- Email: [email protected]
Susceptibility assessment of debris flows using the analytic hierarchy process method − A case study in Subao river valley, China
Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
Contents lists available at ScienceDirect
Journal of Rock Mechanics and Geotechnical Engineering journal homepage: www.rockgeotech.org
Full length article
Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China Xingzhang Chen a, *, Hui Chen a, Yong You b, Jinfeng Liu b a b
School of Environment and Resource, Southwest University of Science and Technology, Mianyang 621010, China Institute of Mountain Hazards and Environment, Chinese Academy of Sciences, Chengdu 610041, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 December 2014 Received in revised form 15 April 2015 Accepted 18 April 2015 Available online xxx
Many debris flows have occurred in the areas surrounding the epicenter of the Wenchuan earthquake. Susceptibility assessment of debris flows in this area is especially important for disaster prevention and mitigation. This paper studies one of the worst hit areas, the Subao river valley, and the susceptibility assessment of debris flows is performed based on field surveys and remote sensing interpretation. By investigating the formation conditions of debris flows in the valley, the following assessment factors are selected: mixture density of landslides and rock avalanches, distance to the seismogenic fault, stratum lithology, ground roughness, and hillside angle. The weights of the assessment factors are determined by the analytic hierarchy process (AHP) method. Each of the assessment factors is further divided into five grades. Then, the assessment model is built using the multifactor superposition method to assess the debris flow susceptibility. Based on the assessment results, the Subao river valley is divided into three areas: high susceptibility areas, medium susceptibility areas, and low susceptibility areas. The high susceptibility areas are concentrated in the middle of the valley, accounting for 17.6% of the valley area. The medium susceptibility areas are in the middle and lower reaches, most of which are located on both sides of the high susceptibility areas and account for 45.3% of the valley area. The remainders are classified as low susceptibility areas. The results of the model are in accordance with the actual debris flow events that occurred after the earthquake in the valley, confirming that the proposed model is capable of assessing the debris flow susceptibility. The results can also provide guidance for reconstruction planning and debris flow prevention in the Subao river valley. Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by Elsevier B.V. All rights reserved.
Keywords: Debris flows Susceptibility assessment Geographic information system (GIS) Analytic hierarchy process (AHP) method Wenchuan earthquake Subao river valley
1. Introduction The Wenchuan earthquake, which occurred on 12 May 2008, caused more than 15,000 geohazards, the majority of which were rock avalanches and landslides (Huang and Li, 2008). Such landslides and avalanches provide a considerable amount of loose materials for debris flows, which are expected to be more frequent and in the active period of the next 20e30 years (Cui et al., 2011). Thus, the susceptibility assessment of debris flows is of great significance to disaster prevention and mitigation in earthquakeprone areas. Assessment factors and methods are vital to assessing the debris flow susceptibility. Li et al. (2006, 2009a) demonstrated that the events of debris flow correlate with particular stages of the basin
* Corresponding author. Tel.: þ86 13881169780. E-mail address: [email protected] (X. Chen). Peer review under responsibility of Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. 1674-7755 Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jrmge.2015.04.003
evolution. The rock hardness determines the ability to provide loose materials and then influences the occurrence frequency of debris flows (Lu et al., 2011). Zhang et al. (2008) ranked the common rocks based on their susceptibility to debris flow in ascending sequence, i.e. basalt, argillaceous limestone, dolomite, slate, Quaternary deposit, sandstone, siltstone, phyllite, and mudstone. Tang and Liang (2008) argued that phyllite and slate were especially prone to forming debris flows in the Beichuan earthquake area. With regard to the assessment method, based on a conditional probability model, Tang (2005) and Zou et al. (2012) analyzed the spatial susceptibility of debris flow in the Nujiang River basin of Yunnan and the Upper Yangtze River basin, respectively. Carrara et al. (2008) developed and compared five models (statistical approach vs. physically based approach) for predicting the debris flow occurrence in an alpine area. They found that all of the statistical models were reliable and robust, while the physically based model had low predictive power. Then, Luca et al. (2011) presented an evolution of gullying susceptibility using bivariate and multivariate statistical models, and found the latter had more predictive power in gullying susceptibility. After the Wenchuan earthquake, many researches were also conducted on the susceptibility and
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
2
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
activities of debris flow in earthquake areas (Liu et al., 2009; Tang et al., 2012; Wang et al., 2013; Liu et al., 2014). The Subao river valley is located in the worst hit area of the Wenchuan earthquake. The Beichuan reverse fault, considered as the seismogenic fault (Burchfiel et al., 2008; Li et al., 2008, 2009b; Xu et al., 2008), passes through the valley. The earthquake induced many landslides and rock avalanches. On 24 September 2008, a rainstorm triggered a group of large debris flows in the valley, causing 11 deaths and significant damages (Tang and Liang, 2008; Tang et al., 2009; You et al., 2010). After that, several debris flows occurred every year in the valley, seriously threatening the postdisaster reconstruction. The valley has become one of the highest susceptibility areas of debris flow. This paper describes the formation conditions of debris flows in the Subao river valley, based on field surveys and remote sensing images. Then, we build a model based on the ArcGIS platform to assess the debris flow susceptibility in the valley area. 2. Study area 2.1. Valley overview The Subao river valley is located to the west of Leigu town, approximately 8 km south of Beichuan county (Fig. 1). The valley covers an area of 72.2 km2, and the river is 16.5 km long. The river, in the middle of the catchment area, flows from west to east. In the valley, the highest point is 2312 m and the relative relief is 1597 m, with a mean relief ratio of 9.28%. The channel is mainly U-shaped, with a width varying from a few meters to several tens of meters. 2.2. Topography Mountains dominate the Subao river valley, which is in the transition area between the front and back of the Longmen Mountains. The valley head is approximately 2312 m wide, falling rapidly to 715 m at the valley mouth. The main channel has a steep gradient with a mean relief ratio of 9.28%. The valley is dominated by steep slopes (Fig. 1). The slope angles are greater than 25 , with some up to 40 e50 or more. The area of steep (25 e35 ) and
steeper (>35 ) slopes is 46.76 km2, accounting for up to 64.76% of total slope area. Table 1 lists the slope statistic data of the Subao river valley acquired from 1:50000 digital elevation model (DEM). 2.3. Geological settings Fig. 2 shows the Beichuan reverse fault passing through the Subao river valley. The strike of the fault plane is almost northeast in direction (Dong et al., 2006). The motion of this fault is responsible for the uplift of the mountains relative to the lowlands of Sichuan Basin to the east. The stratum lithology is controlled by the Beichuan reverse fault. The hanging wall in the northwest is mainly composed of Silurian metamorphic rocks, which originated from regional metamorphism with a foliated texture and numerous cracks. The metamorphic rocks, distributed in the middle and upper catchment areas, are composed of phyllite, slate, sandstone, and other similar rocks. The footwall in the southeast is mainly composed of shallow sea phase sedimentary rocks of the middle Devonian period, including limestone, argillaceous limestone, bioclastic limestone, and siltstone. The footwall also contains a small amount of Cambrian siltstone. The rocks of the footwall disperse farther downstream, where there are steep slopes with hard rocks that are resistant to erosion. According to field surveys and remote sensing images (with a resolution of 0.5 m), it is determined that the Wenchuan earthquake directly induced more than 300 landslides and rock avalanches in the Subao river valley (Fig. 2). Landslides mainly developed in softer rock areas (such as the hanging wall) and rock avalanches mainly occurred in hard rock areas (such as the footwall). These geohazards continue to provide enough loose materials for subsequent debris flows and will last for several decades. 2.4. Climate The Subao river valley has a mild subtropical humid monsoon climate with four distinct seasons and an average annual temperature of 15.6 C. The valley has rich rainfalls with an average annual precipitation of 1399.1 mm. The 24-h maximum precipitation was 101 mm, and the 1-h maximum precipitation was 42 mm. Approximately 80% of the rain falls between June and September. Rainstorms, combined with topographic effects, can result in debris flows in the rainy season. 3. Assessment factors The determination of assessment factors should consider the debris flow formation conditions in the valley. Generally, the formation conditions of debris flow include loose materials, topography and precipitation. 3.1. Loose materials Loose deposits form an important material basis of debris flow events. Generally, the loose materials are determined by the stratum lithology and geological structures. Table 1 Slope statistic data of the Subao river valley.
Fig. 1. The location and topographic features of the Subao river valley.
Slope angle ( )
Area (km2)
<15 1525 2535 >35
7.64 17.81 22.76 24
Total
72.21
Percentage (%) 10.58 24.66 31.52 33.24 100
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
3
considered as an assessment factor. Based on the DEM data, the values of ground roughness in the Subao river valley can be obtained using surface analysis tools in ArcGIS9.3 (Fig. 3d). 3.3. Precipitation
Fig. 2. Geological map and geohazards distribution of the Subao river.
The stratum lithology controls the ability to form loose materials, and affects the formation of debris flows. Tang and Liang (2008) studied the relationship between debris flows and rock types after the Wenchuan earthquake in Beichuan County. They found that the debris flow formation is sensitive to rock types, and the sensitivity from high to low is Silurian phyllite, Cambrian siltstone, and Devonian and Carboniferous limestones. Geological structures, especially the Beichuan reverse fault, have an important effect on producing loose materials. During the Wenchuan earthquake, more loose materials were formed close to the Beichuan reverse fault. Therefore, the stratum lithology and the distance to the Beichuan reverse fault should be considered as two important assessment factors. Landslides and rock avalanches are important transportation ways for loose materials. The landslides and rock avalanches, induced by the Wenchuan earthquake, are of great significance for debris flow formation. These geohazards can supply the debris flows with enough loose materials, which should be considered as an important assessment factor. The geohazards density can represent the loose materials distribution. Based on the 300 geohazards interpreted from a remote sensing image (with a resolution of 0.5 m), we can obtain the geohazards density of the Subao river valley using the kernel density tool in ArcGIS9.3. Therefore, three assessment factors for loose materials are selected, i.e. density of landslides and rock avalanches, distance to the seismogenic fault, and stratum lithology.
3.2. Topography The topography has a direct impact on debris flow formation through its influence on conflux process. Steeper slopes result in a faster speed of flow; both slopes below and above ground surface are prone to form debris flows. It is accepted that slopes steeper than 25 can provide loose materials for debris flows and are conducive to affluxion. In the valley, approximately 64.76% of the total area contains steep slopes (>25 ) and is therefore prone to form debris flows. The hillside angle should be considered as an important assessment factor. In addition to the hillside angle, ground roughness also affects the formation of debris flows. The ground roughness denotes the ratio of surface area to the projective area for a given area. It directly affects the ability of ground confluence and seepage. As one of the topographic factors, the ground roughness of the valley should be
Rainfall, a triggering factor of debris flows, is significant to the formation of debris flows. In the rainy season, rainstorms can directly induce debris flows in the Subao river valley due to the effects of the Wenchuan earthquake. However, the valley only covers an area of 72.2 km2, and the rainfall has no obvious difference in the whole valley. Therefore, the rainfall is very important for debris flow formation, but it is just a background condition and not suitable to be an assessment factor. According to the above analysis, we select five assessment factors. Based on their contribution to debris flow formation, they are density of landslides and rock avalanches (x1), distance to the seismogenic fault (x2), stratum lithology (x3), ground roughness (x4), and hillside angle (x5). Each of them is divided into five grades, as shown in Fig. 3. 4. Assessment method The selected factors are diverse sources of information and difficult to compare to each other. To integrate them into a consistent system for assessment, we define numerical weights of all factors with the analytic hierarchy process (AHP) method. One of the most prominent features of the AHP method is its ability to evaluate quantitative as well as qualitative criteria and alternatives on the same preference scale. The method is widely used in hazard evaluation, including natural hazard assessment (Nefeslioglu et al., 2013), landslide susceptibility evaluation and mapping (Komac, 2006; Nandi and Shakoor, 2009; Kayastha et al., 2013; Shahabi et al., 2014), debris flow risk degree assessment (Yang et al., 2011), single debris flow risk assessment (Tie and Tang, 2006), and decisions for regional debris flow prevention (Ma et al., 2009). In this paper, we weight the assessment factors using the AHP method. The AHP method, pioneered by Saaty in the 1970s, is a structured technique for organizing and analyzing complex decisions based on mathematics and psychology. It is used around the world in a wide variety of decision situations (Vaidya and Kumar, 2006; Saaty, 2008). A detailed description of the AHP method is available in Saaty (1980). The procedure for using the method can be summarized as follows (Saaty, 2008): (1) Define the problem and determine the type of knowledge sought. (2) Structure the decision hierarchy from the top, with the goal being the decision, and the objectives branching down through the intermediate levels to the lowest level. (3) Construct a set of pairwise comparison matrices. Each element in an upper level is compared with the elements in the level immediately below it. (4) Check the consistency of the judgments. The consistency ratio (CR) should be calculated for consistency with the evaluation matrix and the value of the CR should be no higher than 0.1. (5) Use the priorities obtained from the comparisons to weigh the priorities in the level immediately below. Do this for every element. Then, for each element in the level below, add its weighted values and obtain its overall or global priority. The pairwise comparison employs an underlying nine-point recording scale to rate the relative preference on a one-to-one
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
4
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
Fig. 3. Figures of the five assessment factors, each of them divided into five grades. (a) Density of landslides and rock avalanches, (b) Distance to the seismogenic fault, (c) Stratum lithology, (d) Ground roughness, and (e) Hillside angle.
basis of each factor. Table 2 shows a linguistic expression to each corresponding numerical value (Kamp et al., 2008). There are many regional evaluation methods, such as the grey correlation method, the fuzzy comprehensive evaluation method, and the weighted average method. According to the existing research results, the multifactor superposition method is a popular, simple and effective one (Liang et al., 2011; Li et al., 2012), and its calculation method is simple and clear. The calculation result is a summation of the products of factor weight and factor score. The factor weights are determined by the AHP method. The calculation method is expressed as follows:
Y ¼
m X
ðai xi Þ
(1)
i¼1
where Y denotes the summation of the products of factor weight ai and factor score xi. 5. Factors system and assessment model 5.1. Factors system The weights for each factor are calculated by a pairwise comparison matrix. The quality of the comparison is described by the
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7 Table 2 Pairwise comparison nine-point rating scale.
5
Table 4 Assessment factor system of debris flow susceptibility.
Importance Definition
Explanation
Assessment factor
Grades
Weight
Score
Value
1 3 5 7
Contribution to objective is equal Attribute is slightly favored over another Attribute is strongly favored over another Attribute is very strongly favored over another Evidence favoring one attribute is of the highest possible order of affirmation When compromise is needed
Density of landslides and rock avalanches (x1)
Very high High Medium Low Very low <1 km 12 km 23 km 34 km >4 km Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 Very high High Medium Low Smooth 25 25 e35 35 e45 45 e55 >55
0.511
0.458 0.257 0.15 0.087 0.049 0.432 0.263 0.166 0.089 0.05 0.458 0.257 0.15 0.087 0.049 0.458 0.257 0.15 0.087 0.049 0.046 0.082 0.146 0.261 0.465
0.234 0.131 0.076 0.044 0.025 0.114 0.07 0.044 0.024 0.013 0.059 0.033 0.019 0.011 0.006 0.027 0.015 0.009 0.005 0.003 0.002 0.003 0.005 0.009 0.017
9
Equal importance Moderate importance Strong importance Very strong importance Extreme importance
2, 4, 6, 8
Intermediate values
Distance to seismogenic fault (x2)
Stratum lithology (x3)
CR, which is a ratio between the matrix’s consistency index (CI) and random index (RI). A CR value below 0.1 indicates that the comparison matrix is consistent. The pairwise comparison matrix and factor weights are shown in Table 3. The density of landslides and rock avalanches is the most heavily weighted factor, followed by distance to seismogenic fault and stratum lithology. Hillside angle is the least influential parameter, while ground roughness had even less influence. The results verify the view that density of geohazards and distance to seismogenic fault are the most influential event-controlling parameters for debris flows in the Wenchuan earthquake area. The consistency ratio (CR) of the judgment matrix is 0.04, far less than 0.1, which means the consistency of the judgments is very good. The scores for each grade and the weights of the factors are determined by the AHP method, as shown in Table 4. The value is the product of factor weight and score. 5.2. Assessment model Using Eq. (1) and data of Table 4, we build a susceptibility assessment model of debris flows in the study area as follows:
Yi ¼ 0:511x1i þ 0:264x2i þ 0:13x3i þ 0:059x4i þ 0:036x5i
(2)
where the subscript i is the grid number calculated by ArcGIS. The final assessment is based on the spatial analysis of ArcGIS platform. First, the basic data of the study area are converted to raster images of each factor using the spatial analyst tool in ArcGIS9.3. Next, the images are reclassified and assigned the corresponding value of each grade using the reclassify tool. Finally, using the above assessment model, the results are acquired by the raster calculator function of the spatial analyst tool.
Ground roughness (x4)
Hillside angle (x5)
0.264
0.13
0.059
0.036
values into different classes. The method seeks to reduce the variance within classes and maximize the variance between classes to determine the best arrangement of values into different classes. Fig. 4 shows the susceptibility assessment results of debris flows in the Subao river valley. The high susceptibility areas concentrate in the middle of the valley, accounting for 17.6% of the valley area, most of which are on the hanging wall of Beichuan reverse fault. The medium susceptibility areas are in the middle and lower reaches and some in the upper reaches, most of which are located on both sides of high susceptibility areas, accounting for 45.3% of the valley area (Table 5). The remainders belong to low susceptibility areas. The Wenchuan earthquake induced several geohazards in the valley, which provided a considerable amount of loose materials for the subsequent debris flows. Because of the rapid and substantial addition of loose materials, debris flows are much greater and more frequent in the valley. Fig. 4 shows the debris flow catchments (blue points) and non-debris flow catchments (white points) in the Subao river valley.
6. Assessment results The assessment results are plotted in a susceptibility-zoning map of debris flow occurrence, which is divided into three classes using the natural breaks (Jenks) method. The natural breaks (Jenks) method, also called the Jenks optimization method, is a data clustering method designed to determine the best arrangement of
Table 3 Pairwise comparison matrix for calculating factor weights. Attribute
DLA
DSF
SL
GR
HG
Weight
DLA DSF SL GR HG
1 1/3 1/5 1/7 1/9
3 1 1/3 1/5 1/7
5 3 1 1/3 1/5
7 5 3 1 1/2
9 7 5 2 1
0.511 0.264 0.13 0.059 0.036
Note: DLA denotes density of landslides and rock avalanches, DSF denotes distance to the seismogenic fault, SL denotes stratum lithology, GR denotes ground roughness, and HG denotes hillside angle.
Fig. 4. Assessment results of debris flow susceptibility and debris flows occurred after the Wenchuan earthquake in the Subao river valley.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
6
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
significant financial support for this work that could have influenced its outcome.
Table 5 Statistic results of debris flow susceptibility in the Subao river valley. Susceptibility areas
Area (km2)
DFCA (km2)
RDASA (%)
Low susceptibility areas Medium susceptibility areas High susceptibility areas
12.7 32.7 26.8
8.9 17.5 5.7
70.1 53.5 21.3
Total
72.2
32.1
e
Note: DFCA is the debris flow catchment area in the susceptibility area.
Table 5 lists the statistic assessment results of the debris flow susceptibility in the Subao river valley. The ratio of debris flow catchment area in the susceptibility area to the susceptibility area (RDASA) of high susceptibility areas is 70.1%, which means most of the high susceptibility areas formed debris flows. The RDASAs of medium and low susceptibility areas are 53.5% and 21.3%, respectively. The RDASA of low susceptibility areas does not mean that these areas must have formed debris flows because debris flow does not necessarily occur in the whole basin. On the contrary, regardless of whether part of a catchment is in a low susceptibility area or not, the possibility of forming debris flows will increase greatly if all or part of the catchment lies in a high susceptibility area. Fig. 4 and Table 5 show that the debris flow catchments are mainly concentrated in the high and medium susceptibility areas. The assessment results are in accordance with the debris flows that occurred following the 2008 Wenchuan earthquake in the valley. 7. Conclusions The Subao river valley is one of the highest potential occurrence areas of debris flows in the Wenchuan earthquake area. We select five assessment factors, i.e. density of landslides and rock avalanches, distance to the seismogenic fault, stratum lithology, ground roughness, and hillside angle, and weight the factors using the AHP method. On this basis, the susceptibility assessment model of debris flows in the valley is built using the multifactor superposition method. The final assessment is based on spatial analysis of ArcGIS platform using the built model. The assessment results are plotted in a susceptibility-zoning map of debris flow occurrence, which is divided into high susceptibility areas, medium susceptibility areas, and low susceptibility areas using the natural breaks (Jenks) method. The high susceptibility areas concentrate in the middle of the valley. The medium susceptibility areas are in the middle and lower reaches, most of which are located on both sides of the high susceptibility areas. The medium and high susceptibility areas account for 17.6% and 45.3% of the Subao river valley area, respectively. The remainders belong to low susceptibility areas. The assessment results are in accordance with the actual debris flows that occurred after the Wenchuan earthquake in the valley. The results also demonstrate that the method can be used to assess the debris flow susceptibility in an earthquake-prone area. It is suggested that, in the high susceptibility areas, human activities should avoid debris flows and should not carry out control projects against debris flows in the short term. In the medium susceptibility areas, it is suggested that human activities should try to avoid debris flows and engineering control standards should be implemented if necessary. Debris flow assessment must also be conducted in the low susceptibility areas to ensure no threat of debris flows to human activities. Conflict of interest The authors wish to confirm that there are no known conflicts of interest associated with this publication and there has been no
Acknowledgments It is gratefully noted that the project is supported by the National Natural Science Foundation of China (Grant No. 41372301) and the National Science and Technology Support Program (Grant No. 2012BAC06B02). We are also grateful to two anonymous reviewers for their advices and contributions to this paper. References Burchfiel BC, Royden LH, van der Hilst RD, Hager BH, Chen Z, King RW, Li C, Lu J, Yao H, Kirby E. A geological and geophysical context for the Wenchuan earthquake of 12 May 2008, Sichuan, People’s Republic of China. GSA Today 2008;18(7):4e11. Carrara A, Crosta G, Frattini P. Comparing models of debris-flow susceptibility in the alpine environment. Geomorphology 2008;94:353e78. Cui P, Chen XQ, Zhu YY, Su FH, Wei FQ, Han YS, Liu HJ, Zhuang JQ. The Wenchuan earthquake (May 12, 2008), Sichuan province, China, and resulting geohazards. Natural Hazards 2011;56(1):19e36. Dong J, Wei GQ, Chen ZX, Li BL, Zeng Q, Yang G. Longmen Shan fold-thrust belt and its relation to the western Sichuan Basin in central China: new insights from hydrocarbon exploration. AAPG Bulletin 2006;90(9):1425e47. Huang RQ, Li WL. Research on development and distribution rules of geohazards induced by Wenchuan earthquake on 12th May, 2008. Chinese Journal of Rock Mechanics and Engineering 2008;27(12):2585e92 (in Chinese). Kamp U, Growley BJ, Khattak GA, Owen LA. GIS-based landslide susceptibility mapping for the 2005 Kashmir earthquake region. Geomorphology 2008;101(4):631e42. Kayastha P, Dhital MR, Smedt FD. Application of the analytical hierarchy process (AHP) for landslide susceptibility mapping: a case study from the Tinau watershed, west Nepal. Computers & Geosciences 2013;52:398e408. Komac M. A landslide susceptibility model using the analytical hierarchy process method and multivariate statistics in perialpine Slovenia. Geomorphology 2006;74(1):17e28. Li Y, Hu KH, Su FH, Su PC. Debris flow viewed from the basin evolution A case study of Jiangjia Gulley, Yunnan. Journal of Mountain Science 2009a;27(4): 449e56 (in Chinese). Li SL, Lai XL, Yao ZX, Yang Q. Fault zone structures of northern and southern portions of the main central fault generated by the 2008 Wenchuan earthquake using fault zone trapped waves. Earthquake Science 2009b;22(4):417e24. Li Y, Hu KH, Chen XQ, Wei FQ. Characteristic curves and debris flow activity of a valley. Journal of Mountain Science 2006;24(3):320e6 (in Chinese). Li Y, Zhou RJ, Densmore AL, Yan L, Richardson N, Dong SL, ELLIS MA, Zhang Y, He YL, Chen H, Qiao BC, Bolin MA. Surface rupture and deformation of the YingxiuBeichuan fault by the Wenchuan earthquake. Acta Geologica Sinica 2008;82(12):1688e706 (in Chinese). Li ZX, Zhang TB, Gong DX. Exploration of land suitability evaluation based on the three angles of ecology, construction suitability evaluation and city safety a case study of the master planning of Hongya county. Advanced Materials Research 2012:374e7. Liang SY, Wang YX, Wang Y. Risk assessment of geological hazard in Wudu area of Longnan city, China. Applied Mechanics and Materials 2011;39:232e7. Liu CN, Dong JJ, Peng YF, Huang HF. Effects of strong ground motion on the susceptibility of gully type debris flows. Engineering Geology 2009;104(3e4):241e 53. Liu JF, You Y, Chen XQ, Liu JK, Chen XZ. Characteristic and hazard prediction of largescale debris flow of Xiaojia Gully in Yingxiu Town, Sichuan Province, China. Engineering Geology 2014;180:55e67. Lu K, Yu B, Han L, Xie H. A study of the relationship between frequency of debris flow and the lithology in the catchment of debris flow. Advances in Earth Science 2011;26(9):980e90 (in Chinese). Luca F, Conforti M, Robustelli G. Comparison of GIS-based gullying susceptibility mapping using bivariate and multivariate statistics: Northern Calabria, South Italy. Geomorphology 2011;134(3e4):297e308. Ma W, Zhang JN, Tang LS, Liao HR. The decision model for regional debris flow prevention based on analytical hierarchy process. Journal of Catastrophology 2009;24(2):21e4 (in Chinese). Nandi A, Shakoor A. A GIS-based landslide susceptibility evaluation using bivariate and multivariate statistical analyses. Engineering Geology 2009;110(1e2):11e 20. Nefeslioglu HA, Sezer EA, Gokceoglu C, Ayas Z. A modified analytical hierarchy process (M-AHP) approach for decision support systems in natural hazard assessments. Computers & Geosciences 2013;59:1e8. Saaty TL. Decision making with the analytic hierarchy process. International Journal of Services Sciences 2008;1(1):83e98. Saaty TL. The analytic hierarchy process: planning, priority setting, resource allocation. New York, USA: McGraw-Hill; 1980.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7 Shahabi H, Khezri S, Ahmad BB, Hashim M. Landslide susceptibility mapping at central Zab basin, Iran: a comparison between analytical hierarchy process, frequency ratio and logistic regression models. Catena 2014;115:55e70. Tang C, Liang JT. Characteristics of debris flows in Beichuan epicenter of the Wenchuan earthquake triggered by rainstorm on September 24, 2008. Journal of Engineering Geology 2008;16(6):751e8. Tang C, Zhu J, Chang M, Ding J, Qi X. An empirical-statistical model for predicting debris-flow runout zones in the Wenchuan earthquake area. Quaternary International 2012;250:63e73. Tang C, Zhu J, Li WL, Liang JT. Rainfall-triggered debris flows following the Wenchuan earthquake. Bulletin of Engineering Geology and the Environment 2009;68(2):187e94. Tang C. Susceptibility spatial analysis of debris flow in the Nujiang river basin of Yunnan. Geographical Research 2005;24(2):178e86. Tie YB, Tang C. Application of AHP in single debris flow risk assessment. The Chinese Journal of Geological Hazard and Control 2006;17(4):79e84 (in Chinese). Vaidya OS, Kumar S. Analytic hierarchy process: an overview of applications. European Journal of Operational Research 2006;169(1):1e29. Wang J, Ou GQ, Yang S. Study on hazard main-control factors in debris flow-prone areas after Wenchuan earthquake based on CF method. Journal of Sichuan University (Engineering Science Edition) 2013;45(Supp. 1):18e23 (in Chinese). Xu ZQ, Ji SC, Li HB, Hou LW, Fu XF, Cai ZH. Uplift of the Longmen mountains range and the Wenchuan earthquake. Episodes 2008;31(3):291e301. Yang QL, Gao JR, Wang Y, Qian BT. Debris flows characteristics and risk degree assessment in Changyuan Gully, Huairou District, Beijing. Procedia Earth and Planetary Science 2011;2:262e71. You Y, Liu JF, Chen XC. Debris flow and its characteristics of Subao river in Beichuan county after “512” Wenchuan earthquake. Journal of Mountain Science 2010;28(3):358e66 (in Chinese).
7
Zhang J, Wei FQ, Yu SJ, Xie XJ. Susceptibility analysis of debris flow to rocks based on hydrological characteristics of theie weathering products. Chinese Journal of Rock Mechanics and Engineering 2008;27(11):2227e33 (in Chinese). Zou Q, Cui P, Zhang JQ, Xiang LZ. Quantitative evaluation for susceptibility of debris flow in upper Yangtze river basin. Environmental Science & Technology 2012;35(3):159e63.
Xingzhang Chen obtained his Ph.D. from Institute of Mountain Hazards and Environment, Chinese Academy of Sciences. He is now an Associate Professor of Geological Engineering at the Southwest University of Science and Technology in Mianyang, Sichuan Province, China. He has been involved in geohazards (debris flow and landslide) research, consulting and education for more than 5 years. He has authored or co-authored more than 40 scientific papers. He is a member of Geographical Society of China. He has participated in some national projects, including State Key Fundamental Research Program, National Key Technology R&D Program, International Cooperation Project, Knowledge Innovation Project of the Chinese Academy of Sciences. He has also directed or assisted several geohazards assessment projects and the emergency mitigation projects in the Wenchuan earthquake area from the government of Sichuan Province. At present, he is directing a National Natural Science Foundation Project and assisting other two National Natural Science Foundation Projects.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
Contents lists available at ScienceDirect
Journal of Rock Mechanics and Geotechnical Engineering journal homepage: www.rockgeotech.org
Full length article
Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China Xingzhang Chen a, *, Hui Chen a, Yong You b, Jinfeng Liu b a b
School of Environment and Resource, Southwest University of Science and Technology, Mianyang 621010, China Institute of Mountain Hazards and Environment, Chinese Academy of Sciences, Chengdu 610041, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 December 2014 Received in revised form 15 April 2015 Accepted 18 April 2015 Available online xxx
Many debris flows have occurred in the areas surrounding the epicenter of the Wenchuan earthquake. Susceptibility assessment of debris flows in this area is especially important for disaster prevention and mitigation. This paper studies one of the worst hit areas, the Subao river valley, and the susceptibility assessment of debris flows is performed based on field surveys and remote sensing interpretation. By investigating the formation conditions of debris flows in the valley, the following assessment factors are selected: mixture density of landslides and rock avalanches, distance to the seismogenic fault, stratum lithology, ground roughness, and hillside angle. The weights of the assessment factors are determined by the analytic hierarchy process (AHP) method. Each of the assessment factors is further divided into five grades. Then, the assessment model is built using the multifactor superposition method to assess the debris flow susceptibility. Based on the assessment results, the Subao river valley is divided into three areas: high susceptibility areas, medium susceptibility areas, and low susceptibility areas. The high susceptibility areas are concentrated in the middle of the valley, accounting for 17.6% of the valley area. The medium susceptibility areas are in the middle and lower reaches, most of which are located on both sides of the high susceptibility areas and account for 45.3% of the valley area. The remainders are classified as low susceptibility areas. The results of the model are in accordance with the actual debris flow events that occurred after the earthquake in the valley, confirming that the proposed model is capable of assessing the debris flow susceptibility. The results can also provide guidance for reconstruction planning and debris flow prevention in the Subao river valley. Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by Elsevier B.V. All rights reserved.
Keywords: Debris flows Susceptibility assessment Geographic information system (GIS) Analytic hierarchy process (AHP) method Wenchuan earthquake Subao river valley
1. Introduction The Wenchuan earthquake, which occurred on 12 May 2008, caused more than 15,000 geohazards, the majority of which were rock avalanches and landslides (Huang and Li, 2008). Such landslides and avalanches provide a considerable amount of loose materials for debris flows, which are expected to be more frequent and in the active period of the next 20e30 years (Cui et al., 2011). Thus, the susceptibility assessment of debris flows is of great significance to disaster prevention and mitigation in earthquakeprone areas. Assessment factors and methods are vital to assessing the debris flow susceptibility. Li et al. (2006, 2009a) demonstrated that the events of debris flow correlate with particular stages of the basin
* Corresponding author. Tel.: þ86 13881169780. E-mail address: [email protected] (X. Chen). Peer review under responsibility of Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. 1674-7755 Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jrmge.2015.04.003
evolution. The rock hardness determines the ability to provide loose materials and then influences the occurrence frequency of debris flows (Lu et al., 2011). Zhang et al. (2008) ranked the common rocks based on their susceptibility to debris flow in ascending sequence, i.e. basalt, argillaceous limestone, dolomite, slate, Quaternary deposit, sandstone, siltstone, phyllite, and mudstone. Tang and Liang (2008) argued that phyllite and slate were especially prone to forming debris flows in the Beichuan earthquake area. With regard to the assessment method, based on a conditional probability model, Tang (2005) and Zou et al. (2012) analyzed the spatial susceptibility of debris flow in the Nujiang River basin of Yunnan and the Upper Yangtze River basin, respectively. Carrara et al. (2008) developed and compared five models (statistical approach vs. physically based approach) for predicting the debris flow occurrence in an alpine area. They found that all of the statistical models were reliable and robust, while the physically based model had low predictive power. Then, Luca et al. (2011) presented an evolution of gullying susceptibility using bivariate and multivariate statistical models, and found the latter had more predictive power in gullying susceptibility. After the Wenchuan earthquake, many researches were also conducted on the susceptibility and
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
2
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
activities of debris flow in earthquake areas (Liu et al., 2009; Tang et al., 2012; Wang et al., 2013; Liu et al., 2014). The Subao river valley is located in the worst hit area of the Wenchuan earthquake. The Beichuan reverse fault, considered as the seismogenic fault (Burchfiel et al., 2008; Li et al., 2008, 2009b; Xu et al., 2008), passes through the valley. The earthquake induced many landslides and rock avalanches. On 24 September 2008, a rainstorm triggered a group of large debris flows in the valley, causing 11 deaths and significant damages (Tang and Liang, 2008; Tang et al., 2009; You et al., 2010). After that, several debris flows occurred every year in the valley, seriously threatening the postdisaster reconstruction. The valley has become one of the highest susceptibility areas of debris flow. This paper describes the formation conditions of debris flows in the Subao river valley, based on field surveys and remote sensing images. Then, we build a model based on the ArcGIS platform to assess the debris flow susceptibility in the valley area. 2. Study area 2.1. Valley overview The Subao river valley is located to the west of Leigu town, approximately 8 km south of Beichuan county (Fig. 1). The valley covers an area of 72.2 km2, and the river is 16.5 km long. The river, in the middle of the catchment area, flows from west to east. In the valley, the highest point is 2312 m and the relative relief is 1597 m, with a mean relief ratio of 9.28%. The channel is mainly U-shaped, with a width varying from a few meters to several tens of meters. 2.2. Topography Mountains dominate the Subao river valley, which is in the transition area between the front and back of the Longmen Mountains. The valley head is approximately 2312 m wide, falling rapidly to 715 m at the valley mouth. The main channel has a steep gradient with a mean relief ratio of 9.28%. The valley is dominated by steep slopes (Fig. 1). The slope angles are greater than 25 , with some up to 40 e50 or more. The area of steep (25 e35 ) and
steeper (>35 ) slopes is 46.76 km2, accounting for up to 64.76% of total slope area. Table 1 lists the slope statistic data of the Subao river valley acquired from 1:50000 digital elevation model (DEM). 2.3. Geological settings Fig. 2 shows the Beichuan reverse fault passing through the Subao river valley. The strike of the fault plane is almost northeast in direction (Dong et al., 2006). The motion of this fault is responsible for the uplift of the mountains relative to the lowlands of Sichuan Basin to the east. The stratum lithology is controlled by the Beichuan reverse fault. The hanging wall in the northwest is mainly composed of Silurian metamorphic rocks, which originated from regional metamorphism with a foliated texture and numerous cracks. The metamorphic rocks, distributed in the middle and upper catchment areas, are composed of phyllite, slate, sandstone, and other similar rocks. The footwall in the southeast is mainly composed of shallow sea phase sedimentary rocks of the middle Devonian period, including limestone, argillaceous limestone, bioclastic limestone, and siltstone. The footwall also contains a small amount of Cambrian siltstone. The rocks of the footwall disperse farther downstream, where there are steep slopes with hard rocks that are resistant to erosion. According to field surveys and remote sensing images (with a resolution of 0.5 m), it is determined that the Wenchuan earthquake directly induced more than 300 landslides and rock avalanches in the Subao river valley (Fig. 2). Landslides mainly developed in softer rock areas (such as the hanging wall) and rock avalanches mainly occurred in hard rock areas (such as the footwall). These geohazards continue to provide enough loose materials for subsequent debris flows and will last for several decades. 2.4. Climate The Subao river valley has a mild subtropical humid monsoon climate with four distinct seasons and an average annual temperature of 15.6 C. The valley has rich rainfalls with an average annual precipitation of 1399.1 mm. The 24-h maximum precipitation was 101 mm, and the 1-h maximum precipitation was 42 mm. Approximately 80% of the rain falls between June and September. Rainstorms, combined with topographic effects, can result in debris flows in the rainy season. 3. Assessment factors The determination of assessment factors should consider the debris flow formation conditions in the valley. Generally, the formation conditions of debris flow include loose materials, topography and precipitation. 3.1. Loose materials Loose deposits form an important material basis of debris flow events. Generally, the loose materials are determined by the stratum lithology and geological structures. Table 1 Slope statistic data of the Subao river valley.
Fig. 1. The location and topographic features of the Subao river valley.
Slope angle ( )
Area (km2)
<15 1525 2535 >35
7.64 17.81 22.76 24
Total
72.21
Percentage (%) 10.58 24.66 31.52 33.24 100
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
3
considered as an assessment factor. Based on the DEM data, the values of ground roughness in the Subao river valley can be obtained using surface analysis tools in ArcGIS9.3 (Fig. 3d). 3.3. Precipitation
Fig. 2. Geological map and geohazards distribution of the Subao river.
The stratum lithology controls the ability to form loose materials, and affects the formation of debris flows. Tang and Liang (2008) studied the relationship between debris flows and rock types after the Wenchuan earthquake in Beichuan County. They found that the debris flow formation is sensitive to rock types, and the sensitivity from high to low is Silurian phyllite, Cambrian siltstone, and Devonian and Carboniferous limestones. Geological structures, especially the Beichuan reverse fault, have an important effect on producing loose materials. During the Wenchuan earthquake, more loose materials were formed close to the Beichuan reverse fault. Therefore, the stratum lithology and the distance to the Beichuan reverse fault should be considered as two important assessment factors. Landslides and rock avalanches are important transportation ways for loose materials. The landslides and rock avalanches, induced by the Wenchuan earthquake, are of great significance for debris flow formation. These geohazards can supply the debris flows with enough loose materials, which should be considered as an important assessment factor. The geohazards density can represent the loose materials distribution. Based on the 300 geohazards interpreted from a remote sensing image (with a resolution of 0.5 m), we can obtain the geohazards density of the Subao river valley using the kernel density tool in ArcGIS9.3. Therefore, three assessment factors for loose materials are selected, i.e. density of landslides and rock avalanches, distance to the seismogenic fault, and stratum lithology.
3.2. Topography The topography has a direct impact on debris flow formation through its influence on conflux process. Steeper slopes result in a faster speed of flow; both slopes below and above ground surface are prone to form debris flows. It is accepted that slopes steeper than 25 can provide loose materials for debris flows and are conducive to affluxion. In the valley, approximately 64.76% of the total area contains steep slopes (>25 ) and is therefore prone to form debris flows. The hillside angle should be considered as an important assessment factor. In addition to the hillside angle, ground roughness also affects the formation of debris flows. The ground roughness denotes the ratio of surface area to the projective area for a given area. It directly affects the ability of ground confluence and seepage. As one of the topographic factors, the ground roughness of the valley should be
Rainfall, a triggering factor of debris flows, is significant to the formation of debris flows. In the rainy season, rainstorms can directly induce debris flows in the Subao river valley due to the effects of the Wenchuan earthquake. However, the valley only covers an area of 72.2 km2, and the rainfall has no obvious difference in the whole valley. Therefore, the rainfall is very important for debris flow formation, but it is just a background condition and not suitable to be an assessment factor. According to the above analysis, we select five assessment factors. Based on their contribution to debris flow formation, they are density of landslides and rock avalanches (x1), distance to the seismogenic fault (x2), stratum lithology (x3), ground roughness (x4), and hillside angle (x5). Each of them is divided into five grades, as shown in Fig. 3. 4. Assessment method The selected factors are diverse sources of information and difficult to compare to each other. To integrate them into a consistent system for assessment, we define numerical weights of all factors with the analytic hierarchy process (AHP) method. One of the most prominent features of the AHP method is its ability to evaluate quantitative as well as qualitative criteria and alternatives on the same preference scale. The method is widely used in hazard evaluation, including natural hazard assessment (Nefeslioglu et al., 2013), landslide susceptibility evaluation and mapping (Komac, 2006; Nandi and Shakoor, 2009; Kayastha et al., 2013; Shahabi et al., 2014), debris flow risk degree assessment (Yang et al., 2011), single debris flow risk assessment (Tie and Tang, 2006), and decisions for regional debris flow prevention (Ma et al., 2009). In this paper, we weight the assessment factors using the AHP method. The AHP method, pioneered by Saaty in the 1970s, is a structured technique for organizing and analyzing complex decisions based on mathematics and psychology. It is used around the world in a wide variety of decision situations (Vaidya and Kumar, 2006; Saaty, 2008). A detailed description of the AHP method is available in Saaty (1980). The procedure for using the method can be summarized as follows (Saaty, 2008): (1) Define the problem and determine the type of knowledge sought. (2) Structure the decision hierarchy from the top, with the goal being the decision, and the objectives branching down through the intermediate levels to the lowest level. (3) Construct a set of pairwise comparison matrices. Each element in an upper level is compared with the elements in the level immediately below it. (4) Check the consistency of the judgments. The consistency ratio (CR) should be calculated for consistency with the evaluation matrix and the value of the CR should be no higher than 0.1. (5) Use the priorities obtained from the comparisons to weigh the priorities in the level immediately below. Do this for every element. Then, for each element in the level below, add its weighted values and obtain its overall or global priority. The pairwise comparison employs an underlying nine-point recording scale to rate the relative preference on a one-to-one
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
4
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
Fig. 3. Figures of the five assessment factors, each of them divided into five grades. (a) Density of landslides and rock avalanches, (b) Distance to the seismogenic fault, (c) Stratum lithology, (d) Ground roughness, and (e) Hillside angle.
basis of each factor. Table 2 shows a linguistic expression to each corresponding numerical value (Kamp et al., 2008). There are many regional evaluation methods, such as the grey correlation method, the fuzzy comprehensive evaluation method, and the weighted average method. According to the existing research results, the multifactor superposition method is a popular, simple and effective one (Liang et al., 2011; Li et al., 2012), and its calculation method is simple and clear. The calculation result is a summation of the products of factor weight and factor score. The factor weights are determined by the AHP method. The calculation method is expressed as follows:
Y ¼
m X
ðai xi Þ
(1)
i¼1
where Y denotes the summation of the products of factor weight ai and factor score xi. 5. Factors system and assessment model 5.1. Factors system The weights for each factor are calculated by a pairwise comparison matrix. The quality of the comparison is described by the
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7 Table 2 Pairwise comparison nine-point rating scale.
5
Table 4 Assessment factor system of debris flow susceptibility.
Importance Definition
Explanation
Assessment factor
Grades
Weight
Score
Value
1 3 5 7
Contribution to objective is equal Attribute is slightly favored over another Attribute is strongly favored over another Attribute is very strongly favored over another Evidence favoring one attribute is of the highest possible order of affirmation When compromise is needed
Density of landslides and rock avalanches (x1)
Very high High Medium Low Very low <1 km 12 km 23 km 34 km >4 km Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 Very high High Medium Low Smooth 25 25 e35 35 e45 45 e55 >55
0.511
0.458 0.257 0.15 0.087 0.049 0.432 0.263 0.166 0.089 0.05 0.458 0.257 0.15 0.087 0.049 0.458 0.257 0.15 0.087 0.049 0.046 0.082 0.146 0.261 0.465
0.234 0.131 0.076 0.044 0.025 0.114 0.07 0.044 0.024 0.013 0.059 0.033 0.019 0.011 0.006 0.027 0.015 0.009 0.005 0.003 0.002 0.003 0.005 0.009 0.017
9
Equal importance Moderate importance Strong importance Very strong importance Extreme importance
2, 4, 6, 8
Intermediate values
Distance to seismogenic fault (x2)
Stratum lithology (x3)
CR, which is a ratio between the matrix’s consistency index (CI) and random index (RI). A CR value below 0.1 indicates that the comparison matrix is consistent. The pairwise comparison matrix and factor weights are shown in Table 3. The density of landslides and rock avalanches is the most heavily weighted factor, followed by distance to seismogenic fault and stratum lithology. Hillside angle is the least influential parameter, while ground roughness had even less influence. The results verify the view that density of geohazards and distance to seismogenic fault are the most influential event-controlling parameters for debris flows in the Wenchuan earthquake area. The consistency ratio (CR) of the judgment matrix is 0.04, far less than 0.1, which means the consistency of the judgments is very good. The scores for each grade and the weights of the factors are determined by the AHP method, as shown in Table 4. The value is the product of factor weight and score. 5.2. Assessment model Using Eq. (1) and data of Table 4, we build a susceptibility assessment model of debris flows in the study area as follows:
Yi ¼ 0:511x1i þ 0:264x2i þ 0:13x3i þ 0:059x4i þ 0:036x5i
(2)
where the subscript i is the grid number calculated by ArcGIS. The final assessment is based on the spatial analysis of ArcGIS platform. First, the basic data of the study area are converted to raster images of each factor using the spatial analyst tool in ArcGIS9.3. Next, the images are reclassified and assigned the corresponding value of each grade using the reclassify tool. Finally, using the above assessment model, the results are acquired by the raster calculator function of the spatial analyst tool.
Ground roughness (x4)
Hillside angle (x5)
0.264
0.13
0.059
0.036
values into different classes. The method seeks to reduce the variance within classes and maximize the variance between classes to determine the best arrangement of values into different classes. Fig. 4 shows the susceptibility assessment results of debris flows in the Subao river valley. The high susceptibility areas concentrate in the middle of the valley, accounting for 17.6% of the valley area, most of which are on the hanging wall of Beichuan reverse fault. The medium susceptibility areas are in the middle and lower reaches and some in the upper reaches, most of which are located on both sides of high susceptibility areas, accounting for 45.3% of the valley area (Table 5). The remainders belong to low susceptibility areas. The Wenchuan earthquake induced several geohazards in the valley, which provided a considerable amount of loose materials for the subsequent debris flows. Because of the rapid and substantial addition of loose materials, debris flows are much greater and more frequent in the valley. Fig. 4 shows the debris flow catchments (blue points) and non-debris flow catchments (white points) in the Subao river valley.
6. Assessment results The assessment results are plotted in a susceptibility-zoning map of debris flow occurrence, which is divided into three classes using the natural breaks (Jenks) method. The natural breaks (Jenks) method, also called the Jenks optimization method, is a data clustering method designed to determine the best arrangement of
Table 3 Pairwise comparison matrix for calculating factor weights. Attribute
DLA
DSF
SL
GR
HG
Weight
DLA DSF SL GR HG
1 1/3 1/5 1/7 1/9
3 1 1/3 1/5 1/7
5 3 1 1/3 1/5
7 5 3 1 1/2
9 7 5 2 1
0.511 0.264 0.13 0.059 0.036
Note: DLA denotes density of landslides and rock avalanches, DSF denotes distance to the seismogenic fault, SL denotes stratum lithology, GR denotes ground roughness, and HG denotes hillside angle.
Fig. 4. Assessment results of debris flow susceptibility and debris flows occurred after the Wenchuan earthquake in the Subao river valley.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
6
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7
significant financial support for this work that could have influenced its outcome.
Table 5 Statistic results of debris flow susceptibility in the Subao river valley. Susceptibility areas
Area (km2)
DFCA (km2)
RDASA (%)
Low susceptibility areas Medium susceptibility areas High susceptibility areas
12.7 32.7 26.8
8.9 17.5 5.7
70.1 53.5 21.3
Total
72.2
32.1
e
Note: DFCA is the debris flow catchment area in the susceptibility area.
Table 5 lists the statistic assessment results of the debris flow susceptibility in the Subao river valley. The ratio of debris flow catchment area in the susceptibility area to the susceptibility area (RDASA) of high susceptibility areas is 70.1%, which means most of the high susceptibility areas formed debris flows. The RDASAs of medium and low susceptibility areas are 53.5% and 21.3%, respectively. The RDASA of low susceptibility areas does not mean that these areas must have formed debris flows because debris flow does not necessarily occur in the whole basin. On the contrary, regardless of whether part of a catchment is in a low susceptibility area or not, the possibility of forming debris flows will increase greatly if all or part of the catchment lies in a high susceptibility area. Fig. 4 and Table 5 show that the debris flow catchments are mainly concentrated in the high and medium susceptibility areas. The assessment results are in accordance with the debris flows that occurred following the 2008 Wenchuan earthquake in the valley. 7. Conclusions The Subao river valley is one of the highest potential occurrence areas of debris flows in the Wenchuan earthquake area. We select five assessment factors, i.e. density of landslides and rock avalanches, distance to the seismogenic fault, stratum lithology, ground roughness, and hillside angle, and weight the factors using the AHP method. On this basis, the susceptibility assessment model of debris flows in the valley is built using the multifactor superposition method. The final assessment is based on spatial analysis of ArcGIS platform using the built model. The assessment results are plotted in a susceptibility-zoning map of debris flow occurrence, which is divided into high susceptibility areas, medium susceptibility areas, and low susceptibility areas using the natural breaks (Jenks) method. The high susceptibility areas concentrate in the middle of the valley. The medium susceptibility areas are in the middle and lower reaches, most of which are located on both sides of the high susceptibility areas. The medium and high susceptibility areas account for 17.6% and 45.3% of the Subao river valley area, respectively. The remainders belong to low susceptibility areas. The assessment results are in accordance with the actual debris flows that occurred after the Wenchuan earthquake in the valley. The results also demonstrate that the method can be used to assess the debris flow susceptibility in an earthquake-prone area. It is suggested that, in the high susceptibility areas, human activities should avoid debris flows and should not carry out control projects against debris flows in the short term. In the medium susceptibility areas, it is suggested that human activities should try to avoid debris flows and engineering control standards should be implemented if necessary. Debris flow assessment must also be conducted in the low susceptibility areas to ensure no threat of debris flows to human activities. Conflict of interest The authors wish to confirm that there are no known conflicts of interest associated with this publication and there has been no
Acknowledgments It is gratefully noted that the project is supported by the National Natural Science Foundation of China (Grant No. 41372301) and the National Science and Technology Support Program (Grant No. 2012BAC06B02). We are also grateful to two anonymous reviewers for their advices and contributions to this paper. References Burchfiel BC, Royden LH, van der Hilst RD, Hager BH, Chen Z, King RW, Li C, Lu J, Yao H, Kirby E. A geological and geophysical context for the Wenchuan earthquake of 12 May 2008, Sichuan, People’s Republic of China. GSA Today 2008;18(7):4e11. Carrara A, Crosta G, Frattini P. Comparing models of debris-flow susceptibility in the alpine environment. Geomorphology 2008;94:353e78. Cui P, Chen XQ, Zhu YY, Su FH, Wei FQ, Han YS, Liu HJ, Zhuang JQ. The Wenchuan earthquake (May 12, 2008), Sichuan province, China, and resulting geohazards. Natural Hazards 2011;56(1):19e36. Dong J, Wei GQ, Chen ZX, Li BL, Zeng Q, Yang G. Longmen Shan fold-thrust belt and its relation to the western Sichuan Basin in central China: new insights from hydrocarbon exploration. AAPG Bulletin 2006;90(9):1425e47. Huang RQ, Li WL. Research on development and distribution rules of geohazards induced by Wenchuan earthquake on 12th May, 2008. Chinese Journal of Rock Mechanics and Engineering 2008;27(12):2585e92 (in Chinese). Kamp U, Growley BJ, Khattak GA, Owen LA. GIS-based landslide susceptibility mapping for the 2005 Kashmir earthquake region. Geomorphology 2008;101(4):631e42. Kayastha P, Dhital MR, Smedt FD. Application of the analytical hierarchy process (AHP) for landslide susceptibility mapping: a case study from the Tinau watershed, west Nepal. Computers & Geosciences 2013;52:398e408. Komac M. A landslide susceptibility model using the analytical hierarchy process method and multivariate statistics in perialpine Slovenia. Geomorphology 2006;74(1):17e28. Li Y, Hu KH, Su FH, Su PC. Debris flow viewed from the basin evolution A case study of Jiangjia Gulley, Yunnan. Journal of Mountain Science 2009a;27(4): 449e56 (in Chinese). Li SL, Lai XL, Yao ZX, Yang Q. Fault zone structures of northern and southern portions of the main central fault generated by the 2008 Wenchuan earthquake using fault zone trapped waves. Earthquake Science 2009b;22(4):417e24. Li Y, Hu KH, Chen XQ, Wei FQ. Characteristic curves and debris flow activity of a valley. Journal of Mountain Science 2006;24(3):320e6 (in Chinese). Li Y, Zhou RJ, Densmore AL, Yan L, Richardson N, Dong SL, ELLIS MA, Zhang Y, He YL, Chen H, Qiao BC, Bolin MA. Surface rupture and deformation of the YingxiuBeichuan fault by the Wenchuan earthquake. Acta Geologica Sinica 2008;82(12):1688e706 (in Chinese). Li ZX, Zhang TB, Gong DX. Exploration of land suitability evaluation based on the three angles of ecology, construction suitability evaluation and city safety a case study of the master planning of Hongya county. Advanced Materials Research 2012:374e7. Liang SY, Wang YX, Wang Y. Risk assessment of geological hazard in Wudu area of Longnan city, China. Applied Mechanics and Materials 2011;39:232e7. Liu CN, Dong JJ, Peng YF, Huang HF. Effects of strong ground motion on the susceptibility of gully type debris flows. Engineering Geology 2009;104(3e4):241e 53. Liu JF, You Y, Chen XQ, Liu JK, Chen XZ. Characteristic and hazard prediction of largescale debris flow of Xiaojia Gully in Yingxiu Town, Sichuan Province, China. Engineering Geology 2014;180:55e67. Lu K, Yu B, Han L, Xie H. A study of the relationship between frequency of debris flow and the lithology in the catchment of debris flow. Advances in Earth Science 2011;26(9):980e90 (in Chinese). Luca F, Conforti M, Robustelli G. Comparison of GIS-based gullying susceptibility mapping using bivariate and multivariate statistics: Northern Calabria, South Italy. Geomorphology 2011;134(3e4):297e308. Ma W, Zhang JN, Tang LS, Liao HR. The decision model for regional debris flow prevention based on analytical hierarchy process. Journal of Catastrophology 2009;24(2):21e4 (in Chinese). Nandi A, Shakoor A. A GIS-based landslide susceptibility evaluation using bivariate and multivariate statistical analyses. Engineering Geology 2009;110(1e2):11e 20. Nefeslioglu HA, Sezer EA, Gokceoglu C, Ayas Z. A modified analytical hierarchy process (M-AHP) approach for decision support systems in natural hazard assessments. Computers & Geosciences 2013;59:1e8. Saaty TL. Decision making with the analytic hierarchy process. International Journal of Services Sciences 2008;1(1):83e98. Saaty TL. The analytic hierarchy process: planning, priority setting, resource allocation. New York, USA: McGraw-Hill; 1980.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003
X. Chen et al. / Journal of Rock Mechanics and Geotechnical Engineering xxx (2015) 1e7 Shahabi H, Khezri S, Ahmad BB, Hashim M. Landslide susceptibility mapping at central Zab basin, Iran: a comparison between analytical hierarchy process, frequency ratio and logistic regression models. Catena 2014;115:55e70. Tang C, Liang JT. Characteristics of debris flows in Beichuan epicenter of the Wenchuan earthquake triggered by rainstorm on September 24, 2008. Journal of Engineering Geology 2008;16(6):751e8. Tang C, Zhu J, Chang M, Ding J, Qi X. An empirical-statistical model for predicting debris-flow runout zones in the Wenchuan earthquake area. Quaternary International 2012;250:63e73. Tang C, Zhu J, Li WL, Liang JT. Rainfall-triggered debris flows following the Wenchuan earthquake. Bulletin of Engineering Geology and the Environment 2009;68(2):187e94. Tang C. Susceptibility spatial analysis of debris flow in the Nujiang river basin of Yunnan. Geographical Research 2005;24(2):178e86. Tie YB, Tang C. Application of AHP in single debris flow risk assessment. The Chinese Journal of Geological Hazard and Control 2006;17(4):79e84 (in Chinese). Vaidya OS, Kumar S. Analytic hierarchy process: an overview of applications. European Journal of Operational Research 2006;169(1):1e29. Wang J, Ou GQ, Yang S. Study on hazard main-control factors in debris flow-prone areas after Wenchuan earthquake based on CF method. Journal of Sichuan University (Engineering Science Edition) 2013;45(Supp. 1):18e23 (in Chinese). Xu ZQ, Ji SC, Li HB, Hou LW, Fu XF, Cai ZH. Uplift of the Longmen mountains range and the Wenchuan earthquake. Episodes 2008;31(3):291e301. Yang QL, Gao JR, Wang Y, Qian BT. Debris flows characteristics and risk degree assessment in Changyuan Gully, Huairou District, Beijing. Procedia Earth and Planetary Science 2011;2:262e71. You Y, Liu JF, Chen XC. Debris flow and its characteristics of Subao river in Beichuan county after “512” Wenchuan earthquake. Journal of Mountain Science 2010;28(3):358e66 (in Chinese).
7
Zhang J, Wei FQ, Yu SJ, Xie XJ. Susceptibility analysis of debris flow to rocks based on hydrological characteristics of theie weathering products. Chinese Journal of Rock Mechanics and Engineering 2008;27(11):2227e33 (in Chinese). Zou Q, Cui P, Zhang JQ, Xiang LZ. Quantitative evaluation for susceptibility of debris flow in upper Yangtze river basin. Environmental Science & Technology 2012;35(3):159e63.
Xingzhang Chen obtained his Ph.D. from Institute of Mountain Hazards and Environment, Chinese Academy of Sciences. He is now an Associate Professor of Geological Engineering at the Southwest University of Science and Technology in Mianyang, Sichuan Province, China. He has been involved in geohazards (debris flow and landslide) research, consulting and education for more than 5 years. He has authored or co-authored more than 40 scientific papers. He is a member of Geographical Society of China. He has participated in some national projects, including State Key Fundamental Research Program, National Key Technology R&D Program, International Cooperation Project, Knowledge Innovation Project of the Chinese Academy of Sciences. He has also directed or assisted several geohazards assessment projects and the emergency mitigation projects in the Wenchuan earthquake area from the government of Sichuan Province. At present, he is directing a National Natural Science Foundation Project and assisting other two National Natural Science Foundation Projects.
Please cite this article in press as: Chen X, et al., Susceptibility assessment of debris flows using the analytic hierarchy process method L A case study in Subao river valley, China, Journal of Rock Mechanics and Geotechnical Engineering (2015), http://dx.doi.org/10.1016/ j.jrmge.2015.04.003