Investigation of the initiation mechanism of an earthquake- induced landslide during rainfall: a case study of the Tandikat landslide, West Sumatra, Indonesia
© Faris and Fawu; licensee Springer 2014
Received: 21 August 2014
Accepted: 27 August 2014
Published: 3 October 2014
A large earthquake struck Padang Province, West Sumatra, Indonesia, at 17:16 on September 30, 2009. The earthquake had a moment magnitude of Mw 7.6, and triggered landslides in Tandikat, Padang Pariaman Regency. The landslides occurred during rainfall, and originated on mountains mantled with loose pumice, taking many lives. The unfortunate combination of intensive rainfall and strong earthquake probably decreased slope stability. This study seeks to examine the initiation mechanism of earthquake- induced landslides during rainfall, and to develop a new approach to predict pore pressure increase by assuming reciprocal relationships between strain, stiffness, and pore pressure.
In order to assess slope stability, the concept of stiffness degradation was used to predict pore pressure increase due to earthquake. This was achieved by developing empirical formulation based on cyclic triaxial test results. A new procedure based on the “rigid block on quasi plastic layer” assumption was developed to assess slope stability of earthquake-induced landslides during heavy rainfall. Results from cyclic triaxial test experiments showed that effective confining pressure and initial shear stress had considerable influence on increase in pore pressure. Slope stability analysis using actual earthquake acceleration suggest that landslide occurred due to pore pressure build up and the factor of safety decreased rapidly before earthquake acceleration reached its peak.
The results emphasize the high risk of catastrophic earthquake-triggered landslides in tropical regions with high rainfall. It also suggest that landslide with similar mechanism of pore pressure increase are likely to occur on saturated sliding zones during smaller earthquakes.
Studies of initiation and post-failure mechanisms are conducted using different laboratory test methods. Studies of initiation mechanisms commonly use laboratory shearing tests considering “limited displacement” conditions (i.e. triaxial, hollow cylinder torsional shear, direct shear, and simple shear tests). Studies of post-failure mechanisms typically consider “large displacement” tests using ring shear tests.
The post-failure mechanism of volcanic deposits has been examined in several studies. A study of a long run-out landslide in pyroclastic strata by Wang and Sassa () used undrained ring shear apparatus to confirm grain crushing mechanism during shearing. Wang et al. () also evaluated the post-failure mechanism of long run-out pumice material from the Tandikat landslide, using the same apparatus.
Several studies have examined the initiation mechanism of landslides in common volcanic soils, and the role of pore pressure build-up. Hyodd et al. () examined the liquefaction characteristics of several crushable volcanic deposits, and Suzuki and Yamamoto () emphasized the liquefaction characteristics of the Shirashu pyroclastic deposit in Japan, using cyclic triaxial tests on disturbed and undisturbed samples. However, specific research on the dynamic properties of pumice sand and its relation to pore pressure generation is rare.
Researches based on pore pressure models mainly use regular sand for laboratory tests, with very limited effort focused on the dynamic behaviour of volcanic sand, especially pumice sand. Seed et al. () and Lee and Albaisa () used clean sand to study liquefaction, and to develop a pore pressure model based on the number of cyclic loads. Yamazaki et al. (), Sugano and Yanagisawa () and Jafarian et al. () used Toyoura silica sand to derived pore pressure models based on the strain energy concept, using a variety of laboratory tests. Work on the behaviour of pumice during dynamic load has been reported by Marks et al. () and Orense and Pender (). These authors studied the liquefaction characteristics and resistance of crushable pumice soils from North Island, New Zealand, based on undrained cyclic triaxial tests and field test data. They confirmed that pore pressure built up during shearing. Nevertheless, an empirical model for pore pressure generation in such material has not yet been developed.
This study examines the initial mechanism of earthquake-induced landslides during rainfall, by development of a pore pressure model using local pumice sand, and the use of a cyclic triaxial apparatus. The pore pressure model proposed is based on assumption of a reciprocal relationship between strain, stiffness, and pore pressure. This model was then incorporated with groundwater simulation and slope stability analysis to encompass the problem of earthquake-induced landslides during rainfall. To fulfil this purpose, field investigations and laboratory tests using a stress-controlled cyclic triaxial apparatus were conducted to examine the physical behaviour of pumice sand from the Tandikat landslide district.
2.1 Geological setting
The landslides are located in a mountainous area around two volcanoes (Mt. Tandikat and Mt. Singgalang), and are extensively distributed on steep slopes inclined at ~30 to 50 degrees. The slopes are mainly mantled by unconsolidated volcanic deposits that were derived from the nearby mountains. This topographical condition is considered to be an important contributory factor for landslides in Tandikat.
2.2 Seismicity and meteorology
Several high magnitude earthquakes have been recorded in the subduction zone along the west coast of Sumatra in the last few decades. The Mw 9.0 Aceh earthquake of December 26, 2004 caused the catastrophic Indian Ocean tsunami. The Nias earthquake of March 28, 2005, and the South Sumatra earthquake of September 12, 2007, had magnitudes of Mw 8.7 and Mw 8.4, respectively. The most recent large earthquake was the September 30, 2009 Padang earthquake, which had a moment magnitude of 7.6. The epicenter was located offshore, WNW of Padang City, and the hypocenter was located at a depth of 80 km, within the oceanic slab of the Indo-Australian plate. This earthquake has been interpreted as an indication of a higher possibility of an imminent mega-earthquake in this region (Aydan ).
3.1 Field investigation
where ri is the internal radius of the pipe.
Input parameters used in the Green-Ampt infiltration model
Infiltration model parameters
Saturated hydraulic conductivity, ks (mm/h)
Suction head, ψ (mm)
Initial volumetric water content, Δwi
Volumetric water content at saturation, Δws
Slope angle, θ (o)
3.2 Infiltration model
where f(t) = potential infiltration rate at time t, F(t) = cumulative infiltration at time t, ψ = suction head at the wetting front, Δw = volumetric water content deficit, and θ = slope angle.
3.3 Groundwater modelling
Many researches have been conducted to develop numerical model to predict groundwater in unconfined aquifer. Among many methods, Boussinesq equation is most often used to estimate groundwater (Bansal and Das ). The performance of this method is reliable to predict experimental soil flume test (Steenhuis et al. ; Sloan and Moore ). This method generally formulated as a parabolic nonlinear equation, thus linearization process is used to derive analytical solution. Alternatively, finite difference numerical model can be used to utilize the aforesaid equation (Bansal ).
3.4 Laboratory tests
Pumice sand properties obtained from in-situ measurement and laboratory tests
Pumice sand properties
Bulk density (g/cm3)
Dry density (g/cm3)
Relative density (%)
Void ratio, e
Uniformity coefficient (Cu)
Coefficient of curvature (Cc)
Saturated hydraulic conductivity, ks (m/s)
Suction head, ψ (mm)
Volumetric water content deficit, Δw
Internal friction angle, ϕ′ (o)
Full saturation of the specimens was confirmed if Skempton’s B value was greater than 0.95. After full saturation was reached, the specimens were loaded axially into a triaxial cell with differing confining pressures of 20 kPa and 50 kPa, with displacement velocity of 0.7 mm/minute. The undrained strength parameters obtained are summarized in Table 2.
Summary of CTX tests conducted during this study
The test procedure was performed in a manner to simulate the stress condition in the field. At first, the specimens were consolidated with specified confining pressure after full saturated condition has been attained. As an approximation of the irregular motion of earthquake loading, the amplitude of cyclic sinusoidal deviatoric axial stress was taken to be 65% of the maximum magnitude of shear stress induced by the actual earthquake, as proposed by Seed et al. (). The cyclic sinusoidal axial stresses were then applied to the specimens at a rate of 1 Hz until the ultimate failure state was achieved.
3.5 Pore pressure model
where μ is the Poisson’s ratio equal to 0.5 for undrained conditions, and Es, Gs and γs are the secant Young’s modulus, the secant shear modulus, and the shear strain amplitude, respectively. The initial shear modulus, G0, can be referred to Eq. 14 by changing parameter Es to E0.
4Results and discussion
4.1 Result of static triaxial test
Static triaxial tests were conducted under 20 kPa and 50 kPa effective confining pressure to attain the shear strength of the pumice sand, and to understand its basic physical behaviour under stresses. The stress path (Figure 11) shows a large sector of contractive curve, implying the effect of pore pressure increase. It also indicates a low elastic threshold (ET), suggesting that the soil structure contracts easily under low shear stress. As the effective confining pressure decreases, the stress path exhibits dilation behaviour as it approaches the phase transformation line (PTL). This tendency of dilation after passing PTL suggests cyclic mobility behaviour when cyclic loading is applied (Ishihara ). From the test, the internal friction angle (φ′) is equal to 39.0°.
4.2 Typical result of cyclic triaxial test
where Δu is the excess pore pressure, σd is the deviatoric axial stress, and is the effective confining pressure.
4.3 Effect of effective confining pressure on reference cumulative shear strain
4.4 Effect of initial shear stress on reference cumulative shear strain
4.5 Effect of effective confining pressure on stiffness degradation
4.6 Effect of initial shear stress on stiffness degradation
The plot secant shear modulus, Gs vs cumulative shear strain, γt corresponding to initial shear stress ratio, K of 45 kPa effective confining pressure are shown in Figure 23. The effect of initial shear stress, K, on shear stiffness degradation is not clearly observable. The plot shows the same tendency, and insignificant differences of initial shear modulus, G0, at every different values of K. Therefore, the influence of initial shear stress on stiffness degradation is negligible.
5.1 Pore pressure model fitting
In this study, the non-linearity of stress–strain were also implemented by correlating cumulative shear strain, γt and the secant shear modulus, Gs. The influence of effective confining pressure and initial shear stress on shear stiffness degradation was also examined through data obtained from the CTX tests.
The best fit curve showing the relationship of Gs-γt is expressed by Eq. 19. The equation implies that shear strain (γ < 10−5) near zero will transform the equation to G ≈ G0, and the increased cumulative shear strain will gradually decrease the secant modulus.
5.2 Rigid block on quasi-plastic layer and simulation procedure
where γ w is the unit weight of water (9.81 kN/m3), h is the height of phreatic line from the sliding surface obtained by groundwater simulation (Eq. 9) and θ is the slope angle.
where c′ is effective cohesion, φ′ is the effective friction angle, γ is the unit weight of the sliding mass, l is the depth of the sliding mass and kh is the coefficient of horizontal earthquake acceleration.
5.3 Pore pressure simulation and slope stability analysis during actual earthquake
This analysis shows that the slope would fail due to earthquake shaking, even without pore pressure increase. However, because the Tandikat landslide occurred during rainfall and underwent flow mobility, “dry” failure did not occur. Hence, “wet” failure, where the sliding zone reached a fully saturated condition is more realistic. The immediate and rapid failure of the slope before earthquake acceleration reached its peak shows that the saturated slope mass of loose pumice sand needs only slight energy to generate shear strain, which then increased the pore pressure to a critical and catastrophic level. This suggests that earthquake of smaller magnitude than the M7.6 2009.9.30 Padang earthquake can still lead to disaster if the required condition of sliding zone saturation due to rainfall is attained.
Based on field observations, the pumice sand was the main material of the extensive landslide mass. The low density and high porosity of the pumice sand contributed to the slope failure induced by earthquake during rainfall.
The possibility of rainfall saturation of the pumice sand deposit was assessed using the Green-Ampt method. The results suggested that the sliding zone of the pumice sand deposits less than 3 m had a high probability of saturation by rainfall infiltration. High permeability and high water content due to antecedent rainfall facilitated rainfall water percolation into the ground.
Vulnerability of saturated pumice sand to pore pressure increase was confirmed by static and stress-controlled cyclic triaxial tests, which showed contractive behaviour of the pumice deposits, as indicated by excess pore pressure rise at small strains. Immediate pore pressure build-up occurred when fully saturated specimens were tested.
CTX test results showed that effective confining pressure greatly influenced reference cumulative shear strain. The test results showed that reference cumulative shear strain increased linearly with effective confining pressure, suggesting that risk of pore pressure increase during earthquake was greater in saturated shallow pumice sand deposits than in than thicker deposits.
The effect of initial shear stress on reference cumulative shear strain was also examined. The results indicate that soil mass with larger initial shear stress needs larger cumulative shear strain to increase pore pressure ratio to a certain value. This suggests that pore pressure increase during earthquake is more probable on gentle slopes than it is on steep slopes.
The phenomenon of stiffness degradation of pumice sand during cyclic loading was also considered. The effect of effective confining pressure on stiffness degradation was obvious, while the effect of initial shear stress was unclear. These results indicated that the effective confining pressure contributed to the initial shear modulus, G0, which is the initial value of the secant shear modulus Gs. During cyclic loading, the shear modulus decreased rapidly irrespective of the effective confining pressure.
Stability analysis of the Tandikat landslide using a rigid block on a quasi-plastic layer assumption and the actual earthquake acceleration suggested that slope failure occurred due to pore pressure build-up. The factor of safety decreased rapidly before earthquake acceleration peaked. At that time, the energy of the earthquake had not reached its maximum level, suggesting that failure would probably occur on saturated sliding zone even during smaller earthquakes. This finding emphasises the high risk of catastrophic earthquake-triggered landslides in tropical regions with high rainfall.
We appreciate the help of Prof. Dwikorita Karnawati and Dr. T. Faisal Fathani of Gadjah Mada University for providing additional data (SPT and geological logging) used in this study. We are also grateful to Mr. Rahindro Pandhu Mahesworo, S.T, M.T, the Head of Engineering Seismology Data, BMKG Indonesia, for providing the earthquake accelerogram data used in this study. We also express thank to Prof. Barry Roser of Shimane University and Austin O. Chukwueloka for their review of an early draft of the manuscript.
- Aydan Ö: A reconnaissance report on the Pariaman-Padang earthquake of September 30, 2009. Japan Society of Civil Engineers, Japan; 2009.Google Scholar
- Bansal RK (2013) Modelling of groundwater flow over sloping beds in response to constant recharge and stream of varying water level. (in press). [last accessed: June 29, 2014], [http://www.ijm2c.ir/index.php/ijm2c/article/view/108/150]Google Scholar
- Bansal RK, Das SK: An analytical study of water table fluctuations in unconfined aquifers due to varying bed slopes and spatial location of the recharge basin. J Hydrol Eng 2010,15(11):909–917. 10.1061/(ASCE)HE.1943-5584.0000267View ArticleGoogle Scholar
- Chen L, Young MH: Green-Ampt infiltration model for sloping surfaces. Water Resource Res 2006, 42: 1–9.Google Scholar
- EERI (2009) Learning from earthquakes: the MW 7.6 Western Sumatra earthquake of September 30, 2009. [last accessed: April 8, 2013], [http://www.eeri.org/site/images/eeri_newsletter/2009_pdf/Padang-eq-report-NL-insert.pdf]
- Green WH, Ampt GA: Studies on soil physics: 1. Flow of air and water through soils. J Agric Sci 1911, 4: 1–24. 10.1017/S0021859600001441View ArticleGoogle Scholar
- Green RA, Mitchell JK, Polito CP: An energy-based excess pore pressure generation model for cohesionless soils. Proceedings of the John Booker Memorial Symposium. Sydney, New South Wales, Australia, November 16–17, 2000. A.A. Balkema Publishers, Rotterdam; 2000.Google Scholar
- Hsu S-M, Ni C-F, Hung P-F: Assessment of three infiltration formulas based on model fitting on Richard’s equation. J Hydrol Eng 2002,7(5):373–379. 10.1061/(ASCE)1084-0699(2002)7:5(373)View ArticleGoogle Scholar
- Hyodd M, Hyde AFL, Aramaki N: Liquefaction of crushable soils. Géotechnique 1998,48(4):527–543. 10.1680/geot.19126.96.36.1997View ArticleGoogle Scholar
- Ishihara K (1985) Stability of natural deposits during earthquakes. Proceedings, 11th International Conference on Soil Mechanics and Foundation Engineering, San Francisco, Vol. 1. pp 321–376Google Scholar
- Jafarian Y, Towhata I, Baziar MH, Noorzad A, Bahmanpour A: Strain energy based evaluation of liquefaction and residual pore water pressure in sands using cyclic torsional shear experiments. Soil Dynam Earthq Eng 2012, 35: 13–28. 10.1016/j.soildyn.2011.11.006View ArticleGoogle Scholar
- Lee KL, Albaisa A: Earthquake-induced settlement in saturated sands. J Soil Mech Found Div ASCE 1974, 100: 387–406.Google Scholar
- Lee CJ, Sheu SF: The stiffness degradation and damping ratio evolution of Taipei Silty Clay under cyclic straining. Soil Dynam Earthq Eng 2007, 27: 730–740. 10.1016/j.soildyn.2006.12.008View ArticleGoogle Scholar
- Lenart S: The use of dissipated energy at modeling of cyclic loaded saturated soils. Earthquake Engineering: New Research. Nova Science Publishers, Inc., New York; 2008.Google Scholar
- Marks S, Larkin TJ, Pender MJ: The dynamic properties of a pumiceous sand. NZNSEE Bull 1998,31(2):86–100.Google Scholar
- Muñoz-Carpena R, Regalado CM, Álvarez-Benedí J, Bartoli F: Field evaluation of the new Philip–Dunne permeameter for measuring saturated hydraulic conductivity. Soil Sci 2002, 167: 9–24. 10.1097/00010694-200201000-00002View ArticleGoogle Scholar
- Orense RP, Pender MJ (2013) Liquefaction characteristics of crushable pumice sand, Proceeding of the 18th International Conference on Soil Mechanics and Geotechnical Engineering. 2–6 September, Paris. pp 1559–1562Google Scholar
- Petersen M, Harmsen S, Mueller C, Haller K, Dewey J, Luco N, Crone A, Lidke D, Rukstales K (2007) Documentation for the Southeast Asia seismic hazard maps. Administrative report, U.S. Geological Survey. p 67Google Scholar
- Philip JR: Approximate analysis of falling-head lined borehole permeameter. Water Resource Res 1993, 29: 3763–3768. 10.1029/93WR01688View ArticleGoogle Scholar
- Regalado CM, Ritter A, Álvarez-Benedí J, Muñoz-Carpena R: Simplified method to estimate the Green–Ampt wetting front suction and soil sorptivity with the Philip–Dunne falling-head permeameter. Vadose Zone Journal 2005, 4: 291–299. 10.2136/vzj2004.0103View ArticleGoogle Scholar
- Seed HB, Idriss IM, Makdisi F, Banerjee N: Representation of irregular stress time histories by equivalent uniform stress series in liquefaction analyses. University of California, Berkley; 1975.Google Scholar
- Seed HB, Martin PP, Lysmer J: Pore water pressure change during soil liquefaction. J Geotech Eng Div ASCE 1976,102(4):323–346.Google Scholar
- Sipayung SB, Lely QA, Bambang DD, Sutikno A: The analysis of rainfall pattern in Indonesia based on global circulation model (GCM) output. Jurnal Sains Nusantara 2007,4(2):145–154. in Indonesian in IndonesianGoogle Scholar
- Skempton AW: The pore pressure coefficients A and B. Geotechnique 1954, 4: 143–7. 10.1680/geot.19188.8.131.52View ArticleGoogle Scholar
- Sloan PG, Moore ID: Modeling subsurface stormflow on steeply sloping forested watersheds. Water Resources Res 1984, 20: 1815–1822. 10.1029/WR020i012p01815View ArticleGoogle Scholar
- Steenhuis TS, Parlange JY, Sanford WE, HeiligA SF, Walter MF: Can we distinguish Richards’ and Boussinesq’s equations for hillslopes?: The Coweeta experiment revisited. Water Resources Res 1999,35(2):589–593. 10.1029/1998WR900067View ArticleGoogle Scholar
- Sugano T, Yanagisawa E: Cyclic undrained shear behaviour of sand under surface wave stress conditions. 1992.Google Scholar
- Suzuki M, Yamamoto T: Liquefaction characteristic of undisturbed volcanic soil in cyclic triaxial test. 2004.Google Scholar
- Wang FW, Sassa K: Relationship between grain crushing and excess pore pressure generation by sandy soils in ring shear tests. J Nat Disaster Sci 2000,22(2):87–96. 10.2328/jnds.22.87View ArticleGoogle Scholar
- Wang FW, Muhammad Wafid AN, Zhang F: Tandikek and Malalak flowslides triggered by 2009.9.30 M7.6 Sumatra earthquake during rainfall in Indonesia, Geoscience Report. Shimane Univ 2010, 29: 1–10.Google Scholar
- Wu P, Hara M, Hamada JI, Yamanaka MD, Kimura F: Why a large amount of rain falls over the sea in the vicinity of western Sumatra Island during nighttime. J Appl Meteorol Clim 2009, 48: 1345–1361. 10.1175/2009JAMC2052.1View ArticleGoogle Scholar
- Yamazaki F, Towhata I, Ishihara K (1985) Numerical model for liquefaction problem under multi-directional shearing on horizontal plane, Proceeding of Fifth International Conference on Numerical Methods in Geomechanics. 1–5 April 1985, Nagoya, Japan. p 399Google Scholar
- Žlender B, Lenart S: Cyclic liquefaction potential of lacustrine carbonate silt from Julian Alps. Acta Geotechnica Slovenica 2005,2005(1):23–31.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.