Elsevier

Engineering Geology

Volume 287, 20 June 2021, 106104
Engineering Geology

Geological history controlling the debris avalanches of pyroclastic fall deposits induced by the 2009 Padang earthquake, Indonesia: The sequential influences of pumice fall, weathering, and slope undercut

https://doi.org/10.1016/j.enggeo.2021.106104Get rights and content

Highlights

  • The 2009 Padang earthquake (Mw 7.6) induced debris avalanches of pumice fall deposits.

  • A total of 159 landslides were identified in the areas with VIII seismic intensity in MMI.

  • The pumice grains were strongly weathered at the base of the beds to form halloysite.

  • The sliding surfaces were made in the weak halloysite-rich zone.

Abstract

Rapid-moving landslides occurred in many locations during the 2009 Padang earthquake, Sumatra. We interpreted satellite images, performed field surveys, in situ dynamic cone penetration tests, and laboratory analyses for mineralogy and physical and mechanical properties. We found that landslides occurred at 159 locations in an area of 64 km2, and these areas had pumice fall deposits overlying paleosols which were heavily weathered debris flow deposits. These landslides in the investigated area occurred in areas with pumice fall deposits thicker than 350 cm; this depth was probably vital in inducing mechanical instability. The areas had sliding surfaces at the base of the pumice fall deposits, where pumice grains were mixed with the underlying paleosol and had been heavily weathered into halloysite-rich clayey materials by interaction with the percolating water from the ground surface. After the deposition of this pumice fall with mantle bedding, the beds were undercut by subsequent river erosion that loosen their downslope lateral support. This geologic history of pumice fall, weathering, and undercutting is not confined to the affected area but it is common to many volcanic areas, creating the opportunity to predict areas susceptible to earthquake-induced catastrophic debris avalanches.

Introduction

Among the most hazardous types of landslides in volcanic areas are debris avalanches and debris flows of pyroclastic fall deposits induced by rainstorms and earthquakes (Moon and Churchman, 2016). These landslides move very rapidly and have a long runout distance (Hungr et al., 2014). Rain-induced landslides of pyroclastic fall deposits typically occur in Italy (Forte et al., 2019; Santo et al., 2018; Vingiani et al., 2015) and Japan (Towhata et al., 2021; Wang et al., 2019). Earthquake-induced landslides of pyroclastic fall deposits, which are the focus of this paper, have occurred in Japan (Chigira, 1982; Chigira, 2014; Chigira et al., 2014, Chigira et al., 2019; Ishihara and Hsu, 1986; Kawamura et al., 2019; Sugimoto et al., 2012; Tanaka, 1985; Yoshida and Chigira, 2012), El Salvador (Crosta et al., 2005; Evans and Bent, 2004; Ishihara and Hsu, 1986; Jibson et al., 2004) and Indonesia (Wang et al., 2011). Because pyroclastic fall deposits cover wide areas near their origin and experience the same weathering history after deposition, a single rainstorm or an earthquake could trigger clusters of such landslides and devastate wide areas. Chigira and Suzuki (2016) summarized the features of earthquake-induced landslides of pyroclastic fall deposits and pointed out the following: they occur on gentle slopes, which are cataclinal or dip slopes because of the mantle bedding nature of pyroclastic fall deposits; the susceptible slopes are generally undercut; and the sliding surfaces are made of a halloysite-rich layer. Moon and Churchman (2016) also summarized the importance of halloysite in earthquake-induced and rain-induced landslides in volcanic areas. In volcanic regions, however, almost all areas are covered by pyroclastics, and halloysite is a common weathering product of volcanic materials (Aomine, 1966; Lowe, 1986), the research of this nature would contribute to hazard mapping methodology for the earthquake-induced landslides of pyroclastic fall deposits.

The 2009 Padang earthquake occurred 60 km west-northwest of Padang, Sumatra Island, Indonesia, at 5:16 pm on 30 September (local time), with a magnitude of Mw 7.6 (McCloskey et al., 2010) (Fig. 1). The seismic intensity recorded in MMI was VIII in Pariaman (30 km ENE of the epicenter) and Bukittingi (73 km NE) and VII in Padang (60 km ESE) (USGS website, 23 December 2013 access). According to an analysis by the USGS (as cited in EERI, 2009), one strong ground motion record was obtained in Padang, involving approximately 20 s of strong shaking with a peak ground acceleration of 0.3 g. The hypocenter was 81 km deep along the boundary between the Australian and Sunda plates. Sumatra Island is on the Sunda plate, under which the Australian plate is obliquely subducting NNE (McCloskey et al., 2010). The dip slip is accommodated on the subduction interface, and the strike-slip component is accommodated largely by the Sumatran fault. The last historical earthquake that occurred at the plate boundary off of Padang was a 8.7 Mw event in 1797, before the 2009 Padang earthquake occurred in the oceanic slab of the Australian plate (McCloskey et al., 2010). The Sumatran fault, which trends NW-SE through Bukittingi, has caused many earthquakes. Recent earthquakes had Mw 6.4 and 6.3 in 2007, with epicenters near Lake Singkarak (EERI, 2007).

The 2009 Padang earthquake did not cause a tsunami but affected an area with a population of approximately 1.2 million in West Sumatra province of Sumatra which Padang is the capital city; the tremor caused severe damages to buildings in low-lying areas in the coastal zone of downtown Padang and Pariaman, and induced many landslides in hilly areas in Tandikat. The induced landslides were dominated by debris avalanches of pyroclastic fall deposits (Wang et al., 2011): Numerous landslides occurred on valley-side slopes and completely buried valley-bottom villages. There were 1195 total fatalities from this earthquake (EERI, 2009), and more than half of the fatalities were from landslides (Bothara et al., 2010). The landslides destroyed three villages, and there were 130 fatalities in Tandikat village alone (Fig. 2A), which was documented on a monument in the village. The landslides were too fast to allow people to escape (Daily News 5 Oct, 2009). Wang et al. (2011) conducted a field survey and performed geotechnical testing of pumice they found in landslide sites, but the landslide distribution and geological structures and basic causes of the landslides were not clarified. In addition to the debris avalanches described above, the 2009 Padang earthquake induced rockfalls on the walls of the Maninjau caldera, but they are beyond the scope of this paper and do not affect the conclusions of this paper.

Padang, which is 50 km southeast of the 2009 landslide area of Tandikat, receives 4010 mm of annual precipitation, and the average temperature is 27 °C, with a dry season from April to September and a rainy season from October to March (Japan Meteorological Agency, 2021, accessed 6 June 2019). Wang et al. (2011) reported an eyewitness account that described the rain started at 2:00 pm on 30 September and continuing until the earthquake at 5:16 pm and inferred that the rainfall intensity was 20–40 mm/h, which suggests that the total amount was potentially 60 to 130 mm just before the earthquake.

Section snippets

Geologic setting of the study area

The study area was the Tandikat area of Pariaman city located in the southwestern coast of Sumatra Island (Fig. 1). The Maninjau caldera, Singgalang volcano (2877 m), and Tandikat volcano (2438 m) are to the north of this area. The Maninjau caldera is visible in Fig. 1 as a water filled lake known as Lake Maninjau. The Magung River flows from the north to the south between the Maninjau caldera and Singgalang volcano toward the Indian Sea. The downstream portion of the Magung River flows in an

Method

We made a topographic map of an area of 110 km by 110 km with a contour interval of 10 m from ASTER GDEM (ASTER Global Digital Elevation Model) Version 2. We printed it on a scale of 1:25,000 and used it in the field.

We used a SPOT-5 image of France to inventory the landslides induced by the earthquake. The image had an ID of 52,683,510,910,130,337,202 J with 10 m resolution and was taken on 13 October 2009. We observed the false color image and identified landslides, of which scars were easily

Topography

We found 4 groups of river terraces along the midstream and downstream catchments of the Magung River by the interpretation of the ALOS PRISM images; these terraces were named Lh, Lm, Ll1, and Ll2, in order of highest to lowest elevation (Fig. 3, Fig. 4). These terraces and their deposits were not shown in the geologic maps of Pastowo et al. (1996) and Tjia and Muhammad (2008). The Lh surface is widely developed (Fig. 2B) and has been incised and decreased in its area, but its deposits are

Weathering and halloysite formation

All surveyed landslides induced by the 2009 Padang earthquake had sliding surfaces in the mixed zone of pumice and the underlying paleosol, which had been strongly weathered and became clayey materials with halloysite, suggesting that the formation and weathering of these mixed materials were the principal basic causes of the landslides. The mixed zones at the bases of the pumice fall deposits have not been reported in previous papers, even though sometimes they were encountered in the field.

Conclusions

The 2009 Padang earthquake, which occurred on 30 September with Mw 7.6, induced numerous landslides in Sumatra, Indonesia. We performed satellite image interpretation, field investigation, dynamic cone penetration tests in the field, laboratory tests and slope stability analysis. We identified 159 landslides in 64 km2 on satellite images and visited 19 landslides in the field. All visited landslides were pumice fall deposits showing isopachs that centered on the Tandikat volcano. The landslides

Author contribution

Maho Nakano Hosobuchi: Field survey, data analysis, original draft

Masahiro Chigira: Project administration, writing – review and editing, funding aquisition

Chounsian Lim: Field survey

Ibrahim Komoo: Field survey, funding acquisition

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was supported by the SEADPRI of Universiti Kebangsaan Malaysia, Malaysia; the Disaster Prevention Research Institute, Kyoto University, and JSPS KAKENHI, grant number 26282102 (PI Chigira). Discussion in the field with the late Tajul Anuar Jamalludin of Universiti Kebangsaan Malaysia, and Suri Hidayati and Sumaryono Gajam of the Geological Agency of Indonesia, Ministry for Energy and Mineral Resources, was very helpful.

References (47)

  • M. Chigira

    Geological and geomorphological features of deep-seated catastrophic landslides in tectonically active regions of Asia and implications for hazard mapping

    Episodes

    (2014)
  • M. Chigira et al.

    Prediction of earthquake-induced landslides of pyroclastic fall deposits

  • M. Chigira et al.

    Soil-slide avalanches of pyroclastic fall deposits induced by the 2011 off the pacific coast of Tohoku earthquake

  • M. Chigira et al.

    Landslides of pyroclastic fall deposits induced by the 2018 eastern Iburi earthquake with special reference to the weathering of pyroclastics

    Disaster Prevent. Res. Inst. Ann.

    (2019)
  • G.J. Churchman et al.

    Unique but diverse: some observations on the formation, structure and morphology of halloysite

    Clay Miner.

    (2016)
  • EERI Special Earthquake Report

    Western Sumatra Earthquakes of March 6, 2007

    (2007)
  • EERI Special Earthquake Report

    The Mw 7 6 Western Sumatra Earthquake of September 30, 2009

    (2009)
  • S.G. Evans et al.

    The Las Colinas landslide, Santa Tecla: A highly destructive flowslide triggered by the January 13, 2001, El Salvador earthquake

  • G. Forte et al.

    Triggering and predisposing factors for flow-like landslides in pyroclastic soils: the case study of the Lattari Mts. (southern Italy)

    Eng. Geol.

    (2019)
  • O. Hungr et al.

    The Varnes classification of landslide types, an update

    Landslides

    (2014)
  • A. Inoue

    Shape transformation of halloysite particles

    Clay Science

    (2011)
  • K. Ishihara et al.

    Considerations for landslides in natural slopes triggered by earhquakes

    Proc. Japan Soc. Civ. Eng.

    (1986)
  • Japan Meteorological Agency

    ClimatView

    (2021)
  • Cited by (9)

    • Compressibility and permeability of weathered, sensitive volcanic ash (tephra) deposits at the Omokoroa flow slide, New Zealand

      2022, Engineering Geology
      Citation Excerpt :

      The silica-rich tephras at the Omokoroa flow slide have been highly weathered into a residual soil with high concentrations of the secondary clay mineral halloysite (Kluger et al., 2017). Halloysite is known to dominate slip surfaces in Indonesia, Japan, Hong Kong, and New Zealand (Kirk et al., 1997; Moon et al., 2015b; Kameda et al., 2019; Kluger et al., 2020; Hosobuchi et al., 2021). The sensitivity in the sediment column at the Omokoroa flow slide varies between 5 (i.e., low sensitivity) and 55 (i.e., high sensitivity) and was found to correlate with the morphology of halloysite (Kluger et al., 2017) The highest sensitivities were determined for tephra layers dominated by spheroidal halloysite, whereas tubular halloysite was generally found to form soils with low to moderate sensitivity.

    • The origins of transparent and non-transparent white pumice: A case study of the 52 ka Maninjau caldera-forming eruption, Indonesia

      2022, Journal of Volcanology and Geothermal Research
      Citation Excerpt :

      To identify and reconstruct the vertical and lateral variations of the 52 ka eruption deposits, we established 22 observation locations.These locations all encompassed the northern-, western-, southern-, and eastern-deposits (Fig. 1b). However, since LOCs 5, 7, and 9 in the eastern deposit are not well-preserved (i.e., not fresh and have been oxidized intensively, resulting in poorly representative samples for further analyses), and LOCs 8, 19, 20, and 21 corresponds to Malalak tephra (sourced from Tandikat volcano; Hosobuchi et al., 2021), we did not include these locations in the analyses of this study. We observed that the Maninjau caldera-forming eruption deposits consist of a thick succession (up to 250 m) of ignimbrite without any precursory fall deposits (Plinian-free eruption).

    View all citing articles on Scopus
    1

    2–12-6 Takaba, Hitachinaka-shi, Ibaraki-ken, 312–0062, Japan.

    2

    Fukada Geological Institute. 2–13-1 Honkomagome, Bunkyo-ku, Tokyo 113–0021.

    View full text