All eruptions have history. Volcanoes may sometimes appear to erupt out of the blue; the mountain may not even have been recognized as a volcano, before blowing up. But the unexpectedness stems from our lack of knowledge. In the case of monogenetic volcanoes, which only erupt once, the volcanologist will recognize from the cones in the area that there is magma below the surface, waiting for the right time and the right place. Auckland is like that: if an eruption happens within the city limits, do not say there was no warning: the warning is all around you. And if a volcano is deemed inactive and harmless, perhaps this is because our memory does not go back far enough. Naples has both types: one side of the city is build on a volcanic field, and the other side is within striking distance of Vesuvius. Vesuvius has been quiet for 75 years, or approximately a human life time. No one in Naples really knows first-hand what it can do. The volcanic field, Campi Flegrei, hasn’t erupted for over 400 years. But the future feeds off the past, even if that past is many thousands of years ago. To volcanoes, the expression ‘Yesterday is history, tomorrow is a mystery’ is completely wrong. The mystery is the past. Know its history, and the volcano’s future becomes an open book (although admittedly the page numbers may be a tad uncertain). Volcanoes that have forgotten their past are still condemned to repeat it.
Krakatau was such a volcano with a mysterious past. Although the island was known to be volcanic, there had been no eruptions within living memory. In 1881, the mine geologist Rogier Verbeek warned about its peculiar location. It occupied the position where the line of volcanoes on Java, the line of volcanoes on Sumatra, and the line of activity in the Sunda Strait cross. This was considered little more than a curiosity – People did not realize that catastrophe was barely two years in the future, and that the run-in was well underway. Nowadays we look at Krakatau differently: the island has caused not one but two major disasters – the first when the whole island blew up, the second one just recently when its successor, Anak Krakatau, collapsed into the hole left by its infamous predecessor. But before 1883, there was little evidence of dangerous activity, and there were worse mountains in Indonesia to worry about. Krakatau (its Indonesian name) was a signpost on the route between the Indian Ocean and the Java Sea, not an explosion waiting to happen.
That signpost was because the main sailing route used by the Dutch, French and British traders went through the Sunda Strait, between Sumatra and Java. The combination of Malaysia and Indonesia forms an effective barrier to the Indian Ocean. There are two main routes through: the Strait of Malacca between Sumatra and Malaysia, and the Sunda Strait between Sumatra and Java. It was the Portuguese who had control of the former; all other sailors of the 1600’s followed the latter. Nowadays, the Strait of Malacca is preferred because it is deeper and easier to navigate, and it provides a shorter route. The Sunda Strait is harder to navigate due to the many islands, the shallows and the strong currents – but such was the power of the Portuguese. The Dutch traders established themselves just around the corner in 1619, at what is now Jakarta and which they called Batavia. In hindsight, this location may not have been the best choice. The region is swampy, and the European traders were ravaged by tropical diseases to which they had little resistance. The trading ships quickly learned to minimize time in Batavia’s harbour: they would drop off their load, leave, and anchor in the less disease-prone Sunda Strait while waiting for their precious return cargo to get ready. The two best places for this were Prince’s Island (nowadays called Panaitan) at the southern entrance, and Krakatau in the centre of the Strait.
In 1620, soon after the founding of Batavia, a shipyard was started on Krakatau. Apparently this was not a success as I could not find any subsequent mention of this facility! Over the next 2 centuries the island was occupied on at least three occasions, with the shipyard in the 17th century, an attempt at pepper growing in the 18th and apparently a penal colony in the 19th century. (This penal colony is mentioned in just about every web page on Krakatau, using identical wording, but I could not find the origin of the statement. The Dutch penal colonies were mainly used for labour and so would not have suited Krakatau where no labour was needed. It happened in 1809, shortly before the interlude of British rule.)
So Krakatau was well known, and was used, but the failure of subsequent settlements suggests that it was not particularly hospitable. We can guess the reasons for the failures. The island is much further from the coast than the other, settled islands, so it may have been a matter of difficulty of access, also in view of the currents and frequent thunderstorms. Access to fresh water may have been an issue at times. Volcanic activity is not mentioned by anyone as an issue, at least not before the events of 1883.
The Sunda Strait forms a triangular-shaped interruption of Indonesia. Indonesia’s geology is governed by the massive subduction zone to the south, where the Australian plate is diving headlong into the deep, at kangeroo speed. (kangeroos are fast: I remember going for a run and the local kangeroo population joined me, fleeing whatever it was I was running away from!). Sumatra and Java are on the edge of the Sunda Shelf, a piece of continental crust attached to Asia which has dropped below current sea level. (During the ice ages it was at sea level, above water but rather wet.) The subduction generated the usual melt and formed volcanic islands on this Sunda Shelf, some distance from the subduction trench. These volcanic islands slowly merged, helped by the area being pushed up by the descending plate below. Java and Sumatra acquired deltas on their northern side where the sea was quite shallow.
The Sunda Strait failed to take part in this merger activity, although it did have the requisite volcanic islands. What happened? It was a matter of misdirection. A quick glance on the map shows that Sumatra is rotated by about 20 degrees with respect to Java, with the Sunda Strait forming the triangle that links the two. The subduction motion is head-on in Java, but in Sumatra it is oblique, making Sumatra running (with a non-kangaroo definition of running) away from Java. The parallel component of the motion is taken up by the Sumatra fault, which runs along the length of the island. The Strait has been pulled open both by the rotation of Sumatra, and by its movement to the northwest. Which of these is more important is still being debated. The Sumatra fault peters out in the Sunda Strait but does so with some difficulty: it bends south and splits in two. In between the two segments is an area of deep subsidence, at the southern entrance of the Sunda Strait, which reaches 1800 meters in depth and explains the lack of islands west of Krakatau.
Because Java and Sumatra are moving apart, the entire Sunda Strait is a zone of extension. Such zones show subsidence, as in a graben. Therefore, one would expect a graben to run along the length of the Strait, perpendicular to the subsidence trench of the Sumatra Fault. Indeed, deep drilling southeast of Krakatau has found 3 km of sediments, dating to the pliocene (a few million years ago), implying the rocky bottom has sunk by at least that much. But it is not visible on the sea floor, unlike the basin between the two branches of the Sumatra Fault. The copious sediments have completely covered it up.
A line of volcanoes bisects the Sunda Strait, starting from Rajabassa on Sumatra and running via Sebesi, Sebuku and Krakatau to Panaitan Island, north of the termination of the Sumatra Fault. Of these volcanoes, only Krakatau remains active. There is an earthquakes zone along the same line in the Sunda Strait. They have a typical depth of 30-60 km, well above the subduction zone which is some 150 km deep here. The earthquakes reach up to M7 in strength with the strongest earthquakes located towards Panaitan. The last M7 earthquake was in 1943; there has not been a quake stronger than M6.1 in the Sunda Strait since 2000.
The volcanic line suggests that a fault is at play, and the earthquakes support that view. Earthquakes also have shown indications of a second fault, running along the length of the Sunda Strait. This makes sense as this is perpendicular to Sumatra and Java moving apart, and so it could take up the extension from this motion.The deep sediment southeast of Krakatau may show that this fault behaves as a pull-apart basin. It would also fit with Krakatau’s silicic lava, which indicates it may receive its magma from crustal decompression melt, rather than from the oceanic subduction. (It also means that Krakatau is much larger than it appears, as its base must be on the rocky bottom below the deep sediment.) Krakatau lies where the two faults intersect, on the crossing of a line of crustal thinning and a line of volcanic activity, a dangerous (or exciting) combination, the kind of crossing where a 4-way stop sign would be useful.
There are no accurate dates for the other volcanoes on the line, apart from a basaltic flow north of Rajabassa (Sukadana) which is dated to about 1 million years ago. There is still some fumarole activity at Rajabassa. But currently, Krakatau has no equal anywhere around the Sunda Strait.
The main explosions heralding Krakatau’s famous 1883 eruption started on Aug 26, after several months of activity. On Aug 27 the two largest explosions followed, 40 minutes apart; the second one caused the biggest devastation. It destroyed most of Krakatau but half of its highest mountain, Rakata, remained standing. The hole in the sea floor was 250 meters deep. Gravity measurements suggests that the explosion actually made a hole 1 kilometer deep but most of this was filled by low density ejecta from the eruption itself. Interestingly, this hole is to the west of the old island. Whilst the early activity was all on-land, the final explosion took place not on the island but under the water just off its coast. A possible model is that the first of the two explosions was the collapse of the roof of the magma chamber, occurring under the sea to the west of Krakatau. The sea water fell a kilometer down into the heat. The island collapsed into the hole to its side, and this sealed the hole. The water below boiled, but had nowhere to go until the pressure grew large enough to lift the seal. The ensuing explosion rained mud 20 kilometers away, caused the final tsunamis and made a noise heard over a sixth of the earth’s surface. There is a problem with this model which I will come back to.
There were three islands left after the eruption. Together they trace the rim of an old caldera, twice as large as the one formed in 1883. That caldera was already recognized by Rogier Verbeek in 1885 when writing his comprehensive report on the eruption. He speculated about the rim tracing the summit of an older volcano which was destroyed in an ancient explosion. There is at least one other large caldera in the Sunda Strait: gravity measurements have shown evidence for an even larger filled caldera south of Panaitin, with a diameter of 20 km. The date of the formation of the older Krakatau caldera is very uncertain. People who study deposits find it is 60,000 years old. People reading old history say it is 1600 years. Who said that the past was an open book?
Anak Krakatau grew up on the edge of the hole. Interestingly, it made use of the same conduit as the 1883 eruption. The conduit should have been destroyed, but it may have survived because the final explosion, and the excavated hole, was not underneath Krakatau. The edge location made the new cone potentially unstable, as pointed out as early as in 1995 by Deplus. As had been predicted in more than one paper (but still happened unexpectedly – it can be tempting to make the wolf at the door go away by ignoring it), Anak collapsed in 2018, and its remnant is now the smallest of the four islands that make up Krakatau; the summit has been replaced by a salt-water crater lake. To the south is Rakata, which is the high peak from before 1883. To the left is Sertung (Verlaten island – meaning deserted) and to the right is Panjang (Long Island). Rakata and Anak together coincide with old Krakatau – the other two islands were already separate from it. (Another small island, called ‘Polish Hat’, disappeared in the 1883 eruption.) Since its collapse, Anak Krakatau is no longer visible from the main land: it is hidden behind Panjang which is now taller. This will remain so until the rebuilding program begins. There has been no indication of any volcanic activity since the collapse, apart from continuing discolouration of the sea on the west and south which may be debris washing into the sea.
Now we know the location, the geology and the current state. With the present covered, it is time to go back to its mysterious past. What do we know about Krakatau’s history?
Read all about the history in part II
Albert, May 2019