Knowledge based landslide susceptibility mapping in the Himalayas
© Pathak. 2016
Received: 6 January 2016
Accepted: 4 May 2016
Published: 12 May 2016
Landslides are common geological hazard occurring in the mountainous region. The Himalayan belt is prone to landslide disasters, which is directly linked to the prosperity and development of the area. The present study was carried out around the Chamoli-Joshimath area, which is situated in the Northernmost-belt of the Garhwal Himalaya, India. A strategic road connecting Tibet which also links the famous Hindu temples Badrinath and Kedarnath traverses the area. The main purposes of the present study is to delineate the landslide susceptible zones in the area so that it could be helpful towards landslide disaster risk reduction and to highlight the applicability of knowledge based susceptibility mapping method in the Himalayas.
The area comprises low-to-high grade metamorphic rocks as well as carbonate rocks such as limestone and dolomite. In the study area, most of the landslides occur along the road and river sections, and in the thrust or fault zones. The landslide zones are strongly controlled by the Main Central Thrust and other faults and the resulting geomorphic condition. Most of the unstable slopes are prone to plane and wedge failures. There are many active and dormant landslides (covered by vegetation) in the area. The active landslides are due to reactivation of pre-existing ones.
The predicted landslide susceptible zones are in good agreement with the historical landslide locations, which is good indication that knowledge based landslide susceptibility mapping can be successfully applied in the Himalayas provided the causative factors are thoroughly understood.
Landslides are important geological events in many parts of the world. Himalaya is extremely vulnerable to natural disasters due to its geology, steep slopes, high relief and monsoon climates. The active tectonics in the Himalaya is responsible for the generation of faults, crushed zones, and several sets of joints that make the rocks weak, resulting in steep hill slopes susceptible to sliding (e.g. Dadson et al. 2004; Kirby and Whipple 2012; Chen et al. 2015a, 2015b).
Satellite images and aerial photos are among important sources of data for mapping landslides, geology, geomorphology, lineaments, and thrust/faults. The lineament analysis using satellite images have been applied by many authors (e.g. Ali and Pirasteh 2004; Mostafa and Bishta 2005). Remote sensing data is commonly used for analysis and prediction of mass movement through the landslide inventories, susceptibility mapping, hazard zonation, change detection, and landslide monitoring. Most ground surface changes related to the movement of a pre-existing landslide can be identified through the use of temporal aerial photographs or high resolution satellite sensor’s optical imagery (Rosin and Hervas 2005). Detecting landslides and monitoring their activity is of great relevance for disaster prevention, preparedness and mitigation in hilly areas. Landslide monitoring using remote sensing is a powerful method to point based ground surveying techniques (Keaton and DeGraff 1996; Pradhan 2010).
The landslide related parameters involved in landslide activity or triggering are the detection of anomalous patterns from contour map, surface drainage, slope profile and correlation with contour map, overburden thickness, geological structure, past movement evidences, rate of movements and so on. In this situation, the use of small scale aerial photograph together with the satellite images has significant role while studying landslides. While large-scale satellite data could serve to detect landslide bodies, smaller scale data will help in detection of parameters triggering it or characterizing it. Thus the interpretation of aerial photo together with satellite images can generate relevant landslide-related information that can help in landslide understanding and further study. The traditional visual photo-interpretation methods for landslide mapping and monitoring is still equally important even with the availability of latest generation space-borne digital imagery (Soeters and van Westen 1996).
The main purposes of the present study is to delineate the landslide susceptibility zones in the area so that it could be helpful towards landslide disaster risk reduction and to highlight the applicability of knowledge based susceptibility mapping method in the Himalayas through proper understanding of the various causative factors responsible for landslide in the study area. Present study utilized the remote sensing and GIS techniques in order to delineate landslide susceptible area. The geomorphologic and structural information has been given due importance in the analysis. Statistical methods are common approach in delineating the susceptibility zones but the present study successfully applied the knowledge based approach. This is especially important and useful when the researcher has thorough knowledge of the study area which enables strong control on the analysis. This study has opened up the possibility of utilizing this method in the Himalayas, including in Nepal Himalaya.
The field area lies about 250 km northward from Dehradun, the capital city of Uttaranchal State, and is connected by the National Highway, reaching up to Badrinath, the famous Hindu temple. The road is often disrupted in winter due to heavy snowfall and frequent landslides occurring in the rainy season. Since the area is falling under the Lower and Higher Himalayan region, the topography varies sharply, ranging from about 1000 m around the river valleys to around 4000 m, forming the peaks.
The area experiences subtropical climate with hot dry season around April-June, rainy season from July-September, and winter season from October-March. Snow fall during Jan-March is quite common in the area lying above the altitude of 2100 m. However, even at the altitude around 1300 m, snow fall can be observed during these months for short interval of time. The area is mostly covered by the forest area, followed by the rocky and barren land, cultivated area and settlements. The type and thickness of soils differs in steep slopes, valley side slopes, valleys and terraces. The barren, steep slope is consisting of shallow and loose textured soil, while densely vegetated areas are having comparatively thick soil that is rich in organic matters. The terraces consist of granular soils and are rich in moisture.
Geology of the study area
The study area mainly comprises of Almora Goup, Tejam Group, Ramgarh Group and Damtha Group of rocks. In general, the study area lies southwards from the Vaikrita Thrust (MCT), belonging to the Lesser Himalaya that mainly consists of dolomitic limestones, limestone, quartzites, slate, schist and gneiss. Several thrust traverses the area.
Two important tectonic discontinuities of regional significance lying in the area are Vaikrita Thrust (VT) and Munsiari Thrust (MT) that are considered tectonically active (Valdiya et al. 1999). The VT is passing through the Joshimath area and the MT is passing through the Helang area in the study area. Several other faults and thrusts are branching/associated with these two major thrusts (Fig. 2). Birhi Thrust is the one passing through the southern part of the study area and the Gulabkoti Thrust lies south of MT. Most of these thrusts are dipping towards north. Lineaments are very common in the study area and plays vital role in the slope failure. There are several sets of joints/fractures and the intersecting joints are forming the wedge and hence resulting in the wedge failure.
Differential weathering and erosion of various rock types has resulted in such relief variation. The low relief area is basically consisting of weaker rocks like slate and phyllite, while quartzite, gneiss and dolomitic limestones give rise to higher relief with sharp crested ridge because of relatively resistant to weathering and erosion.
The scarp slopes formed by limestone can be seen around the Pipalkoti village and in Birhi Ganga valley. Presence of steep scarps, deep narrow valleys, springs, straight course of river suggest that the area is still in its youthful stage of geomorphic cycle.
The actual landslide mapping can be done through visual analysis of aerial photographs, satellite images, topographic maps, geologic maps, field observations and the use of historic data. Satellite and aerial photo interpretation has been extensively carried out in the present study, which are the most common tools used for the detection and classification of landslides.
The aerial photographs taken during 1970’s were partly available for the study area. Likewise, the satellite images used were IRS LISS3 (23.5 m spatial resolution) and PAN (5.8 m spatial resolution) of 1999, IRS LISS4 (5.8 m spatial resolution) of 2004 and LANDSAT ETM image (28 m spatial resolution). Since the field area lies in the mountainous zone, most of the images were party covered either by ice or by cloud/fog. Therefore, it was necessary to use different images to extract the necessary information, otherwise, LISS4 image would have been a good choice.
Visual image interpretation of the satellite images and aerial photo were carried out for identification of the major geological lineaments, thrust/faults; demarcate different geomorphic units; spatial location and size of the landslides and preparation of geomorphic map. The verification of the landslides extracted from the remote sensing data along the major rivers (Alakananda, Birhi Ganga and Dhauli Ganga) revealed that most of the current landslides are the reactivation by pre-existing ones.
Digital image processing was carried out through several stages like geo-referencing, resolution merge (IRS Liss3 and PAN) to obtain output pixel size of 5.8 m and image enhancement (contrast enhancement, histogram equalization, filtering etc.). Image enhancement is especially useful to have better visualization of the images as well as for more clarity of the different geological structures and geomorphologic features. The major thrust and faults (roughly trending in east-west direction) as well as lineaments could be extracted from the remote sensing data. The regional structures like thrusts were initially identified and extracted from the LANDSAT image, which was later refined with the help of PAN image.
Contour and drainage lines were digitized from the topographic map and the geological map at hard copy was scanned and digitized. Different thematic layers like landslide, slope, lithology, lineament, structure, geomorphology, and drainage were prepared in GIS. During the field study, observations were taken mostly along the road to Badrinath that is basically constructed along the north-south flowing Alakananda River, along the Dhauli Ganga section, and along the Birhi Ganga river section. A preliminary landslide susceptible map was developed, which was verified in the field. This preliminary landslide susceptibility map was prepared from the secondary data, extracted from topographic map and remote sensing data (aerial photograph and satellite images). The proper knowledge of field condition was lacking at this stage.
An extensive field work was carried out to verify the preliminary susceptible map and additional data were collected. The field visit was carried out for about 21 days in the month of February-March. The area is mostly accessible through the motorable road except the Birhi Ganga section, which had to be covered by foot. Verification was made by comparing the susceptibility condition as predicted on the preliminary susceptibility map with the real field condition. This is particularly important to identify the dominant factor for the occurrence of landslides in the study area. This helped to revise the ranks and weights assigned to different thematic maps and its classes. The collected field data were lithology, geological structure, discontinuity, distribution of soil on slope, its type and thickness, landslide and its relationship with other factors (like geology, structure, slope, human activities, geomorphology etc.).
Various models have been applied to landslide susceptibility and hazard mapping (Lee 2007; Guzzetti et al. 1999; Pradhan and Lee 2009; Chung and Fabbri 2003; Remondo et al. 2003; Van Westen et al. 2003; Dahal et al. 2012; Bonham-Carter 1994). The knowledge based approach has been followed in the present study.
Weight and ranks for themes and classes
Geomorphology (Weight =7.0)
Highly dissected denudo-structural hill
Moderately dissected denudo-structural hill
Low dissected denudo-structural hill
Lithology (Weight = 9.5)
Crystalline limestone with thin bands of slate
Dolomite/crystalline limestone with bands of slate
Dolomitic limestone with slate
High grade meta. rocks (gneiss, garnet schist)
Limestone and quartzite
Limestone and slate
Slate with quartzitic bands
Slate and dolomitic limestone
Slate with dolomite
Slate with quartzitic bands
Slope in degree (Weight = 10.0)
Lineament Density (Weight = 5.5)
Drainage Density (Weight = 6.5)
Debris Thickness (Weight = 7.5)
Thick (>5 m)
Moderate (1–5 m)
Thin (<1 m)
Proximity to Fault (Weight = 8.5)
Fault - 75 m buffer
Thrust - 200 m buffer
S = output score
Wi = weight for each themes, and
Sij = rank for each class
Thus prepared map was validated with the landslides dataset and an acceptable model was developed, which could better represent the landslide susceptibility condition in the study area.
Results and discussion
The geomorphology can be described in terms of drainage and landform. These parameters plays important role in explaining the major geomorphic units of any area.
The drainage density map prepared from the drainage database provides information on the active soil erosion areas (where drainage density is high). It is observed that the northern part of the study area (east of Joshimath) is having high drainage density. Likewise, parts of central (around Helang and Darmi village) and Pipalkoti in the south western part of the study area are represented by medium drainage density.
The following types of landforms are observed in the study area:
Though the study area is still under youthful stage of development, prominent fluvial landforms are not developed. However, alluvial fans of smaller dimension are present in the study area. These are basically developed at the confluence of tributary stream and major river. The terraces are developed around the confluence of Birhi and Alakananda (Birhi village). Further, because of bursting of the Gauna Lake (landslide dammed lake) in 1971, huge amount of sediments had been transported, which resulted in development of new terraces that are more than 5 m in thickness.
Structure has played important role in developing some specific landforms in the study area. Anticlinal hills and synclinal valleys are good examples of structural landforms. Further, the triangular facets has been developed on the fault scarp. Cuesta and hogback are other structural landforms.
Thin to thick layers of colluvial materials derived from either landslide or weathering are deposited on the slope. The thickness of such materials is considerable along the left bank of Dhauli Ganga. Terrace cultivation is widely practiced in these materials wherever the slope is relatively gentle and the deposited materials are suitable for the cultivation.
Patalganga landslide is situated along the Patalganga valley at a distance of 61 km from Badrinanth. The Munsiari thrust is passing through this landslide. The height of the crown and the depth of the toe of the landslide from the road are 80 m and 40 m respectively. Extensive toe erosion by Patalganga River can be observed. The main rock types in the slide zone are schist, dolomite, slate and phyllite. The Pakhi Landslide situated at 62 km southwards from Badrinath is another landslide that has affected around 50 m of road section. The main exposed rocks are dolomite interbedded with slate.
Pagna Landslide is located upstream of the Birhi Ganga River (on the right bank), near the Pagna village (Fig. 10b). The extensive toe cutting by the river is further triggering the landslide. The river has shifted towards the opposite bank due to the slide materials, which are up to 5 m x 4 m x 2 m sized boulders of dolomite and dolomitic limestones.
Assigning weight to thematic layers and rank to the classes
The slope classes 25–30, 30–35 and 35–40 occupies almost equal area (around 14 % by each class). The steepest area (>60° slope) and most gentle area (<15° slope) occupies respectively, around 4 % and 10 % of the total study area. Similarly, the thickness of debris on the slope governs the amount of sediment production during the slope movement and hence is responsible for the degree of damages. The slope with thick pile of debris is more hazardous due to the chances of deeper failure plane when the sediments get saturated.
Landslide susceptibility map
Landslide susceptibility maps provide an indication of where landslides are most likely to occur in the future. Identification of the sites where there is a likelihood of occurrence of landslide events is the main task of landslide susceptibility mapping. Although landslides are often caused by single triggering events, such as heavy rains or human activity, they depend on several primary factors that make slopes susceptible to failure, such as geometry, lithological, structural and hydrographic characteristics. The landslide susceptibility map can be used as background information on the possibility of occurrence of landslides within the study area and in identification of the elements at risk within the high landslide susceptibility zones. Such information is the pre-requisite in landslide disaster risk reduction activities.
The high susceptibility area lies in the northern part of the study area, along the Alakananda River, and along Patal Ganga River valley (including Darmi village) and in the Birhi Ganga River valley (Birhi and Gauna village area).
The very high susceptibility class occupies minimum area while the slide density is maximum in this class. Maximum area is occupied by the low and moderate susceptibility classes. The higher slide density with higher susceptibility class justifies the model.
Landslides in the study area pose serious environmental and social problems. Several disturbed zones, dissected hills and hollows along the slopes are present, and are the main reasons for mass wasting process in the area. The study shows that most of the landslides occur along the road section, river section and in the thrust/fault zones. Most of the fresh landslides are the reactivation of pre-existing ones. The old landslides are covered by vegetation that, in some cases, makes the recognition of the phenomena difficult. These old, dormant slides can be reactivated by triggering factors like rainfall and man-made disturbance.
The landslide susceptibility is controlled by the occurrence of highly fractured rocks, as observed in quartzites and gneisses. The occurrence of landslide zones along Birhi Ganga section is mainly resulted due to the presence of Birhi Ganga Fault. Several old to active and small scale to large scale landslides can be observed in this section. The section is basically consisting of limestone, dolomite and slate. The Dhauli Ganga River is almost running along the Vaikrita Thrust. This is the reason huge amount of slide materials (colluvium) have been deposited along the left bank of the river. These materials are forming the potential sites of debris flow at the availability of discharging fluid, at the places where slope is rather steeper.
Construction/widening of roads have initiated a large number of slides in the area and the widening of the road to Badrinath that is aligned along the left bank of Alakananda River, is affected by the bank cutting resulting in the slope failure. Further, when the road crosses thrust zones, we can see landslides on huge scale. Remote sensing data is very useful to locate, identify and demarcate the different geologic features like lineament, thrust/fault, geomorphology.
The landslide susceptibility mapping of the study area, carried out by integrating different thematic layers like lithology, geomorphology, slope, lineament, drainage density, proximity to thrust/fault and debris thickness shows that the high susceptibility area lies in the northern part of the study area, along the Alakananda River, and along Patal Ganga River valley (including Darmi village) and in the Birhi Ganga River valley (Birhi and Gauna village area). The landslide susceptibility model well reflects the real field condition and also the model is well validated. The main purpose of the present study has been accomplished through the preparation of landslide susceptibility map through the application of knowledge based susceptibility mapping method in the Himalayas through proper understanding of the various causative factors responsible for landslide in the study area. Thus, the present study successfully applied the knowledge based landslide susceptibility mapping in the Garhwal Himalaya, which can be replicated to the other parts of Himalaya including Nepal Himalaya.
The author is thankful to Dr. P.K. Champati ray, Indian Institute of Remote Sensing (IIRS), Dehradun, India, for his valuable suggestion in carrying out the research work. Prof. R.C. Lakhera, formerly at IIRS is thanked for his guidance in aerial photo interpretation. The author is grateful to IIRS for providing the facilities, including satellite imageries required for the present study, under the CSSTEAP-IIRS program.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Ali, SA, and S Pirasteh. 2004. Geological applications of Landsat Enhanced Thematic Mapper (ETM) data and Geographic Information System (GIS): mapping and structural interpretation in south-west Iran, Zagros Structural Belt.International Journal of Remote Sensing 25(21):4715–4727.View ArticleGoogle Scholar
- Bonham-Carter, GF. 1994. Geographic Information Systems for Geoscientists: Modeling with GIS. Pergamon, Oxford.Google Scholar
- Chen, R-F, C-W Lin, Y-H Chen, T-C He, and L-Y Fei . 2015a. Detecting and Characterizing Active Thrust Fault and Deep-Seated Landslides in Dense Forest Areas of Southern Taiwan Using Airborne LiDAR DEM. Remote Sensing 7: 15443–15466. doi:10.3390/rs71115443.View ArticleGoogle Scholar
- Chen, Y-W, JBH Shyu, and C-P Chang. 2015b. Neotectonic characteristics along the eastern flank of the Central Range in the active Taiwan orogen inferred from fluvial channel morphology. Tectonics 34:2249–2270. doi:10.1002/2014TC003795.View ArticleGoogle Scholar
- Chung, CJF, and AG Fabbri. 2003. Validation of spatial prediction models for landslide hazard mapping. Natural Hazards 30:451–472.View ArticleGoogle Scholar
- Dadson, SJ, N Hovius, H Chen, WB Dade, J-C Lin, M-L Hsu, C-W Lin, M-J Horng, T-C Chen, J Milliman, and CP Stark. 2004. Earthquake-triggered increase in sediment delivery from an active mountain belt. Geology 32(8):733–736. doi:10.1130/G20639.1.View ArticleGoogle Scholar
- Dahal, RK, S Hasegawa, NP Bhandary, PP Poudel, A Nonomura, and R Yatabe. 2012. A replication of landslide hazard mapping at catchment scale. Geomatics, Natural Hazards and Risk. doi:10.1080/19475705.2011.629007.Google Scholar
- Guzzetti, F, A Carrara, M Cardinali, and P Reichenbach. 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study. Geomorphology 31:181–216.View ArticleGoogle Scholar
- Kanungo, DP, and S Sharma. 2014. Rainfall thresholds for prediction of shallow landslides around Chamoli-Joshimath region, Garhwal Himalayas, India. Landslides 11:629–638. doi:10.1007/s10346-013-0438-9.View ArticleGoogle Scholar
- Keaton, J, and J DeGraff. 1996. Surface observation and geologic mapping. In: Turner K, Schuster R (eds) Landslides Investigation and Mitigation. Transportation Research Board Sp Rep 247, Washington, DC, 178–230.Google Scholar
- Kirby, E, and KX Whipple. 2012. Expression of active tectonics in erosional landscapes. Journal of Structural Geology 44:54–75. doi:10.1016/j.jsg.2012.07.009.View ArticleGoogle Scholar
- Lee S. 2007. Comparison of landslide susceptibility maps generated through multiple logistic regression for three test areas in Korea. Earth Surface Processes and Landforms 32:2133–2148.View ArticleGoogle Scholar
- Mostafa, ME, and AZ Bishta. 2005. Significance of lineament patterns in rock unit classification and designation: a pilot study on the Gharib-Dara area, northern Eastern Desert. Egyptian Journal of Remote Sensing 26(7):1463–1475.View ArticleGoogle Scholar
- Pradhan, B. 2010. Remote sensing and GIS-based landslide hazard analysis and cross-validation using multivariate logistic regression model on three test areas in Malaysia. Advances in Space Research 45:1244–1256.View ArticleGoogle Scholar
- Pradhan, B, and S Lee. 2009. Delineation of landslide hazard areas on Penang Island, Malaysia, by using frequency ratio, logistic regression, and artificial neural network models. Environment and Earth Science. doi:10.1007/s12665-009-0245-8.Google Scholar
- Remondo, J, A González, J Ramón, A Cendrero, A Fabbri, and CJF Chung. 2003. Validation of landslide susceptibility maps: examples and applications from a case study in Northern Spain. Natural Hazards 30:437–449.View ArticleGoogle Scholar
- Rosin PL, and J Hervas. 2005. Remote sensing image thresholding methods for determining landslide activity. International Journal of Remote Sensing 26(6):1075–1092.View ArticleGoogle Scholar
- Soeters R, and CJ van Westen. 1996. Slope instability recognition, analysis and zonation. In: Turner K, Schuster R (eds) Landslides Investigation and Mitigation. Transportation Research Board Sp Rep 247, Washington, DC, 129–177.Google Scholar
- Valdiya, KS 1998. Dynamic Himalaya. Hyderabad: Universities Press.Google Scholar
- Valdiya, KS. 1980. Geology of Kumaon Lesser Himalayas. India: Wadia Inst Him Geol.Google Scholar
- Valdiya, KS, SK Paul, T Chandra, SS Bhakuni, and RC Upadhyay. 1999. Structure and lithology of Himadri (Great himalaya) between Kali and Yamuna rivers, central Himalaya. Himalayan Geology 20:1–15.Google Scholar
- van Westen, CJ, N Rengers, and R Soeters. 2003. Use of geomorphological information in indirect landslide susceptibility assessment. Natural Hazards 30:399–419.View ArticleGoogle Scholar