Bulletin of Volcanology

, Volume 74, Issue 7, pp 1581–1609 | Cite as

Recent explosive eruptions and volcano hazards at Soputan volcano—a basalt stratovolcano in north Sulawesi, Indonesia

  • Kushendratno
  • John S. Pallister
  • Kristianto
  • Farid Ruskanda Bina
  • Wendy McCausland
  • Simon Carn
  • Nia Haerani
  • Julia Griswold
  • Ron Keeler
Research Article

Abstract

Soputan is a high-alumina basalt stratovolcano located in the active North Sulawesi-Sangihe Islands magmatic arc. Although immediately adjacent to the still geothermally active Quaternary Tondono Caldera, Soputan’s magmas are geochemically distinct from those of the caldera and from other magmas in the arc. Unusual for a basalt volcano, Soputan produces summit lava domes and explosive eruptions with high-altitude ash plumes and pyroclastic flows—eight explosive eruptions during the period 2003–2011. Our field observations, remote sensing, gas emission, seismic, and petrologic analyses indicate that Soputan is an open-vent-type volcano that taps basalt magma derived from the arc-mantle wedge, accumulated and fractionated in a deep-crustal reservoir and transported slowly or staged at shallow levels prior to eruption. A combination of high phenocryst content, extensive microlite crystallization and separation of a gas phase at shallow levels results in a highly viscous basalt magma and explosive eruptive style. The open-vent structure and frequent eruptions indicate that Soputan will likely erupt again in the next decade, perhaps repeatedly. Explosive eruptions in the Volcano Explosivity Index (VEI) 2–3 range and lava dome growth are most probable, with a small chance of larger VEI 4 eruptions. A rapid ramp up in seismicity preceding the recent eruptions suggests that future eruptions may have no more than a few days of seismic warning. Risk to population in the region is currently greatest for villages located on the southern and western flanks of the volcano where flow deposits are directed by topography. In addition, Soputan’s explosive eruptions produce high-altitude ash clouds that pose a risk to air traffic in the region.

Keywords

Basalt Explosive volcanism Petrology Remote sensing Eruption forecasting Indonesia 

Introduction

Soputan (1°6.90′ N and 124°44.2′ E) is one of ten active volcanoes that overlie a west-dipping subduction zone, which underlies North Sulawesi and the Sangihe Islands. Soputan has a summit altitude of 1,800 m and a volume of approximately 60 km3. It lies just outside the southern margin of the Quaternary Tondono caldera (Fig. 1). Little detail is available in the international literature about the volcanoes of North Sulawesi and their eruptions. Here we present the results of an investigation of recent eruptions of Soputan, a 50–51 % SiO2 basalt volcano that is unusual by virtue of its stratovolcano morphology, extrusion of summit lava domes, and explosive vulcanian eruptions.
Fig. 1

Index maps showing study area (red square) on tectonic map of Indonesia (a) and location of Soputan volcano on a shaded-relief digital elevation model of North Sulawesi (b). Subduction zones are indicated by lines with teeth on over-riding plates in (a). Dual subduction zones with opposite polarity are present in the subsurface at the east and west margins of Molukka Sea (between North Sulawesi and Halmahera), although shallow thrusts with reverse polarity are mapped at the surface (Hamilton 1979). Soputan volcano and other stratovolcanoes in the area are indicated by red triangles in (b). The elongate oval depression in the center of North Sulawesi is the Quaternary Tondono Caldera, which contains small inactive post-caldera volcanoes and an active geothermal system. It is ringed by weakly welded and intensely dissected ignimbrite sheets with ages of 2.0, 1.3 and 0.1 Ma (Lécuyer 1990)

Soputan volcano lies in a populous and administratively complex region. It overlaps a series of political-administrative subdistricts of the Minahasa–North Sulawesi Province. The nearest city is Amurang (17 km to the west) although dozens of small villages occupy valleys surrounding the volcano. The summit of the volcano and the upper slopes above about 900 m elevation are within a National Park with no permanent population. Consequently, the principal risk from small eruptions is to visitors to the park, to rock miners who frequent the slopes on the west and southwest and to workers in coconut plantations on the south and west slopes. Larger and more explosive eruptions pose hazards from pyroclastic flows, lahars, and tephra fall to villages, also mainly on the south and southwest, as described in the final section of this paper. Data from the Smithsonian Institution’s Global Volcanism Project (Lee Siebert, 2009, written communication) and from LandScan 20071 (Dobson et al. 2000) indicate the population within 10 km of Soputan is approximately 70,000 and within 30 km, approximately 500,000.

Explosive eruptions have become increasingly frequent at Soputan in recent years, with eight Volcano Explosivity Index (VEI) 2 and 3 eruptions since 2003 (Table 1). The more recent of these eruptions resulted in destruction of an active summit lava dome that had been growing since 1991. Collapse of this dome began in 2005–2006 and continued in 2007–2008. The 2006–2007 eruptions produced a new summit crater that was subsequently breached by pyroclastic flows and by a crater-rim collapse, which produced a small debris avalanche. Merapi-type dome-collapse pyroclastic flows as well as Vulcanian eruptions (Morrissey and Mastin 2000) occur at Soputan, the latter more common in 2007–2011 owing to an open vent. Three of the recent eruptions produced high-altitude (>10 km) ash clouds.
Table 1

Eruptive history of Soputan Volcano, based on “Volcano Database, Ambang, Awu, Colo, Lokon, Karangetang, Mahawu, Ruang, Soputan and Tangkoko, North Sulawesi” (http://www.vsi.esdm.go.id; CVGHM 2009), Smithsonian Global Volcanism Program Weekly Activity Reports (http://www.volcano.si.edu/reports), and Siebert et al. (2010)

Year

Repose period (years)

Eruption

1785–1786

 

VEI 2

1819

33

VEI 2

1833–1838

14

VEI 1–2 eruptions from Soputan’s summit crater

1845

7

8 February: VEI 2 summit eruption

1890

45

VEI 2 summit eruption

1901

11

4 February: VEI 2 Flank eruption (parasitic crater) and phreatic eruption

1906

5

17 September: VEI 2 flank eruption with lava flow and Aesoput parasitic tephra cone formed

1907

1

5–23 June: VEI 2 eruption from Aesoput crater

1908–1909

1

June 1908–June 1909: VEI 2 Aesoput eruptions produced mainly lava flows

1910

1

Details unknown

1911–1912

1

November 1911–April 1912: VEI 2 eruptions from Aesoput generated lava flows

1913

1

April–July: VEI 2 eruptions from Aesoput

1915

2

April–June: VEI 2 eruptions. Aesoput dome was formed and lava flows to southeast

1917

2

November VEI 2 eruption from Aesoput

1923–1924

6

27th November 1923–18th January 1924: VEI 2 eruptions from Aesoput

1947

23

22nd–27th August: VEI 2 eruption from Aesoput

1953

6

November: VEI 2 eruption from Aseoput1

1966–1967

13

May 1966–November 1967: VEI 3 eruption at Soputan summit crater, lava dome formed, filled crater and flowed to the west

1968

1

July–August: VEI 1 eruption at summit

1970

1

February: VEI 2 eruption at summit

1971

1

19 May: VEI 1 eruption at summit

1973

2

VEI 2 summit eruption

1982

9

26 August, 17–18 September, 9–10 November: VEI 2 and 3 eruptions from summit

1984

2

25–26 May, 31 August: VEI 2 and 3 eruptions at summit crater

1985

1

19th May: VEI 2 eruption at summit crater

1989

4

22th April: VEI 2 eruption at summit crater

1991

2

22–28 May, 8–12 May, 4–12 October, 4–13 December: VEI 2 and 3 explosive eruptions produced scoria cone and extensive tephra followed by growth of new summit lava dome

1992

1

6 February-9 April, 6–7 June, 16–27 September, 18 October-3 November: VEI 2 eruptions produce lava dome and flows within the summit crater; volume, 19.6 Mm3

1995

3

May: VEI 1–2 eruption produces black eruption column to a height about 1,000 m above the crater. Lava dome continues growing at the summit

2000–2003

5

May 2000–September 2003: VEI 1–2 eruption produces black column to 1,000 m above the volcano in May. Lava dome growth continues at the summit through 2003 as well as Strombolian activity and dome collapse pyroclastic flows and surges in July 2003

2004

1

13–19 October and 8–14 December: VEI 2–3 explosive eruption cause ash plumes, tephra fall and rock avalanches, ash plume to 10.7 altitude reported by Darwin VAAC on 12 December. Lava dome growth continues and approaches crater rim level

2005

1

20–26 April: lava fountains and low level ash plumes altitude; 21–27 December: phreatic and Strombolian eruptions, dome collapse pyroclastic flows, and ash plumes to 6 km altitude

2006

1

13–19 December: collapse of lava dome to the east, accompanied by surge, and dome-collapse pyroclastic flows and tephra

2007

<1

20–26 June: diffuse, low-level ash plumes; 8–14 August: low-level ash plumes and lava flow to the east, followed by dome-collapse pyroclastic flows to the east and north

24–30 October: VEI 3 explosive eruption with pyroclastic flows to north and west; collapse of west crater rim produces debris avalanche; new lava flow in crater and on upper northwest slope, ash plumes to 13.7 km altitude. Summit lava dome created between 1991 and 2007 destroyed by end of October eruptions

2008

<1

6 June: VEI 3 explosive eruptions with Vulcanian columns to 13.7 km altitude; column-collapse (St. Vincent type) pyroclastic flows containing breadcrusted bombs, and ash deposits 4 cm thick at 5 km northwest. 18–24 June: additional smaller eruptions; 1–7 October: explosive eruption with plume to 7.6 km, additional St. Vincent type pyroclastic flows

2011

3

3 July: VEI 2 Strombolian open-vent eruption produced 6-km-high ash plume and pyroclastic flows to 4 km down the west flank. Ash deposits 5 cm thick extended 13 km east from the summit and ash clouds temporarily closed the Sam Ratulangi international airport in Manado. Remote sensing data showed that no new lava dome had formed and an open vent structure remained at the summit

Aesoput is a parasitic cinder cone located 1 km northeast of Soputan’s summit (Fig. 3). Ash plumes described as “low level” have reported altitudes of <5 km altitude (Soputan summit is 1.8 km)

Geological setting and eruptive history

Soputan is located in a tectonically complex region of Neogene volcanism in North Sulawesi. Volcanic activity in the region is a consequence of dual subduction of the Molukka Sea, west beneath North Sulawesi and east beneath Halmahera (Fig. 1). The Quaternary geology of North Sulawesi is dominated by the Tondano caldera, a 15 by 30 km NE-SW elongate caldera formed by large-volume explosive eruptions and caldera collapse episodes at approximately 2.0, 1.3, and 0.1 Ma that produced extensive ignimbrites of dacite to rhyolite compositions (Lécuyer 1990). Soputan volcano lies at the southwest margin of the caldera and, unlike Lokon volcano on the northwest caldera margin, Soputan lacks andesitic to dacite magmas.

Subsequent to the main 2.0 Ma caldera-forming eruption and prior to the eruptions at modern Soputan, dacitic to basaltic eruptions took place at several now deeply eroded volcanic centers north of the volcano. Although poorly known, a wide range in composition (basalt to dacite) and in mode of eruption (lava flows to pumiceous pyroclastic flows) at these eroded volcanic centers (Kartadinata et al. 1998) suggests that these centers were derived from the caldera’s magmatic system and are not directly related to the Soputan basalts, which are geochemically distinct from caldera magmas.

Inactive craters form a series of summits just east of the main Soputan summit and mark a local vent migration from east to west (Kartadinata et al. 1998). The first recorded eruptions at the location of Soputan’s current summit took place in 1785 and 1786 (Table 1). Additional summit eruptions took place in 1833, 1845, and 1890. In 1906, eruptive activity shifted to a parasitic cinder cone named Aesoput, located 1 km northeast of the summit and on a trend back toward the inactive craters. Activity at Aesoput continued with lava-flow eruptions during the period 1907–1953. Activity then shifted back to Soputan’s summit with eruptions of lava and tephra in 1966–1991. The 1991 eruption produced a large tephra cone southwest of the summit and the summit crater in which the 1991–1995 and subsequent lava dome grew. The summit dome partially collapsed in 1995 then was rebuilt and continued to grow until it became unstable and collapsed completely during the 2005–2008 eruptions.

Between 1785 and 2011, Soputan erupted at least 33 times, alternating between mainly explosive and mainly effusive activity (Table 1). The eruptions caused loss of property but no casualties. According to historical records, the longest permissive quiescent period was 45 years and the shortest was less than 1 year. From the historical data in Table 1, the average repose period since 1785 is 7 years, but since 1982, the average is only 2 years (Fig. 2). However, the longer repose periods were during the earlier part of the record, before the Soputan observatory was established in the 1980s near the village of Maliku on the volcano’s northwest flank and small eruptions prior to that time may have not made it into the official record. Duration of individual eruptions ranged from a few days to about a month, although the main explosive phases of the most recent eruptions have typically lasted less than 1 day.
Fig. 2

Eruption dates and preceding repose periods of Soputan Volcano. In 2007 and 2008, Soputan erupted two times each year

Methods

Geologic investigations of 2006–2008 eruptive products and deposits

Our geologic investigations of the 2006–2008 eruptions focused on interpretations of satellite imagery, several days of field reconnaissance, and detailed observations made during the eruptions by the Indonesian Center for Volcanology and Geologic Hazard Mitigation (CVGHM) staff at the Soputan Observatory at Maliku and at the North Sulawesi Regional Observatory at Kakaskasen (KKVO). A nearly cloud-free electro-optical Quickbird satellite image from 19 May 2008 (Fig. 3a) provided the basis for a photo-geologic map of the 2006–2008 eruptive products (Fig. 3b, c). Deposits were assigned eruption dates on the basis of satellite data and observations by observatory staff. We conducted reconnaissance field investigations at Soputan in the springs of 2004 and 2007 and again on 7 and 11 March 2008 and on 15–16 March 2009. This field work allowed us to confirm and improve our photo-geologic interpretations, to verify mapped contact locations using handheld GPS, and to examine deposits and collect samples. Because of ongoing low-level seismicity and possibility of additional eruptions during 2008 and 2009, we restricted field time in proximal areas, and we kept in contact with CVGHM staff at nearby observatories, who provided status reports on seismic activity to our field team. During preparation of this paper another eruption took place at Soputan on 3 July 2011. A brief summary of that eruption based on reports from the CVGHM quick response team and on our remote sensing observations is included in Table 1 and minimum areas and volumes of the 2006–2008 deposits are given in Table 2.
Fig. 3

a Quickbird image of Soputan Volcano, provided by Digital Globe, Inc. Width of the image is approximately 5 km. This panchromatic image was collected on 19 May 2008 and has a native resolution of approximately 60 cm. The image is presented at approximately the same scale as the photo-geologic maps (b, c), but the image is not geo-rectified, consequently no scale or grid is given. The prominent light-gray deposit extending west-northwest from the crater is the 26 October 2007 debris avalanche. The dark valley-filling deposit extending off the west edge of the image (in the northwest quadrant) is a lahar deposit composed of remobilized materials from the 2007 flows and the large dark cone at lower center is tephra produced by an eruption in 1991. A series of observations were used to construct photo-geologic maps showing the distribution of the recent flow deposits from 2006 to 2007 (b) and through 2008 (c). RIN and SOP refer to seismic stations referred to in the text

Table 2

Minimum areas and bulk volumes of 2006–2008 flow deposits from Soputan

Date

Product

Measured area (×1,000 m2)

Estimated thickness (m)

Estimated bulk volume (×1,000 m3)

Dec. 2006

PF

320

2.0

640

Surge

200

0.002

4

 Total

   

644

Aug. 2007

PF

111

2.0

222

PF

190

2.0

380

LF

56

1.0

56

 Total

   

658

Oct. 2007

PF

75

1.0

75

PF

296

1.0

296

PF

242

1.0

242

LF

100

2.0

200

LF

6

4.0

2

DA

846

1.0

846

 Total

   

1,661

June 2008

PF

2,057

3.0

6,171

Oct. 2008

PF

759

1.0

759

LF

180

1.0

180

 Total

   

939

Areas are measured from photo-geologic maps (Fig. 3). Thickness estimates are based on field measurements. The volume estimates are minima, due to unmapped small pyroclastic flow and tephra fall deposits related to each of the eruptions, which were too thin or indistinct to reliably assigned to specific eruptions

PF pyroclastic flow, LF lava flow, DA debris avalanche

Field checking of our photo-geologic maps indicated that the locations, extents and areas of the deposits from the 2006–2008 eruptions of Soputan could be determined using a combination of photo-geologic mapping using satellite images and field mapping. We then calculated areas of the flow deposits from 2006 to 2008 using Geographic Information System analysis of our photo-geologic maps (Fig. 3). We estimated average thicknesses of deposits from field observations and calculated approximate deposit volumes (Table 2). The photogeologic mapping method used here may not be accurate as a detailed field mapping campaign; however, given the high level of activity at the volcano, it could be accomplished quickly and with minimal risk. In addition, frequent satellite imaging allowed us to map the extent of eruptive deposits that were buried within months by products of subsequent eruptions.

In May 2004 Soputan’s summit consisted of a 100-m high tephra rampart forming the walls of the crater that contained the 1991–2004 basalt lava dome. During a traverse around the rim of the summit crater we observed that the lava dome had largely filled the crater and we forecast that it would soon collapse, probably over the lower southern rim. Because of time and safety constraints, it was impossible to conduct detailed observations of the dome, and because of fume only areas of the dome adjacent to tephra rampart were visible (Fig. 4a). However, we were able to determine that the ∼200-m-diameter dome had surface features similar to those of more silica-rich compositions. These included multiple flow lobes, a broadly convex upper surface and a marginal breccia composed of coarse (1- to 5-m-scale) blocks, some with prismatic joints (Fig. 4a). Collapse of the lava dome to the south and southeast began in early December 2006, and during the eruption that ensued on 14 December 2006 an ash cloud reached 12-km altitude and pyroclastic flows extended 2.4 km to the south and southeast along the Londola Pinamangkidan valley (Fig. 3). Collapse of the lava dome was accompanied by a small directed blast and surge of hot gas and ash, which downed trees and killed vegetation over an area of approximately 125,000 m2 southeast of the main descent path of the pyroclastic flows.
Fig. 4

a Southeast margin of Soputan’s summit dome on 23 May 2004 and tephra rampart (left and in background) that made up the summit crater rim at the time. The dome had grown to an elevation that exceeded the rim. The visible area of the dome was composed of flow lobes (such as the darker body at the upper right with a relatively smooth upper surface) with a steep blocky flow breccia margin. Blocks in lower right of image are 1–3 m across. b Debris avalanche deposit from 25 to 26 October 2007. Note the reddish, oxidized color of the matrix and the abundant rounded non-juvenile boulders and blocks from the former summit dome and crater. The path of the avalanche is just visible in the photograph, ascending into the clouds and toward the summit in the background of the image. The prominent block in the lower center of image is about 2 m across

Collapse of the lava dome continued during the eruptions of August 2007, and produced block-and-ash-type pyroclastic flows directed to the south and northwest. Then on the night of 25–26 October 2007, additional collapse produced block and ash pyroclastic flows that descended to the west and swept parts of the upper north-facing slopes of the volcano. The October eruption also produced an ash cloud that reached 13.7-km altitude and a distinct reddish-tan colored landslide deposit (Fig. 4b). Rock miners arrived near the western terminus of these deposits at about 0600 local time on the morning of 26 October. They noted the presence of both the landslide and pyroclastic-flow deposits, thereby constraining emplacement of both to the night of October 25–26. The landslide deposit forms a distinct mappable unit composed mainly of non-juvenile rock debris from the summit, along with rare juvenile breadcrust blocks. The matrix of the avalanche deposit consists of sandy altered rock fragments, which give rise to the reddish-tan color. Average thickness of the deposit is only about a meter, although outsize boulders protrude many meters above its surface. This deposit extends 2.7 km west from the summit (Fig. 3) and has a volume of 850,000 m3. Non-juvenile blocks and outsized boulders range to more than 10 m in diameter (Fig. 5). The landslide deposit overlies the pyroclastic flow deposits, thereby constraining its emplacement to late during the night or early morning of 25–26 October. We infer that the initial phase of the eruption left a steep western segment of the former summit crater wall, which then collapsed (along with remaining parts of the lava dome) to produce the deposit. Because of the mainly non-juvenile character and inferred origin, we refer to the landslide deposit as a debris avalanche deposit.
Fig. 5

Seven-meter long prismatically jointed boulder in debris avalanche deposit from 25 to 26 October 2007 avalanche deposit; derived from the former summit dome. The boulder is draped by a younger 2008 pyroclastic deposit

We conclude that the eruption of 25–26 October began with dome-collapse down the north and west flanks, producing block-and-ash pyroclastic flows and eroding new channels through the former crater rim and leaving a steep and unstable western crater wall. This early phase of the eruption removed much of the 1991–2007 lava dome and exposed the conduit, which led to a Vulcanian eruption that was accompanied by column-collapse pyroclastic flows with breadcrusted blocks of juvenile basalt. Late during the night or morning of 25–26 October, ground shaking and steepening of the slopes by explosive disruption triggered failure of the over-steepened western sector of the crater wall, which collapsed down the steep west flank of the volcano to form the debris avalanche.

Small lahar deposits were observed downslope and south of the front of the October 2007 pyroclastic flow and debris avalanche deposits. Another prominent lahar forms a dark channel-filling unit, which extends off the west margin of the Quickbird image in Fig. 3. This western lahar formed as a result of rainfall that mobilized ash and debris from the pyroclastic flows and the debris avalanche of the 2007 eruptions.

By the close of the 26 October 2007 eruption, most of the 1991–2007 summit lava dome was destroyed, thereby unplugging the volcano’s conduit and resulting in the mainly explosive eruptions that followed in 2008 and 2011. During the waning stage of the 2007 eruption on 27–30 October, small lava flows accumulated in the new crater, one of which flowed to the northwest and then collapsed down the steep north flank to produce a blocky pyroclastic-flow deposit (Fig. 3).

The magmatic phases of the 2008 eruptions were also of the open-vent Vulcanian–St. Vincent type, although, in addition to breadcrust blocks, they contain abundant boulders (some prismatically jointed), derived from the remains of the former summit lava dome or from the 2007 crater lava flows (Fig. 6). On the basis of greatest extent of the pyroclastic flow deposits (5.5 km along the Royongan Papang valley, Fig. 3), maximum heights of the convective columns (13.7 km), distribution of ash to >15 km (Kristianto and Loeqman 2008), and high levels of seismic energy release (Fig. 10), the 5–6 June 2008 eruption and the 25–26 October 2007 eruptions are considered the largest of the 2004–2008 eruptions.
Fig. 6

Pyroclastic flow deposit of 5–6 June 2008 (a) taken shortly after the eruption by Agus Solihin of CVGHM (see also Hadisantono et al. 2008; Suantika 2008; Santosa 2008) and photo of western flank of Soputan (b) taken in March 2010 by Kushendratno showing the 1991 tephra cone and 2008 lava flow and pyroclastic flow and the new summit crater (steam filled) where just a few years before there had been a summit lava dome

Remote sensing of eruption plumes and sulfur dioxide emissions

Satellite remote sensing provided key observations of eruption cloud heights and release of sulfur dioxide (SO2) during the recent eruptions. Altitudes of eruption clouds and dispersion of ash from several of the eruptions were determined using reports from the Darwin Volcanic Ash Advisory Center (VAAC; www.bom.gov.au/info/vaac/) utilizing satellite data, such as the Multi-functional Transport Satellite (MTSAT) image for a plume from the 25 October 2007 eruption (Fig. 7) and ground observations by CVGHM. Plume altitudes are as follows: 14 December 2006, 12 km; 14 August 2007, 4.6 km, 25–26 October 2007, 13.7 km; 6 June 2008, 13.7 km; 6 October 2008, 7.6 km; and 3 July 2011, 6 km. The summit of Soputan has an altitude of 1,800 m.
Fig. 7

MTSAT image from 25 October 2007 (02:57 UTC) showing the steam and ash plume extending to approximately 12 km (40,000 ft) altitude. Plumes reached a maximum height of 13.7 km later in the eruption. Image from the Japanese Meteorological Agency and the Darwin Volcanic Ash Advisory Center. Total SO2 emission by 05:08 (UTC) on 25 October 2007 was estimated as 5–7 kilotons

We utilized a series of Ozone Monitoring Instrument (OMI) images to estimate the tonnage of SO2 emitted in each of Soputan’s recent eruptions. Examples are shown in Fig. 8. The 27 December 2005 OMI image was collected shortly after peak activity, which consisted of a period of violent Strombolian activity that took place between 0400 and 0450 UTC. The 14 August 2007 and 25 October 2007 images were collected relatively late during these eruptions. The image of 7 October 2008 was collected well after peak activity but still shows the full extent of the westward drifting plume at altitude.
Fig. 8

Ozone Monitoring Instrument (OMI) images from the Aura satellite, showing SO2 plumes from eruptions of Soputan volcano (triangle). Images were acquired on a 27 December 2005 at 0520 UTC, b 14 August 2007 at 0600 UTC, c 25 October 2007 at 0505 UTC, and d 7 October 2008 at 0530 UTC and show mid-tropospheric (TRM) SO2 columns retrieved by the operational OMI SO2 algorithm (Yang et al. 2007). Units are Dobson Units (DU) where 1 DU = 0.0285 g m−2 of SO2

Because of the timing of OMI measurements (overpass at ∼1:45 p.m. local time) with respect to peak activity during these eruptive episodes, OMI’s reduced sensitivity to SO2 at low altitudes, and expected rapid SO2 oxidation to sulfate aerosol in the tropical atmosphere, we consider the OMI estimated SO2 tonnages to be minima for the total emissions during each eruptive episode. However, as each image was collected late with respect to peak activity, they can be used in a relative sense for comparison to one another and to the 2006 eruption of Merapi, as shown in Fig. 9. From this comparison, it is evident that the June 2008 eruption was accompanied by the largest magmatic gas emission, and that overall, despite the small magmatic volumes of the eruptions, SO2 emissions from Soputan are high relative to other Indonesian volcanoes.
Fig. 9

Total daily SO2 burdens measured by OMI over Indonesia (latitude 12° S–7° N, longitude 95–132° E) from 6 September 2004–1 November 2008. Eruptions of Soputan are denoted by star symbols and show that Soputan was one of the strongest SO2 sources in Indonesia in this period. Notable spikes in SO2 due to other eruptions are also indicated, with Indonesian volcanoes indicated in orange. Note that for non-Indonesian eruptions (i.e., Manam (PNG), Anatahan (CNMI), Soufriere Hills (Montserrat), and Rabaul (PNG)), the SO2 tonnages correspond to the SO2 measured as the volcanic clouds drifted over Indonesia after long-range transport, and not the total erupted SO2 mass. The gray curve shows ultraviolet (UV) reflectivity, which increases with meteorological cloud coverage in the region

Seismic data

In 2004, CVGHM and the Volcano Disaster Assistance Program began a cooperative project to build a new regional volcano observatory for North Sulawesi and the Sangihe Islands and to upgrade monitoring instrumentation throughout the region. The KKVO, located in the town of Kakaskasen, Tomohon Province (between Lokon-Empung and Mahawu volcanoes, Fig. 1), receives real-time telemetered seismic data from a network of seismometers on Soputan and nine other active volcanoes in North Sulawesi and the Sangihe Islands (Ambang, Lokon, Empung, Mahawu, Klabat, Tongkoko, Ruang, Karangetang, and Awu). Seismic data are acquired and processed at KKVO and are transmitted by a Very Small Aperture Terminal satellite system or by Internet to CVGHM headquarters in Bandung using USGS Earthworm software (Bittenbinder et al. 1994).

Because of access and maintenance considerations and because the networks are designed principally as operational systems for early warning, only two short-period stations were installed at Soputan in March 2007. The stations are a single-component L4 1-Hz seismometer, SOP, located 1.4 km to the northeast of the summit of Soputan, and a three-component borehole L22 2-Hz seismometer, RIN, located 2.5 km to the northeast of the summit (Fig. 3). Changes were made to these stations the following year, including the reorientation of the vertical seismometer SOP and a 6 dB increase in the gain of station RIN. Both these changes affect our estimates of the Real-Time Seismic Amplitude Measurement (RSAM; Endo and Murray 1991): RSAM for SOP is underestimated by an unknown amount in 2007 because it was not perfectly level; and RSAM for RIN appears to be larger by a factor of two in 2008 because of the gain change. The majority of the seismic analyses were done using station SOP because it has the most data available during the period of interest.

We restrict our seismic analyses to the available digital data and to the first-order interpretations that can be used in eruption forecasting. RSAM graphs are used to define the times of peak energy release and to identify periods of precursory, eruption and post-eruption seismicity (Fig. 10). Coda duration magnitudes, Md, were determined using the formula Md = 1.86 log(t) − 0.85, where t is time in seconds. Digital data were also examined to characterize the event types (e.g., volcano tectonic (VT), hybrid, low frequency (LF), explosion, tremor, and events associated with rockfalls and pyroclastic flows) (Fig. 11). These characteristics and patterns of precursory and post-eruption seismicity provide insights into the types of seismic signals and duration of precursors that we expect as warnings to future eruptions.
Fig. 10

Ten-Minute RSAM averages for stations SOP (blue) and RIN (red) for each of the four eruptions in 2007–2008 (ad) and of background seismicity in March 2010 (e). Vertical scale is logarithmic. Onsets of VT swarms, LFs, emission signals, tremor and continuous eruptions are marked with arrows. Seismicity following each eruption consists primarily of LFs, emission and debris flow signals and continues for days to weeks. RSAM provides a quantitative understanding of the energy release during the eruption and is particularly useful when events cannot be located, magnitudes cannot be determined, or the seismicity is dominated by tremor, emissions, eruptions or debris flow events

Fig. 11

Examples of seismic signals recorded at Soputan on station SOP. The left column (af) shows the uncorrected velocity time series, and the right column (gl) shows the corresponding spectrogram. All spectrograms are scaled the same based on the power spectrum amplitude range (blue is low and red is high). a, g Example of a broadband high frequency VT from the precursory VT swarm in June 2008. b, h Example of a LF earthquake during the October 2008 eruption. c, i are Example of eruption tremor from the October 2008 eruption. Tremor at Soputan is generally low frequency but not narrow band, and at times during the October 2008 eruption, the tremor had a dominant frequency around 2.6 Hz with several overtones that glide within about a 0.5 Hz of the original frequency. d, j Examples of a signal that starts low frequency and then becomes broadband and an emission, small explosion or pyroclastic flow. These signals can occur both before and after an eruption. e, k Example of a rockfall or pyroclastic flow signal from the August 2007 eruption. f, l Example of the onset of an eruption from October 2008. These signals often start low frequency and become more broadband and higher frequency within the first minute of the onset

Since March 2007, the local seismicity between eruptions at Soputan exhibits characteristics common at ‘open’ volcanic systems (defined here as volcanic systems with relatively frequent eruptions at intervals typically measured in months to years). The local events consist primarily of small LF earthquakes and rockfall signals, with occasional small VT earthquakes. While the activity level can change day to day, it changes slowly and within a limited amplitude range (Fig. 10e).

Soputan erupted four times between March 2007 and March 2010. For each of these four eruptions, there was a short seismic precursory period immediately preceding the eruption, that consisted primarily of high frequency VT earthquakes, with maximum Md of 2.5 (Figs. 10 and 12). This VT dominated precursory period is distinctly different from the background seismicity in both earthquake rate and magnitudes. The precursory seismicity lasted less than 8 h in August 2007, 13.5 h in October 2007, 44.5 h in June 2008 (Fig. 12), and 47.5 h in October 2008. In each case, smaller LF earthquakes occurred along with the VT earthquakes. For three of the eruptions, the VT seismicity dominated the initial seismic energy release until the first large explosion. However, in one case, during the 32 h leading up to the largest eruption in August 2007, the seismicity became dominated by LF earthquakes, small explosions, tremor and seismicity produced by pyroclastic flows and/or rockfalls (Fig. 11). Energy release rates as determined from RSAM (which does not differentiate between event types) accelerated before each initial eruption (Figs. 10 and 12).
Fig. 12

Helicorder plot and corresponding 10-minute RSAM for the eruption of 5-6 June 2008 on station SOP. Each line represents an hour of data, and the colors change for each line for easier viewing. The first 25.5 hours show the increasingly intense precursory VT swarm. On June 6, around 01:46 UTC these events quickly transitioned into longer duration and lower frequency explosion signals, and by 03:00 Soputan was in a sustained eruption of explosions, pyroclastic flows and tremor. The sustained eruption was large enough that the associated seismic ground velocities exceeded the dynamic range of the instrument and lasted for over 17 hours. Seismicity remained elevated above background levels for about a week with small explosions, LFs, VTs, and tremor. Calibration spikes are visible on June 5 at 04:18 and 22:18, those on June 6 are obscured by the eruption. RSAM rapidly increases when the events transition from VTs to the eruption, and returns to pre-eruption values within a week of the eruption

Seismicity during the four eruptions consists of at least several hours of low frequency tremor and events related to explosions and pyroclastic flows. Frequency content changes are evident in the spectral domain; however, we cannot always ascertain the true mechanisms (i.e., pyroclastic flow verses rockfall) with the current seismic network. Post-eruption seismicity returns to background at a slower rate (over days to months) than the precursory seismicity ramps up to eruption. The post-eruption seismicity is dominated by signals related to rockfall and debris flow events and by low-frequency earthquakes. Largest of the eruptions in terms of seismic energy release is the June 2008 event, and the longest based on the duration of the eruption seismicity was the October 2008 eruption. Overall, the key message from the seismicity is that a relatively open pathway for magma ascent is currently present at Soputan, and that the volcano is capable of explosive eruptions with less than 24 h of seismic warning.

Geochemistry and petrology

Samples were collected from juvenile breadcrust blocks in the 2007 and 2008 pyroclastic-flow deposits. Additional samples were collected from lava flows erupted in 1911–1912, 1991–1995, 2000–2001, and 2004. Locations for geochemical samples are given in Table 3. Samples were analyzed at Washington State University using standard X-ray fluorescence methods for major elements and inductively coupled plasma mass spectrographic methods for trace elements. Geochemical data for Soputan samples are listed in Table 4 and geochemical analyses for other volcanoes in the North Sulawesi–Sangihe arc are given in the ESM. All of the Soputan samples are relatively evolved low-Mg high-alumina basalts (HAB) with 19–20 wt.% Al2O3 and total iron approaching 10 wt.%. Despite an age range of 100 years, our samples are remarkably homogeneous in composition, with an average SiO2 of 50.73 ± 0.14 wt.% and an average atomic Mg/(Mg + Fe+2) ratio of 0.53 ± 0.01. The Soputan basalts also have small amounts of normative quartz and average K2O abundances of ∼0.30 wt.%, making them low-K tholeiites according to the criteria of Peccerillo and Taylor (1976).
Table 3

Descriptions and locations of samples collected for this report

PTS

Description

N. Latitude

E. Longitude

S52304-4

2003 lava

1°7.3′

124°44.5′

S52304-5

1991–95 lava flow

1°7.3′

124°44.5′

S31407-2

Aseput lava flow, 1911–1912

1°7.355′

124°44.500′

S31407-5

Aseput lava flow, 1911–1912

1°7.400′

124°44.500′

S030708-2

Juvenile block in 2007 pyroclastic flow

1°7.360′

124°44.321′

S030708-4

2004 lava-flow front

1°7.146′

124°44.669′

S030708-6

Juvenile breadcrust block in 2007 pyroclastic flow deposit

1°7.000′

124°44.733′

S030708-7

2007 lava flow exposed in channel below 2007 pyroclastic flow

1°7.033′

124°44.783′

S030808-2

Lava block in 2007 pyroclastic flow deposit; from summit dome

1°7.342′

124°43.951′

S030808-2b

Non-Juvenile block in November 2007 pyroclastic flow deposit; from former summit dome

1°7.342′

124°43.951′

S031108-1

Juvenile breadcrust block in November 2007 pyroclastic flow deposit

1°7.182′

124°42.609′

S031108-2

Non juvenile lava block in November 2007 pyroclastic flow deposit; from former summit lava dome

1°7.185′

124°42.878′

S031108-4

2000–2001 lava flow

1°7.322′

124°42.922′

S031108-6

Juvenile breadcrust block in 2007 pyroclastic flow deposit

1°7.429′

124°42.949′

Table 4

Whole-rock major-element and trace-element analyses of samples from Soputan volcano

 

S052304-4

S052304-5

S031407-2

S031407-5

S030708-2

S030708-4

S030708-6

S030708-7

S030808-2

S030808-2b

S031108-1

S031108-2

S031108-4

S031108-6

Average

SD

2003 lava

1991–1995 lava

1911–1912 lava Aseoput

1911–1912 lava Aseoput

2007 PF J. block in PF

2004 lava

2007 PF J. block in PF

2007 PF J. block in PF

2007 PF J. block in PF

N.J. block in PF

2007 PF

Non-juvenile block in PF

2000–2001 lava

2007 PF J. block in PF

Normalized major elements (wt.%)

SiO2

51.02

50.83

50.58

50.76

50.68

50.58

50.60

50.81

50.70

50.96

50.68

50.51

50.77

50.70

50.73

0.14

TiO2

0.84

0.83

0.85

0.85

0.83

0.83

0.83

0.84

0.83

0.83

0.83

0.83

0.86

0.84

0.84

0.01

Al2O3

19.71

19.69

19.52

19.65

19.83

19.83

19.79

19.54

19.75

19.67

19.67

19.71

19.52

19.59

19.68

0.10

Fe2O3

1.19

1.20

1.19

1.15

1.18

1.20

1.21

1.21

1.19

1.20

1.20

1.22

1.22

1.20

1.20

0.02

FeO

8.49

8.52

8.46

8.22

8.42

8.57

8.65

8.60

8.49

8.52

8.58

8.73

8.69

8.55

8.53

0.13

MnO

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.20

0.00

MgO

5.27

5.49

5.74

5.61

5.47

5.44

5.50

5.52

5.51

5.29

5.47

5.60

5.46

5.56

5.50

0.12

CaO

10.18

10.20

10.46

10.53

10.39

10.41

10.20

10.25

10.35

10.27

10.36

10.27

10.21

10.36

10.32

0.11

Na2O

2.68

2.63

2.58

2.61

2.60

2.54

2.60

2.62

2.59

2.64

2.59

2.53

2.62

2.59

2.60

0.04

K2O

0.31

0.31

0.31

0.31

0.30

0.29

0.30

0.30

0.30

0.31

0.30

0.29

0.32

0.30

0.30

0.01

P2O5

0.12

0.12

0.11

0.11

0.11

0.11

0.11

0.11

0.11

0.11

0.11

0.11

0.13

0.12

0.11

0.00

Total

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

 

Mg/Mg + Fe+2

0.53

0.53

0.55

0.55

0.54

0.53

0.53

0.53

0.54

0.53

0.53

0.53

0.53

0.54

0.53

0.01

CIPW normative minerals

Quartz

0.6

0.3

0.0

0.1

0.2

0.3

0.1

0.3

0.2

0.6

0.2

0.1

0.3

0.2

0.3

0.19

Plagioclase

63.5

63.3

62.6

63.1

63.6

63.3

63.4

62.8

63.3

63.2

63.1

63.0

62.7

62.9

63.1

0.30

Orthoclase

1.8

1.8

1.8

1.8

1.8

1.7

1.8

1.8

1.8

1.8

1.8

1.7

1.9

1.8

1.8

0.05

Nepheline

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.00

Diopside

7.4

7.3

8.6

8.7

7.7

7.5

7.0

7.8

7.7

7.7

7.9

7.2

7.7

8.0

7.7

0.48

Hypersthene

23.0

23.7

22.9

22.8

23.3

23.6

24.1

23.7

23.5

23.0

23.5

24.4

23.7

23.6

23.5

0.46

Wollastonite

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.00

Olivine

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.00

Ilmenite

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

1.6

0.02

Magnetite

1.7

1.7

1.7

1.7

1.7

1.7

1.8

1.8

1.7

1.7

1.7

1.8

1.8

1.7

1.7

0.02

Apatite

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.3

0.02

XRF (ppm)

Ni

9

12

15

12

10

10

10

10

12

9

10

12

11

9

11

1

Cr

4

6

7

6

4

6

3

4

6

5

4

6

6

5

5

1

Sc

27

28

29

31

27

27

26

28

27

29

27

28

29

29

28

1

V

256

246

262

261

252

256

256

263

251

256

252

260

263

261

257

5

Ba

59

58

57

59

59

61

59

62

56

55

59

59

66

57

59

3

Rb

5

5

5

6

5

5

6

5

6

5

5

6

5

6

5

0

Sr

306

300

302

305

303

301

305

300

299

302

302

301

303

302

302

2

Zr

35

35

33

33

33

34

35

35

33

35

34

34

35

34

34

1

Y

17

16

16

17

16

16

16

16

16

16

16

17

16

16

16

1

Nb

0.3

1.7

0.4

0.6

0.5

0.7

0.6

0.9

0.4

0.6

0.5

0.4

0.4

0.4

0.6

0.4

Ga

16

16

15

16

15

15

16

16

17

16

16

15

18

15

16

1

Cu

84

87

86

83

88

87

88

88

87

90

86

52

88

92

85

10

Zn

74

72

71

73

75

73

74

74

72

71

74

72

74

70

73

1

Pb

2

3

2

3

2

2

1

3

1

2

2

2

2

2

2

1

La

2

3

3

4

5

5

6

2

6

4

7

6

0

3

4

2

Ce

10

6

6

6

7

6

7

10

4

9

9

10

8

4

7

2

Th

1

1

1

1

1

1

2

0

1

2

1

1

1

1

1

0

Nd

10

7

7

6

8

7

6

6

6

9

7

7

9

5

7

1

U

0

0

2

0

0

1

3

2

1

0

0

2

0

1

1

1

ICP-MS (ppm)

La

2.64

2.77

2.61

2.66

2.67

0.07

Ce

7.15

7.49

7.12

7.25

7.25

0.17

Pr

1.16

1.21

1.14

1.14

1.16

0.03

Nd

6.12

6.29

6.06

6.05

6.13

0.11

Sm

1.95

2.16

2.00

2.10

2.05

0.09

Eu

0.84

0.88

0.84

0.84

0.85

0.02

Gd

2.49

2.61

2.52

2.55

2.54

0.05

Tb

0.46

0.48

0.46

0.48

0.47

0.01

Dy

3.05

3.09

2.92

3.06

3.03

0.08

Ho

0.64

0.66

0.61

0.65

0.64

0.02

Er

1.75

1.83

1.72

1.74

1.76

0.05

Tm

0.25

0.27

0.25

0.26

0.26

0.01

Yb

1.57

1.63

1.58

1.62

1.60

0.03

Lu

0.26

0.27

0.25

0.26

0.26

0.01

Ba

56.4

59.7

57.1

56.5

57.4

1.6

Th

0.38

0.38

0.37

0.37

0.38

0.00

Nb

0.73

0.75

0.71

0.72

0.73

0.02

Yb

15.7

16.4

15.8

16.0

16.0

0.3

Hf

0.99

1.05

0.99

1.02

1.01

0.03

Ta

0.05

0.05

0.05

0.05

0.05

0.00

Yb

0.10

0.10

0.10

0.11

0.10

0.00

Pb

2.49

2.50

2.49

2.54

2.50

0.02

Rb

4.98

5.21

4.86

5.00

5.01

0.15

Cs

0.40

0.38

0.31

0.37

0.37

0.04

Sr

300

302

303

299

301

2

Sc

24.9

25.7

25.7

26.9

25.8

0.8

Zr

29.4

30.9

29.6

29.9

29.9

0.7

Major element abundances, volatile-free with Fe2O3 = 0.123 FeO (total)

PF pyroclastic flow, J juvenile, N.J. non-juvenile blocks in 2007 pyroclastic flow from former summit dome, SD standard deviation, XRF X-ray fluorescence, ICP-MS inductively coupled plasma mass spectrographic methods

All of the samples are crystal rich with plagioclase and olivine phenocrysts and abundant small phenocrysts and microphenocrysts of pigeonite, consistent with their tholeiitic compositions. In this paper, we use the term phenocrysts for crystals with long dimensions in thin section of >0.1 mm, microphenocrysts for generally equant crystals with long dimensions of less than 0.1 mm (100 μm) and microlites for equant or needle shaped crystals with average dimensions in thin section less than about 0.01 mm (10 μm). The Soputan basalts also contain relatively common megacrysts and xenocrysts of olivine and augite, as well as gabbroic (olivine–augite–plagioclase) glomerocrysts. The largest of these are on the order of 1 cm in diameter.

Two types of olivine are present; most are subhedral, rounded, or embayed, cluster at Fo60 in composition and have reacted with melt to produce pigeonite and oxide rims (Table 5; Figs. 13 and 14a, b). A second, smaller proportion is euhedral to subhedral, not rimmed by pigeonite, and clusters at Fo70 composition (Fig. 14c). Augite occurs only in glomerocrysts and as separate subhedral crystals that are likely derived from xenocrysts. The phenocryst phases show only modest amounts of compositional zoning. Plagioclase phenocrysts are euhedral to subhedral, and some crystals show weak oscillatory zoning. Plagioclase cores are calcic, ranging from An85-95 and they have prominent rims (Figs. 13 and 15) that range from An55–75, overlapping at the low-An end with compositions of groundmass microlites. The pigeonite in these samples is typically euhedral to subhedral and occurs as small phenocrysts or microphenocrysts in the groundmass and surrounding resorbed olivine grains (Fig. 14). A small proportion of the plagioclase phenocrysts in all samples has irregular glass-inclusion-rich zones, and all of the silicate phases contain sparse glass inclusions. A single oxide phase (titanomagnetite) is present and typically occurs as abundant subhedral phenocrysts, and as microphenocrysts in the groundmass. Large titanomagnetite (to 1 mm) crystals also occur in some of the coarse grained gabbroic glomerocrysts; no chrome-rich spinel (chromite) or sulfides were identified.
Table 5

Electron microprobe analyses of minerals and glass in Soputan basalt samples S031108-1 and S030808-2

 

Numbera

SiO2

TiO2

Al2O3

FeO

MnO

MgO

NiO

CaO

Na2O

K2O

P2O5

S

Cl

Total

Te

Fs

Fo

Fa

Wo

En

Ca-Ol

An

Ab

Or

H2O*

Pyroxene

 Augite xenocrysts

 Average

17

51.20

0.51

2.83

9.64

0.32

15.38

 

19.69

0.26

    

99.84

 

12.4

  

42.0

45.6

     

 SD

0.46

0.06

0.39

0.44

0.04

0.27

 

0.38

0.02

    

0.49

 

0.8

  

0.9

0.6

     

 Sub-calcic augite rim on augite xenocryst

1

50.37

0.86

3.61

12.58

0.48

14.77

 

16.65

0.27

    

99.60

 

19.6

  

36.0

44.4

     

 Pigeonite phenocrysts and microphenocrysts

 Average

42

52.97

0.32

0.93

20.24

0.63

20.71

 

4.24

0.09

    

100.13

 

32.1

  

8.7

59.2

     

 SD

0.41

0.06

0.28

0.40

0.05

0.43

 

0.44

0.04

    

0.59

 

0.8

  

0.9

0.9

     

 Pigeonites with En~65

 Average

7

53.52

0.28

1.15

19.21

0.56

23.33

 

2.14

0.06

    

100.26

 

30.0

  

4.3

65.7

     

 SD

0.33

0.03

0.10

0.17

0.04

0.08

 

0.11

0.00

    

0.25

 

0.3

  

0.2

0.2

     

 Pigeonites with En~75

 Average

4

53.63

0.27

1.61

16.38

0.56

26.08

 

1.93

0.03

    

100.50

 

23.1

  

3.9

73.0

     

 SD

0.40

0.05

0.20

0.14

0.04

0.28

 

0.07

0.00

    

0.55

 

0.2

  

0.2

0.3

     

Olivine

 Resorbed olivine

 Average

26

36.21

  

31.61

0.60

30.16

0.13

0.25

     

98.96

0.7

 

62.4

36.6

  

0.4

    

 SD

0.38

  

0.97

0.04

0.84

0.11

0.05

     

0.55

0.05

 

1.3

1.3

  

0.1

    

 Euhedral-subhedral olivine

 Average

24

37.51

  

25.76

0.50

35.15

0.18

0.21

     

99.33

0.6

 

70.3

28.8

  

0.3

    

 SD

0.33

  

0.61

0.04

0.53

0.10

0.06

     

0.40

0.05

 

0.8

0.7

  

0.1

    

Plagioclase

 High-Ca cores

 Average

6

44.35

 

36.07

0.69

0.00

0.10

 

19.25

0.62

0.01

   

101.10

       

94.4

5.5

0.0

 

 SD

0.29

 

0.20

0.20

0.01

0.05

 

0.30

0.08

0.01

   

0.20

       

0.8

0.7

0.1

 

 Cores of smaller phenocrysts (An80-90)

 Average

15

46.36

 

34.06

0.83

0.01

0.10

 

17.51

1.57

0.03

   

100.47

       

85.8

14.0

0.2

 

 SD

1.32

 

1.29

0.39

0.02

0.09

 

1.27

0.20

0.03

   

3.69

       

2.1

2.1

0.1

 

 Interior rims (An70-75)

 Average

4

51.02

 

31.13

0.99

0.00

0.11

 

14.47

3.20

0.08

   

101.01

       

71.1

28.4

0.5

 

 SD

0.08

 

0.28

0.04

0.01

0.01

 

0.25

0.11

0.03

   

0.34

       

1.0

1.0

0.2

 

 Mid-rims (An60-65)

 Average

4

53.51

 

29.27

0.91

0.00

0.14

 

12.67

4.15

0.15

   

100.81

       

62.2

36.9

0.9

 

 SD

0.16

 

0.15

0.05

0.02

0.01

 

0.09

0.07

0.03

   

0.18

       

0.3

0.4

0.2

 

 Edges and microlites (An<60)

 Average

7

55.03

 

28.39

1.00

0.01

0.14

 

11.43

4.63

0.16

   

100.80

       

57.1

41.9

1.0

 

 SD

0.54

 

0.34

0.12

0.02

0.01

 

0.43

0.12

0.02

   

0.30

       

1.5

1.5

0.1

 

Glass

 Matrix glass

 Average

7

57.34

1.95

13.31

13.23

0.23

2.49

 

7.00

2.28

1.12

0.48

0.002

0.146

99.57

          

0.4

 SD

0.75

0.09

0.46

0.51

0.03

0.32

 

0.38

0.55

0.09

0.05

0.002

0.029

0.89

          

0.9

 G.I. in resorbed Fo60 olivine

 Average

3

54.12

0.93

15.33

13.82

0.28

2.92

 

8.24

2.45

0.51

0.20

0.069

0.149

99.02

          

1.0

 SD

0.15

0.01

0.16

0.16

0.04

0.07

 

0.10

0.04

0.02

0.00

0.004

0.003

0.20

          

0.2

G.I. in resorbed, unrimmed Fo70 olivine

 Average

5

53.93

1.28

15.94

13.18

0.27

3.29

 

7.70

2.47

0.58

0.20

0.038

0.155

99.03

          

1.0

 SD

1.69

0.16

1.21

0.42

0.04

0.43

 

0.54

0.12

0.10

0.04

0.018

0.006

0.22

          

0.2

 G.I. in augite xenocryst

 Average

5

65.87

0.87

18.15

3.38

0.12

0.89

 

4.28

1.45

0.42

0.24

0.030

0.190

95.90

          

4.1

 SD

0.73

0.07

0.60

0.76

0.03

0.15

 

0.40

0.10

0.04

0.01

0.009

0.009

0.66

          

0.7

All analyses by JEOL Hyperprobe electron microprobe at Washington State University, utilizing 15 keV accelerating voltage, 30 nA beam current, and 3 μm spot diameter, except for glass analyses, which were done at a minimum of 10 μm spot diameters and utilized a runtime correction procedure to correct alkali loss during the analyses

SD standard deviation, Fs ferrosilite component in pyroxenes, Wo wollastonite component in pyroxenes, En enstatite component in pyroxenes, Te tephroite component in olivine, Fo forsterite component in olivine, Fa fayalite component in olivine, Ca-Ol calcic-olivine component in olivine, An anorthite component in plagioclase, Ab albite component in plagioclase, Or orthoclase component in plagioclase, H2O* approximate water abundance by difference from 100 wt.% in glass analyses, G.I. glass inclusion

aNumber of analytical points used in average and standard deviation

Fig. 13

Backscattered electron images of Soputan basalt sample S0310801-1 showing crystal-rich character, resulting from growth of small phenocrysts and microlites of pigeonite and plagioclase in the groundmass, as illustrated in (a) and (c). The light gray matrix between the microlites and vesicles in (c) is volcanic glass. As illustrated in (a), plagioclase phenocrysts are euhedral and homogeneous except for relatively thin edge zones (darker gray boundaries, which reflect lower Ca and lower mean atomic number). Augite is present as a xenocrystic phase, typically in glomerocrysts with olivine and plagioclase, as shown in (b). Symbols: ol olivine, pl plagioclase, mt magnetitie, pig pigeonite, aug augite, ves vesicles. Scale bars are given in the lower right of each image

Fig. 14

Backscattered electron images of olivine crystals in Soputan basalt sample S0310801-1 showing a typical Fo∼60 olivine with glass inclusion and pigeonite reaction rim, b extreme replacement of Fo∼60 olivine by pigeonite, and c subhedral Fo∼70 olivine without reaction rim. Abbreviations as in Fig. 13, except gl glass

Fig. 15

Mineral compositions in samples S031108-1 and SO 30808-2 from Soputan volcano, recast into enstatite (En)–wollastonite (Wo)–ferrosilite (FS) components for pyroxenes, forsterite (Fo)–fayalite (Fa) components for olivine, and anorthite (An)–orthoclase (Or) components for plagioclase

Matrix glass constitutes up to about 20 % by volume in Soputan samples and is variably devitrified. In quenched parts of breadcrust blocks, matrix glass is well preserved and is andesitic in composition with ∼57 wt.% SiO2. Microprobe totals for the matrix glass approach 100 % and S and Cl are near detection limits, indicating extensive near-surface degassing. Glass inclusions have variable SiO2 contents, ranging from 54 to 65 wt.% SiO2. Maximum measured S is ∼0.070 wt.% (1,400 ppm as SO2) in an olivine-hosted inclusion within a resorbed Fo∼60 olivine crystal of xenolithic origin and maximum Cl is ∼0.20 wt.% (2,000 ppm) in an augite-xenocryst-hosted inclusion, which also has a low oxide total of ∼96 % and suggests a maximum of ∼4 wt.% H2O + CO2 in the parent melt for these xenoliths. In contrast, analyzed glass inclusions in the olivine phenocrysts all have relatively high totals, indicating relatively low volatile contents during crystallization of the Soputan basalt.

The petrography and phase chemistry of the Soputan basalts show several unusual features. First, the extent of crystallization is high with 40–50 vol.% phenocrysts of plagioclase, olivine, pigeonite and titanomagnetite (on a vesicle-free basis) and ∼80 vol.% total crystal content when microlites are included in the totals (Table 6; Figs. 13 and 14). Such a high crystal content played an important role in controlling eruptive behavior, as explained below. Second, the quartz-saturated tholeiitic basalt composition, combined with the high degree of crystallization resulted in extensive crystallization of pigeonite, which is present both as a reaction product of olivine and as a phenocryst and microphenocryst phase. Third, the modest range of zoning indicates crystallization from a magma that changed little in composition during the extensive crystallization. And finally, the presence of coarse-grained olivine and augite xenocrysts and olivine–clinopyroxene–plagioclase glomerocrysts, as well as two distinct compositions of olivine, suggest entrainment of crystals from a coarse-grained gabbroic crystal mush. Each of these features gives insight into the origin and ascent history of the magma, as discussed below.
Table 6

Modal data for samples of lavas and pyroclastic flow deposits from Soputan volcano

Sample no.

Note

Gm

Pl

Ol

Pig

Aug

Ox

Ph

Ves

Points

Gl

S031407-2

1911–1912 lava

53.6

37.6

5.4

2.4

0.3

0.7

46.4

5.0

1,000

n.d.

S031407-5

1911–1912 lava

52.8

37.0

5.4

3.4

0.4

1.1

47.2

5.6

1,000

n.d.

S052304-5

1991–1995 lava

54.0

37.5

5.8

1.7

0.2

0.9

46.0

9.0

1,000

n.d.

S052304-4

2003 lava

58.7

32.3

5.8

1.9

0.2

1.1

41.3

5.2

1,037

n.d.

S030708-4

2004 lava

55.6

37.4

3.2

2.9

0.3

0.7

44.5

9.9

1,000

n.d.

S030708-2a

2007 bc

47.4

32.3

11.51

3.3

3.81

1.6

52.5

24.1

1,000

n.d.

S030708-6

2007 bc

56.2

36.4

4.1

2.2

0.1

1.0

43.8

20.3

1,000

n.d.

S030708-7

2007 lava

54.8

37.4

5.2

1.5

0.2

0.9

45.2

10.0

1,000

n.d.

S030808-2a

2007 bk

48.8

35.0

6.3

4.2

4.51

1.2

51.2

13.8

1,000

15–20

S030808-2b

2007 nj-bk

57.0

35.6

5.2

1.2

0.1

1.0

43.0

5.3

1,000

n.d.

S031108-1

2007 bc

58.7

30.7

6.1

3.4

0.3

0.8

41.3

12.9

1,000

20–25

Abundances given in volume percent, recalculated vesicle free

Abbreviations: Gm phases with grain sizes of <0.1 mm (mostly plagioclase, pigeonite, and oxides), Pl plagioclase, Ol olivine, Pig pigeonite, Aug augite, Ox oxide (titanomagnetite), Ph phenocryst total, Ves vesicles, Gl matrix glass, bc breadcrust block, bk juvenile block, nj-bk non-juvenile block

aThe thin sections of samples S030708-2 and S030808-2 contain olivine and augite megacrysts or glomerocrysts which account for the high abundances of these phases in the modes. Abundance of matrix glass (gl) estimated from scanning electron microbeam images for samples S030808-2 and S031108-1, not determined (n.d.) in other samples

Discussion

Eruption magnitudes

Consistent with eruptive size as inferred from ash column heights, SO2 emissions, and seismic energy release, the June 2008 eruption was the largest of the period 2004–2008, with a cumulative minimum bulk volume of flow deposits of approximately 6,000,000 m3 (Table 2). We note that Hadisantono et al. (2008) divided the pyroclastic flows into three classes depending on the size of boulders and attempted to take into count the depths of pre-eruption drainages and original thickness of deposits from depths of tree-trunk burial. They estimated a considerably larger cumulative volume of about 17,000,000 m3 for the June 2008 deposits, which we take as a maximum. The October 2007 eruption was the second largest, with a minimum flow-deposit volume of about 1,700,000 m3.

Tephra-fall deposits are not included in our volume estimates. Isopach mapping of the June 2008 tephra deposit (Hadisantono et al. 2008; Kristianto and Loeqman 2008) shows 3.5 cm of ash at 6 km and 1 mm at 15 km in the downwind direction and a volume of <1,000,000 m3. This suggests that tephra accounts for <15 % of the recent Vulcanian eruption products of Soputan, although initial reports suggest that a 2011 eruption produced a larger fraction of tephra. Overall, the volumes of deposits from the recent Soputan eruptions are modest, especially given the relatively high altitudes of some of the ash columns. The eruptive volume estimates (7–18 Mm3), heights of eruptive columns (4.6–13.7 km) and minimum magma volumes estimated from SO2 emissions (Table 7 and SO2 modeling below) indicate that the June 2008 eruption had a VEI of 3, and the October 2007 eruption had a VEI of 2 to 3; the December 2006 and August 2007 eruptions were VEI 2 events, and initial reports suggest a VEI 2 for the July 2011 eruption.
Table 7

Comparison of SO2 emissions calculated by the petrologic method to observed emissions from the OMI satellite (Fig. 8)

Eruption

Mass erupted melt (mt)

Max.SO2 by PM (kt)

OMI SO2 emission (kt)

December 2006

640

1.3

5

August 2007

660

1.3

8

October 2007a

840

1.7

7

June 2008

6,200

12.3

17

October 2008

940

1.9

11

All calculations assume: 50 vol.% melt fraction in the magma that is completely degassed during ascent and eruption, deposit density of 2.0, magma density of 2.6–2.8 g cm−3, and 1,500 ppm SO2 in melt

PM petrologic method

aThe October 2007 figures exclude the debris avalanche

Origin of Soputan magmas and controls on eruptive behavior

The restricted range in composition of the Soputan basalts stands in marked contrast to other stratovolcanoes in the North Sulawesi-Sangihe arc (Fig. 16), such as Lokon volcano, located 25 km to the north at the opposite side of the Tondono caldera (Fig. 1). The rare-earth elements (REE) are trace elements that behave similarly during magmatic processes and therefore provide a powerful tool for evaluating partial melting, fractionation and other processes affecting magmas during generation and transport. With respect to trace-element compositions, the Soputan basalts are distinct from all of the other young volcanic rocks we have analyzed from North Sulawesi (Fig. 17). In addition to having the lowest overall REE abundances, they are depleted in the light-rare earths (La through Eu), a characteristic of basalts derived from depleted mantle sources. As Eu+2 is taken up preferentially over Eu+3 and the other middle REE in plagioclase and excluded from pyroxene, the slight enrichment in Eu relative to Sm and Gd seen in Fig. 17 suggests either a minor amount of plagioclase accumulation, fractionation of pyroxene under reducing conditions or low degree partial melting of a reduced, plagioclase-free clinopyroxene-rich source. Consistent with the latter points, we note that it is in reduced, dry magmas where pigeonite typically occurs (e.g., Mid-Ocean Ridge Basalts (MORB), Skaergaard intrusion, Iceland tholeiites, and arc tholeiites).
Fig. 16

Total alkali-silica volcanic rock classification diagram of Le Bas et al. (1986) comparing compositions of basalts from Soputan to samples from other North Sulawesi volcanoes. Soputan samples (X symbols) are from Table 5. The + are Soputan ∼1900 ad lava flow and ad 1994–1995 dome samples of de Hoog, et al. (2001). Other points are from analyses reported in the ESM

Fig. 17

Chondrite-normalized diagram showing abundances of rare-earth elements in basalts from Soputan (black) compared with abundances from Lokon volcano (red), Klabat (green), and Tondono caldera (blue). Chondrite abundances used for normalization are from Sun and McDonough (1989)

Additional insights are gained by plotting abundances of a wider suite of trace elements, normalized and ordered in a manner that reflects their relative compatibility in mantle-derived melts, as well as their relative mobility in hydrous phases (Fig. 18). Such “spidergrams” provide a graphical means to evaluate a variety of geochemical processes that affect magmas, ranging from fluid fluxing of mantle source areas for basalts to crystal fractionation during transport of through the crust (Pearce 1984).
Fig. 18

Spidergram showing MORB-normalized abundances of trace and minor elements for samples from North Sulawesi, arranged in an order that reflects their relative compatibility in mantle derived melts and their mobility in hydrous fluids. MORB abundances used for normalization are from Sun and McDonough (1989). Colors as in Fig. 14. Abbreviations: Sop. Soputan volcano, Lok. Lokon volcano, Klab. Klabat volcano, Ton. Tondono caldera, bas basalt, and. andesite, rhyodac. rhyodacite, rhy. rhyolite, obs. rhyolite obsidian

Figure 18 shows prominent depletions of Nb and Ta characteristic of supra-subduction zone magmas worldwide and related to fluid-fluxing of the mantle wedge (Pearce 1984; Kelemen et al. 1993; Baier et al. 2007). Among the North Sulawesi volcanics, the Soputan basalts are distinguished by the lowest overall abundances of incompatible trace elements and, as noted above, by relatively flat REE patterns. They also have prominent positive Sr anomalies, which along with the positive Eu anomalies, indicate some combination of plagioclase accumulation, fractionation of pyroxene under reducing conditions or low degree partial melting of a reduced, plagioclase-free clinopyroxene-rich source, as noted previously. Depletions in the more compatible elements Ba, Sr, P, and Ti in the rhyodacites and rhyolites of the Tonodono caldera and the overall increase in incompatible trace elements with increasing SiO2 are characteristics of fractionation involving plagioclase, apatite, oxides and mafic silicate phases. However, the lack of more evolved compositions at Soputan and absence of LREE and other incompatible-element enrichment suggests that the Soputan basalts were derived from a separate source and did not interact with Tondono magmas.

The Soputan basalts have an average δ34S of +5.1 ± 1.4 ‰ (de Hoog et al. 2001). After considering a range of potential isotopic enrichment processes, de Hoog et al. (2001) concluded that the elevated δ34S in basalts from Soputan was the result of fluxing of the mantle wedge by fluids derived from subducted sediments. We therefore conclude that the parent magmas for Soputan were generated through a process involving fluid fluxing of the supra-subduction mantle wedge, a widely accepted model for generation of arc basalts (e.g., see Kimura and Yoshida (2006) and references therein). However, the relatively evolved ferrobasalt bulk composition of the Soputan basalt, with Mg/Mg + Fe+2 of 0.53 (Table 4) that yields calculated equilibrium olivines of Fo81 are clearly not compatible with mantle peridotites, which have Fo≥90 olivines. Consequently, we suggest that more magnesian HAB parent magmas for Soputan were generated in the fluid fluxed supra-subduction zone mantle wedge, but these magmas subsequently accumulated in a relatively large reservoir at intermediate to deep crustal levels where they evolved to low-Mg HAB’s by fractionation of olivine, augite, oxides and then plagioclase. This model is consistent with the experimentally constrained models proposed by Gust and Perfit (1987) for high-alumina ferrobasalts in the Aleutian arc and more generally by Sisson and Grove (1993) to explain the origin of the common low-Mg HAB’s seen in arcs worldwide. The uniformity of the Soputan composition also suggests that it may have been close to that of a reaction boundary liquid near the base of the crust.

The Soputan basalts are notable not only for their uniform composition and eruptive style, but also for their abundant phenocrysts (40–50 %), and number of saturating crystalline phases (plagioclase, olivine, pigeonite, augite, and magnetite). Consequently, a guide to physical conditions after the phenocryst assemblage formed but prior to eruption can be obtained by thermodynamic modeling of equilibrium pressure and water contents for magmas of the Soputan composition that are ∼45 % crystalline and contain plagioclase>olivine≈pigeonite>augite>magnetite (Fig. 19).
Fig. 19

Pressure-water content diagram for the average Soputan basalt crystallized to 55 wt.% liquid, contoured for mineral abundances, and showing the theoretical field of Soputan crystallization (grey area), as calculated using MELTS (Ghiorso and Sack 1995; Asimow and Ghiorso 1998). Approximate depth equivalents of pressures are shown on the right side of the diagram, assuming an average crustal density of 2.85 g cm−1. Cyan curves are plagioclase (15, 20, 25, and 30 wt.%), green curves are olivine (0 and 5 wt.%), black curves are augite (5, 10, and 15 wt.%), and red curves are low-Ca pyroxene (0 and 10 wt.%), which is pigeonite except within the area bounded by the grey dashed line where it is orthopyroxene. Black hexagons indicate water contents and pressures where phase assemblages and compositions were determined by MELTS with temperatures ranging from 943 to 1,165°C; MELTS derived temperatures for the hexagons in the gray field are 1,075 and 1,085°C. Not plotted are magnetite and exsolved H2O. Ferric-ferrous ratio of the bulk magma was set to that of a liquid at the fayalite–magnetite–quartz buffer curve +1 log unit. A fixed bulk composition was used except for varying H2O. Note the near disappearance of clinopyroxene as the abundance of plagioclase, and low-Ca pyroxene and olivine increase with decreasing pressure and water content

The thermodynamic modeling program MELTS predicts a narrow range of low pressure and water-poor (but not dry) conditions (shaded grey area in Fig. 19) where the phenocryst abundances observed in Soputan basalts are satisfied. Based on the MELTS results, the range in temperature for this field is approximately 1,040 to 1,100°C. For comparison, the equilibrium temperature calculated using QUILF (Andersen et al. 1993) for the average composition of the pigeonite phenocrysts and microlites and the sub-calcic augite rim on an augite xenocryst (Table 5) is 1,024 ± 27°C, overlapping with the MELTS results. We acknowledge that there are uncertainties in using MELTS for this type of analysis (e.g., limitations of the experimental database, inability to handle CO2, limits on distinction of orthopyroxene and pigeonite, and assumption of equilibrium despite petrographic evidence of modest mineral zoning in the Soputan samples). Nevertheless, MELTS provides a good general guide to phase stability and has the advantage over single-phase thermobarometers for these crystal-rich assemblages in considering the system as a whole. Based on these results, it appears that the Soputan basalts either staged at very shallow levels prior to eruption, or they ascended sufficiently slowly through the uppermost crust that they behaved as though they staged at shallow levels.

The MELTS results also suggest that Soputan magmas did not have high bulk H2O concentrations (if they had, plagioclase would be less abundant than augite in the equilibrium assemblage); however, we acknowledge that there are few calibrating experiments on basalts with 1–3 wt.% H2O, so this inference must be taken with caution. We note that back arc basin basalts (BABB) can reach ∼2.2 wt.% H2O, but, based on a survey of analyses in the Petrologic Database of the Ocean Floor (PetDB; http://www.petdb.org, Lehnert et al. 2000), virtually none have Al2O3 of ∼19 wt.%. Consequently, the Soputan basalt composition was probably wetter than common BABB, but probably not by much. A relatively dry magma (for an arc) is also consistent with the LREE depleted and low-K, low-Rb character of the Soputan magmas since all estimated slab fluids include appreciable LREE and alkalis.

In summary, the Nb and Ta depletions, uniform compositions, petrography and the modeling of phase equilibria suggest that the Soputan basalt originated as a relatively dry melt produced by melting of the arc-mantle wedge and then accumulated and fractionated in a large and uniform reservoir, likely near the base of the crust (Fig. 20). The basalt crystallized abundant phenocrysts during slow ascent or while staged at shallow crustal levels and it crystallized abundant microlites prior to eruption, increasing dramatically in viscosity. At 2 wt.% H2O, PMELTS (Ghiorso et al. 2002) indicates that the melt density at liquidus temperature would have ranged from 2.67 g cm−1 at 10 kbar to 2.56 g cm−1 at 2 kbar, i.e., close to the density of the crust and consistent with slow ascent. The two factors that are most influential in causing explosive eruptions are gas pressure and magma viscosity. In the discussion above, we outline an ascent history that accounts for the high viscosity of the Soputan basalt. Below we consider the question of gas pressure.
Fig. 20

Diagrammatic cross section of the Tondono caldera region showing the location of Soputan volcano outside the south caldera margin and contrasted to Lokon-Empung volcano, which is located just inside the caldera

First, it is evident from their distinct geochemisty that magmas and fluids derived from the nearby Tondono caldera played no role in Soputan’s eruptions. Ascending Soputan magmas bypassed the still-active caldera magma system, unlike the case for the active Lokon-Empung volcanic center located along the northwestern caldera margin (Fig. 20).

We test for the presence of a separate fluid phase in the Soputan magma by comparing expected SO2 emissions per unit of erupted basalt magma to measured emission data. We do so by assuming complete degassing of magma with a melt fraction of ∼50 vol.% and containing 1,500 ppm SO2 (the maximum elemental sulfur in Soputan melt inclusions reported by de Hoog et al. (2001) was 760 ppm, equivalent to 1,520 ppm as SO2). The resulting theoretical emissions by this petrologic method are less than the minimum SO2 emissions observed by the OMI satellite by factors of 1.4 to 6 (Table 7). Such underestimates by the petrologic method are common in arc magmas (Wallace 2001, 2004) and suggest the separation of a sulfur-rich fluid phase in the magma prior to eruption. A similar conclusion was reached by de Hoog et al. (2001), who suggested losses of more than 90 % of the sulfur to the fluid phase on the basis of comparison of S abundances in bulk rocks and melt inclusions and the common presence of vapor bubbles in the melt inclusions.

Taking the petrologic and SO2 data together, we suggest that the reason this basalt volcano erupts explosively is a consequence of the basalt being both highly viscous due to a high phenocryst and microlite content and due to its becoming gas saturated at a late stage. As a distributed separate fluid phase, the volatile components (mainly H2O, CO2, and SO2) increase buoyancy and pressure and enable eruption of the otherwise dense and viscous crystal-rich basalt. The phenocryst-rich character of the Soputan basalts, combined with near-surface degassing-driven microlite growth in the shallow conduit, resulted in the high effective viscosities required for lava dome growth and Vulcanian eruptions—features rare with basalts. Moreover, the record of eruptions at Soputan (Table 1) and the geologic history of alternating basalt lava flows, basalt tephra accumulations and basalt pyroclastic flows indicates that variation in the dynamics of magma transport (e.g., presence or absence of a dome plugging the top of the conduit and variable load pressure related to central versus parasitic vent localities) played roles in controlling gas release and explaining why this basaltic system formed a stratovolcano instead of a lava shield. Overall, Soputan enforces the lesson that the composition of magma alone does not always dictate eruptive style and explosivity—gas and crystal content may exert more important controls.

Explosive basalt eruptions are typically attributed to high gas contents at shallow levels at other volcanoes as well. For example, Cerro Negro volcano, Nicaragua has produced high ash columns in addition to lava flows. In 1992, an explosive eruption of Cerro Negro sent an eruptive column to 7-km altitude. The explosivity of the 1992 eruption was attributed to the retention of high CO2 in the magma to shallow crustal levels (Roggensack et al. 1997). In an even more extreme case, high gas contents are thought to be responsible for ancient plinian eruptions of basalt magma at Masaya caldera, Nicaragua, which is thought to have had eruptive columns to ∼50 km and eruption rates as high as 2 × 105 m3 s−1 (Williams 1983). In contrast for Soputan, high crystallinity appears to have exerted the dominant control on explosivity.

Low-Mg HAB are common in arcs worldwide and are now generally attributed to fluid-fluxed melting of the mantle wedge above subduction zones to produce hydrous (e.g., 6 wt.% H2O) high-Mg HAB parent magmas, which evolve to low-Mg HAB by crystal fractionation of olivine and clinopyroxene within the arc crust (Sisson and Grove 1993; Grove et al. 2003). Because of the high water contents, evolution of a separate hydrous fluid phase is inevitable, which lowers magma density and increases magmatic pressure. This is turn results in rapid ascent and common eruption of low-Mg HAB at arc volcanoes. Even though our modeling suggests that the Soputan basalt was relatively dry and ascended slowly, enhanced groundmass crystallization at shallow levels combined with shallow vesiculation and high viscosity was enough to drive its explosive eruptions.

Revised hazard assessment and long-term forecast

On the basis of the frequency and characteristics of its recent eruptions and new insights regarding its magmatic system, we update the hazard assessment of Suratman (1990) and we present a forecast for future eruptions at Soputan. The current open-vent structure and frequent eruptions indicate that Soputan will likely erupt again, perhaps repeatedly, during the next 10 years.2 As long as the volcano lacks a summit lava dome, eruptions are likely to be explosive, and most will be in the VEI 2–3 range, with a small chance of a much larger VEI 4 eruption. Even in the case of growth of a new summit dome, collapse of such a dome will produce dangerous Merapi-type pyroclastic flows. Because of the open-vent structure and aseismic magma ascent from depth, seismic precursors to eruptions in the next decade are likely to be of short duration, e.g., a few days to a week of elevated shallow seismicity preceding eruptions. It is significant to note that such short warning characterized other periods of eruptive activity in Soputan’s past. In describing early historical eruptions, Junghuhn (1853) noted that “earthquakes are usually observed 2–3 days before an eruption takes place” at Soputan.

The summit of the volcano and upper slopes above about 900 m elevation are within a National Park, with no permanent population. Consequently, the principal risk from small eruptions is to visitors to the park, to rock miners who frequent the upper slopes on the west and southwest and to coconut farmers on the south and west slopes. Larger and more explosive eruptions pose hazards from pyroclastic flows, lahars and tephra fall to villages mainly on the south and southwest (Tombatu subdistrict). Hazards to people in the region are greatest for villages located on the southern and southwestern flanks of the volcano. These include: Silian Satu and Silian Dua (9–10 km), Tombatu (10 km), Lobu and Ranoketangatas (12 km), Kuyanga and Mundung (10 km), Winorangian (9 km), and Molompar, Tolombukan, and Liwutung (10 km). Villages on the southeast (e.g., Ratahan) and northeast (e.g., Noogan, Raringis, Ampreng, Touure, Tonsewer, and Pinabetengan Kanonang) are also within 10 km, and the towns of Langowan, Tompaso, and Kawankoan are within 12 km of the summit. These areas may be subject to tephra fall, but they are protected from flows from all but the largest eruptions of Soputan by high elevation terrain east of the current summit. In addition to the ground hazards, ash clouds from Soputan’s explosive eruptions pose a hazard to regional and international air traffic.

A simplified event tree (Fig. 21) summarizes our semi-quantitiative assessment of probabilities for sizes and impacts of explosive eruptions for the next 5 to 10 years. We recognize the value of more quantitative treatments, which include statistical analysis of uncertainties (e.g., Marzocchi et al. 2004). However, limitations of the available dataset for Soputan preclude the effective use of such techniques. Instead, we estimate probabilities based on the authors’ knowledge of previous eruptive activity at Soputan and characteristics of the deposits. The frequent eruptions of the past decade, the current open-vent character and continuous steam and gas plumes from the volcano suggest a high (∼90 %) probability of additional eruptions in the next 10 years. VEI <3 eruptions are most likely with a cumulative probability of about 80 %, whereas, larger VEI 3 eruptions (like that of June 2008) are assigned a cumulative probability of about 15 %. We believe there is a small (∼5 %) chance of a much larger (VEI 4) eruption, which would produce higher altitude ash clouds and widespread ash fall. Pyroclastic flow impacts of VEI < 3 eruptions would likely be restricted to within 5 km of the summit on the north, west and south flanks. Areas more than a few kilometers to the east are protected by a topographic divide (Figs. 3 and 22). Eruption-induced lahars pose a modest threat under all scenarios. However, as shown separately in Fig. 22, there is a higher probability of lahars in the short term and in the case of extreme amounts of rainfall because of the presence of voluminous poorly consolidated and still un-vegetated flow and fall deposits from the recent eruptions on the upper flanks of the volcano. Such lahars could extend to >5 km from the summit, but as with pyroclastic flows, they would be restricted by topography to the north, west, and south flanks (Fig. 22). Such lahars could occur at any time there is unusually high rainfall, with or without eruptions.
Fig. 21

Event tree with forecast for explosive eruptions and impacts at Soputan volcano for the 10-year period 2011–2021. Percentages in the boxes are cumulative probabilities that take into account the preceding paths. Low explosivity (i.e., VEI 0-1 eruptions) are not included in this analysis

Fig. 22

LAHARZ model results for the lahar inundation of 100,000 (dark orange), 300,000 (light orange), 1,000,000 (light yellow), and 3,000,000 m3 (dark yellow). The hazard map scenario assumes a VEI 3 eruption that distributes 10,000,000 m3 material over a 6- to 10-km2 area as a 1-m thick pyroclastic deposit and/or tephra fall. Source areas above any one drainage could provide 100,000–3,000,000 m3 material for lahar entrainment during a heavy rainfall event. The tan polygon surrounding the Soputan edifice represents an energy cone hazard zone for pyroclastic flows, lava flows, projectiles, and tephra fall. The tan polygon also serves as an approximate hazard zone for pyroclastic flows, surges, and lava flows for VEI of ≤3 eruptions

We evaluate potential lahar inundation zones using the computer program LAHARZ (Iverson et al. 1998; Schilling 1998). Results for a series of different source volumes between 100,000 m3 and 3,000,000 m3 are shown in Fig. 22. This range overlaps volumes of eruptive products from 2006 to 2008. The eruptive products of this period consist mostly of pyroclastic flow deposits, which were concentrated in the headwaters of the drainages on the west and south flanks (Londola Pinamangkidan, Royong Lowian, and the valley leading to Ranoketangtua; Fig. 22). However, the tephra fall deposits from the 2006 to 2008 eruptions are relatively thin (the largest produced a deposit of 3.5 cm thick at a distance of 6 km downwind to the west, and 1 mm at 15 km; Hadisantono et al. 2008; Kristianto and Loeqman 2008). Consequently, we regard our LAHARZ results to represent a conservative model of inundation areas. However, we caution that larger inundation areas could result from much larger tephra-producing eruptions, analogous to those of 1991, if combined with intense rainfall. We also note that tephra from the July 2011 eruption was more widespread, with 5 cm thickness of tephra extending 13 km northeast from the volcano (Table 1). Although extensive lahars have not been reported, the 2011 eruption demonstrates the potential for larger tephra-producing eruptions at Soputan.

The impacts column of the event tree in Fig. 21 uses a distance of 10 km, selected on the basis of distance to the closest villages. Our estimates of impacts are weighted by the effects of the 2006–2008 eruptions, but we also take into account low-probability high-impact events, such as the 1991 eruption, which produced a large tephra cone high on the northern flank (Kartadinata et al. 1998). Low-explosivity eruptive activity, such as sustained low-rate dome growth and lava effusion are common at Soputan, but pose little hazard except in areas close to summit, or when such activity leads to dome collapse and surges. Due to their limited run-out and relatively low hazard at Soputan, lava flows are not explicitly included in the event tree, but are expected under any of the eruption scenarios within a 5-km radius of the summit.

An alternative scenario for the coming decades at Soputan assumes that the eruptions of 2007–2011 depleted the reservoir of gas-rich basalt magma, such that slower magma ascent will result in gas loss and groundmass crystallization and a new period of lava flow and dome growth will ensue. Because of the current topography of the summit crater, with breeches to the west and north, we anticipate this would result in episodic dome-collapse pyroclastic flows and lava flows, principally in these sectors. Over the long term, continued lava eruptions would eventually heal the breeches to the north and west enabling growth of a new lava dome at the summit and effectively plugging the conduit. However, as in 2007, this situation would eventually give way to additional summit collapse and conduit unroofing. The duration of the last cycle of explosive eruptions to summit lava dome growth to explosive eruptions took 16 years (1991–2007).

Footnotes

  1. 1.

    Population estimates were utilized from the LandScan 2007™ High Resolution global Population Data Set copyrighted by UT-Battelle, LLC, operator of Oak Ridge National Laboratory under Contract no. DE-AC05-00OR22725 with the US Department of Energy. The US Government has certain rights in this Data Set. Neither UT-Battle, LLC nor the US Department of Energy, nor any of their employees makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of the data set.

  2. 2.

    The forecast presented here was made prior to and validated by the eruption of 3 July 2011. We anticipate that it will remain valid for the next 5–10 years, although it should be re-evaluated with each eruption, as changes in the morphology of the volcano may affect the forecast.

Notes

Acknowledgments

The authors acknowledge the Indonesian Geological Agency and its Center for Volcanology and Geologic Hazard Mitigation (Surono, Director) and the USAID Office of Foreign Disaster Assistance for their support of our work in North Sulawesi. SAC acknowledges funding from NASA through grants NNX09AJ40G (Aura Validation), NNX10AG60G (Atmospheric Chemistry Modeling and Analysis Program), and NNX11AF42G (Aura Science Team). We benefitted substantially from constructive and informative reviews by William E. Scott, Thomas W. Sisson, Fidel Costa, and Rüdiger Escobar-Wolf. We especially thank reviewer Sisson for his assistance with MELTS modeling of the Soputan basalt composition and his insights into HAB petrogenesis.

Supplementary material

445_2012_620_MOESM1_ESM.docx (50 kb)
ESM 1(DOCX 49.7 kb)

References

  1. Andersen DH, Lindsley DH, Davidson PM (1993) QUILF: a PASCAL program to assess equilibria among Fe Mg Mn Ti oxides, pyrosenes, olivine and quartz. Comput Geosci 19:1333–1350CrossRefGoogle Scholar
  2. Asimow PD, Ghiorso MS (1998) Algorithmic modifications extending MELTS to calculate subsolidus phase relations. Am Mineral 83:1127–1131Google Scholar
  3. Baier J, Audetat A, Keppler H (2007) The origin of the negative niobium tantalum anomaly in subduction zone magmas. Earth Planet Sci Lett 267:290–300. doi:10.1016/j.epsl.2007.11.032 CrossRefGoogle Scholar
  4. Bittenbinder AN, Bogaert BM, Johnson C, Malone SD (1994) Earthworm; a modular distributed processing approach to seismic network processing. Seism Res Lett 65:51Google Scholar
  5. CVGHM (2009) Volcano Database, Ambang, Awu, Colo, Lokon, Karangetang, Mahawu, Ruang, Soputan and Tangkoko, North Sulawesi. Indonesian Center for Volcanology and Geologic Hazard Mitigation (CVGHM) database and report, English translation, Feb 2009: 121 p. Available at http://www.vsi.esdm.go.id
  6. de Hoog JCM, Taylor BE, van Bergen MJ (2001) Sulfur isotope systematics of basaltic lavas from Indonesia: implications for the sulfur cycle in subduction zones. Earth Planet Sci Lett 189:237–252CrossRefGoogle Scholar
  7. Dobson JE, Bright EA, Coleman PR, Durfee RC, Worley BA (2000) LandScan: a global population database for estimating populations at risk. Photogramm Eng Remote Sens 66:849–857Google Scholar
  8. Endo ET, Murray T (1991) Real-time seismic amplitude measurement (RSAM); a volcano monitoring tool. Bull Volcanol 53:533–545CrossRefGoogle Scholar
  9. Ghiorso MS, Sack RO (1995) Chemical mass transfer in magmatic processes. IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid-solid equilibria in magmatic systems at elevated temperatures and pressures. Cont Mineral Petrol 119:197–212CrossRefGoogle Scholar
  10. Ghiorso MS, Hirschmann MM, Reiners PW, Kress VC III (2002) The pMELTS: an revision of MELTS aimed at improving calculation of phase relations and major element partitioning involved in partial melting of the mantle at pressures up to 3 GPa. Geochem Geophys Geosyst 3(5). doi:10.1029/2001GC000217
  11. Grove TL, Elkins-Tanton LT, Parman SW, Chatterjee N, Müntener O, Gaetani GA (2003) Fractional crystallization and mantle-melting controls on calc-alkaline differentiation trends. Cont Mineral Petrol 145:515–533. doi:10.1007/s00410-003-0448-z CrossRefGoogle Scholar
  12. Gust DA, Perfit MR (1987) Phase relations of a high-Mg basalt from the Aleutian Island Arc: implications for primary island arc basalts and high-Al basalts. Cont Mineral Petrol 97:7–18CrossRefGoogle Scholar
  13. Hadisantono RD, Mulyana R, Purwoto, Ridwan I (2008) Gejala perubahan pada erupsi G. Soputan, 6 Juni 2008. Vulkanologi dan Bencana Geologi 3:6–12Google Scholar
  14. Hamilton W (1979) Tectonics of the Indonesian Region. USGS Prof. Paper 1078Google Scholar
  15. Iverson RM, Schilling SP, Vallance JW (1998) Objective delineation of lahar-inundation hazard zones. Bull Geol Soc Am 110:972–984CrossRefGoogle Scholar
  16. Junghuhn F (1853) Java, Deszelfs Gedaante, Bekleeding en Inwendige Structuur, pt. 2. P.N. Van Kampen, AmsterdamGoogle Scholar
  17. Kartadinata MN, Irawan W, Solihin A, Mulyana AR (1998) Geologic Map of Soputan volcano, North Sulawesi. Volc Surv Indonesia Geol MapGoogle Scholar
  18. Kelemen PB, Shimizu N, Dunn JT (1993) Relative depletion of niobium in some arc magmas and the continental crust: partitioning of K, Nb, La and Ce during melt/rock reaction in the upper mantle. Earth Planet Sci Lett 120:111–133CrossRefGoogle Scholar
  19. Kimura J-I, Yoshida T (2006) Contributions of slab fluid, mantle wedge and crust to the origin of Quaternary lavas in the NE Japan arc. J Petrol 47:2185–2232. doi:10.1093/petrology/egl041 CrossRefGoogle Scholar
  20. Kristianto, Loeqman (2008) Erupsi G. Soputan, 6 Oktober 2008. Bull Vulkanologi dan Bencana Geologi 3:1–5Google Scholar
  21. Le Bas MJ, Le Maitre RW, Streckeisen A, Zanettin B, IUGS Subcommission on the Systematics of Igneous Rocks (1986) A chemical classification of volcanic rocks based on the total alkali-silica diagram. J Petrol 27:745–750Google Scholar
  22. Lécuyer F (1990) Relation entre le volcanisme et la tectonique active dans la region de Tondano au Nord de Sulawesi (Indonesie). Ph.D. dissertation no. 183, Universite Blaise Pascal, 159 pGoogle Scholar
  23. Lehnert K, Su Y, Langmuir C, Sarbas B, Nohl U (2000) A global geochemical database structure for rocks. Geochem Geophys Geosyst 1. doi:10.1029/1999GC000026
  24. Marzocchi W, Sandri L, Gasparini P, Newhall C, Boschi E (2004) Quantifying probabilities of volcanic events: the example of volcanic hazard at Mount Vesuvius. J Geophys Res 109:B11201. doi:10.1029/2004JB003155, 18 pCrossRefGoogle Scholar
  25. Morrissey MM, Mastin LG (2000) Vulcanian eruptions. In: Sigurdsson H, Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia of volcanoes. Academic, San Diego, pp 463–475Google Scholar
  26. Pearce JA (1984) Role of the sub-continental lithosphere in magma genesis at active continental margins. In: Hawkesworth CJ, Norry M (eds) Continental basalts and mantle xenoliths. Shiva, Nantwich, pp 230–249Google Scholar
  27. Peccerillo A, Taylor SR (1976) Geochemistry of Eocene calc-alkaline volcanic rocks from the Kastamonu area, northern Turkey. Cont Mineral Petrol 58:63–81CrossRefGoogle Scholar
  28. Roggensack K, Hervig RL, McKnight SB, Willliams SN (1997) Explosive basaltic volcanism from Cerro Negro volcano: Influence of volatiles on eruptive style. Science 277:1639–1642CrossRefGoogle Scholar
  29. Santosa I (2008) Kegiatan Sosialisasi bahaya G. Soputan, Sulawesi Utara, Juni 2008. Bull Vulkanologi dan Bencana Geologi 3:25–28Google Scholar
  30. Schilling SP (1998) LAHARZ: GIS programs for automated mapping of lahar-innundation hazard zones. U.S. Geological Survey Open-File Report 98-638Google Scholar
  31. Siebert L, Simkin T, Kimberly P (2010) Volcanoes of the World. University of California Press, BerkeleyGoogle Scholar
  32. Sisson TW, Grove TL (1993) Temperatures and H2O contents of low-MgO high-alumina basalts. Cont Mineral Petrol 113:167–184CrossRefGoogle Scholar
  33. Suantika G (2008) Evalusi kegempaan punggungan maya tahun 1999–2008 kaitannya terhadap kegiatan vulkanik Gunungapi Soputan di Sulawesi Utara. Bull Vulkanologi dan Bencana Geologi 3:29–36Google Scholar
  34. Sun SS, McDonough WF (1989) Chemical and isotopic systematics of oceanic basalts; implications for mantle composition and processes. In: Saunders AD and Norry MJ (eds) Magmatism in the Ocean Basins 42. Geol. Soc. Lond, pp. 313–345Google Scholar
  35. Suratman (1990) G. Soputan—hazard map and assessment report. Volc Surv Indonesia Bull 146Google Scholar
  36. Wallace P (2001) Volcanic SO2 emissions and the abundance and distribution of exsolved gas in magma bodies. J Volcanol Geotherm Res 108:85–106CrossRefGoogle Scholar
  37. Wallace P (2004) Volatiles in subduction zone magmas: concentrations and fluxes based on melt inclusion and volcanic gas data. J Volcanol Geotherm Res 140:217–240CrossRefGoogle Scholar
  38. Williams SN (1983) Plinian airfall deposits of basaltic composition. Geology 11:211–214CrossRefGoogle Scholar
  39. Yang K, Krotkov NA, Krueger AJ, Carn SA, Bhartia PK, Levelt PF (2007) Retrieval of large volcanic SO2 columns from the Aura Ozone Monitoring Instrument (OMI): comparison and limitations. J Geophys Res 112:D24S43. doi:10.1029/2007JD008825 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag (outside the USA) 2012

Authors and Affiliations

  • Kushendratno
    • 1
  • John S. Pallister
    • 2
  • Kristianto
    • 1
  • Farid Ruskanda Bina
    • 1
  • Wendy McCausland
    • 2
  • Simon Carn
    • 3
  • Nia Haerani
    • 1
  • Julia Griswold
    • 2
  • Ron Keeler
    • 4
  1. 1.Geological Agency, Indonesian Ministry of Petroleum and MineralsCenter for Volcanology and Geologic Hazard Mitigation (CVGHM)BandungIndonesia
  2. 2.U.S. Geological Survey and USAID Office of Foreign Disaster Assistance, Volcano Disaster Assistance Program (VDAP), Cascades Volcano ObservatoryVancouverUSA
  3. 3.Michigan Technological UniversityHoughtonUSA
  4. 4.U.S. Geological SurveyRestonUSA

Personalised recommendations