- Email: [email protected]
Critical acceleration as a criterion in seismic landslide susceptibility assessment
Geomorphology 217 (2014) 15–22
Contents lists available at ScienceDirect
Geomorphology journal homepage: www.elsevier.com/locate/geomorph
Critical acceleration as a criterion in seismic landslide susceptibility assessment Xiao-Li Chen a,⁎, Chun-Guo Liu b, Lu Yu a, Chuan-Xiang Lin c a b c
Key Lab of Active Tectonics and Volcano, Institute of Geology, China Earthquake Administration, Beijing, 100029, China China Earthquake Networks Center, Beijing, 100045, China Beijing Institute of Geology, Beijing 100120, China
a r t i c l e
i n f o
Article history: Received 23 July 2013 Received in revised form 1 April 2014 Accepted 7 April 2014 Available online 18 April 2014 Keywords: Earthquake-triggered landslides Newmark's method Critical acceleration Potential landslide assessment The Wenchuan earthquake
a b s t r a c t The 2008 Wenchuan earthquake in China triggered approximately 200,000 landslides. This study examined a detailed landslide inventory for the event. Based on Newmark's method, the correlation between critical acceleration and landslide occurrence was analyzed for the Beichuan region in Sichuan Province, where various kinds of geohazards occurred due to the earthquake. The results indicate that critical acceleration is a good and reliable criterion to assess slope stability, and that slope gradient and material component are important factors influencing landslide occurrences. It was found that external forces behave differently in different directions. Critical acceleration in horizontal direction is more important for assessing the stability of steeper slopes. This knowledge will help in the understanding of the mechanism of earthquake-triggered landslides and facilitate the combined use of critical acceleration and landslide distribution maps for determining peak ground acceleration in a region without abundant seismic instrument records. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Earthquake-triggered landslides have been a major cause of casualties and damages in recent earthquake events. The 2008 Wenchuan earthquake in China triggered numerous landslides, causing about 20,000 deaths (Yin et al., 2009). Such large and widely distributed landslides can be prevented neither by current mitigating countermeasures nor by regular monitoring to predict rainfall-triggered landslides, although research on the zoning of earthquake-triggered landslides has been conducted in various countries (e.g., Guzzetti et al., 1999; Rodríguez et al., 1999; Miles and Keefer, 2001a, 2001b; Bommer et al., 2002; National Institute for Land and Infrastructure Management, Ministry of Land, Infrastructure, Transport and Tourism, 2004; Jibson and Michael, 2009). Many studies about earthquake-triggered landslides have focused on the relationship between landslide distribution and factors contributing to slope failures, although they are generally complex and difficult to assess with confidence. As a factor of landslides triggered by earthquake, peak ground acceleration (PGA) has been proved to be strongly correlated with landslide density: with the increase of PGA, landslides increase (Keefer, 1984; Harp and Jibson, 1996; Wang et al., 2003, 2007, 2008; Kieffer et al., 2006; Meunier et al., 2007; Chen et al., 2010; Qi et al., 2010). ⁎ Corresponding author. Tel.: +86 10 62009056. E-mail address: [email protected] (X.-L. Chen).
http://dx.doi.org/10.1016/j.geomorph.2014.04.011 0169-555X/© 2014 Elsevier B.V. All rights reserved.
Many methods have been developed to assess the stability of slopes. Newmark's method is the one that has been widely and successfully applied in seismic landslide hazard assessment (Jibson et al., 2000; Miles and Keefer, 2001a, 2001b; Jibson and Michael, 2009). Wilson and Keefer (1985) used Newmark's method to model the dynamic behavior of landslides on natural slopes and to assess seismic slope stability for a broad region in the Los Angeles area. Based on this physical method, Jibson et al. (2000) proposed a method to map the probabilities of seismic slope failure, which yielded reasonable results for the 1994 Northridge earthquake in Oat Mountain, California. Kaynia et al. (2010) also developed a real-time mapping system for earthquaketriggered landslides based on this method. In the studies above, the permanent-displacements calculated from different empirical prediction models were used to predict the stability of natural slopes during an earthquake event. Nevertheless, the equations used for displacement predictions were calibrated using data from specific regions, and their applications to other regions with different geological or climatic conditions will increase the uncertainty of the results. The 2008 Wenchuan earthquake, which originated in the Longmen Shan fault zone at the eastern margin of the Tibetan Plateau, was the largest seismic event in China in more than 50 years. The earthquake triggered more than 60,000 destructive landslides that caused about one-third of the total number of fatalities (Huang and Fan, 2013). Many studies so far have dealt with different aspects of the landslides triggered by the event, and the landslide inventory has been gradually improved (Huang and Li, 2009; Yin et al., 2009; Qi et al., 2010; Dai
16
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
et al., 2011; Xu et al., 2013). However, few studies have related the critical acceleration of slopes to the landslides. Based on Newmark's method, we have calculated the critical acceleration of the natural slopes in the Beichuan region which suffered severe landslide damage due to the Wenchuan earthquake and correlated the critical acceleration values with the landslide distribution. We propose to use critical acceleration and landslide distribution maps as references to determine peak ground acceleration in a region without abundant seismic instrument records. 2. Geological setting and landslides of the study area The Wenchuan earthquake triggered numerous landslides from rock falls of a few m3 to rock avalanches of tens of millions m3. The most detailed landslide inventory is provided by Xu et al. (2013), who analyzed nearly 200,000 landslides triggered by the earthquake distributed in an area of about 110,000 km2. This study presents a good data set for deeper and more detailed landslide research. Landslides triggered by earthquakes are clustered along causative seismic faults (Keefer, 1984, 2002; Khazai and Sitar, 2003; Wang et al., 2003, 2007; Jibson et al., 2004; Wen et al., 2004). This is the case with the Wenchuan earthquake (Huang and Li, 2008, 2009; Chen et al., 2010; Qi et al., 2010; Xu and Li, 2010; Dai et al., 2011). The distribution of the large landslides is particularly correlated with that of the coseismic displacements (Chen et al., 2012). We divided the region where most of the large landslides (≧50,000 m2 in area) are concentrated into four zones along the Longmen Shan fault: Yingxiu, Gaochuan, Beichuan and Qingchuan from southwest to northeast (Chen et al., 2012). The chosen study area is mainly distributed in the Beichuan zone (Figs. 1 and 2). The study area is about 40 km long and 16 km wide. The seismogenic Yingxiu–Beichuan fault cuts through the study area from southwest to northeast and forms an obvious rupture zone during the event
(Fig. 2). According to the latest landslide inventory provided by Xu et al. (2013), there are more than 7600 landslides with a total affected area around 60 km2 in the study area. These landslides have different volume sizes in the form of various failure types, and most of them are in a small or moderate scale. Statistics show that more than 70% of the landslides are with an area of b 5000 m2, and more than 20% of the landslides are with an area of b1000 m2. Although there are only 214 landslides greater than 50,000 m2, they occupy more than 23 km2, which is 38% of the total affected area. On the whole, landslides are clustered along the Yingxiu–Beichuan fault, and are distributed unevenly on both sides of the rupture zone (Fig. 2). More than 80% of the landslides are located on the hanging wall, showing sharp effects of the wall. Landslide–area ratio (LAR), which is expressed as a percentage of the area affected by landslide activity, is around 7.9% on the hanging wall side, whereas it is 2.1% on the footwall side. As to the large landslides with a plane area greater than 50,000 m2, there are 190 landslides on the hanging wall side compared to 24 landslides on the footwall side. Some notable landslides like the Tangjiashan landslide and the Chengxi landslide also occurred in this study region. The Tangjiashan landslide formed the largest landslide-dammed lake which blocked the upper portion of the Jianjiang River, 5 km from Beichuan County Town. The Chengxi landslide is 4.8 × 106 m3 in volume, which killed 1600 people and destroyed half of the old area of the Beichuan County Town (Yin et al., 2009). The study region has exposures of the strata from the Sinian to Quaternary periods, while there is a lack of Jurassic to Cretaceous sequences (Fig. 2, Table 1). Almost all bedrocks appearing in this region are weathered and deformed. Rocks found here mainly include sandstone, mudstone, and shale. Alluvial deposits are developed along the river courses. More than 62% of the landslides are present in the Cambrian and Sinian strata, which outcrop along the seismogenic fault (Fig. 2).
Fig. 1. Distribution of large landslides during the Wenchuan earthquake. F1: Wenchuan–Maowen fault, F2: Yingxiu–Beichuan fault, F3: Guanxian–Anxian fault, (1) Yingxiu segment; (2) Gaochuan segment; (3) Beichuan segment; (4) Qingchuan segment.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
17
Fig. 2. Landslide distribution map and geological strata in the study area (modified after CGS (2001)). For the codes of the geological strata, see Table 1.
According to the rock engineering standard by General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Ministry of construction of the People's Republic of China (1995) the rocks in the study area can be classified into three types: II (relatively hard), III (medium hardness) and IV (relatively soft). The main component of Type II is carbonate rocks, i.e. limestone and conglomerate, which are mostly situated on the footwall side of the Yingxiu– Beichuan rupture zone. Type III mainly consists of sandstones and slates and is situated on the hanging wall side. Type IV rocks are weathered rocks and loose deposits in Quaternary. Relief in the study region gradually decreases eastward. On the west side of the Yingxiu–Beichuan fault, the elevation is commonly above 1000 m, while on the east side it is lower than 1000 m. With respect to the epicenter region, topography in this study area is less steep. The distribution of slope gradient indicates that more than 62% of
Table 1 Simplified geologic strata system of the study area. Modified after Qi et al. (2010). Sequence
Symbol
Lithology
Quaternary Triassic Permian Carboniferous Devonian Silurian Ordovician Cambrian Sinian
Q T P C D S O
Alluvium and loose deposit Sandstone, limestone, and slate Limestone intercalated slate Limestone, marble, and sandstone Quartzose sandstone Sandstone and phyllite intercalated with limestone Limestone, marble, and phyllite Metamorphic sandy conglomerate and limestone Metamorphic sandstone and metamorphic limestone
Z
the study region is occupied by slopes with an angle of 20° to 40°, and the slopes with an angle of above 40° only cover 2% of the area (Fig. 3). In general, the steeper and higher slopes have higher susceptibility to landslide activity. Statistical studies of the relationship between slope gradient and landslides in this region show the results similar to those from other reported cases of earthquakes: LAR increases with increasing slope. It is also found that more than 65% of landslides occurred in the area with slopes of 20° to 40°, the highest LAR value is around 11% in the segment with slopes of 30° to 40°. 3. Method 3.1. Methodology Newmark's (1965) method models a landslide as a rigid friction block that slides on an inclined plane. The block has a known critical acceleration, ac, which is simply the threshold base acceleration required to overcome shear resistance and initiate sliding (Fig. 4). Critical acceleration of a potential landslide block is a simple function of the static factor of safety and landslide geometry, expressed as: ac ¼ ð F S −1Þg sin α
ð1Þ
where ac is the critical acceleration in terms of g, the acceleration of Earth's gravity; FS is the static factor of safety; and α is the angle from the horizontal direction where the center of mass of the potential landslide block first moves, which can generally be approximated as the slope angle. FS is expressed as: FS ¼
c0 tan φ0 mγw tan φ0 þ − γt sin α tan α γ tan α
ð2Þ
18
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
Fig. 3. Slope map of the study area.
where φ′ is the effective friction angle, c′ is the effective cohesion, α is the slope angle, γ is the material unit weight, γw is the unit weight of water, t is the slope-normal thickness of the failure slab, and m is the proportion of the slab thickness that is saturated. When data sets describing the topography, geology, and shear strength of an area or region can be procured, critical acceleration of natural slopes then can be calculated.
3.2. Data preparation In Newmark's method, slope data and geological data are two key inputs. Slope in degrees in an area can be produced from a digital elevation model (DEM). In this study, the SRTM 90-m DEM (http://srtm.csi. cgiar.org/SELECTION/inputCoord.asp) was used to produce the slope map of Fig. 3. The 1:200,000-scale geologic maps (CGS, 2001) of the region were also used to assign material properties. After the geological maps were digitized and rasterized at a grid spacing of 90 m using the ArcGIS software, they were combined with the slope map to yield the critical acceleration grids.
Although material strength parameters play an important role in determining slope stability, it is not practical to test the parameters for every kind of the rocks in a large region. Therefore, selecting representative shear-strength values for geologic units seems a good solution. As noted before, the rocks in this study region can be classified into three types. For simplicity, the rock material shear-strength parameters were assigned as shown in Table 2, according to the rock engineering standards used in China (General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China and Ministry of construction of the People's Republic of China, 1995) and a related reference (Jibson et al., 2000).
4. Results Based on the topographical and geological maps, the factor of safety for the study area can be obtained using Eq. (2) and a map of critical acceleration is to be developed using Eq. (1). Fig. 5 shows the constructed map of critical acceleration for the study area. It reveals that more than 5% of the slopes have critical acceleration values less than 0.3 g, 21% of the slopes have values between 0.4 and 0.6 g, 44% have values between 0.6 and 0.8 g, and 30% slopes have values greater than 0.8 g. Commonly, a greater ac value means that a slope requires a greater force to overcome its stability. When the ac value is low, sliding occurs more easily on a slope. The critical acceleration distribution map corresponds well to the actual landslides triggered by the Wenchuan earthquake. The fit appears to be very good: most of the triggered landslides are lying in the higher acceleration areas (warmer colored areas in Table 2 Rock strength data for the study area.
Fig. 4. Sliding-block model used for Newmark analysis. Modified after Jibson et al. (2000).
Rock type
c′ (MPa)
φ′ (°)
γ (kN m−3)
II III IV
0.022 0.015 0.005
35 25 10
27.0 22.0 15.0
c′ (MPa): effective cohesion; φ′ (°): effective friction angle; γ (kN m−3): material unit weight.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
19
Fig. 5. Distribution of landslides superimposed on the critical acceleration map.
Fig. 5). Similar critical acceleration studies have been done to the Ms 7.0 Lushan earthquake on 20 April, 2013 (Chen et al., 2013). The result also confirms that Newmark's method is effective in assessing the susceptibility of earthquake-triggered landslides. Besides the calculation of critical acceleration ac parallel to the slope surface, the critical acceleration in horizontal direction (ac_h) is
calculated using Eq. (3): ac
h
¼ ac cos α:
ð3Þ
Fig. 6 shows the horizontal critical acceleration distribution map for the study area.
Fig. 6. Distribution of landslides superimposed on the horizontal critical acceleration map.
20
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
160 140
Area (km2)
120 100 80 60 40
Parallel Critical Acceleration
20
Horizontal Critical Acceleration
0 <0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
>=0.9
Critical Acceleration (g) Fig. 7. Comparison of horizontal critical acceleration and parallel critical acceleration.
Analysis on the area occupied by critical accelerations in both horizontal and parallel directions shows that unstable areas (where ac or ac_h b 0.6 g) increases by around 15% when ac_h is used instead of ac (Figs. 6 and 7). More landslides are located in the higher probability areas (warmer colored areas in Fig. 6), where slopes are steeper. The monotonous decline of cosα with increasing α suggests that steeper slopes are more sensitive to the force in the horizontal direction than gentle slopes. 5. Discussion In Newmark's method, when the geometry and material strength of a natural slope are known, the force or acceleration needed to trigger landsliding can be determined. Obviously, this method only takes account of factors belonging to the slope itself while ignoring other factors such as seismic magnitude or the distance from the fault or the epicenter. In this study as well as other previous studies on the regional seismic hazard analysis, ac is calculated separately in each grid cell which is connected to each other in the physical space. Therefore, the accuracy of the
results depends on the cell size and the material parameters being adopted in the calculation. In our study, material strength data are simplified by classifying rocks into three types, although they are composed of more than 10 formations. This simplification may result in imprecise outcomes in some areas. Moreover, considering the principle of this method, there seems to be no difference in the critical acceleration value between landslides with different plane areas, while the depth of landslides is found to greatly affect the critical acceleration in practical calculations. For the same location, deeper landslides need lower critical acceleration to initiate sliding. Correlation of the critical acceleration with landslide distribution suggests that the critical acceleration can be used as a reference for determining the peak ground acceleration. Although hundreds of seismic instrument records of main shock were obtained during the 2008 Wenchuan earthquake, near field seismic data are not optimal for further study. Within or near the study area of this paper, only three seismic stations recorded the peak ground motions (Fig. 8, Table 3). According to the seismic intensity distribution map of the Wenchuan earthquake (China Earthquake Administration, 2008), the seismic intensity value for most of the study area was greater than XI (Fig. 8),
Fig. 8. Distribution of landslides with seismic intensity.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22 Table 3 Peak ground acceleration recorded in the study area (Li et al., 2008). Seismic station
Distance from the rupture (km)
E–W (gal)
N–S (gal)
U–D (gal)
Jiangyou Hanzeng Wudu
1.6 13.0 17.0
493.4 347.0 470.2
548.6 512.5 519.6
214.4 448.1 210.8
which corresponds to the peak ground motion greater than 708 gal (Li et al., 2008). Our study on the critical acceleration of the Beichuan region indicates that most of the study area has high PGA values above 0.6 g (Fig. 5), meaning that slopes were relatively stable before the earthquake. However, a great amount of landslides occurred in this area during the earthquake, revealing that the area has experienced strong ground motion that can cause failures on previously stable slopes. Based on the distribution of critical acceleration and landslides, it is estimated that ground motion in this region should be at least 0.6 or 0.7 g. These observations support the delimitation of the seismic intensity distribution map issued by China Earthquake Administration (China Earthquake Administration, 2008). Because of the limited amount and accuracy of the data used, detailed seismic ground motion cannot be obtained at present. Nevertheless, our study has indicated the possibility of using critical acceleration and landslide distribution maps as references for estimating the peak ground acceleration of an earthquake event. 6. Summary While some other researchers used permanent displacements to predict the potential for earthquake-triggered landslides on natural slopes, we have used the critical acceleration for such assessments. It was useful to compare the constructed critical acceleration map and the landslide distribution map because it proves that Newmark's method is feasible and practicable in seismic hazard analysis. In addition, combined analysis of critical acceleration maps and earthquaketriggered landslide distribution maps can play a key role in determining the peak ground acceleration in a region without abundant seismic records. Slope angle and material components are important factors in landslide occurrences. Slopes with harder rocks and lower gradients are likely to stay stable during seismic events. Critical acceleration in a horizontal direction affects the stability of steeper slopes more than that of gentler slopes. This knowledge will help in the understanding of the various mechanisms of landslides triggered by earthquakes. It also suggests that, in steep mountainous areas, more attention should be given to the peak ground motion in the horizontal direction for assessing landslide susceptibility. Acknowledgments This research was supported by the National Key Basic Research Program of China (Grant No. 2013CB733205) and the National Key Technology R & D Program (Grant No. 2012BAK15B01-03). The authors would like to express thanks to Dr. Chong Xu for allowing as to use his landslide inventory in the study area. Deep appreciation goes to anonymous referees and editor Professor Takashi Oguchi for their helpful comments. Thanks are also given to Ms. Hannah Boughton for her kind help to polish the English of the manuscript. References Bommer, J.J., Carlos, E., Rodríguez, C.R., 2002. Earthquake-induced landslides in Central America. Eng. Geol. 63, 189–220.
21
China Earthquake Administration, 2008. Ioseismal map of the Wenchuan Earthquake. http://www.cea.gov.cn/manage/html/8a8587881632fa5c0116674a018300cf/_content/ 08_08/29/1219980517676.html. China Geological Survey (CGS), 2001. Regional geological map of Sichuan Province (1: 200, 000), Geological Press. Chen, X.L., Zhou, B.G., Ran, H.L., Yamamoto, Y., Hyodo, M., 2010. Geohazards induced by the Wenchuan earthquake. In: Williams, A.L., Pinches, G.M., Chin, C.Y., McMorran, T.J., Massey, C.I. (Eds.), Geologically Active. Taylor & Francis Group, London, pp. 77–84. Chen, X.L., Ran, H.L., Yang, W.T., 2012. Evaluation of factors controlling large earthquakeinduced landslides by the Wenchuan earthquake. Nat. Hazards Earth Syst. Sci. 12, 3645–3657. Chen, X.L., Yuan, R.M., Yu, L., 2013. Applying the Newmark's model to the assessment of earthquake-triggered landslides during the Lushan earthquake. Seismol. Geol. 35, 661–670 (in Chinese, with English Abstr.). Dai, F.C., Xu, C., Yao, X., Xu, L., Tu, X.B., Gong, Q.M., 2011. Spatial distribution of landslides triggered by the 2008 Ms 8.0 Wenchuan earthquake, China. J. Asian Earth Sci. 40, 883–895. General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Ministry of construction of the People's Republic of China, 1995. Standard for Engineering Classification of Rock Masses. Standards Press of China, Beijing (in Chinese). Guzzetti, F., Carrara, A., Cardinali, M., Reichenbach, P., 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy. Geomorphology 31, 181–216. Harp, E.L., Jibson, R.W., 1996. Landslides triggered by the 1994 Northridge, California earthquake. Bull. Seismol. Soc. Am. 86 (1B), s319–s332. Huang, R.Q., Li, W.L., 2008. A study on the development and distribution rules of geohazards triggered by “5.12” Wenchuan earthquake. Chin. J. Rock Mech. Eng. 27, 2585–2592 (in Chinese, with English Abstr.). Huang, R.Q., Li, W.L., 2009. Analysis on the number and density of landslides triggered by the 2008 Wenchuan earthquake, China. J. Geol. Hazards Environ. Preserv. 20 (3), 1–7 (in Chinese, with English Abstr.). Huang, R.Q., Fan, X.M., 2013. The landslide story. Nat. Geosci. 6, 325–326. Jibson, R.W., Harp, E.L., Michael, J.A., 2000. A method for producing digital probabilistic seismic landslide. Eng. Geol. 58, 271–289. Jibson, R.W., Harp, E.L., Schulz, W., Keefer, D.K., 2004. Landslides triggered by the 2002 M7.9 Denali Fault, Alaska, earthquake and the inferred nature of the strong shaking. Earthquake Spectra 20, 669–691. Jibson, R.W., Michael, J.A., 2009. Maps showing seismic landslide hazards in Anchorage, Alaska. U.S. Geological Survey Scientific Investigations Map 3077, scale 1:25,000, 11-p. pamphlet. Kaynia, A.M., Skurtveit, E., Saygili, G., 2010. Real-time mapping of earthquake-induced landslides. Bull. Earthq. Eng.. http://dx.doi.org/10.1007/s10518-010-9234-2. Keefer, D.K., 1984. Landslides caused by earthquakes. Geol. Soc. Am. Bull. 95, 406–421. Keefer, D.K., 2002. Investigating landslides caused by earthquakes — a historical review. Surv. Geophys. 23, 473–510. Khazai, B., Sitar, N., 2003. Evaluation of factors controlling earthquake-induced landslides caused by Chi–Chi earthquake and comparison with the Northridge and Loma Prieta events. Eng. Geol. 71, 79–95. Kieffer, D.S., Jibson, R., Rathje, E.M., Keith, K., 2006. Landslides triggered by the 2004 Niigata Ken Chuetsu, Japan, earthquake. Earthquake Spectra 22, S47–S73. Li, X.J., Zhou, Z.H., Huang, M., Wen, R.Z., Yu, H.Y., Lu., D.W., Zhou, Y.N., Cui, J.W., 2008. Preliminary analysis of strong-motion recordings from the magnitude 8.0 Wenchuan, China, earthquake of 12 May 2008. Seismol. Res. Lett. 79, 844–854. Meunier, P., Hovius, N., Haines, J.A., 2007. Regional patterns of earthquake-triggered landslides and their relation to ground motion. Geophys. Res. Lett. 34. http://dx.doi.org/ 10.1029/2007GL031337 L20408. Miles, S.B., Keefer, D.K., 2001a. Seismic Landslide Hazard for the Cities of Oakland and Piedmont, California. (http://pubs.usgs.gov/mf/2001/2379/oakpamph.pdf). Miles, S.B., Keefer, D.K., 2001b. Seismic Landslide Hazard for the City of Berkeley, California. (http://pubs.usgs.gov/mf/2001/2378/berkpamph.pdf.). National Institute for Land and Infrastructure Management, Ministry of Land, Infrastructure, Transport and Tourism, 2004. A Study on Methodology for Assessing the Potential of Slope Failures During Earthquakes [Report], Japan (in Japanese). Newmark, N.M., 1965. Effects of earthquakes on dams and embankments. Geotechnique 15, 139–160. Qi, S.W., Xu, Q., Lan, H.X., Zhang, B., Liu, J.Y., 2010. Spatial distribution analysis of landslides triggered by 2008.5.12 Wenchuan earthquake, China. Eng. Geol. 116, 95–108. Rodríguez, C.E., Bommerb, J.J., Chandlerb, R.J., 1999. Earthquake-induced landslides: 1980–1997. Soil Dyn. Earthq. Eng. 18, 325–346. Wang, H.B., Sassa, K., Xu, W.Y., 2007. Analysis of a spatial distribution of landslides triggered by the 2004 Chuetsu earthquakes of Niigata Prefecture, Japan. Nat. Hazards 41, 43–60. Wang, W.N., Wu, H.L., Nakamura, H., Wu, S.C., Ouyang, S., Yu, M.F., 2003. Mass movements caused by recent tectonic activity: the 1999 Chi–chi earthquake in central Taiwan. Island Arc 12, 325–334. Wang, Y.S., Luo, Y.H., Ji, F., Huo, J.J., Wu, J.F., Xu, H.B., 2008. Analysis of the controlling factors on geo-hazards in mountainous epicentre zones of the Wenchuan earthquake. J. Eng. Geol. 16, 759–763 (in Chinese, with English Abstr.). Wen, B.P., Wang, S.J., Wang, E.Z., Zhang, J.M., 2004. Characteristics of rapid giant landslides in China. Landslides 4, 247–261. Wilson, R.C., Keefer, D.K., 1985. Predicting areal limits of earthquake-induced landsliding. In: Ziony, J.I. (Ed.), Evaluating Earthquake Hazards in Los Angeles Region — An Earthscience Perspective. U.S. Geological Survey, Professional Paper, 1360, pp. 317–345.
22
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
Xu, Q., Li, W.L., 2010. Distribution of large scale landslides induced by the Wenchuan earthquake. J. Eng. Geol. 18, 818–826 (in Chinese, with English Abstr.). Xu, C., Xu, X.W., Yao, Q., Wang, Y.Y., 2013. GIS-based bivariate statistical modeling for earthquake-triggered landslides susceptibility mapping related to the 2008
Wenchuan earthquake China. Q. J. Eng. Geol. Hydrogeol.. http://dx.doi.org/10.1144/ qjegh2012-006. Yin, Y.P., Wang, F.W., Sun, P., 2009. Landslide hazards triggered by the 2008 Wenchuan earthquake, Sichuan, China. Landslides 6, 139–152.
Contents lists available at ScienceDirect
Geomorphology journal homepage: www.elsevier.com/locate/geomorph
Critical acceleration as a criterion in seismic landslide susceptibility assessment Xiao-Li Chen a,⁎, Chun-Guo Liu b, Lu Yu a, Chuan-Xiang Lin c a b c
Key Lab of Active Tectonics and Volcano, Institute of Geology, China Earthquake Administration, Beijing, 100029, China China Earthquake Networks Center, Beijing, 100045, China Beijing Institute of Geology, Beijing 100120, China
a r t i c l e
i n f o
Article history: Received 23 July 2013 Received in revised form 1 April 2014 Accepted 7 April 2014 Available online 18 April 2014 Keywords: Earthquake-triggered landslides Newmark's method Critical acceleration Potential landslide assessment The Wenchuan earthquake
a b s t r a c t The 2008 Wenchuan earthquake in China triggered approximately 200,000 landslides. This study examined a detailed landslide inventory for the event. Based on Newmark's method, the correlation between critical acceleration and landslide occurrence was analyzed for the Beichuan region in Sichuan Province, where various kinds of geohazards occurred due to the earthquake. The results indicate that critical acceleration is a good and reliable criterion to assess slope stability, and that slope gradient and material component are important factors influencing landslide occurrences. It was found that external forces behave differently in different directions. Critical acceleration in horizontal direction is more important for assessing the stability of steeper slopes. This knowledge will help in the understanding of the mechanism of earthquake-triggered landslides and facilitate the combined use of critical acceleration and landslide distribution maps for determining peak ground acceleration in a region without abundant seismic instrument records. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Earthquake-triggered landslides have been a major cause of casualties and damages in recent earthquake events. The 2008 Wenchuan earthquake in China triggered numerous landslides, causing about 20,000 deaths (Yin et al., 2009). Such large and widely distributed landslides can be prevented neither by current mitigating countermeasures nor by regular monitoring to predict rainfall-triggered landslides, although research on the zoning of earthquake-triggered landslides has been conducted in various countries (e.g., Guzzetti et al., 1999; Rodríguez et al., 1999; Miles and Keefer, 2001a, 2001b; Bommer et al., 2002; National Institute for Land and Infrastructure Management, Ministry of Land, Infrastructure, Transport and Tourism, 2004; Jibson and Michael, 2009). Many studies about earthquake-triggered landslides have focused on the relationship between landslide distribution and factors contributing to slope failures, although they are generally complex and difficult to assess with confidence. As a factor of landslides triggered by earthquake, peak ground acceleration (PGA) has been proved to be strongly correlated with landslide density: with the increase of PGA, landslides increase (Keefer, 1984; Harp and Jibson, 1996; Wang et al., 2003, 2007, 2008; Kieffer et al., 2006; Meunier et al., 2007; Chen et al., 2010; Qi et al., 2010). ⁎ Corresponding author. Tel.: +86 10 62009056. E-mail address: [email protected] (X.-L. Chen).
http://dx.doi.org/10.1016/j.geomorph.2014.04.011 0169-555X/© 2014 Elsevier B.V. All rights reserved.
Many methods have been developed to assess the stability of slopes. Newmark's method is the one that has been widely and successfully applied in seismic landslide hazard assessment (Jibson et al., 2000; Miles and Keefer, 2001a, 2001b; Jibson and Michael, 2009). Wilson and Keefer (1985) used Newmark's method to model the dynamic behavior of landslides on natural slopes and to assess seismic slope stability for a broad region in the Los Angeles area. Based on this physical method, Jibson et al. (2000) proposed a method to map the probabilities of seismic slope failure, which yielded reasonable results for the 1994 Northridge earthquake in Oat Mountain, California. Kaynia et al. (2010) also developed a real-time mapping system for earthquaketriggered landslides based on this method. In the studies above, the permanent-displacements calculated from different empirical prediction models were used to predict the stability of natural slopes during an earthquake event. Nevertheless, the equations used for displacement predictions were calibrated using data from specific regions, and their applications to other regions with different geological or climatic conditions will increase the uncertainty of the results. The 2008 Wenchuan earthquake, which originated in the Longmen Shan fault zone at the eastern margin of the Tibetan Plateau, was the largest seismic event in China in more than 50 years. The earthquake triggered more than 60,000 destructive landslides that caused about one-third of the total number of fatalities (Huang and Fan, 2013). Many studies so far have dealt with different aspects of the landslides triggered by the event, and the landslide inventory has been gradually improved (Huang and Li, 2009; Yin et al., 2009; Qi et al., 2010; Dai
16
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
et al., 2011; Xu et al., 2013). However, few studies have related the critical acceleration of slopes to the landslides. Based on Newmark's method, we have calculated the critical acceleration of the natural slopes in the Beichuan region which suffered severe landslide damage due to the Wenchuan earthquake and correlated the critical acceleration values with the landslide distribution. We propose to use critical acceleration and landslide distribution maps as references to determine peak ground acceleration in a region without abundant seismic instrument records. 2. Geological setting and landslides of the study area The Wenchuan earthquake triggered numerous landslides from rock falls of a few m3 to rock avalanches of tens of millions m3. The most detailed landslide inventory is provided by Xu et al. (2013), who analyzed nearly 200,000 landslides triggered by the earthquake distributed in an area of about 110,000 km2. This study presents a good data set for deeper and more detailed landslide research. Landslides triggered by earthquakes are clustered along causative seismic faults (Keefer, 1984, 2002; Khazai and Sitar, 2003; Wang et al., 2003, 2007; Jibson et al., 2004; Wen et al., 2004). This is the case with the Wenchuan earthquake (Huang and Li, 2008, 2009; Chen et al., 2010; Qi et al., 2010; Xu and Li, 2010; Dai et al., 2011). The distribution of the large landslides is particularly correlated with that of the coseismic displacements (Chen et al., 2012). We divided the region where most of the large landslides (≧50,000 m2 in area) are concentrated into four zones along the Longmen Shan fault: Yingxiu, Gaochuan, Beichuan and Qingchuan from southwest to northeast (Chen et al., 2012). The chosen study area is mainly distributed in the Beichuan zone (Figs. 1 and 2). The study area is about 40 km long and 16 km wide. The seismogenic Yingxiu–Beichuan fault cuts through the study area from southwest to northeast and forms an obvious rupture zone during the event
(Fig. 2). According to the latest landslide inventory provided by Xu et al. (2013), there are more than 7600 landslides with a total affected area around 60 km2 in the study area. These landslides have different volume sizes in the form of various failure types, and most of them are in a small or moderate scale. Statistics show that more than 70% of the landslides are with an area of b 5000 m2, and more than 20% of the landslides are with an area of b1000 m2. Although there are only 214 landslides greater than 50,000 m2, they occupy more than 23 km2, which is 38% of the total affected area. On the whole, landslides are clustered along the Yingxiu–Beichuan fault, and are distributed unevenly on both sides of the rupture zone (Fig. 2). More than 80% of the landslides are located on the hanging wall, showing sharp effects of the wall. Landslide–area ratio (LAR), which is expressed as a percentage of the area affected by landslide activity, is around 7.9% on the hanging wall side, whereas it is 2.1% on the footwall side. As to the large landslides with a plane area greater than 50,000 m2, there are 190 landslides on the hanging wall side compared to 24 landslides on the footwall side. Some notable landslides like the Tangjiashan landslide and the Chengxi landslide also occurred in this study region. The Tangjiashan landslide formed the largest landslide-dammed lake which blocked the upper portion of the Jianjiang River, 5 km from Beichuan County Town. The Chengxi landslide is 4.8 × 106 m3 in volume, which killed 1600 people and destroyed half of the old area of the Beichuan County Town (Yin et al., 2009). The study region has exposures of the strata from the Sinian to Quaternary periods, while there is a lack of Jurassic to Cretaceous sequences (Fig. 2, Table 1). Almost all bedrocks appearing in this region are weathered and deformed. Rocks found here mainly include sandstone, mudstone, and shale. Alluvial deposits are developed along the river courses. More than 62% of the landslides are present in the Cambrian and Sinian strata, which outcrop along the seismogenic fault (Fig. 2).
Fig. 1. Distribution of large landslides during the Wenchuan earthquake. F1: Wenchuan–Maowen fault, F2: Yingxiu–Beichuan fault, F3: Guanxian–Anxian fault, (1) Yingxiu segment; (2) Gaochuan segment; (3) Beichuan segment; (4) Qingchuan segment.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
17
Fig. 2. Landslide distribution map and geological strata in the study area (modified after CGS (2001)). For the codes of the geological strata, see Table 1.
According to the rock engineering standard by General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Ministry of construction of the People's Republic of China (1995) the rocks in the study area can be classified into three types: II (relatively hard), III (medium hardness) and IV (relatively soft). The main component of Type II is carbonate rocks, i.e. limestone and conglomerate, which are mostly situated on the footwall side of the Yingxiu– Beichuan rupture zone. Type III mainly consists of sandstones and slates and is situated on the hanging wall side. Type IV rocks are weathered rocks and loose deposits in Quaternary. Relief in the study region gradually decreases eastward. On the west side of the Yingxiu–Beichuan fault, the elevation is commonly above 1000 m, while on the east side it is lower than 1000 m. With respect to the epicenter region, topography in this study area is less steep. The distribution of slope gradient indicates that more than 62% of
Table 1 Simplified geologic strata system of the study area. Modified after Qi et al. (2010). Sequence
Symbol
Lithology
Quaternary Triassic Permian Carboniferous Devonian Silurian Ordovician Cambrian Sinian
Q T P C D S O
Alluvium and loose deposit Sandstone, limestone, and slate Limestone intercalated slate Limestone, marble, and sandstone Quartzose sandstone Sandstone and phyllite intercalated with limestone Limestone, marble, and phyllite Metamorphic sandy conglomerate and limestone Metamorphic sandstone and metamorphic limestone
Z
the study region is occupied by slopes with an angle of 20° to 40°, and the slopes with an angle of above 40° only cover 2% of the area (Fig. 3). In general, the steeper and higher slopes have higher susceptibility to landslide activity. Statistical studies of the relationship between slope gradient and landslides in this region show the results similar to those from other reported cases of earthquakes: LAR increases with increasing slope. It is also found that more than 65% of landslides occurred in the area with slopes of 20° to 40°, the highest LAR value is around 11% in the segment with slopes of 30° to 40°. 3. Method 3.1. Methodology Newmark's (1965) method models a landslide as a rigid friction block that slides on an inclined plane. The block has a known critical acceleration, ac, which is simply the threshold base acceleration required to overcome shear resistance and initiate sliding (Fig. 4). Critical acceleration of a potential landslide block is a simple function of the static factor of safety and landslide geometry, expressed as: ac ¼ ð F S −1Þg sin α
ð1Þ
where ac is the critical acceleration in terms of g, the acceleration of Earth's gravity; FS is the static factor of safety; and α is the angle from the horizontal direction where the center of mass of the potential landslide block first moves, which can generally be approximated as the slope angle. FS is expressed as: FS ¼
c0 tan φ0 mγw tan φ0 þ − γt sin α tan α γ tan α
ð2Þ
18
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
Fig. 3. Slope map of the study area.
where φ′ is the effective friction angle, c′ is the effective cohesion, α is the slope angle, γ is the material unit weight, γw is the unit weight of water, t is the slope-normal thickness of the failure slab, and m is the proportion of the slab thickness that is saturated. When data sets describing the topography, geology, and shear strength of an area or region can be procured, critical acceleration of natural slopes then can be calculated.
3.2. Data preparation In Newmark's method, slope data and geological data are two key inputs. Slope in degrees in an area can be produced from a digital elevation model (DEM). In this study, the SRTM 90-m DEM (http://srtm.csi. cgiar.org/SELECTION/inputCoord.asp) was used to produce the slope map of Fig. 3. The 1:200,000-scale geologic maps (CGS, 2001) of the region were also used to assign material properties. After the geological maps were digitized and rasterized at a grid spacing of 90 m using the ArcGIS software, they were combined with the slope map to yield the critical acceleration grids.
Although material strength parameters play an important role in determining slope stability, it is not practical to test the parameters for every kind of the rocks in a large region. Therefore, selecting representative shear-strength values for geologic units seems a good solution. As noted before, the rocks in this study region can be classified into three types. For simplicity, the rock material shear-strength parameters were assigned as shown in Table 2, according to the rock engineering standards used in China (General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China and Ministry of construction of the People's Republic of China, 1995) and a related reference (Jibson et al., 2000).
4. Results Based on the topographical and geological maps, the factor of safety for the study area can be obtained using Eq. (2) and a map of critical acceleration is to be developed using Eq. (1). Fig. 5 shows the constructed map of critical acceleration for the study area. It reveals that more than 5% of the slopes have critical acceleration values less than 0.3 g, 21% of the slopes have values between 0.4 and 0.6 g, 44% have values between 0.6 and 0.8 g, and 30% slopes have values greater than 0.8 g. Commonly, a greater ac value means that a slope requires a greater force to overcome its stability. When the ac value is low, sliding occurs more easily on a slope. The critical acceleration distribution map corresponds well to the actual landslides triggered by the Wenchuan earthquake. The fit appears to be very good: most of the triggered landslides are lying in the higher acceleration areas (warmer colored areas in Table 2 Rock strength data for the study area.
Fig. 4. Sliding-block model used for Newmark analysis. Modified after Jibson et al. (2000).
Rock type
c′ (MPa)
φ′ (°)
γ (kN m−3)
II III IV
0.022 0.015 0.005
35 25 10
27.0 22.0 15.0
c′ (MPa): effective cohesion; φ′ (°): effective friction angle; γ (kN m−3): material unit weight.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
19
Fig. 5. Distribution of landslides superimposed on the critical acceleration map.
Fig. 5). Similar critical acceleration studies have been done to the Ms 7.0 Lushan earthquake on 20 April, 2013 (Chen et al., 2013). The result also confirms that Newmark's method is effective in assessing the susceptibility of earthquake-triggered landslides. Besides the calculation of critical acceleration ac parallel to the slope surface, the critical acceleration in horizontal direction (ac_h) is
calculated using Eq. (3): ac
h
¼ ac cos α:
ð3Þ
Fig. 6 shows the horizontal critical acceleration distribution map for the study area.
Fig. 6. Distribution of landslides superimposed on the horizontal critical acceleration map.
20
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
160 140
Area (km2)
120 100 80 60 40
Parallel Critical Acceleration
20
Horizontal Critical Acceleration
0 <0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
>=0.9
Critical Acceleration (g) Fig. 7. Comparison of horizontal critical acceleration and parallel critical acceleration.
Analysis on the area occupied by critical accelerations in both horizontal and parallel directions shows that unstable areas (where ac or ac_h b 0.6 g) increases by around 15% when ac_h is used instead of ac (Figs. 6 and 7). More landslides are located in the higher probability areas (warmer colored areas in Fig. 6), where slopes are steeper. The monotonous decline of cosα with increasing α suggests that steeper slopes are more sensitive to the force in the horizontal direction than gentle slopes. 5. Discussion In Newmark's method, when the geometry and material strength of a natural slope are known, the force or acceleration needed to trigger landsliding can be determined. Obviously, this method only takes account of factors belonging to the slope itself while ignoring other factors such as seismic magnitude or the distance from the fault or the epicenter. In this study as well as other previous studies on the regional seismic hazard analysis, ac is calculated separately in each grid cell which is connected to each other in the physical space. Therefore, the accuracy of the
results depends on the cell size and the material parameters being adopted in the calculation. In our study, material strength data are simplified by classifying rocks into three types, although they are composed of more than 10 formations. This simplification may result in imprecise outcomes in some areas. Moreover, considering the principle of this method, there seems to be no difference in the critical acceleration value between landslides with different plane areas, while the depth of landslides is found to greatly affect the critical acceleration in practical calculations. For the same location, deeper landslides need lower critical acceleration to initiate sliding. Correlation of the critical acceleration with landslide distribution suggests that the critical acceleration can be used as a reference for determining the peak ground acceleration. Although hundreds of seismic instrument records of main shock were obtained during the 2008 Wenchuan earthquake, near field seismic data are not optimal for further study. Within or near the study area of this paper, only three seismic stations recorded the peak ground motions (Fig. 8, Table 3). According to the seismic intensity distribution map of the Wenchuan earthquake (China Earthquake Administration, 2008), the seismic intensity value for most of the study area was greater than XI (Fig. 8),
Fig. 8. Distribution of landslides with seismic intensity.
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22 Table 3 Peak ground acceleration recorded in the study area (Li et al., 2008). Seismic station
Distance from the rupture (km)
E–W (gal)
N–S (gal)
U–D (gal)
Jiangyou Hanzeng Wudu
1.6 13.0 17.0
493.4 347.0 470.2
548.6 512.5 519.6
214.4 448.1 210.8
which corresponds to the peak ground motion greater than 708 gal (Li et al., 2008). Our study on the critical acceleration of the Beichuan region indicates that most of the study area has high PGA values above 0.6 g (Fig. 5), meaning that slopes were relatively stable before the earthquake. However, a great amount of landslides occurred in this area during the earthquake, revealing that the area has experienced strong ground motion that can cause failures on previously stable slopes. Based on the distribution of critical acceleration and landslides, it is estimated that ground motion in this region should be at least 0.6 or 0.7 g. These observations support the delimitation of the seismic intensity distribution map issued by China Earthquake Administration (China Earthquake Administration, 2008). Because of the limited amount and accuracy of the data used, detailed seismic ground motion cannot be obtained at present. Nevertheless, our study has indicated the possibility of using critical acceleration and landslide distribution maps as references for estimating the peak ground acceleration of an earthquake event. 6. Summary While some other researchers used permanent displacements to predict the potential for earthquake-triggered landslides on natural slopes, we have used the critical acceleration for such assessments. It was useful to compare the constructed critical acceleration map and the landslide distribution map because it proves that Newmark's method is feasible and practicable in seismic hazard analysis. In addition, combined analysis of critical acceleration maps and earthquaketriggered landslide distribution maps can play a key role in determining the peak ground acceleration in a region without abundant seismic records. Slope angle and material components are important factors in landslide occurrences. Slopes with harder rocks and lower gradients are likely to stay stable during seismic events. Critical acceleration in a horizontal direction affects the stability of steeper slopes more than that of gentler slopes. This knowledge will help in the understanding of the various mechanisms of landslides triggered by earthquakes. It also suggests that, in steep mountainous areas, more attention should be given to the peak ground motion in the horizontal direction for assessing landslide susceptibility. Acknowledgments This research was supported by the National Key Basic Research Program of China (Grant No. 2013CB733205) and the National Key Technology R & D Program (Grant No. 2012BAK15B01-03). The authors would like to express thanks to Dr. Chong Xu for allowing as to use his landslide inventory in the study area. Deep appreciation goes to anonymous referees and editor Professor Takashi Oguchi for their helpful comments. Thanks are also given to Ms. Hannah Boughton for her kind help to polish the English of the manuscript. References Bommer, J.J., Carlos, E., Rodríguez, C.R., 2002. Earthquake-induced landslides in Central America. Eng. Geol. 63, 189–220.
21
China Earthquake Administration, 2008. Ioseismal map of the Wenchuan Earthquake. http://www.cea.gov.cn/manage/html/8a8587881632fa5c0116674a018300cf/_content/ 08_08/29/1219980517676.html. China Geological Survey (CGS), 2001. Regional geological map of Sichuan Province (1: 200, 000), Geological Press. Chen, X.L., Zhou, B.G., Ran, H.L., Yamamoto, Y., Hyodo, M., 2010. Geohazards induced by the Wenchuan earthquake. In: Williams, A.L., Pinches, G.M., Chin, C.Y., McMorran, T.J., Massey, C.I. (Eds.), Geologically Active. Taylor & Francis Group, London, pp. 77–84. Chen, X.L., Ran, H.L., Yang, W.T., 2012. Evaluation of factors controlling large earthquakeinduced landslides by the Wenchuan earthquake. Nat. Hazards Earth Syst. Sci. 12, 3645–3657. Chen, X.L., Yuan, R.M., Yu, L., 2013. Applying the Newmark's model to the assessment of earthquake-triggered landslides during the Lushan earthquake. Seismol. Geol. 35, 661–670 (in Chinese, with English Abstr.). Dai, F.C., Xu, C., Yao, X., Xu, L., Tu, X.B., Gong, Q.M., 2011. Spatial distribution of landslides triggered by the 2008 Ms 8.0 Wenchuan earthquake, China. J. Asian Earth Sci. 40, 883–895. General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Ministry of construction of the People's Republic of China, 1995. Standard for Engineering Classification of Rock Masses. Standards Press of China, Beijing (in Chinese). Guzzetti, F., Carrara, A., Cardinali, M., Reichenbach, P., 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy. Geomorphology 31, 181–216. Harp, E.L., Jibson, R.W., 1996. Landslides triggered by the 1994 Northridge, California earthquake. Bull. Seismol. Soc. Am. 86 (1B), s319–s332. Huang, R.Q., Li, W.L., 2008. A study on the development and distribution rules of geohazards triggered by “5.12” Wenchuan earthquake. Chin. J. Rock Mech. Eng. 27, 2585–2592 (in Chinese, with English Abstr.). Huang, R.Q., Li, W.L., 2009. Analysis on the number and density of landslides triggered by the 2008 Wenchuan earthquake, China. J. Geol. Hazards Environ. Preserv. 20 (3), 1–7 (in Chinese, with English Abstr.). Huang, R.Q., Fan, X.M., 2013. The landslide story. Nat. Geosci. 6, 325–326. Jibson, R.W., Harp, E.L., Michael, J.A., 2000. A method for producing digital probabilistic seismic landslide. Eng. Geol. 58, 271–289. Jibson, R.W., Harp, E.L., Schulz, W., Keefer, D.K., 2004. Landslides triggered by the 2002 M7.9 Denali Fault, Alaska, earthquake and the inferred nature of the strong shaking. Earthquake Spectra 20, 669–691. Jibson, R.W., Michael, J.A., 2009. Maps showing seismic landslide hazards in Anchorage, Alaska. U.S. Geological Survey Scientific Investigations Map 3077, scale 1:25,000, 11-p. pamphlet. Kaynia, A.M., Skurtveit, E., Saygili, G., 2010. Real-time mapping of earthquake-induced landslides. Bull. Earthq. Eng.. http://dx.doi.org/10.1007/s10518-010-9234-2. Keefer, D.K., 1984. Landslides caused by earthquakes. Geol. Soc. Am. Bull. 95, 406–421. Keefer, D.K., 2002. Investigating landslides caused by earthquakes — a historical review. Surv. Geophys. 23, 473–510. Khazai, B., Sitar, N., 2003. Evaluation of factors controlling earthquake-induced landslides caused by Chi–Chi earthquake and comparison with the Northridge and Loma Prieta events. Eng. Geol. 71, 79–95. Kieffer, D.S., Jibson, R., Rathje, E.M., Keith, K., 2006. Landslides triggered by the 2004 Niigata Ken Chuetsu, Japan, earthquake. Earthquake Spectra 22, S47–S73. Li, X.J., Zhou, Z.H., Huang, M., Wen, R.Z., Yu, H.Y., Lu., D.W., Zhou, Y.N., Cui, J.W., 2008. Preliminary analysis of strong-motion recordings from the magnitude 8.0 Wenchuan, China, earthquake of 12 May 2008. Seismol. Res. Lett. 79, 844–854. Meunier, P., Hovius, N., Haines, J.A., 2007. Regional patterns of earthquake-triggered landslides and their relation to ground motion. Geophys. Res. Lett. 34. http://dx.doi.org/ 10.1029/2007GL031337 L20408. Miles, S.B., Keefer, D.K., 2001a. Seismic Landslide Hazard for the Cities of Oakland and Piedmont, California. (http://pubs.usgs.gov/mf/2001/2379/oakpamph.pdf). Miles, S.B., Keefer, D.K., 2001b. Seismic Landslide Hazard for the City of Berkeley, California. (http://pubs.usgs.gov/mf/2001/2378/berkpamph.pdf.). National Institute for Land and Infrastructure Management, Ministry of Land, Infrastructure, Transport and Tourism, 2004. A Study on Methodology for Assessing the Potential of Slope Failures During Earthquakes [Report], Japan (in Japanese). Newmark, N.M., 1965. Effects of earthquakes on dams and embankments. Geotechnique 15, 139–160. Qi, S.W., Xu, Q., Lan, H.X., Zhang, B., Liu, J.Y., 2010. Spatial distribution analysis of landslides triggered by 2008.5.12 Wenchuan earthquake, China. Eng. Geol. 116, 95–108. Rodríguez, C.E., Bommerb, J.J., Chandlerb, R.J., 1999. Earthquake-induced landslides: 1980–1997. Soil Dyn. Earthq. Eng. 18, 325–346. Wang, H.B., Sassa, K., Xu, W.Y., 2007. Analysis of a spatial distribution of landslides triggered by the 2004 Chuetsu earthquakes of Niigata Prefecture, Japan. Nat. Hazards 41, 43–60. Wang, W.N., Wu, H.L., Nakamura, H., Wu, S.C., Ouyang, S., Yu, M.F., 2003. Mass movements caused by recent tectonic activity: the 1999 Chi–chi earthquake in central Taiwan. Island Arc 12, 325–334. Wang, Y.S., Luo, Y.H., Ji, F., Huo, J.J., Wu, J.F., Xu, H.B., 2008. Analysis of the controlling factors on geo-hazards in mountainous epicentre zones of the Wenchuan earthquake. J. Eng. Geol. 16, 759–763 (in Chinese, with English Abstr.). Wen, B.P., Wang, S.J., Wang, E.Z., Zhang, J.M., 2004. Characteristics of rapid giant landslides in China. Landslides 4, 247–261. Wilson, R.C., Keefer, D.K., 1985. Predicting areal limits of earthquake-induced landsliding. In: Ziony, J.I. (Ed.), Evaluating Earthquake Hazards in Los Angeles Region — An Earthscience Perspective. U.S. Geological Survey, Professional Paper, 1360, pp. 317–345.
22
X.-L. Chen et al. / Geomorphology 217 (2014) 15–22
Xu, Q., Li, W.L., 2010. Distribution of large scale landslides induced by the Wenchuan earthquake. J. Eng. Geol. 18, 818–826 (in Chinese, with English Abstr.). Xu, C., Xu, X.W., Yao, Q., Wang, Y.Y., 2013. GIS-based bivariate statistical modeling for earthquake-triggered landslides susceptibility mapping related to the 2008
Wenchuan earthquake China. Q. J. Eng. Geol. Hydrogeol.. http://dx.doi.org/10.1144/ qjegh2012-006. Yin, Y.P., Wang, F.W., Sun, P., 2009. Landslide hazards triggered by the 2008 Wenchuan earthquake, Sichuan, China. Landslides 6, 139–152.