Proto Krakatau

Historical and archaeological evidence now suggests that the European Dark Ages might have been triggered by a massive volcanic eruption. The global consequences to humanity of a similar super eruption are much more complex and far reaching than has ever been considered according to Ken Wohletz. Such an event can fundamentally and permanently change human history. There are numerous potentially active calderas that can erupt in the near future with a similar magnitude. – B.H.

Modern history has its origins in the tumultuous 6th and 7th centuries. During this period agricultural failures and the emergence of the plague contributed to: (1) the demise of ancient super cities, old Persia, Indonesian civilizations, the Nasca culture of South America, and southern Arabian civilizations; (2) the schism of the Roman Empire with the conception of many nation states and the re-birth of a united China; and (3) the origin and spread of Islam while Arian Christianity disappeared. In his book, Catastrophe An Investigation into the Origins of the Modern World, author David Keys explores history and archaeology to link all of these human upheavals to climate destabilization brought on by a natural catastrophe, with strong evidence from tree-ring and ice-core data that it occurred in 535 AD. With no supporting evidence for an impact-related event, I worked with Keys to narrow down the possibilities for a volcanic eruption that could affect both hemispheres and bring about several decades of disrupted climate patterns, most notably colder and drier weather in Europe and Asia, where descriptions of months with diminished sun light, persistent cold, and anomalous summer snow falls are recorded in 6th-century written accounts. Writings from China and Indonesia describe rare atmospheric phenomena that possibly point to a volcano in the Indonesian arc. Although radiocarbon dating of eruptions in that part of the world are spotty, there is strong bathymetric and volcanic evidence that Krakatau might have experienced a huge caldera eruption. Accordingly, I encouraged a scientific expedition to be led by Haraldur Sigurdsson to the area. The expedition found a thick pyroclastic deposit, bracketed by appropriate radiometric dates, that suggests such a caldera collapse of a “Proto-Krakatau” did occur perhaps in the 6th century. Bathymetry indicates a caldera some 40 to 60 km in diameter that, with collapse below sea level, could have formed the Sunda Straits, separating Java from Sumatra, as suggested by ancient Javanese historical writings. Such a caldera collapse likely involved eruption of several hundred cubic kilometers of pyroclastic debris, several times larger than the 1815 eruption of Tambora. This hypothetical eruption likely involved magma-seawater interaction, as past eruptions of Krakatau document, but on a tremendous scale. Computer simulations of the eruption indicate that the interaction could have produced a plume from 25 to >50 km high, carrying from 50 to 100 km3 of vaporized seawater into the atmosphere. Although most of the vapor condenses and falls out from low altitudes, still large quantities are lofted into the stratosphere, forming ice clouds with super fine (<10 micrometer) hydrovolcanic ash. Discussions with global climate modelers at Los Alamos National.

Laboratory led me to preliminary calculations that such a plume of ash and ice crystals could form a significant cloud layer over much of the northern and southern hemispheres. Orders of magnitude larger than previously studied volcanic plumes, its dissipation and impact upon global albedo, the tropopause height, and stratospheric ozone are unknown but certainly within possibilities for climate destabilization lasting years or perhaps several decades. If this volcanic hypothesis is correct, the global, domino-like effects upon epidemics, agriculture, politics, economics, and religion are far-reaching, elevating the potential role of volcanism as a major climate control, and demonstrating the intimate link between human affairs and nature.


Tree-ring data from Keith Briffa (CRU, Univ. of East Anglia), corroborated by European data compiled by Mike Baillie (Queen’s Univ., Belfast), shows clear evidence of a 535 AD climate perturbation, and it is now known worldwide. The origins of this event are likely to have been volcanic since ice core from Greenland and Antarctica show sulfuric acid spikes during this time interval, and for the Byrd core (Antarctica), it is the largest in the last 2000 years. David Keys and I worked with Claus Hammer (Niels Bohr Institute) to reinvestigate the GRIP core from Greenland that Clausen et al. (1997) had already identified.

Although asteroid/comet impact remain as potential causes, I focus on a volcanic source located near the equator. Of over 100 potential equatorial volcanoes considered, I found best corroborating evidence in Indonesia, where 6th Century geo-political discontinuity is well documented. Tephra dates are very useful, but there can be pitfalls. For example, some published dates for Rabaul that looked like a fit turned out to be erroneous. It was a translation of the Javenese “Pustaka Raja Purwa” (The Book of Ancient Kings) that alerted David Keys and I to a massive eruption of Krakatau during the 338th year of the Shaka Calendar, which is known to likely be misaligned to the western calendar date of AD 416. We spent considerable time identifying and translating text that could be demonstrated as not having been “contaminated” by post 1883 writings, and in that text is the ancient Javanese tradition of the separation of Java from Sumatra in a cataclysm that fits a description of a large explosive eruption. In other words, a caldera collapse that produced the modern Sunda Straits.

A pre-1883 British Admiralty chart of Sunda Straits shows shallow (~10 m) depths of sea floor in straits, and islands of Krakatau, Bezee, Sebooko, Thwart way, and mountains near Katimbang (Mt. Rajah Bassa); these islands may be vestiges of volcanic vents surrounding the flanks of Proto-Krakatau, the predecessor of the present Krakatau (Krakatoa) volcano. Connecting these vents outlines a caldera with a surface diameter of ~50 km centered in the Sunda Straits about 20 km NE of present day Krakatau.

Drawing a west-to-east cross section of hypothetical caldera collapse, shows how Sunda Straits might have formed during the 6th century eruption of Proto Krakatau.

Reconstructing the Eruption

If one assumes that the average amount of collapse over the entire caldera area averaged ~100 m, then the volume of collapse was approximately 200 km3. This volume is a rough estimate of the amount of magma evacuated from the magma chamber at a depth of several km below the Sunda Straits, a chamber that might have held over a 2000 km3 of magma at the time of eruption (a cylindrical body 50 km in diameter and several km thick).

The 6th century eruption likely started with widespread tumescence of the ground over an area of a thousand square kilometers around Proto Krakatau, occurring over a period of years and reaching a magnitude of several meters or more near Proto Krakatau. This rising of the ground was likely to have proceeded so slowly that people in the area might not have noticed it, but they would have noticed the ever increasing occurrence of small earthquakes that perhaps were felt every few weeks in the year before the giant eruptions and reached nearly continuous shaking in the weeks before the eruption.

The following illustrations were generated by the volcanic eruption simulator, Erupt3, and they depict cross-sectional views of the Sumatra-Java island arc structure, built up by volcanism over thousands of years, consisting of layers of lava, pumice, and ash (red, magenta, and green colors). Proto-Krakatau is shown as a cone-like structure forming a land area connecting Sumatra to the west and Java to the east.

Analyzing the Eruption

While the preceding eruption simulations give a good qualitative picture of how the eruption progressed, we desire more detailed information, specifically regarding the physical parameters of the eruption. Some of these parameters can be constrained by considering the scale of the eruption portrayed above, while others must be calculated. Using the results of supercomputer simulations that solve mathematical equations that express the physical behavior of the eruption, we obtain results that are useful for atmospheric and sound wave models.

Eruption Duration: From the size of the assumed caldera, we showed above that ~200 km3 of magma was erupted. For eruptions of this magnitude, scaling of smaller historical eruptions indicates mass discharge rates of 1 billion kg per second or more. This flux is equivalent to one-one thousandths of a km3 per second or 3.6 km3 per hour. All other parameters being equal, the eruption would have taken at least 34 hours, but owing to waxing and waning fluxes during the eruption, the cataclysmal parts of the eruption might have lasted over 10 days. In fact, such eruptions might occur in day-long pulses, occurring over a period of years.


Eruption Products: Most of magma was fragmented by the tremendous forces of the eruption into pieces of pumice and ash. Especially during the Phreatoplinian eruption, ash was the preferred form of expelled magma. This ash is composed of tiny fragments of rock and glass shards, ranging from about 1 micrometer to a couple of millimeters in diameter. During the Phreatoplinian eruption, as much as 50% of this ash could have been composed of fragments less than 50 micrometers in diameter. Such tiny fragments have exceedingly long resident times in the atmosphere, able to be kept aloft for months by normal atmospheric turbulence. Assuming that 75% of the total 125 km3 of magma were involved with the Phreatoplinian eruption, perhaps as much as 30 km3 of fine ash particles were put into global circulation. Sulfur from the magma likely condensed on ash particles as sulfuric acid droplets, but its abundance is not known.


Water Vapor: Along with the pumice and ash, a lot of water was vaporized and injected into the atmosphere and stratosphere, especially during the Phreatoplinian phase of the eruption, perhaps volumetrically the most important of the cataclysmal phases. For optimum water vaporization, molten magma contains enough heat to vaporize an approximately equal volume of water. Assuming that 75 of the total 125 km3 of magma interacted with water, 75 km3 were vaporized, which upon expansion to atmospheric pressure may have occupied nearly 100,000 km3 of the atmosphere. To be sure, much of this vapor would have condensed an fallen out as ash-clogged rain with in hours of the eruption, but perhaps as much as half of it was carried around the world by stratospheric winds. This vapor would have condensed to form ice crystals, and these ice crystals would disperse in the rarified air to form stratus clouds, darkened by entrained ash. Along with the water vapor from the sea chlorine is also carried into the stratosphere, a component having an important effect on stratospheric chemistry.

Jet and Plume Structure: The jet of ash, pumice, and gas emerged from the vent at supersonic speeds of 650 m/s but decelerated to slow buoyant rise speeds of several tens of meters per second after reaching 10 to 15 km into the atmosphere. The buoyant plume had sufficient energy to continue rising to nearly 50 km before it became neutrally buoyant and began spreading laterally. The width of the jet and plume during the Plinian eruption phase may have ranged from several km at its base to many as many as 40 km near its apex, just below the level where it spread out as a giant anvil-shaped cloud. Ash concentrations within the plume decreased upwards because of admixture of ambient air into the plume as it rose, but these concentrations probably stayed at levels of 1 part in a million by volume. This is a very conservative estimate of total plume volume over the course of the eruption, giving a value of well over 100 million km3 and likely reaching several orders of magnitude.



Collapse of 100 m over 50-km diameter caldera can involve eruption of ~200 km3 of magma

Scaling of similar-sized eruptions predicts mass fluxes of ~109 kg s-1



Eruption of 200 km3 of magma at 109 kg s-1 requires at least ~60 hours of continuous eruption

Continuous eruption not likely, so duration may have lasted from a week to over one month


Dominantly pumice and ash

For phreatoplinian eruption up to 50 wt-% of the tephra can be fragmented to fine ash (< 63 mm)

If 75 vol-% of the erupted products were phreatoplinian, 75 km3 of fine ash were injected into the atmosphere



Dominantly H2O

For strong water/magma interaction, a volume of water vaporized is nearly equal to the volume of magma in the interaction

If 75 vol-% of the erupted products experience strong water/magma interaction then 150 km3 of seawater were vaporized, creating up to 200,000 km3 of water vapor in the atmosphere

Likely a large portion (50 vol-% or more) condensed an precipitated from low altitudes

Remaining vapor likely formed ice crystals in the stratosphere

Plume Structure

Erupted at speeds up to 650 m s-1, the eruption column may have exceeded 50 km in height before reaching neutral buoyancy

From numerical models, tephra volume concentrations after mixing with the atmosphere are 10-6 or less

If the fine-ash component remained in the plume with water vapor that formed ice crystals, then the total plume volume may have reached several tens of million of km3

Atmospheric Wonderings

Consider the large volume of the ash and vapor plume (10 80 million km3).

The earths tropopause surface area is ~5.2×108 km2.

Eruption plume could have produced a cloud layer from 20 to 150 m thick over the entire globe.

With the plume source near the equator, both north and south hemispheres would be affected.

Such a large magnitude volcanic plume has never been considered for CGMs, and its only analog might be the ejecta plume of the K-T impact.

Application of nuclear winter models (soot) for the K-T impact indicate such a burden of particles in the stratosphere would cause collapse of the troposphere.

In addition the large volume of water vapor may produce huge stratospheric ice clouds, leading to destruction of the ozone. 

by Ken Wohletz, Los Alamos National Laboratory