The New Decade Volcano Program; #6, Bali


A week later than posted on Volcano Cafe, here is number 6 on the guys proposed list! Introducing the dangers of Bali…

Romantic Paradise Destination – The New Decade Volcano Program #6, Bali

Sunset from the 3,148 m high summit of Gunung Agung. The peak in the distance is G. Abang, a remnant of a far loftier peak, Ancestral Batur. (WikimediaCommons, photo by Mrllmrll).

It has often been pointed out that the deadliest volcano is the one you did not know about. This is our dilemma. When you try to identify the potentially most dangerous ones, by necessity you have to go out on a limb to find those that are not well known nor well studied and there is always the chance to end up with egg on your face. But in this we are not alone. As an example, it was long thought that a particularly heavy layer of volcanic dust in ice core samples dated to c. 3650 BP belonged to Thera. Only recently has most of this been identified as belonging to the far larger, contemporaneous, 100km3 DRE Aniakchak eruption in the Aleutians, Alaska.

When it comes to large volcanic eruptions, one of the more striking features is the Sunda Arc that runs from Sumatra via Java and the Sunda Strait through the Lesser Sunda Islands. Sumatra is home to the Toba caldera, source and result of the largest volcanic eruption in the past 100 kA. Recently, a vast body of magma underlying Java was discovered, one that feeds that islands prodigious volcanic activity. But of the southern part of this arc; the Sunda Strait and the Lesser Sunda Islands, little is known. Yet this part of the Sunda Arc is home to two of the largest volcanic eruptions of the past 1,000 years; Rinjani (~1257 AD, <80 km3 DRE) and Tambora (1815, 33 – 41 km3 DRE). Sufficient to say, was there a repeat of either of those eruptions today, the islands hosting these giants are home to some 4½ million people each and neither such VEI 7 blast would be survivable. As both had “mega colossal” eruptions recently geologically speaking, neither is a good candidate for another one in the foreseeable future. But on the premise that a similar magmatic feed into a similar geological setting will most likely result in similar volcanic activity, let’s take a closer look! Lightning did after all strike twice here within the past millennium!

The Sunda Arc from Eastern Java through the Sunda Strait and the Lesser Sund Islands where the Australian plate subducts under the Sunda Plate at a rate given as 6-7 cm per year, relatively high. Note that for Sumatra and Western Java, the subducting plate is the Indian plate. The names of the Islands is green, active volcanic complexes, calderas and the larger stratovolcanoes are denoted in red. Sangenes (yellow) is thought to be extinct.

From a birds-eye view, this area is characterised by the formation of very large stratovolcanic cones with a prominence in excess of 3 km (eg Raung, ancestral Catur, Ancestral Batur. Agung, Rinjani, Tambora and the partly submarine Sangeang Api), volcanic complexes (eg. Biau, Buyan-Bratan and Batur) and 10-15 km calderas (eg. Biau, Bedegul, Batur). It all comes together on Bali, tropical island paradise and the place to go for a romantic holiday. Apart from the 1963 VEI 5 (5.3) eruption of Gunung Agung, little is known about the volcanism of Bali.


With a population of 4,225,000 as of January 2014, Bali is home to most of Indonesia’s Hindu minority which according to the 2010 Census constituted 84.5% of the island’s population. Just over a quarter of a century ago, the economy was mainly based on agriculture. Before the 2003 terrorist bombings, over 80% of the economy was tourism-related and Bali had become the richest of all Indonesian territories. Annual tourism is in excess of eight million with five being Indonesian and the remaining three international. To crown it all, Bali was host to the 2013 Miss World pageant.

The main tourist locations are concentrated to the South; Sanur on the east coast which once was the only tourist location, Kuta with its beach and close to the Ngurah Rai International Airport, Ubud in the centre of the island and the newer development Nusa Dua and Pecatu. Kuta, the main tourist location, lies 55 km from the centre of the Buyan-Bratan volcanic complex, 61 km from the centre of the Batur Caldera and 60 km from the peak of the 3,031 m high Gunung Agung, The town of Ubud is basically at half that distance while Dempasar, the capital with over 800,000 inhabitants, is within 50 km of all three.

The crust beneath Bali Island is about 18 km thick and has seismic velocities similar to those of oceanic crust (Curray et al, 1977). The depth of the Benioff Zone beneath the Batur Volcano is 165 km, which has been computed by multiple linear regression analyses (Hutchison, 1976). The depth of the seismic zone beneath the arc reaches to approximately 650 km depth between Java and Flores. The oldest widely exposed rocks are lower Tertiary shallow marine sediments, which are intruded and overlain by plutonic and related volcanic rocks in a zone only slightly south of the present-day volcanic arc (Bemmelen, 1949). The rocks of the Sumatra to Bali sector range from tholeiitic through calc-alkaline to high-K calc-alkaline series.

Geologic map of Bali Island. Note the extent of the “Buyan-Bratan and Batur Tuffs and Lahar deposits” and compare with the previous map of Bali showing the main settlements. (After Purbo-Hadiwidjojo, 1971)

Volcanism in Bali is concentrated to three areas, the Buyan-Bratan volcanic complex which formed roughly 100,000 years ago but holds several young stratovolcanic cones to the SSW, the Batur Caldera which formed <100,000 to 25,000 years ago and has the highly active stratovolcanic cone of Batur. Both areas contain large lakes within the caldera perimeters. Finally, there is Gunung Agung which had a powerful VEI 5 eruption as recently as 1963. However, the eruptive record of Agung extends no further back than to the 1808 VEI 2 eruption and that of Batur to a VEI 2 eruption in 1804. Being located just south of the Equator, the tropical climate and vegetation quickly covers whatever volcanics that have been deposited. This may create a false sense of security.

Buyan-Bratan Volcanic Complex

The Buyan-Bratan volcanic complex is also known as the Bedegul caldera, Bratan caldera, Catur or Tjatur caldera. The southern caldera wall has disappeared beneath a superimposed field of young, heavily vegetated stratovolcanoes including Gunung Batukaru (2,276 m), Adeng (1,826 m), Pohen (2,063 m), Sengayang (2,087 m), Lesung (1,865 m), Tapak (1909 m). Although the ancestral volcano is known as Mt Catur, the location of today’s Catur (2,096m) on the NE calera rim argues that it may not be a volcano even if it is sometimes referred to as being one.

The age of the 6 x 11 km Bedegul caldera which formed when ancestral Mount Catur collapsed is unknown although it must be substantially older than ~30,000 years and possibly even hundreds of thousands of years. The field of young stratovolcanoes to the SW, the Byan-Bratan Volcanic Complex, is heavily vegetated, thus the latest period of activity remains unknown but has been tentatively placed hundreds or thousands of years ago (Wheller, 1986). Two of those stratovolcanoes, Tapak and Lesung must have formed after the last large eruption of the nearby Batur Caldera as they not covered by deposits of its youngest dacitic pumice eruptions. As this has been dated to 20,150 years ago, these stratovolcanoes with prominences of 625 and 669 m respectively as measured from the surface of Lake Beretan must therefore be less than this age. Inside the caldera, geothermal activity is exploited at the Buyan-Bratan geothermal power plant and there are at least a dozen hot springs in the area.

The municipality of Beretan is a major Hindu enclave and contains a Shiivaite temple, the Bratan Bali. (Indonesia Tourism)

The outline of the remaining caldera walls suggest that there may have been two events; the first forming the 9 to 10 km diameter Western part with the stratovolcanic cone of Tapak forming subsequently near the centre, the second forming the smaller 5.5 to 6 km diameter Eastern part. Very tentatively and assuming that the calderas were formed by the subsequent collapse of those edifices following a major eruption, also assuming that the ancestral volcanoes were similarly steep to the nearby Mount Agung, we can make an educated guess at the size of those eruptions. Ancestral Catur (Catur A) would have been about 3,300 m high (a.s.l.) and the caldera bottom, allowing for subsequent infill, would have been about 600 to 800 m deep as measured from the remaining walls. This yields a figure on the order of 52 + 16 = 78 km3 or borderline VEI 7 for the larger caldera, Catur A. Catur B would have been about 2,400 m a.s.l. and the caldera ~500-600m deep as measured from the remaining walls prior to infill. This results in figures of 11.3 + 4.7 = 16 km3 or a small to medium-sized VEI 6 eruption. Please note that this is speculation on my part! No doubt better-informed readers will hasten to correct my assumptions from a position of superior knowledge!

The southernmost stratovolcano of the Buyan-Bratan volcanic complex is the 2,276 m high Batukaru, which means "coconut shell" in Balinese. It has a prominence of ~1,500 m as its edifice can be traced to just below the 800 m topographic isoline. (Bali Foto Galerija)

Apart from the already mentioned Gunung Tapak (1909 m), the volcanic field subsequent to the caldera forming event(-s) includes at least another five major stratovolcanoes – Batukaru (2,276 m), Adeng (1,826 m), Pohen (2,063 m), Sengayang (2,087 m), Lesung (1,865 m). There is no information on any eruptive activity but as previously stated, due to the tropical climate and vegetations, all we can definitely state is that there has been no activity in the past two to three hundred years as there is no historical record of any. With at least two of them being younger than ~20,000 years, the likelihood is that all have been active recently, geologically speaking. What their presence does suggest however, is that the original magmatic system of ancestral Catur (Catur A & B) has been well and truly destroyed and that if in the future, there is renewed volcanic activity in the Buyan-Bratan volcanic complex, this will be from one or more of these young stratovolcanoes and most likely not greater than VEI 3, possibly a very small VEI 4 eruption in the sense that the eruption of Eyjafjallajökull in 2010 counts as one. As an example, at Tapak there are at least five layers of scoria separated by four layers of paleosoil, indicative of at least five periods of extended eruptive activity separated by four periods of repose. (Watanabe et al:2010). Watanabe and his co-authors repeatedly lament the fact that while Batur Caldera nowadays is relatively well studied, almost no research whatsoever (apart from their own exploratory field study, author’s note) seems to have been undertaken of the less easily accessible Buyan-Bratan Caldera and volcanic complex.

Batur Caldera

Ancestral Batur was an approximately 4,000-meter high stratovolcano, nearly a kilometre higher than present-day Agung (3,148 m), which had an enormous eruption in prehistoric times to form the outer, 10×13.8 km caldera around 29,300 BP which today contains a caldera lake, Danau Batur. The inner 7½ km caldera was formed at about 20,150 BP.

Gunung Batur (1,717 m.a.s.l., prominence 700 m) is a small stratovolcano in north-central Bali and its most active. It has several craters and remains active to this day. The first historically documented eruption of Batur was in 1804 and it has erupted over 20 times in the last two centuries (VEI 1 – 2). Larger eruptions occurred in 1917, 1926 and 1963. Clinopyroxene from the 1963 eruption of Batur record crystallisation depths between 12 and 18 km, whereas clinopyroxene from the 1974 eruption show a main crystallisation level between 15 and 19 km. Furthermore, plagioclase melt thermobarometry indicates the existence of shallow level magma reservoirs with depths between 2 and 4 km for the 1963 eruption and between 3 and 5 km for the 1974 event (Geiger:2014). This suggests the existence of a very large and rather deeply lying primary or lower magma chamber as well as a moderately substantial upper magma chamber.

The term “Batur” often refers to the entire caldera, including Gunung Abang, Bali’s third-highest peak, which is situated along the rim. Batur is a popular trekking mountain among tourists, as its peak is free from forest cover, offers spectacular views and is easily accessible.

The substantial lava field from the 1968 eruption (Batur III, VEI 2) that began on Jan 23rd and ended on Feb 15th 1968. (Martin Moxter)

Batur has produced vents over much of the inner caldera, but a NE-SW fissure system has localized the Batur I, II, and III craters along the summit ridge. Historical eruptions have been characterized by mild-to-moderate explosive activity (Strombolian?) sometimes accompanied by effusive emissions of basaltic lava flows from both summit and flank vents which have reached the caldera floor and the shores of Lake Batur in historical time.

The Batur caldera formed in two stages. Through radiocarbon dating, we have a relatively good idea of when. The first and larger of these is associated with the 84 km3 dacitic ignimbrite known as the “Ubud Ignimbrite” which in locations is over 120 m thick. About 29,300 years BP, Ancestral Batur had a “mega-colossal” VEI 7 eruption which caused a steep-walled depression about 1 km deep and over ten km in diameter. The second ignimbrite, the 19 km3 dacitic “Gunungkawi“ Ignimbrite”, erupted about 20,150 years BP from a large crater in the area of the present-day lake. The second eruption triggered a second collapse, which created the central 7½ km diameter circular caldera, and formed a basin structure. Both the Ubud and Gunungkawi Ignimbrites are of a similar dacitic composition although the latter is more mafic, white to red in main with less than 10% dark grey to black dacitic pumice clasts. In the case of the second of these ignimbrite, two different cooling layers were identified. The lower, thus first ejected, is finely grained and welded, hence it was far hotter. In places, it is between 5 and 20 m thick. The upper, coarser, partially welded and hence “cooler” unit has suffered much erosion but is in places up to between 50 and 70 metres thick. The calculated volume of erupted material for the Ubud (84 km3) and Gunungkawi (19 km3) Ignimbrites coincide with and are proportional to the size of related collapses of Caldera I (80 km3) and Caldera II (18 km3).

After these eruptions, there were two further ignimbrite-producing eruptions, both mainly intra-caldera. The Batur Ignimbrite is a densely welded dacitic ignimbrite, typically 50 – 200 m thick, which at one point overflows the caldera rim to form 30 to 70 m thick layers of non-welded ignimbrite. The Blingkang Ignimbrite is a non-welded to moderately welded intra-caldera ignimbrite deposit between 5 to 15 metres thick. Sparse charcoal clasts scattered in this sheet give an age of 5,500 ± 200 years B.P. The thick phreatomagmatic and surge deposits which are found below the ignimbrite indicate that this was preceded by phreatomagmatic eruptions. In addition to these four sequences, basaltic to basaltic andesite lavas and pyroclastic deposits are inter-layered with and underlie the ignimbrite sequences, particularly in the southern slope of the caldera.

In spite of the frequently erupting modern Gunung Batur with its moderately sized eruptions, this caldera cannot yet be said to have shot its bolt due to the implied existence of a very large magma reservoir, one that was apparently not destroyed by the caldera-forming eruptions. Both the Batur and Buyan-Bratan calderas illustrate a recurring theme where first a very large stratovolcanic edifice is built after which there is a substantial VEI 7 ignimbrite-forming eruption followed by the formation of a dacitic to andecitic dome complex after which a large, ignimbrite-forming VEI 6 eruption follows. Even if one of these volcanic complexes almost certainly is no longer capable of such large eruptions and the other probably not in the foreseeable future, there remains one gigantic stratovolcano on Bali, one that has dimensions of 8 x 11 km as measured at the 1200-m isoline, 2,000 m above which its somewhat truncated summit towers.

Gunung Agung

Gunung Agung photographed from about 60 km to the south during an overflight of the main tourist areas of Bali. Except for the extreme bottom of the picture, the entire plain visible is covered in ignimbrite deposits from the Batur I eruption of about 29,000 BP (WikiMedia Commons)

Located in the eastern part of Bali, Mt Agung is a young basaltic to andesitic composite volcano. Bordered to the east by the inactive or extinct volcanic cone Seraja, to the south by an ancient volcanic complex and to the NW by a valley that separates it from the Batur volcanic complex, Agung goes all the way down to the Indian Ocean to the NE and through a long unimpeded decline over the Buyan-Bratan and Batur ignimbrites and lahar deposits to the SW and WSW, all the way to the capital Denpasar and beyond. South of Agung, there are older Tertiary volcanic deposits as well as remnants of coral reefs. The present-day volcano is surrounded by older Quarternary andesitic and basaltic-andesitic lavas and pyroclastic deposits, something that has prompted the conclusion that Agung overlies an older caldera formation (S. Self et al:1979).

The eruptive record of Agung goes back only to 1808 when the volcano had a VEI 2 eruption. Since that date, Agung erupted again in 1821 (uncertain) and 1843, both VEI 2 eruptions after which it remained dormant for 120 years until the great eruption of 1963. Prior to 1808 is a big unknown, although the relative symmetry of the mountain, the state of its upper slopes as well as a comparison with similar volcanoes suggests that Agung would have erupted relatively frequently.

On February 18th 1963, locals reported hearing a loud explosion after which a dark eruption cloud rose over Agung. The first explosions were probably phreatic or phreatomagmatic. On February 24th, highly viscous lava oozed over the northern slope, 0.5-0.8 km wide and 30-40 m in height. It was moving so slowly that it took 18 to 20 days to reach 500 m a.s.l after travelling some 7 km down from the peak. This works out at a speed of about 4 mm per second or 14 m per hour. The volume of lava erupted was estimated to be on the order of 0.05 km3. After this, the eruption continued with a combination of effusive and explosive events.

On March 17th came the main eruption. The eruption cloud reached 8-10 km above the volcano but the lower portions fell down the slopes as nuees ardentes that travelled with a speed of about 60 km/hour up to 12-15 km from the crater down the valleys to the south and east. From this description, it seems the eruption was peléean. The pyroclastic flows destroyed many villages around the volcano and caused the deaths of many people living near the river valleys. Estimates are that 820 people were killed by the pyroclastic flows, 163 people were killed by ashfall and volcanic bombs and a further 165 people were killed by lahars.

A comparison between the lavas erupted by Batur and Agung reveals that while Batur tends to erupt more trachytic magmas, the magmatic system of Agung produces more evolved magmas. The comparison between the historic and modern lavas indicates that, at present and probably, the Batur caldera does not produce the types of more evolved magma required to cause ignimbrite-forming eruptions. Unfortunately, it seems the lavas of Agung have not been similarly analysed and the data is based on a single eruption, that of 1963. (H. Geiger:2014)

For the 1963 Agung eruption, results from clinopyroxene melt thermobarometry suggest dominant crystallisation levels between 18 and 22 km depth. Plagioclase melt thermobarometry indicates the existence of shallow level magma reservoirs, with depths between 3 and 7 km for the 1963 eruption, located around the boundary between the (upper) sedimentary and the oceanic type mid- to lower crust. The deep magma storage regions notably coincide with lithological boundaries in the crust and mantle beneath Bali, at the boundary between MOHO and crust, while the shallow reservoirs are consistent with recent geophysical studies that point to regional shallow level magma storage. An along-arc comparison reveals this trend to be characteristic of Sunda arc magma storage systems. According to Harri Geiger, the author, the result “highlights the utility of a thermobarometric approach to detect multi-level systems beneath recently active volcanic systems.” (Geiger: 2014)


As was remarked at the beginning; a similar magmatic feed into a similar geological setting will most likely result in similar volcanic activity. This premise is further substantiated by the conclusion presented by Geiger, that the deep magma storage regions notably coincide with lithological boundaries in the crust and mantle and that this is a characteristic of the Sunda Arc. The conclusions that can be inferred from these observations are:

  • Very large caldera-forming, ignimbrite depositing eruptions VEI 6 to 7 are a characteristic of Lower Sunda Arc volcanism
  • The location of the deep magma reservoirs is such that these are not likely to be destroyed by the caldera-forming eruptions unlike those at other locations (e.g. Roccamonfina, Mt Mazama, Aniakchak)
  • Bali contains no less than three such volcanic systems of which the currently inactive Buyan-Bratan Volcanic complex is in a phase of stratovolcanic dome construction, the Batur Caldera is in the process of rebuilding a main stratovolcanic edifice while the Agung system is meandering towards the end of that phase
  • All three volcanic systems pose potential hazards to the Balinese population and require further studies as well as systematic monitoring
  • Of the three, the greatest danger is posed by the Agung system and at present, there is insufficient data to rule out a very large, caldera-forming and or ignimbrite depositing eruption

For these reasons, Bali is our proposed number six on the New Decade Volcano program.


Acknowledgement: I am indebted to Shérine France for finding and bringing Watanabe et al 2010 and Geiger 2014 to my attention.

Igan S. Sutawidjaja, “Ignimbrite Analyses of Batur Caldera, Bali, based on C14 Dating”, Jurnal Geologi Indonesia, Vol. 4 No. 3 September 2009: 189-202.

K. Watanabe, T. Yamanaka, A. Harijoko, C. Saitra and I W. Warmada; ”Caldera Activities in North Bali, Indonesia”, Journal of S.E. Asian Applied Geology, 2010

O. Reubi & I. A. Nicholls, “Structure and Dynamics of a Silicic Magmatic System Associated with Caldera-Forming Eruptions at Batur Volcanic Field, Bali, Indonesia”, Victorian Institute of Earth and Planetary Sciences, 2005.

H. Geiger, “Characterising the Magma Supply System of Agung and Batur Volcanoes on Bali, Indonesia”, Uppsala 2014


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s