Analysis of geological and geomorphological characteristics of landslides triggered by 2004 Chuetsu earthquake in Japan
Surangani BANDARA,1* Satoru OHTSUKA,2 Tsukasa IWABE, 3 Yasuyuki MIYAKI, 4 and Koichi ISOBE5
1 Nagaoka University of Technology (1603-1 Kamitomioka, Nagaoka, Niigata 940-2188 Japan) 2 Nagaoka University of Technology (1603-1 Kamitomioka, Nagaoka, Niigata 940-2188 Japan)
3 Kumamoto National College of Technology (2659-2 Suya Nishigoshi-machi, Kikuchi-Gun, Kumamoto, Japan) 4 Nagaoka University of Technology (1603-1 Kamitomioka, Nagaoka, Niigata 940-2188 Japan)
5 Hokkaido University (Kita 13, Nishi 8,Kita-ku,Sapporo,Japan)
*Corresponding author. E-mail: s125076@stn.nagaokaut.ac.jp
One of the most significant effects of the 23rd October 2004 Chuetsu, Mid Niigata prefecture earthquake (M6.8) was the triggering of the thousands of landslides over a wide-ranging area, including surface failure (), shallow landslide () and deep-seated landslide (). A great number of houses collapse in town of Kawaguchi by this earthquake. A very large number of landslides occurred in the upland village of Yamakoshi, destroying the entire village. Slope failures and landslides trigged by the Mid Niigata prefecture earthquake have been previously analyzed from the various viewpoints such as topography, geology and forest or field vegetation however geological and geomorphological characteristics of landslides have not been studied sufficiently. The purpose of this study was to identify correlations between landslide occurrences with geologic and geomorphologic conditions using two indexes based on Geographical Information System (GIS). Total collapse area and landslide occurrence ratio (LOR), defined as the percentage of the area affected by landslides were used to analyze the relations of the distribution of these landslides using various parameters, such as distance from an earthquake source (epicenter fault line), topographic parameters (slope steepness, slope morphology) and geological units and distance from rivers and ponds.
Key words: Landslide occurrence ratio (LOR), Shallow landslide, Deep-seated landslide, and Surface failure
1.INTRODUCTION
At 17:56 Japanese Standard Time (JST) on 23rd October 2004, a catastrophic earthquake with magnitude of 6.8 stuck Mid-Niigata Prefecture in Japan. It caused many landslides and slope failures in the Chuetsu region of Niigata prefecture. The hypocenter of the main shock was located at
; , in a depth of .
The strong ground motion, which is triggered by the earthquake, caused a vast damage in mountainous areas including Ojiya city, Nagaoka city and Yamakoshi village due to the occurrence of several landslides. The earthquake claimed 40 lives and injured 4,496 people while 2,770 houses and buildings were damaged (Fire and Disaster Management Agency of Japan 2004).
Study on the spatial distribution of landslides which are triggered by an earthquake is a vital subject for understanding which areas could be more
prone to landsliding in imminent tremors. In this regard, many researchers have analyzed the correlations of landslide occurrence with slope steepness, distance from the earthquake source, and rock types or geology respectively (e.g., Keefer, 2000; Parise and Jibson, 2000; Wang et al., 2007).
Landslides which triggered by the 1989 Loma Prieta earthquake (Mw 6.9) have been studied by Keefer.
Keefer (2000) has mapped 1280 landslides and pointed out that the landslides which induced by an earthquake are possibly correlated with slope steepness, distance from a source, and rock type.
Harp and Jibson (1996) have identified more than 11,000 landslides which had been triggered by the 1994 Northridge earthquake (Mw 6.7), California.
Accordingly, the most common types of landslide triggered by the earthquake were highly disrupted, shallow falls and slides of rock and debris. Parise and Jibson (2000) have described landslide morphologies by computing simple morphometric
Table 1 Summary of landslide collapse area Category Number of
landslides
Total collapse area(m2)
Average collapse area(m2)
Deep 141 2,710,406 19,223
Shallow 609 1,253,243 2,058
Surface 4504 2,621,340 582
parameters (i.e. area, length, width, aspect ratio, slope angle) and statistically quantified and ranked the susceptibility of each geologic unit to seismically induced landslides.
Wang et al. (2002,2003a,b) and Chigira et al.
(2003) have identified nearly 10,000 landslides, which triggered by the 1999 Chi–Chi earthquake (Mw 7.5), from SPOT images. According to the findings, the distribution of landslides revealed a significant correlation with epicentral distance and the rock type. Moreover, the geological features of deep-seated landslides and relatively smaller slides were described locally in Taiwan.
Qi et al. (2010) referring to a spatial database of landslides which covers 11 countries severely damage by Wenchuan earthquake with area of about pointed that the distribution of landslides which are triggered by the earthquake have mainly depended on the distance to the causative faults and slope gradient.
Basharat et al. (2013) have analyzed the relationship of the distribution of the mass movements triggered by the 2005 Kashmir earthquake (Mw 7.6) with several parameters such as distance from an earthquake source (epicenter and fault), slope steepness, slope aspect, elevation and geological units. The results revealed that the mass movement concentration principally depends on the distance from the earthquake source whereas, the topographic parameters and geological units plays subsidiary roles in the distribution of mass movements.
2.GEOLOGICAL SETTING OF THE STUDY AREA
The study area includes Yamakoshi village and its vicinity (Fig.1). Yamakoshi village, where many landslides have occurred is located on the Higashiyama hills and its adjacent alluvial plain, where the Shinano River runs from South West to North East then turns to North West to North after merging with the Uono River. The Imo River flowing from North to South bisects the central part of the hills. The epicenter area consists with many anticlines and synclines. The axes from West to East are Higashiyama anticline, Konpira syncline, Toge anticline, Kajikane syncline, and Komatsuguru anticline respectively (Fig.2) (Chigira and Yagi, 2006). Higashiyama mountain area has been regarded as one of the most landslide-prone zones in Japan. Rocks and strata of the study area were classified into seven units: igneous rock, sandstone, mudstone, siltstone, sand, gravel and alternating beds of mudstone and sandstone.
3.METHODOLOGY
Landslides classifications based on morphology, substantial, mechanism of initiation, or other
Fig. 2 Distribution of earthquake-triggered landslides
Fig. 1 Index map
principles have been proposed by many authors.
Varnes (1978) and Keefer (1984) were classified landslides which induced by earthquake in to 14 categories based on the principles and terminology.
In this paper, we classified the landslides in to three different types according to their depth of failure such as surface failure (), shallow landslide () and deep-seated landslides (). Analysis was performed for three different types of landslides which triggered by the 2004 Chuetsu earthquake. ArcGIS 9.3 was employed to analyze the correlation of landslides with respect to the distance from the earthquake source fault line, topographic parameters (slope angle, slope morphology), rocks and strata of the study area and distance from rivers and ponds. In the study area, the Landslide Occurrence Ratio (LOR), which is defined as the percentage of the area affected by landslides, and total collapse area, were used to determine correlations between landslide occurrence and geologic, geomorphologic conditions.
4.EARTHQUAKE CHARACTERISTICS AND DISTRIBUTION OF LANDSLIDES
The main shock of earthquake in the Mid Niigata prefecture occurred at 17:56 (JST) on October 23rd 2004, with a 6.8 magnitude and many aftershocks including four aftershocks of magnitude 6 or greater.
Table 1 depicts the summary of the total collapse area of three different kinds of landslides. Even though the number of deep-seated landslide is small, the total collapse area is nearly equal to surface failure collapse area.
5. STATISTICAL ANALYSIS OF LAND SLIDE DISTRIBUTION
5.1 Collapse area and LOR in view point of distance from epicenter fault line
Figure 3 indicates distance definition, which used to calculate LOR and total collapse area with respect to the distance from epicenter fault line. It may be seen clearly that the occurrence ratio of surface failure reaches the highest at from the epicenter fault line and drops off at
-ve ve
Epicenter
Hypocenter
Fig. 3 Distance definition
ve Positive Negative -ve
0 0.5 1 1.5 2 2.5 3 3.5 4
0 1 2 3 4 5 6 7 8
-6.5 -6 -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1
-0.5 0.5
1 1.5 2 2.5 3 3.5 4 4.5 LOR(%)
Collapse area X 105(m2)
Distance from source fault line(km)
Surface failures LOR(%)
Area
Fig. 4 Surface failure occurrence ratio and collapse area in viewpoint of epicenter fault line
0 0.5 1 1.5 2 2.5 3 3.5 4
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
-6 -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1
-0.5 0.5
1 1.5 2 2.5 3 3.5 4 4.5 LOR(%)
Collapse areaX105(m2)
Distance from source fault line(km) Shallow landslides
Area LOR(%)
Fig. 5 Shallow landslide occurrence ratio and collapse area in viewpoint of epicenter fault line
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0 1 2 3 4 5 6 7 8 9
-6 -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0.5 1 1.5 2 2.5 3 3.5 LOR(%)
Collapse area X 105(m2)
Distance from source fault line(km) Deep seated landslides LOR(%)
Area
Fig. 6 Deep-seated landslide occurrence ratio and collapse area in viewpoint of epicenter fault line
and steadily decreases from . However, LOR values do not, as expected display a relative simple negative correlation in the negative distance zone.
Figure 5 outlines shallow LOR and collapse area with respect to epicenter fault line. It can be seen that the LOR increases sharply until reached its maximum at then after monotonically decreases. As same as surface failure, here also dramatically drops off at about , then sharply increases until . Deep-seated LOR reaches the highest near the epicenter. However, LOR is relatively higher at away from the epicenter.
5.2 Collapse area and LOR in viewpoint of slope angle
The slope angle was extracted based on the
Digital Elevation Model (DEM) at resolution and reclassified at intervals of . In general, steeper and higher slopes have higher susceptibility for landslide occurrences, even when
the slope failures are not triggered by an earthquake.
Figure 7 shows surface failure occurrence ratio and total collapse area with respect to slope angle.
Almost of total collapses have occurred within the category, however, the occurrence ratio shows an increasing trend till the slope angle reaches its maximum in the category of , and then decreases a little. The absence of landslides in steeper slope categories is mainly due to the small number of such cells. This result implies that slope angle is strongly correlated with surface failures which induced by the earthquake. Figure 8
0 1 2 3 4 5 6 7 8
0 5 10 15 20 25
5 10 15 20 25 30 35 40 45 50 55 60 65
Collapse area X 105(m2)
LOR
Slope angle(0) Surface failures
Surface failiure LOR(Surface)
Fig. 7 Surface failure occurrence ratio and collapse area in viewpoint of slope angle
0 2 4 6 8 10 12
0 1 2 3 4 5 6 7 8 9
5 10 15 20 25 30 35 40 45 50 55 60 65
Collapse area X 105(m2)
LOR
Slope angle(0)
Deep-seated landslide
Deep LOR(Deep)
Fig. 9 Deep-seated landslide occurrence ratio and collapse area in viewpoint of slope angle
0 1 2 3 4 5 6 7 8
0 10 20 30 40 50 60 70 80 90 100
-6.5 -6.0 -5.5 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
LOR(%)
Slope angle occurance ratio(%)
Distance from source fault line
65° 60° 55° 50°
45° 40° 35° 30°
25° 20° 15° 10°
5° LOR
Fig. 10 Total slope angle distribution and surface failure occurrence ratio according to the distance from source fault line
0 0.5 1 1.5 2 2.5 3
0 0.5 1 1.5 2 2.5 3
5 10 15 20 25 30 35 40 45 50 55 60 65 52Collapse area X 10(m)
LOR
Slope angle(0)
Shallow landslides Shallow
LOR(Shallow)
Fig. 8 Shallow landslide occurrence ratio and collapse area in viewpoint of slope angle
0 1 2 3 4 5 6 7 8
0 10 20 30 40 50 60 70 80 90 100
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
LOR(%)
Geolog occurance ratio(%)
Distance from sourc fault line(km) Others
Alternating beds of mudstone and sandstone Sand
Siltstone Mudstone LOR(Deep)
Fig. 14 Landslide occurrence ratio and collapse area in viewpoint of geology types
shows shallow LOR and total collapse area with respect to slope angle. The percentage share of total collapse occurrences in the category is higher as . Shallow landslide occurrence ratio reached the highest at 0 slope angle.
Generally, almost all the deep-seated landslides have occurred at steep slopes with the slope angle of . Meanwhile, more than of the deep-seated landslides have an original slope angle of (Fig.9). In respect to ordinary landslides, deep-seated landslides are more prone to be occurred at steeper slopes. Figure 10 shows surface failure occurrence ratio with respect to distance from source fault line and slope angle. Accordingly, the surface failure occurrence ratio is very low in negative distance zone. This may be due to the higher ratio of gentle slope in that area (Bandara and Ohtsuka., 2014).
5.3 Collapse area and LOR in viewpoint of distance from rivers and ponds
The detailed study area has been selected
considering the highest concentration of landslides.
The area consist of Imo river where step farming paddy fields and fish ponds for carp feeding have formed regional settings and increased the risk of slope failure exposures due to penetration of surface water in to soils or rocks.
An enormous landslide has occurred on the left bank of the Imo River displacing a huge soil mass, which blocked the river in the Yamakoshi village. In this analysis, we assume that the ground water level is very high near rivers and ponds. It can be seen from Figures 11, 12 and 13 respectively that several landslides have occurred close to rivers and ponds.
5.4 Collapse area and LOR in viewpoint of types of rocks and strata
The study area is underlain by Miocene to Quaternary sedimentary rocks. It consists of a siltstone, sandstone, sand, alternating beds of mudstone and sandstone, mudstone, igneous rock and gravel. Figure 14 represents deep-seated LOR and area of each rocks types and strata against the distance from the epicenter fault line. It can be seen that mudstone unit increases until from epicenter fault line .It does not exist from
Fig. 11 Surface failure occurrence ratio and collapse area in viewpoint of distance from rivers and ponds
viewpoint of distance from rivers and ponds
Fig. 12 Shallow landslide occurrence ratio and collapse area in viewpoint of distance from rivers and ponds
8 100
Fig. 13 Deep-seated landslide occurrence ratio and collapse area in viewpoint of distance from rivers and ponds
to and again starts to appear from . Furthermore, it can be seen that area of siltstone increases from to . Compared with other two types of landslides, rocks types and strata are significant for occurrence of the landslides with large scale of volume like deep-seated landslides.
6. GEOLOGICAL STRUCTURE PREDICTION OF THE STUDY AREA
Naturally, beds and rocks are deposited horizontally (more or less) but due to the gravity and other external forces, different bedding plane may dip in any direction. We can determine the dip of a bedding plane using geologic data obtained from a detailed map using it’s contour lines and symbols.
The orientation of a tilted bedding plane can be described by two angular measurements. Strike is an imaginary horizontal line on a plane measured from the North. It is usually given in degrees. Dip means the angle of the ‘slope’ to the plane. The data is recorded on a map by a geologist using a line drawn according the angle of the strike.
A dip is indicated by a short line perpendicular to the strike. Taken together, the strike and dip of a tilted bedding plane describes its spatial orientation.
Dip and strike are representing together with T-shaped mark in a map but it is complicated (Fig.15). Hence, in this study, we introduced a three-dimensional gradient vector for strike formation where θ is tilt angle, is tilt direction
and is normal vector Eqs.
Normal vector e represent by using tilt angle and tilt direction (Fig.16).
1
2 3 4
Fig. 17 Observation points and predict point Landslide area
Strike and dip
φ
N
Z e
θ θ
N
Dip
Strike
Fig. 15 Strike and dip of the bedding plane e
θ
Bedding pl ane
Fig. 16 Components of the normal vector ponents of Bedd
ing pl ane
Table 2 represents the correlation between the true value and predicted value with vector representation and without respectively. And also compares the average value of the inner product of true and predicted values. Average value of inner product shows a significant correlation coefficient for tilt angle and tilt direction.
For the geological structure prediction an arbitrary point, we have considered direction, number of observation points and distance relationship of the observation points (by using weight function). In this analysis, we considered four observation points in each direction (Fig.17). In order to consider distance to the observation point, we introduced ‘Z’ as a weighting function. Tilt direction, tilt angle and normal vector were calculated using Eqs.2a and 2b. Finally, verified the validation of the (direction – dip) geological structure of the target area. For the verification we used 596 individual observation points within the target area. In this analysis, we classified hill slopes
by considering inner product of (two-dimensional unit vector in both cases) strata inclination direction and terrain slope direction of the slope. Figure 18 indicates the classification of hill slopes.
Figure 19 represents the relationship of LOR and collapse area with respect to slope morphology.
According to the obtained results, the impact of slope morphology for the occurrences of surface failure is very low. LOR of shallow and deep-seated landslides in dip slopes is higher than reverse dip slopes. However, pre-exist landslides influence for occurrence of shallow and deep-seated landslides is higher than slope morphology.
7. DISCUSSION
In this analysis, LOR was defined to express the influence of landslide occurrence: which is expressed as a percentage of the area affected by landslide activity. Landslides were categorized according to their depth. Variation of LOR with the distance from earthquake source was analyzed with respect to epicenter fault line. According to the Fig.10, light and dark colour solid bars depicts the gentle slopes () and steep slopes ( ) occurrence ratio in the study area respectively.
In the study area, slopes steeper than occupied
Dip slope Reverse dip slope Horizontal dip
α α α α: 0°~ 180°
0°< α ≤ 85° Dip slope
85°< α ≤105° Horizontal dip slope 105° α ≤180° Reverse dip slope
Bedding plane Slope
Fig 18 Classification of hill slopes
Stratum in clinatio
n dire ction ope
Slo pe di
rect ion
Table 2 Comparison between true value and vector prediction value
Correlation coefficient Vector representation Yes No
φ 0.798 0.777
θ 0.802 0.812
The average value of the inner product
0.967 0.957
Fig. 19 Landslide occurrence rate and collapse area in view point slope morphology
a smaller area whereas slopes ranging from to covered a much larger area. Most of the deep seated and shallow landslides have occurred on slope angles between However, of surface failures have occurred on slopes between . In Chuetsu earthquake most commonly, landslides have occurred on slope angle lower than . In this analysis, we classified landslide slopes in to three as dip slope, revers dip slope and horizontal dip slope (considering the strata inclination direction and terrain slope direction), and examined the LOR with respect to slope morphology.
Dip slopes are more prone to the occurrence of shallow landslides. Surface failures showed the highest occurrence ratio in horizontal dip slopes.
However, the difference of LOR in dip slopes and revere dip slopes is insignificant. It could be seen that deep-seated landslides occurrence ratio was higher in dip slopes than other slopes.
8.CONCLUSION
Analysis of the landslides triggered by the 2004 Chuetsu earthquake provides valuable insights to the characteristics of seismically triggered landslides in Chuethsu region of Niigata prefecture. Even though the most common type of landslides recorded in the study area was surface failure on steep slopes, this event have triggered 141 deep-seated landslides and 609 shallow landslides also. Most of the shallow landslides and surface failures have occurred within the positive distance zone. It can also be seen from Fig. 10 that the occurrence ratio of surface failure is very high in the steep slope. This result implies that the slope angle is strongly correlated with the occurrence of surface failures induced by earthquake.
Influence of preexisting landslides, slope morphology and geology are not significant for occurrence of surface failure. According to the Fig.
14, siltstones and mudstones units have played an imperative role in the generation of deep-seated landslides which are far from the epicenter. It can be seen from Fig. 19 that most of the deep seated and shallow landslides have been occurred due to the reactivation of existing landslides. Dip slopes have more pronounced (shallow and deep seated)
landslide occurrence ratio compared to the other slopes.
REFERENCES
Bandara, K.M.S. and Ohtsuka, S. (2014): Spatial distribution of landslides triggered by the 2004 Mid Niigata prefecture earthquake in Japan, The 14th International conference of the International Association for Computer Methods and Advances in Geomechanics, Kyoto, Japan,Vol.,pp..
Basharat, M., Rohn, J., Baig, M.S. and Khan, M.R. (2013):
Spatial distribution analysis of mass movements triggered by the 2005 Kashmir earthquake in the Northeast Himalayas of Pakistan, Geomorphology, Vol. 206, pp. 203-214.
Chigira, M., Wang, W.N., Furuya, T. and Kamai, T. (2003):
Geological causes and geomorphological precursors of the Tsaoling landslide triggered by the 1999 Chi-Chi Earthquake, Taiwan, Engineering Geology 68, pp.259–273.
Chigira, M. and Yagi, H.(2006):Geological and geomorphologi cal characteristics of landslides triggered by the 2004 Mid Niigta prefecture earthquake in Japan Engineering Geology, 82 , pp. 202–221.
Harp, E.L. and Jibson, R.W. (1996): Landslides triggered by the 1994 Northridge, California, earthquake, Bulletin of the Seismological Society of America 86, pp. 319–332.
Keefer, D.K. (1984): Landslides caused by earthquakes Geolo gical Society of America Bulletin 95, pp. 406–421.
Keefer, D.K., (2000): Statistical analysis of an earthquake-induc ed landslide distribution – the 1989 Loma Prieta, California
event, Engineering Geology 58, pp. 213–249.
Parise, M. and Jibson, R.W. (2000): A seismic landslide susc eptibility rating of geologic units based on analysis of characteristics of landslides triggered by the 17 January, 1994 Northridge, California earthquake, Engineering Geology 58, pp. 251–270.
Qi, S., Xu, Q., Lan, H., Zhang, B. and Liu, J. (2010): Spatial distribution analysis of landslides triggered by 2008.5.12 Wenchuan Earthquake, China, Engineering Geology 116, pp. 95–108.
Varnes, D.J. (1978): Slope movement types and processes, Land slide Analysis and Control, National Research Council, Transportation Research Board, Washington, pp.11–13.
Wang, W.N., Furuya, T. and Chigira, M. (2003): Geomorpho logical precursors of the Chiu- fen- erh-shan landslide triggered by the Chi-chi earthquake in Central Taiwan, Engineering Geology 69, pp. 1–13.
Wang, W.N., Wu, H.L., Nakamura, H., Wu, S.C., Ouyang, S.
and Yu, M.F. (2003): Mass movements caused by recent tectonic activity: the 1999 Chi-Chi earthquake in central Taiwan, The Island Arc 12, pp. 325–334.
Wang, W.N., Nakamura, H., Tsuchiya, S. and Chen, C.C.
(2002): Distributions of landslides triggered by the Chi-Chi earthquake in Central Taiwan on September 21, 1999, Journal of the Japan Landslide Society 38 (4), pp.318–326.
Wang, H.B., Sassa, K. and Xu, W.Y. (2007). Analysis of a spatial distribution of landslides triggered by the 2004 Chuetsu earthquakes of Niigata Prefecture, Japan, Natural Hazards 41(1), pp. 43–60.
