In my last post I introduced the model for a new eruption mechanism/style. I will be referring to these events as big basalt blasts, this is just the silly preliminary name, not its definitive one I hope.
So how did it work? I will briefly summarize. First a magma reservoir drains through a lateral eruption or intrusion, lowering magmatic pressure allows improved contact between magma and water heating up the hydrothermal system through convection to very high temperatures, if the hydrothermal system is at the surface shallow steam explosions happen and finally one of these explosions can uncork the superheated water at depth flashing into steam in a large blast. In the heights of the phreatoplinian eruption some juvenile basaltic magma gets incorporated into the explosion, but often very minor amounts.
I have been able to confirm 3 such events since 1900, in the last post I talked about 2. The third event was actually the one that made me realize there was an eruption mechanism that didn’t seem to be well understood, this is the one this post will be about.
It was produced by a volcano that inspires deep respect to the volcanic community, a volcano in the Philippine island of Luzon. This is the story of how Taal gained much of its terrifying reputation and the aforementioned model is going to help us see beyond the surface to what was happening in the belly of the beast.
King of basalt calderas, its collapse is 24 x 15 km by far the largest of any basaltic volcano, and it is similar to those produced by upper end VEI 7 eruptions. The depression is filled by Taal Lake, formerly known as Lake Bongbong, the name by which the volcano was also called, though the true original name of the volcano was Polo. The central part of the volcano forms a square island in the lake, known as Volcano Island, with 4 short rift zones roughly perpendicular to each other. It rises to only 300 m high which is lower than some of the surrounding land. Its summit hosts the Main Crater of Taal, in the past often a spectacular display of hydrothermal activity and coloured lakes.
Barely anything is known about its history. Taal was first born as a silicic system, age unknown. By 5000 years ago it was already erupting basaltic andesite and basalt, and it seems to have been that way ever since. Many volcanoes start as silicic and then become more primitive, mafic. This may seem counterintuitive, the reason probably is that the volcano receives a high supply of primitive magma that cannot evolve as fast as it gets it. The result is that the evolved magma is gradually flushed out until not much is left.
Taal is Located within the Macolod Corridor, which is a rift related to rollback of the Manila Trench. As the trench retreats away from Luzon it creates a stretching of the island accommodated by the corridor. Rifts undergo riftings, and these need magma to fill the new space, usually provided by a volcano, this volcano here being the monster of the lake.
There are 2 ways Taal can send magma into the Macolod Corridor, to the southwest, where it can rift about 15 km away from the volcano, and to the northeast, where it can rift 70 km or more. Taal rifted to the southwest in January this year and in 1911, you will see that both events are remarkably similar except on how they ended. This is the account of 1911, a disaster from which much can be learned for the future.
Back to 1911
It all started the night of January 27, at about the same hour earthquakes start to be registered at Manila and a huge column of smoke rises from Taal. This is probably an initial intrusion below the summit (Main Crater) of the volcano. Earthquakes rapidly start to ramp up in frequency and intensity, the intrusion has started to propagate laterally to the southwest, beneath the valley of Pansipit River. Many earthquakes are felt in Manila over the following days.
This is a dyke intrusion, a small rifting you might say. How do we know? Well, first a line of fractures formed running down away from Taal, the 2 largest fractures are the faults facing each other at the sides of Pansipit River while the block in between dropped in places up to 3 meters into a graben. Where the graben meets the coast the sea invaded inland. The southwestern coast of Lake Taal rose significantly above the lake, this is the typical pattern of dyke deformation, a central graben surrounded by an area of uplift. A sword-like body of magma inflates pushing the sides outward and upward, the land right above is stretched and drops down.
More evidence is the subsidence of Volcano Island, by the time the events were over the entire island had subsided between 1 and 3 meters. There may not have been interferograms back then but with attentive observation and the knowledge required it is easy to see what was happening. The reservoir of Taal was draining into a dyke. Volcano Island was sinking as magma pressure below decreased. There is no other explanation for this subsidence because as we will see the later eruption was relatively small and ejected mostly lithics.
Let’s use the model shall we? We have a deflating reservoir, pressure lowers allowing contact between water and magma, this heats up the hydrothermal system, so how was this going? The morning of January 29 a group of explorers and excursionists climbs to the Main Crater where they watch explosions of mud and rocks. They had decided to stay overnight but noticed the increasing earthquakes and growing explosions so they withdrew back to the shore, wise choice. Sadly many were not as cautious.
The hydrothermal system has become a ticking bomb, much like Fernandina on June 11. At 11 PM explosions grow much more violent, a tall plume forms stricken by lightning. At 2 AM January 30 the system uncorks in an explosion heard as far as 500 km away, people report an earthquake at that time, close to the lake it was felt like three upward blows from the ground, yet it is not visible from the records. Another phantom earthquake?
A furious base surge races down and outwards, it cuts, debarks and uproots trees. It blows houses to bits. On impact with the lake it pushes a tsunami onto the shores. Almost every living being within 15 km to the west is killed. The eruption is short, just about an hour. It is not too voluminous, 0.075 km3 (it was measured freshly fallen so I doubt it’s an underestimate). It is rather small, there is barely any geologic evidence of it left behind, recent studies have failed to find any deposits of this eruption. And yet more than 1300 people are dead and the blast creates a scene of absolute destruction. The damage seems too excessive, how is this possible?
First what is a base surge? The concept was born with nuclear tests, it consists of a wall of material that is pushed outwards by the pressure of the blast and then breaks over. At Taal it was directed more to the west, the eruption vent was located near the east wall of the crater so that it probably projected the weak alluvium of the crater floor westward into a giant wave.
The survivors describe they heard detonations with brilliant flashes followed by a rain of mud. The mud burned on contact with the skin, in the reports I have been reading they attribute it to chemical substances and to acid. This does make sense, after all the material in the surge is described as fluid mud and it is likely to have been colder than the boiling point of water, yet the burns were much deeper than those of scalding water when inspected. A hydrothermal system can be highly acidic due to all the volcanic gasses going through and reacting, this eruption mechanism is after all a gigantic hydrothermal explosion. Much of the base surge may have been acid mud.
Taal shows us big basalt blasts are capable of producing extremely hazardous eruptions with modest volume. They represent the highest degree of water involvement there can be in a volcanic eruption so the ejected material is very muddy, often accretionary lapilli is present, base surges and high lithic content are all characteristic of these explosions. Taal also shows us that a caldera collapse is not necessary for these kind of eruptions, if the reservoir drops below a certain pressure threshold, collapse or not, a explosive events can happen. It doesn’t appear there was any collapse in 1911.
Other eruptions of Taal
This year’s eruption in January was very similar to what happened back in 1911: the River Pansipit valley dropped into a graben, Volcano Island subsided, but it didn’t culminate in a large explosion, why? There is a short interval between 1911 and 2020, maybe the space created in the rift was not enough to deflate Taal as much as needed. Other factors may have been involved but that one seems relevant.
I think that one previous historical eruption of Taal may as well be grouped as a big basalt blast, the events of 1749-1754. I said rifting events can go both ways through the Macolod Corridor, the last time it went southwest was this year, the last time it went northeast was in 1749. The eruption of 1749 formed a graben system running from the NE shore of Taal lake to Laguna de Bay at Calamba. A strong seismic swarm was also felt in the line from Taal to Talim Island and the Mountains of Antipolo, beyond Lake Bay. This suggests a dyke that may have propagated across the entire Macolod Corridor, more than 70 km long. Though with the scarce information available figuring out the exact path of the dyke may be too ambitious!
The 1749 eruption may itself have evolved into a depressurization explosion. For 3 days ashfall was so thick that people had to use lights at noon. Whether this was the case or not I think the rifting event lengthened for 5 years, this is something very common for large scale rifting events (like the Krafla Fires of Iceland). A large white plume always existed at the volcano until 1754 when an enormous and exceptional eruption took place that I find very hard to explain without some equally exceptional triggering process like rifting, a caldera collapse or depressurization-driven eruptions (a big basalt blast)..
The 1754 eruption went almost non-stop for 7 months with multiple paroxysms and times of lower activity. While there is no good volume estimate available, the event is likely to have been a VEI 5+, descriptions are quite dramatic. Fr. Buencuchillo spent the last and worst days of the eruption at a convento 17 km upwind from Taal, this is his account:
“Between 3 and 4 o’clock in the afternoon of the said 29th, it began to rain mud and ashes at Caysasay and this rain lasted three days. The most terrifying circumstance was that the whole sky was shrouded in such darkness that we could not have seen the hand placed before the face, had it not been for the sinister glare of incessant lightnings. Nor could we use artificial lights as this was extinguished by the wind and copious ashes which penetrated everywhere. All was horror during those three days, which appeared rather like murky nights and we did not occupy ourselves which anything but see to it that the natives swept off the roofs the large quantities of ashes and stones which kept accumulating upon them and threatened to bring them down upon us, burying us alive beneath their weight”
Predicting the catastrophe
This series has looked closely into 3 eruptions belonging to the same eruption style, all of them had clear signs of an increasingly unstable hydrothermal system, yet in 2 of them people did not evacuate, the other situation was just lucky for the volcano was in an uninhabited area.
So I find it necessary to write here what signs should someone look for that might indicate an eruption of these characteristics brewing. It can take place in ANY volcano either silicic or basaltic or bimodal where a dyke intrusion or flank eruption is taking place that reduces significantly the pressure of the magma reservoir. It is more common in basaltic systems and it is one of the main ways that can turn them from gentle red to killer grey volcanoes, but nothing impedes their silica-rich brothers from partaking, only that in their case a lot of fresh magma is going to get involved in the eruption (so not exactly the same eruption style).
The hydrothermal system needs to be close to the surface but knowing whether this is or isn’t the case may not always be possible. I think the red flag should be increased steaming, mud explosions, any sign of hydrothermal unrest. The activity will probably ramp up over time. By the time this is confirmed the eruption may not be too far off so its best to understand well the situation before the red flag shows up, know your volcano and your hydrothermal system, look for dyke intrusions and their volume, be constantly alert for the tiniest puff of steam the volcano might throw and then catastrophe may be avoided. Or well, at least the human lives part, you can’t move a city after all, you can’t move Man… Oops I won’t spoil the next article.
Next post we are looking at potential candidates to repeat the story and going into the worst nightmares of destruction.
On the eruption of 1911, with photographs of the time as well as summaries of Taal historic activity:
More on the events of 1749-1754 (Spanish):