Regional level landslide inventory maps of the Shyok River watershed, Northern Pakistan
- First Online:
- 1.3k Downloads
A preliminary landslide inventory map has been prepared for the Shyok River watershed in northern Pakistan. The document is intended for use as a screening tool in route selection of engineered infrastructure of this mountainous area preparatory to undertaking site-specific investigations of possible landslides that could adversely impact such routes. Remote sensing techniques and GIS software can be utilized to prepare a reconnaissance-level landslide inventory and gross susceptibility maps at relatively low cost. This watershed was selected as a test location for regional landslide inventory mapping because of the high density of deep-seated bedrock landslides. Topographic recognition keys were employed on a stitched shaded topographic map of the study area using 40-m contour topographic maps and 30-m resolution ASTER DEM data. The goal of these efforts is to delineate slopes with anomalous topography, often indicative of past land slippage. Most of the landslide features are expressive of composite and complex landslides, extending into the parent bedrock units. The results of this mapping were then compared with historic data and regional level landslide susceptibility maps. This effort is intended to serve as a reconnaissance-level inventory map to provide qualitative, first-order information about the probable existence of bedrock landslides and related mass wasting features, which are subject to field verification.
KeywordsRemote sensing Landslide inventory Bedrock Mass wasting Shyok River
Rugged mountains exert a controlling influence in triggering rock avalanches and bedrock landslides all over the world (Hewitt 2002). Reconnaissance-level landslide susceptibility maps are those carried out without ground checking for verification, using remote sensing techniques and morphometric analogs of common landforms associated with recent landsliding. These products are intended to serve as preliminary evaluations, carried out prior to any significant development (Guzzetti et al. 1999; Balteanu et al. 2010; Ahmed and Rogers 2014). Landslide inventory maps can be prepared for any region based on anomalous topographic signatures (Varnes 1978; Hansen 1984; Carrara et al. 1991; Hutchinson 1968; Dikau et al. 1996; Cruden and Varnes 1996; Van Den Eeckhaut et al. 2009, 2011; Ahmed and Rogers 2014). Mass wasting events in excess of five contour interval lengths along the slope usually leave noticeable changes in surface topography (Ahmed and Rogers 2014). These patterns can be recorded, characterized, and mapped with the aid of remote-sensing imagery, such as aerial photographs, and digital elevation models (DEMs) data (Rib and Liang 1978; McCalpin 1984; Pike 1988). Various types of landslides exert unique and distinct surface expression, depending upon their respective slope movement, failure extent, depth, and size (Varnes 1978; Rogers/Pacific 1994; Cruden and Varnes 1996).
The scale of the inventory maps is influenced by the size of the study area, the resolution of available data, and financial resources. Small-scale landslide inventories (<1:100,000) can be prepared from the analysis of stereopair aerial photographs and hillshade topographic maps (Cardinali et al. 1990; Brabb 1991; Rogers/Pacific 1994; Guzzetti et al. 2006; Ahmed and Rogers 2014). Satellite imagery and shaded DEMs of varying resolutions are increasingly available world-wide, and can be utilized as significant tools in discerning the presence of past land slippage.
An understanding of these anomalous topographic features aids interpreters in identifying and reconstructing recent and prehistoric landslide features in the steep mountain valleys of the Karakoram and Himalaya ranges of Pakistan. According to Hewitt (2002), the geometry of the opposing slopes, run-out paths, and pattern of interfluves (higher ground between two rivers or former river terraces formed by earth flows and bisected by fluvial erosion; Whittow 1984) are significant variables controlling the shape and surface geomorphology of rock avalanches in this region. These mountains are renowned for the incidence of snow avalanches, rock avalanches, debris flows, rockfalls, and landslide dams (Hewitt 1982, 2002; Goudie et al. 1984; Rogers and Ahmed 2013; Ahmed et al. 2015). Roughly 5 % of the ~70,000 km2 of northern Pakistan was examined for mass wasting features (Hewitt 2002). The total area and volume of these rockslides varies from 2 to ~40 km2, and 1.5 × 106 to 20 × 106 m3, respectively (Hewitt 2002, 2011). Hewitt (2002) found more than 40 rock avalanches have impacted the main river channels in the Karakorum and Himalaya region, creating large impoundments of water. Many of these slide dams have been partially breached, while the remnant lakes continue to accumulate sediment for many kilometers upstream.
Many of the rockslide deposits in the Karakorum and Himalaya ranges are complex in nature and can be difficult to recognize in the field if they do not become significantly disaggregated with downslope translation. Environmental factors, such as weathering, erosion by glaciers and rivers, or deposition behind landslide dams, can also complicate their recognition. Most of the larger landslides are influenced by structural geologic features, such as bedding, foliation, jointing, sheared zones, and faults (Hewitt 2002; Ahmed and Rogers 2014). Many of the landslide and glacier dams, as well as glacial surges, have been documented in eastern part of the Karakoram and Nanga Perbat Haramosh Massif (NPHM) Ranges in the Upper Indus River Basin (Hewitt 1982, 2002).
The main focus of this study was to identify anomalous topographic expression (i.e., parallel and converging drainage patterns, divergent contours, crenulated contours, and isolated knobs, etc.) felt to be reflective of large-scale mass movements in the Shyok River Basin. The proposed methodology was then validated by comparing historically recorded landslides in the basin (Hewitt 1982, 1998, 2002; Korup et al. 2010; Hewitt et al. 2011) with those mapped “blind,” as part of this regional effort. In every case the historic landslides were positively identified.
Overview of the study area
The majority of hill slopes in northern Pakistan exceed 45°, with at least 500 m of elevation gain. The severity of the slopes aids the frequency of landslides (Hewitt 2002). The slopes within the Shyok watershed have spawned a dense array of bedrock landslides, which appear to have occurred more or less uninterrupted throughout the Holocene. The river channel exhibits slight meandering or braided flow where the valleys spread out and widen (e.g., at Nubra and Khaplo). This tendency increases after the river passes through steep-sided gorges. Devastating flooding has been reported more than 1200 km downstream of the river, due to the catastrophic failures of natural dams, between 1926 and 1932 (Hewitt 1982, 2002). The valley bottoms are filled with fine, well-sorted, silty and sandy sediments accumulated in reservoirs trapped behind temporary landslide dams. Breached landslide dams have been identified at Haldi, Hum Bluk Hushe, Litak Hushe, Kunis-Ghah, and Shyok. Each of the outbreak floods associated with these breaches have deposited elevated, but discontinuous, outbreak flood terraces (Shroder and Bishop 1998; Hewitt 2002).
Geologic and tectonic settings
Significant rockslides have been observed across the entire basin, without any particular preference to bedrock type or formation (Hewitt 2002). Most of the largest rockslides are believed to be seismically triggered, based on historic records and descriptions (Cockerill 1902; Mason 1914; Keefer 1984; Hewitt et al. 2011). A number of mega-rockslide avalanches (area >10 km2) have been documented along the main channel and its tributaries, including those at Masher-brum, Saltoro, and Khaplo (Hewitt 1998, 2002).
Reprocessed advanced spaceborne thermal emission and reflection radiometer (ASTER) global digital elevation model 2 (GDEM2) tiles with 30-m pixel resolution of the Shyok River watershed were downloaded from an open source website (http://www.jspacesystems.or.jp/ersdac/GDEM/E/index.html). These GDEM tiles were then georeferenced to UTM Zone 43 N, using ENVI 4.8 (Environment for Visual Information Solutions software). The ASTER GDEM2 tiles appear to offer a better option for a regional level of studies over the older shuttle radar topography mission (SRTM) data (90 m, 3 arc s). This is likely ascribable to its higher resolution, which allows more accurate measurements in steep mountainous terrain due to higher radar reflectivity (Tachikawa et al. 2009; Tachikawa et al. 2011). The Shyok watershed was extracted using the hydrology module in ArcGIS10. The shape file of the Shyok River was then used to clip the hillshade map of the study area using spatial analyst tools.
The topographic map sheets employing 40-m contour intervals at 1:200,000 scale were obtained from an online source, "Topographic Maps of the World" (www.mapstor.com), and overlaid with the hillshade map to create a seamless shaded topographic map of the region. The study area was then subdivided into small sections using ArcGIS 10 and printed on large sheets (122 × 92 cm) to allow easy viewing from any azimuth, which helps delineate landslide features. The mapped landslides and related features were digitized in ArcGIS 10 and geo-referenced with the original hillshade map.
A comparative examination of slopes is of the utmost importance in identifying evidence of mass wasting. Such features are laterally restricted, because landslides tend to translate downslope, disrupting topographic patterns controlled by bedrock structure, such as benches and cliffs (Rogers 1980; Rogers/Pacific 1994; Doyle and Rogers 2005). The screening process begins with an examination of topographic patterns, searching for anomalies and inconsistencies which do not appear on adjacent slopes or changes which do not appear to be typical of underlying lithologic and structural contacts (Terzaghi 1950; Rogers 1980; Cruden and Varnes 1996; Doyle and Rogers 2005; Crozier 2010; Hart et al. 2012). In some cases, deranged or parallel drainage patterns allow large features, such as detachment complexes and landslides, to be easily discerned (Doyle and Rogers 2005).
The most common topographic anomalies associated with landslides are briefly described below (also see Fig. 3).
Contours that curve upslope adjacent to contours that curve downslope. These often suggest the removal of material (deflation) from the upper portion of a landslide and deposition (inflation) in the lower portion of the slope.
Contours that exhibit waviness or scalloping, not otherwise associated with the underlying geologic material and/or structure. Crenulated contours are often diagnostic of terrain underlain by repeated sequences of superposed land slippage, especially flowage of disaggregated debris.
Arcuate headscarp evacuation areas
Arcuate headscarp evacuation areas are steep-sided, curvilinear features at the upslope boundary of a landslide, formed by translation of sliding material downslope.
Isolated topographic benches
Isolated topographic benches are relatively broad, flat areas which often form below the headscarp evacuation area formed by back-rotation of slumped masses or infilling of pull-apart grabens.
Extended topographic ridges or isolated topographic knobs
These features are often formed by deep-seated translational movement of slump blocks or translational slides causing separation of a ridge’s caprock.
Sudden up- or down-slope turns in hillside contours
Sudden up- or down-slope turns in hillside contours result from sharp changes in topography that do not appear related to the underlying geologic structure or material changes. They are often caused by downslope movement of an isolated portion of the hillside, parallel to the slope’s natural fall-line.
In steep mountainous areas, the ASTER-derived GDEM2 tiles with 30 m resolution are generally suitable for the identification of landslides greater than five times the contour interval (Ahmed and Rogers 2014). For example, if the contour interval is 40 m, the slide feature would need to perturb at least five consecutive contours to be discerned. If a translational slide occurred on a natural slope inclined about 26° that would necessitate an elevation differential of at least 200 m, which would have a minimum length of about 500 long on a map (Ahmed and Rogers 2014).
Results and discussions
The application of anomalous topographic protocols allowed the identification and reconstruction of many landslide features in the steep mountain valleys of the Karakoram and Himalaya. According to Hewitt (2002), the geometry of the (impact) opposite slopes, run-out paths, and pattern of interfluves, are major controlling factors shaping the surface morphology of rock avalanches in this region.
The Haldi prehistoric rockslide avalanche is another example of an overtopping interfluve type landslide. The source rock was located approximately 2000 to 2500 m above the valley floor. The rock avalanche debris initially filled in the Saltoro River valley, blocking the channel. Subsequent avalanche debris passed over the valley fill and rode up the opposing valley slopes (nearly 500 m high), overtopping it and then moving down the far slope (Hewitt 2002). Some portion of this debris also blocked the Shyok River Valley (see Fig. 9a, b).
Topographic expression is a powerful tool in identifying slopes that have been modified by mass wasting processes. The use of topographic mapping protocols allows an inexpensive means to screen large mountainous regions for bedrock landslide features. This reconnaissance-level inventory identified over 600 bedrock landslide features more than 500 m long within the Shyok River watershed. The majority of these landslide features exhibit characteristics typical of composite landslides developed in the parent bedrock units. Most of the bedrock landslide features we identify are more than 1 km in length and tended to be structurally controlled when they were initially triggered.
The historic rockslide events identified in the Hushe and Khaplo areas of the Shyok River watershed (i.e., Haldi, Litak, and Kunis-Gwah) also exhibited anomalous topography, which served as useful analogs for mapping similar landslide-related features across the subject watershed. The documented historic events appear to validate the procedure employed here, which is intended to tentatively identify similar landslide-related features (>500 m in length). These maps are intended to serve as “guides” for more detailed analyses of specific projects sites, such as power transmission corridors, pipelines, structures, highways, tunnels, dams, and powerhouses.
The authors are thankful to the Natural Hazards Mitigation Institute at the Missouri University of Science and Technology, Rolla, MO, USA, for their support of this study. This research was funded by a scholarship grant from the University of Engineering and Technology, Lahore, Pakistan.
- Ahmad I, Jan MQ and Dipietro A. (2003) Age and tectonic implications of granitoid rocks from the Indian Plate of northern Pakistan in Singh S, (ed.), Granitoids of the Himalayan collisional belt: Journal of the Virtual Explorer 11: paper 01Google Scholar
- Ahmed MF, Rogers JD, Ismail HE (2015) Historic Landslide Dams along the Upper Indus River, Northern Pakistan. Nat Hazards Rev 16(3). doi:10.1061/(ASCE)NH.1527-6996.0000165
- Brabb EE (1991) The world landslide problem. Episodes 14(1):52–61Google Scholar
- Cardinali M, Guzzetti F, Brabb EE (1990) Preliminary map showing landslide deposits and related features in New Mexico. US Geological Survey Open File Report 90/293, 4 sheets, scale 1:500,000Google Scholar
- Carrara M, Cardinali M, Guzzetti F (1991) Uncertainty in assessing landslide hazard and risk. ITC J 2:172–183Google Scholar
- Cockerill GK (1902) Byways of Hunza and Nagar. Geogr J 60:98–112Google Scholar
- Cruden DM, Varnes DJ (1996) Landslide types and processes. In: Turner AK, Schuster RL (eds) Landslides, investigation and mitigation, special report 247. Transportation Research Board, Washington, pp 36–75Google Scholar
- Dikau R, Brunsden D, Schrott L, and Ibsen ML (Eds.) (1996) Landslide Recognition Identification, Movement and Causes. Wiley, ChichesterGoogle Scholar
- Goudie AS, Brunsden D, Collins DN, Derbyshire E, Fergu-son, RI, Hashnet Z, Jones DKC, Perrott FA, Said M,Waters RS, Whalley WB (1984) The geomorphology of the Hunza Valley, Karakoram Mountains, Pakistan. In: MillerK Ed. International Karakoram Project. Cambridge Univ. Press Cambridge pp. 33–411Google Scholar
- Glade T (2001) Landslide hazard assessment and historical landslide data: an inseparable couple? In: Glade T, Albini P & F Frances (Hrsg.): the use of historical data in natural hazard assessments: Springer, Berlin. pp 153–167Google Scholar
- Hansen A (1984) Landslide hazard analysis. In: Slope instability, edited by: Brunsden, D, Prior DB. Wiley, New York. pp 523–602Google Scholar
- Hewitt K (1982) Natural dams and outburst floods of the Karakoram Himalaya, in Glen J ed. Hydrological aspects of Alpine and High Mountain areas. International Hydrological Association Publication 138:259–269Google Scholar
- Hewitt K (2002) Styles of rock-avalanche depositional complexes conditioned by very rugged terrain, Karakoram Himalaya Pakistan in Evans SG, Degraff JV, eds. Catastrophic landslides: effects, occurrence, and mechanisms. Boulder, Colorado, Geological Society of America, Reviews in Engineering Geology XV:345–377Google Scholar
- Kazmi AH, Jan MQ (1997) Geology and tectonics of Pakistan, Graphic Publishers, p 554. ISBN: 9698375007Google Scholar
- McCalpin J (1984) Preliminary age classification of landslides for inventory mapping: Proceedings 21st annual engineering geology and soils engineering symposium. University Press, Moscow, pp 99–111Google Scholar
- Negi SS (1991) Himalayan rivers, lakes, and glaciers. Indus Publishing Company New Delhi, p 182Google Scholar
- Rib HT, Liang T (1978) Recognition and identification, In: Schuster, RL, Krizek RJ (eds.), Landslide Analysis and Control. Transportation Research Board Special Report No. 176, National Academy Sciences, Washington, pp 34–80Google Scholar
- Rogers JD (1980) Factors affecting hillslope profile: current topics in geomorphology, Berkeley UC, unpublished manuscript. UC Water Resources Center Archives Riverside CA, p 67Google Scholar
- Rogers J, David (1998) Topographic expression of deep-seated bedrock landslide complexes. Notes accompanying evaluation and mitigation of seismic hazards, University of California Extension, Los Angeles, p 13Google Scholar
- Rogers/Pacific Inc. (1994) Report Accompanying Map of Landslides and Other Surficial Deposits of the City of Orinda, CA. Consultant’s report for the City of Orinda Public Works Department, p 141Google Scholar
- Rogers JD, Ahmed, MF (2013) Discussion of risk factors for triggering of rockslide avalanche dams in Pakistan, Afghanistan, and Tajikistan and mitigation strategies. In: Proceedings 47th US Rock Mechanics/Geomechanics Symposium, American Rock Mechanics Association, Paper 13–591, San Francisco CAGoogle Scholar
- Searle MP (1991) Geology and tectonics of the Karakoram Mountains. New York, John Wiley 5:231–236Google Scholar
- Tachikawa T, Kaku MA, Iwasaki A (2009) ASTER GDEM validation. Presentation at the 35th ASTER Science Team Meeting, Kyoto, JapanGoogle Scholar
- Tachikawa T, Kaku M, Iwasaki A, Gesch D. Oimoen M, Zhang Z, Danielson J, Krieger T, Curtis B, Haase J, Abrams M, Crippen R, and Carabajal C (2011) ASTER Global Digital Elevation Model Version 2-Summary of Validation Results. Report to the ASTER GDEM Validation Team, Tokyo, Japan, pp 15–24Google Scholar
- Terzaghi K (1950) Mechanism of Landslides. In: Paige S, Ed. application of geology to engineering practice (Berkey Volume). The Geological Society of America, pp 83–123Google Scholar
- Varnes DJ (1978) Slope movement types and processes, in Schuster RL, and Krizek RJ, eds Landslides: Analysis and control. National Research Council, Washington, DC, Transportation Research Board, National Academy Press, Special Report 176:11–33Google Scholar
- Varnes DJ (1984) Landslide hazard zonation: a review of principles and practice. Commission of Landslides of the IAEG UNESCO, Natural Hazards No. 3, p 61Google Scholar
- Whittow J (1984) Dictionary of physical geography. London: Penguin, p 275Google Scholar
- WP/WLI (International Geotechnical Societies UNESCO Working Party for World Landslide Inventory) (1993) The Multilingual Landslide Glossary. Bi-Tech Publishers, Richmond, British Columbia Canada, p 59Google Scholar