting that these are the highest ridges of the old caldera rim. The lower lying land, perhaps some tens of meters above sea, is all gone. That includes two small islands that existed 3 kilometers to the south, on the southern caldera rim. The destruction suggests that it was caused by more than the explosion (which would have gone for the steep, high ridges) . The tsunami may have contributed to the lost land, especially if it was much higher than we assumed.
This is from the same Sentinel image as above, but using different filters. It covers a larger area of about 20 km across. The bright spot right of the centre is the remnant of Hunga Tonga. The linear stripes are floating debris.
Photos suggest something like 1-2 cm of ash on Tonga, 60 km from the epicentre of the blast. That information is not enough information to calculate the total volume of ash! For that you need to know thickness at several distances from the eruption. But we can do a comparison. Hekla in 1108 AD covered about 10,000 km2 within the 2-cm isopach (contour of ash thickness). Note that that was with a strong southerly wind: the ash was almost entirely to the north, with 2 cm thickness out to 150 km away but only in a narrow cone. Assuming equal distribution in all directions, the 2-cm isopach of Hunga Tonga would be similar in area. Hekla ejected 2.5 km3 of tephra. We are in VEI-5 territory, but with very large uncertainty!
Comparing to to Ilopongo 540 AD, where the same isopach covers an area of around 150,000 km2 shows that we are very definitely not talking VEI-7! Taking that eruption (107 km3 tephra including ignimbrite) and scaling it to our area gives 8 km3 for Hunga Tonga. Those two numbers, 2.5 and 8km3 give us a ballpark, but no more than that. We don’t know enough yet.
A report from Tonga itself speaks of 10 cm of ash. That would push the numbers up to over 10 km3 of ash. However, I expect this refers to maximum thickness (wind blown) rather than average.
There have been claims that this eruption was ash-poor for its size. Before we make that conclusion, we need to know much more about the amount of ash. These estimations point at a VEI-5 tephra class, but it is based on a number guessed from one long range photo!
The tsunami was far worse than first reports suggested. The early videos and stories showing a dangerous but manageable wave appear to have come from sheltered areas. The capital was not exposed to the largest waves. But elsewhere on the archipelago, waves of 5-10 meters high have now reported. A few small islands were wiped off the map. This explains the size of the long-range tsunami, with 0.5-1 meter commonly across the Pacific. The height of a tsunami depends a lot on the shallowing sea near the coast, and on bay resonances. A 10-meter wave at Tonga would be down to 10 cm at the distance of Alaska (this includes a correction for the curvature of the Earth – people adamant that the Earth is flat will see a smaller tsunami). That is indeed roughly what was reported.
The scary thing is that a VEI-5 managed a Pacific-wide tsunami. Imagine what a large VEI-6 could do in similar circumstances!
Let’s discuss three aspects: the explosion, the tsunami and the air pressure wave. After I will talk about the destruction of the islands of Hunga Tonga before the bottom line: how did this happen, and can it happen again?
This was an underwater explosion. This is clear from the fact that the area above sea level around the centre of activity had already been removed b the explosion of Jan 14, itself possibly VEI 4. The sea had covered this area.
There are of course examples of underwater explosions. We even had made some ourselves. Here is one we made earlier: the Hardtack Umbrella event, in hindsight a stupendously stupid polluting idea to use nuclear weapons in the sea. It was an 8 kton explosion. detonated 50 meters below water. On the video, note the wave engulfing the ship.
The wave is caused by displacement of water. The big explosion vaporises the water around it. This expands by a lot (roughly taking up 1000 times its original volume). If a globe of 100 meter radius evaporates, it creates a bubble of 1 kilometer in radius. (The Hardtack explosion did not reach anywhere near this size.) The bubble expands supersonically – faster than the speed of sound in water (1500 m/s). The water can’t get out of the way that fast, and a shock wave forms. The bubble continues to expand but it overshoots. Once it gets too large, the pressure in the bubble is now lower than that in water, and the bubble begins to collapse. You will get a pulsating behaviour where the bubble expands, collapses, expands, etc, all within a fraction of a second. There are many youtube videos showing this effect. Each collapse causes an overpressure of water at the central point, and this causes a jet of water suddenly spraying up, as seen in the video above. And if the explosion is at the bottom of the sea, this jet can also bring up debris from the sea floor.
The description above is for explosions where the bubble is smaller than the depth underwater. This is also the case for those youtube videos on experiments of underwater explosion (if only for reasons of safety!). In the case of Hunga Tonga, this wasn’t true. The bubble became larger than the depth, and exploded into the air on its first expansion, with a speed far exceeding the sound speed in air. This caused the huge bang heard as far as Alaska (though regrettably not by our local correspondent motsfo): it was a sonic boom amplified by the fact that the speed of sound in water is so much higher than that in air.
On reaching the freedom of the atmosphere, the overpressure in the bubble dropped instantly and water saw a chance. It pushed back into the hole, with the water flowing in from all directions and colliding in the centre at the speed of sound. This violent collision threw a big jet of water kilometers into the air. It also caused the water bulge which started the tsunami. The bulge, tens of meters high, began to travel outward.
The speed of a tsunami depends on the depth of the water: it goes as the square root of the depth. The wave took around 10 hours to arrive at Alaska. That gives a speed of 250 m/s. This speed is reached for an average depth of roughly 6 km. This is indeed the depth of much of the Pacific.
The wave travels at this enormous speed through the ocean, but at very small amplitude, only a few centimeters in height. A wave a few hundred kilometers long and a few centimeter high is difficult to detect! But when the wave approaches the coast, things change. It slows down by a lot. When the sea becomes 100 meter deep, the speed is down by a factor of 7. The wave from behind is still traveling at the old speed. The water now piles up. This piling up causes the tsunami.
The same effect occurred at Hunga Tonga itself, but in reverse. Initially the wave was slow, as the water around Hunga Tonga is not deep. It accelerated as it reached deeper water. The other Tonga islands are sitting on a separate underwater ridge, so here the wave slowed down again and the wave height build up. The height it reached would have depended very much on local conditions, the presence of a reef, etc. But the wave would have come suddenly. As it began as a water bulge, there was no initial withdrawal of the water. There was no real warning.
There has been discussion whether the long distance tsunami really was a water wave. The alternative would be a meteo-tsunami, formed by the air pressure wave. The timing settles this question: the arrival in Alaska agreed with speed of the ocean wave, whilst the air pressure wave had arrived saveral hours earlier. The tsunami around the Pacific really did come from this single-point failure, the Hunaa Tonga eruption. It will have been amplified by local conditions, such as the shape of a bay, but it originated from the sea.
But what about the reports of waves in the Atlantic? That is rather far away, and America forms a bit of an obstacle: it takes a big wave to cross that! In the Caribbean, waves of some tens of centimeters were reported. That could be within range of what can be caused by the air pressure wave. And it is not unique: after Karakatau 1883, the tsunami reached far and wide with an arrival time consistent with the known depth of the oceans. But in the Caribbean it arrived much too early. It was unexplained at the time, and a different cause was suggested, perhaps some local event such as an earthquake. Perhaps this too was the air pressure, a meteotsunami. Meteotsunamis are often caused by wind, but that can be excluded here.
The explosion was heard far and wide. It arrived at Tonga still as a sonic boom, and it was deafening. That would have been within minutes of the explosion, perhaps five minutes before any ocean waves arrived.
It is an interesting fact that closer to the eruption, nothing may have been heard. The sound wave was emitted in all directions. But the speed of sound is larger at sea level than at higher altitudes. It depends on temperature (not pressure): the sound speed is higher in warmer air. As you go up in the atmosphere, the temperature drops and sound slows down.
Imagine carrying a very big vertical stick (representing the crest of one wave of the sound where the sound is moving horizontally; it may be easier to visualize as water wave on its side). If the top moves slower than the bottom, then the stick will not remain vertical but will start to bend back. The sound wave travels at a right angle to the stick, so now it begins to travel away from the ground. Indeed, when St Helens blew, people tens of kilometers away (where they should not have been) heard nothing. As far as sound was concerned, it was a deadly quiet zone. The same is true for lightning storms. You hear the thunder for nearby storms, but not for ones you see on the horizon. You are in their silent shadow zone.
Sound carries further if the air at ground (or sea) level is colder than air higher up. In that case the wave hugs the ground and can be heard farther away. You really do hear better in winter!
Once the sound wave travels upward and away from the ground, it eventually reaches the tropopause. But here the opposite happens. The temperature, and thus the sound speed, now increase with altitude. The temperature is lowest at the tropopause, and it gets warmer again higher up. So the sound wave does not enter the stratosphere but it reflects back to the surface.
This reflection can happen close to the ground if there is a temperature inversion. This is called troposphere ducting and it allows (sometimes) thunderstorms to be heard to hundreds of kilometers distance. That is not enough for us. By ducting the sound at the tropopause, if it is loud enough it can carry many thousands of kilometers. In the right conditions it bends back to earth, to startle the unsuspected listener with the sounds of a distant battle. In the past, it was often compared to cannon fire. Of course that is not something we are too familiar with nowadays!
The people on Tonga startled by the bang, heard something that may have come back from the stratosphere. So did people on Fiji, in Australia, New Zealand and even Alaska. Hunga Tonga, I think, now holds the record for the most distant sound heard by ear by people. Someone in Alaska is the world record holder!
The audible sound was accompanied by a pressure wave, a rise and fall in air pressure, This wave was measured all around the earth. In Manchester it arrived at 7pm UT, some 15 hours after the eruption, and lasted 20 minutes. A quick computation shows that it traveled at roughly 350 m/s, i.e. at the speed of sound. Krakatoa had done the same, and its pressure wave was measured all over the world.
There was a second pass of the wave in Manchester at 2am. This was the wave that had traveled the other way around the world.
The next day, the wave re-appeared at 7am, followed by its counter-rotating counterpart at 2pm. By this time it had gone around the world 1.5 times. After Krakatoa 1883, in some places the pressure wave was detected 7 times. Imagine a sound so loud that it was heard worldwide for five days! (Of course we can’t hear these pressure waves.) We are not there yet with Hunga Tonga but who knows.
The figure shows the Manchester wave of Jan 17 with bumps at 7am and 2pm. There are better detections available – but it is nice to see one next door.
The Manchester wave carried an excess pressure of 1.5 mbar. This was similar to what was reported at various places after Krakatoa. The sound wave appears to carry a similar amount of energy to the sound wave of Krakatoa, even though that was a significantly larger eruption. I believe this is due to this eruption being a little below the water surface. This amplified the pressure wave in a way similar to tsunamis: as the pressure wave entered the air and slowed down, it piled up and this increases the amplitude. This was (and perhaps still is?) a pressure tsunami.
Destruction of Hunga Tonga
Explosions under water create a far more destructive pressure wave than one in air.
This wave expands out into the water. The water is incompressible and just passes the wave on. When it reaches something solid, it becomes hugely damaging, as the wave has the weight of water behind it. A torpedo acts this way: the shock wave in water punctures the hull of a ship. (It also does considerable more damage as the wave passes through the structure of the ship, much more so than an explosion in air would.)
Such a pressure wave was set by the original explosion at Hunga Tonga. But this was within a confined space. The old caldera wall, 3 kilometer wide, is still there. This shock wave would have bounced off the shallow bottom and the steep wall and reflected back towards the centre. Perhaps this is the cause of the wide destruction of Hunga Tonga. The two small island on the southern rim are gone, even though they were some distance from the explosion. I imagine that this supersonic shock wave shook the old crater rim to pieces below sea level. This acted together with a water wave perhaps as high as 50 meters and the explosion above the water. Only the toughest parts of the caldera ring, protected below water by a shelf and above water by its height, survived.
The power of water
That leaves the question how this eruption happened. Why was it so large and so loud?
We are talking about water getting in contact with magma, as driving the explosion. It has been presented here as a collapse of the roof of a magma chamber, water rushing in and instantly flashing to vapour. That picture has the basics but it is too simple.
When water is brought in touch with lava, there are explosions. But there isn’t one big explosion. For instance, when lava entered the Kilaueau lake (remember that one?), the lake boiled away but it did not cause a big explosion. That is because of insulation. The water touching the lava becomes vapour. It now forms a layer between the water and the lava, and this layer insulates the water. The rest of the water does not boil until the vapour has moved out of the way and the process begins again. Lava itself also insulates well, and once the surface has lost heat to the water, it takes time to heat up again. This is not the way to get a big explosion.
What you need is to confine the water, so nothing can escape. You also need to put it under pressure, and give it time to heat up. The pressure raises the boiling temperature so now the water can be superheated. But don’t put it under too much pressure, as it can no longer boil (see the post on black smokers).
How is this done? For Hunga Tonga, the process started with the Jan 14 eruption. It destroyed the central part of the island to below sea level, and left the conduit open and under water. Water flowed in, but did not do too much except make a big steam cloud, for reasons explained above. Now the conduit collapsed. Suddenly the water was caught between a rock and a hot place, and under considerable pressure from the mountain above. I imagine this happened a few hundred meters below sea level, but no more than 500 meters because the rock would have carried too much weight for the water to boil. The water could not escape, and it began to heat up. Eventually some water boiled and vaporized. The expansion as it vaporized increased the pressure: remember that water is not compressible, and does not respond well to being pushed. The extra pressure broke the blockage. This dropped the pressure, and therefore the boiling temperature came down. All of the liquid water suddenly found itself above the boiling temperature, flashed into vapour (cooling down quite a bit in the process) and suddenly the conduit was 1000 times too small for its contents. And up it went.
The same process could happen on the flank as well: the flank breaks open, water goes in, flank slides down which closes the hole, and it is time to get yourself to a safe distance. It is now a ticking grenade.
This model explains why Hunga Tonga’s destruction happened a day after an earlier large eruption which took the cone top below sea level. It explains why Krakatoa’s huge boom came after three or four big booms over the previous day. These kind of events require priming. And finally, it explains why Krakatoa’s magma chamber survived. (We know that it did because Anak Krakatau formed in the same location.) The roof of its magma chamber was fine: all the action was in the conduit. It predicts that while the initial explosion happened under water, it immediately reached the surface: the vapour bubble grew larger than the depth of the explosion which was only 100 to 500 meters below the sea.
This model it makes it less likely that another such explosion will happen soon. Grenades don’t go off twice. It is possible there is a second grenade there waiting to be primed. For now, however, we should look very carefully at similar locations elsewhere around the world. We know similar sea level volcanic islands, and have discussed them here at VC. We will be much more cautious about them in the future. This kind of explosion is not common. It may not happen every century. But once burned, twice shy.
This won’t be the end of Hunga Tonga. One day the island will reform, and the cycle begins again. But that will take time. In the mean time, few people will ever be able to say to have been on this island. That also is similar to Krakatau: if we had know it would cease to exist, they would have documented it better. We know remarkably little about it. You may want to look at the story below, perhaps the best remaining photographic record of this phase of Hunga Tonga. I think the person is too old to qualify for a Darwin award, but what a story to tell, walking on the biggest bomb on earth. Almost nothing in the photos and video still exists.
Albert Zijlstra, Jan 2022