A review of the 2005 Kashmir earthquake-induced landslides; from a remote sensing prospective

Journal of Asian Earth Sciences 118 (2016) 68–80

Contents lists available at ScienceDirect

Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes

Review

A review of the 2005 Kashmir earthquake-induced landslides; from a remote sensing prospective Muhammad Shafique a,⇑, Mark van der Meijde b, M. Asif Khan a,c a

National Centre of Excellence in Geology, University of Peshawar, Peshawar, Pakistan Faculty of Geo-Information Science and Earth Observation (ITC), University of Twente, The Netherlands c Karakorum International University, Gilgit, Pakistan b

a r t i c l e

i n f o

Article history: Received 20 October 2015 Received in revised form 30 December 2015 Accepted 5 January 2016 Available online 6 January 2016 Keywords: Landslide Remote sensing 2005 Kashmir earthquake Topographic attributes Geology

a b s t r a c t The 8th October 2005 Kashmir earthquake, in northern Pakistan has triggered thousands of landslides, which was the second major factor in the destruction of the build-up environment, after earthquakeinduced ground shaking. Subsequent to the earthquake, several researchers from home and abroad applied a variety of remote sensing techniques, supported with field observations, to develop inventories of the earthquake-triggered landslides, analyzed their spatial distribution and subsequently developed landslide-susceptibility maps. Earthquake causative fault rupture, geology, anthropogenic activities and remote sensing derived topographic attributes were observed to have major influence on the spatial distribution of landslides. These were subsequently used to develop a landslide susceptibility map, thereby demarcating the areas prone to landsliding. Temporal studies monitoring the earthquakeinduced landslides shows that the earthquake-induced landslides are stabilized, contrary to earlier belief, directly after the earthquake. The biggest landslide induced dam, as a result of the massive Hattian Bala landslide, is still posing a threat to the surrounding communities. It is observed that remote sensing data is effectively and efficiently used to assess the landslides triggered by the Kashmir earthquake, however, there is still a need of more research to understand the mechanism of intensity and distribution of landslides; and their continuous monitoring using remote sensing data at a regional scale. This paper, provides an overview of remote sensing and GIS applications, for the Kashmir-earthquake triggered landslides, derived outputs and discusses the lessons learnt, advantages, limitations and recommendations for future research. Ó 2016 Elsevier Ltd. All rights reserved.

Contents 1.

2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Earthquake induced landslides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Remote sensing for earthquake induced landslides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3. Landslide susceptibility mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Kashmir earthquake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Tectonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Environmental settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Earthquake induced landslides in Kashmir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Landslide inventories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Type of landslides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Hattian Bala landslide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Impact of causative factors on landslide distribution and density. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Tectonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

⇑ Corresponding author. E-mail address: [email protected] (M. Shafique). http://dx.doi.org/10.1016/j.jseaes.2016.01.002 1367-9120/Ó 2016 Elsevier Ltd. All rights reserved.

69 69 70 70 71 71 71 71 71 72 73 74 74 74

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

4. 5.

3.4.3. Topography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Landuse and landcover. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.5. Anthropogenic factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.6. Impact of succeeding monsoons rains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Landslide susceptibility mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions and recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction The Kashmir earthquake (8th October 2005) was the most devastating natural hazard, in the history of Pakistan, killing about 90,000 people, leaving millions homeless, and causing economic loss of about 5 billion US$ (ADB and WB, 2005). One of the distinct characteristics of the Kashmir earthquake, were the widespread slope failures, mostly along the Neelum, Jhelum and Kunhar valleys (Fig. 1). The Kashmir earthquake-induced landslides were spread throughout the affected area of >7500 km2 (Owen et al., 2008) and resulted in 1000 direct fatalities and many more indirectly (Kamp et al., 2008). 1.1. Earthquake induced landslides Earthquakes are recognized as one of the major triggers for landslides. Earthquakes in a rough terrain can produce hundreds to thousands of landslides in a short time that can have devastating impacts on human lives and economy. Earthquake-induced ground shaking, causes short duration disturbances in the balance of

69

75 75 76 76 76 77 78 78

forces, in a slope that eventually leads to slope failure. Intensity and spatial distribution of earthquake-induced landslides is influenced by the earthquake magnitude, causative fault characteristics, site specific amplification of seismic shaking and variation in physical, meteorological, topographical and anthropogenic factors (Gorum et al., 2013; Jibson et al., 2004; Meunier et al., 2008, 2013; Parkash, 2013; Sato et al., 2007; Sepulveda et al., 2005; Tatard and Grasso, 2013). Amplification of earthquake-induced ground shaking attributed to topography, soil and impedance contrast also trigger landslides (Bozzano et al., 2008; Harp and Jibson, 2002). Physical factors comprised of underlying geology, which determine the magnitude and material of landsliding, and topographic attributes like elevation, slope, curvature, drainage and aspect of terrain, control the frequency and spatial distribution of landsliding (Gorum et al., 2011; Kamp et al., 2008; Korup, 2010). Meteorological factors comprised of climate, magnitude and intensity of rainfall control the temporal occurrences of landslides. Anthropogenic factors include deforestation, excessive grazing and toe-excavation of slopes for roads, houses, mining or construction material. Evaluating the impact of these seismic, physical,

Fig. 1. Simplified tectonic map of the study area (northern Pakistan). NNW trending fold is defined as the Hazara-Kashmir Syntaxis, which refolds the major thrusts including the Panjal Thrust (PT) and the Main Boundary Thrust (MBT). Muzaffarabad Thrust, the causative fault of the Kashmir earthquake cuts across the core of the syntaxis and joins with the PT-MBT at the western margin of the syntaxis (after Searle and Khan, 1996; Hussain et al., 2009). The Kashmir earthquake induced landslide inventory is after Basharat et al. (2014).

70

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

meteorological and anthropogenic factors on intensity and spatial distribution of earthquake-induced landslides are important to develop and implement the strategies for landslide management. Landslides triggered by earthquakes were recognized as early as 1789 BC in China and 373 BC in Greece (David, 1994). Recent examples of earthquake-induced landslide include, the 1999 Chi Chi earthquake in Taiwan (Weissel and Stark, 2001), the 2005 Kashmir earthquake (Basharat et al., 2014; Kazmi et al., 2013), the 2008 Wenchuan earthquake in China (Chigira et al., 2010; Gorum et al., 2011) and the 2010 Haiti earthquake (Gorum et al., 2013). All these earthquakes, triggered hundreds to thousands of landslides with devastating impacts on human lives and economy. However, often the damages caused by the earthquake-induced landslides, directly or indirectly, are merged in the overall earthquake-induced damages and therefore their long term impact on society and infrastructure is underestimated. For instance, landslides induced by the 2005 Kashmir earthquake directly caused 30% of the total earthquake-induced causalities (Petley et al., 2006) and the Wenchuan earthquake-induced landslides have killed more than 20,000 people (Tang et al., 2011). The effects of earthquake-induced landslides can remain persistent for years after the event in the affected area with more intensified landslides, soil erosion and sediment transport in streams and rivers (Hovius et al., 2011; Kazuo et al., 2009; Parkash, 2013; Tang et al., 2011). Furthermore, earthquake-induced landslides are found to modify the geomorphology of the affected area, by reducing the overall topography due to erosion of more earth mass, than being pushed to the surface by the earthquake due to uplifting (Parker et al., 2011). Therefore, mapping, analysing and monitoring of the earthquake-induced landslides is important for mitigating the devastating impacts of landslides. 1.2. Remote sensing for earthquake induced landslides Satellite images and aerial photos, with their extensive spatial coverage, fine resolution and repeated acquisition, are frequently and effectively used to map, analyse and monitor earthquakeinduced landslides, from local to regional scale (Gorum et al., 2011; Owen et al., 2008; Weissel and Stark, 2001; Xu, 2014). The inventory of earthquake-induced landslides, often derived from remote sensing images, is the preliminary requirement for landslide hazard analysis, vulnerability and risk assessment. An inventory of earthquake-induced landslides should include information about the location, spatial extent, source, type of mass movements, date of occurrences, number of landslides and any other relevant characteristics (Guzzetti et al., 2012). However, a large number of these images based inventories are insufficient, because either they have an incomplete coverage of the affected area, record only location of the landslides, ignore numerous small landslides, or mapped landslides as points instead of polygons (Harp et al., 2011). Harp et al. (2011) proposed a criteria for selecting remote sensing images that should be used for developing an inventory of earthquake-induced landslides, i.e. (a) the selected image must provide coverage of the entire earthquake affected area, (b) the spatial resolution of the imagery should allow mapping of landslide, of even a few meters across, (c) has stereo coverage to overlay on a Digital Elevation Model (DEM) to acquire a 3D perspective of the area and, (d) should be acquired closely before and after the triggering seismic event. To develop a landslide inventory, several image classification techniques are in practice. Often visual image classification techniques, like variation in tone and texture, are used on aerial photos and fine resolution satellite imageries (e.g. QuickBird, IKONOS, WorldView) to map landslides (Chini et al., 2011; Nichol et al., 2006; Saba et al., 2010; Xu, 2014; Zhang et al., 2014). There are, however, significant challenges with these images e.g. high cost,

incomplete spatial coverage of the affected area, orthorectification issues, and time consuming, specifically for studies at a regional scale (Wasowski et al., 2011). For mapping earthquake-induced landslides at a regional scale, optical remote sensing imagery at moderate resolution such as from MODIS, Landsat, ASTER and SPOT-5 are effectively applied (Gorum et al., 2014; Lodhi, 2011; Nefeslioglu et al., 2012; Ouzounov and Freund, 2004; Owen et al., 2008; Ray et al., 2009; Sato et al., 2007). For these moderate resolution images, digital image classification techniques are applied to map landslides (Chini et al., 2011; Lodhi, 2011). Semi-automated techniques are developed and applied to promptly develop a landslide inventory following a major earthquake (Siyahghalati et al., 2014). Such inventories can be used for quick assessment of the intensity and distribution of landslides. However, remote sensing derived landslide inventories can be prone to errors, such as features like barren area, rock quarries, excavations, recent fills appear similar to landslides on images and therefore, can possibly, be mapped as earthquake-induced landslides (Harp et al., 2011). Therefore, ground validation of such landslide inventories is strongly recommended to minimize the uncertainty through adjusting the landslides outlines, including the minor and new landslides and stating the field observations. A Digital Elevation Model (DEM) is often used to estimate the volume and temporal variation of landslides by calculating the elevation difference between cross sections, derived from successive DEMs (Chen et al., 2014; Coe et al., 1997; Guzzetti et al., 2009; Singhroy, 2009; Singhroy and Molch, 2004; Van Westen and Getahun, 2003; Zhao et al., 2012). 1.3. Landslide susceptibility mapping The findings from landslide inventories and monitoring are subsequently used for susceptibility mapping (Kamp et al., 2008). According to Dai et al. (2002) landslide susceptibility mapping requires knowledge of (a) causative factors that make the slope susceptible to failure, such as topographic attributes, geology, geomorphology, vegetation cover, drainage pattern and (b) triggering variables such as earthquakes, heavy rains, floods and slope disturbances due to anthropogenic activities. Information on causative factors at a regional scale, are often derived from remote sensing images and DEMs, and used in relation to the landslide inventory, to evaluate their influence on the intensity and spatial distribution of the landslides. For a landslide susceptibility map, the study area is classified in zones with varying degree of susceptibility caused by landslides (Varnes, 1984). The landslide susceptibility maps are further used to formulate strategies for mitigating the impacts of landslide. However, the uncertainties in data input, mainly in landslide inventory and causative factors, might lead to under- or over estimation of the existing hazard in the region, and the issue is exacerbated especially at a regional scale. The potential role of remote sensing for mapping of co-seismic landslides, motivated the global space agencies, to sign the International Charter ‘‘Space and Major Disasters” (www.disasterscharter.org) with the aim to provide fine resolution remote sensing images for free to the charter’s members. However, in the majority of the seismic events, these freely available fine resolution images are provided mainly for the severely devastated areas, to detect structural damages for effective rescue, and hence might have limited applications for assessing the magnitude and intensity of earthquake-induced landslides over the entire affected area. Nevertheless, for the image covered areas, these images can be effectively used to develop rapid landslide inventory to assess the magnitude of hazard and subsequently plan for effective response activities. This paper provides a review of remote-sensing and GIS applications for mapping, monitoring and susceptibility assessment of landslides triggered by the 2005 Kashmir earthquake, in

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80 Table 1 Stratigraphic set up of the Kashmir-Hazara region of northern Pakistan (Hussain et al., 2004; Latif et al., 2008). Formations

Lithology

Age

Poorly or unconsolidated alluvium/colluvium deposits Murree Formation

Gravels, sand and clay

Quaternary to recent Miocene

Margala Hill limestone Hangu-Patala Formations Panjal Formation Abbottabad/Muzaffarabad Formation Mansehra orthogenesis Tanawal Formation Hazara Formation Salkhala Formation

Mudstone, Siltstone, Sandstone, Shale Limestone Sandstone, shale, limestone, Greenschist (metabasalt), pelite and marble Dolomite, sandstone, quartzite limestone Granite, doerite dykes Quartzite, quartzose Schist Slate, shale, siltstone, limestone Pelitic schist, limestone

Eocene Paleocene Permian Cambrian Cambrian Precambrian

northern Pakistan, and highlights challenges and recommendations for future research. 2. The Kashmir earthquake

71

structures (e.g. Hazara-Kashmir syntaxis, Muzaffarabad anticline) has rendered these lithologies to jointing and fracturing that have further weakened their mechanical strength. These fragile geological conditions are aggravated due to ruggedness of the terrain characterized by high-relief mountains, deep incised valleys and steep slopes. The ASTER DEM derived elevation, of the earthquake affected area ranges from 614 m to 3789 m above sea level (asl) and the terrain slope ranges from 0° to 65°. The study area has a subtropical highland climate. The rough terrain, causes variation in climatic pattern in the area. The annual precipitation at Muzaffarabad is about 1500 mm, with more than one third falling during the monsoon season in the months of July and August (DCR, 1998). At Muzaffarabad (Fig. 1b), the mean maximum and minimum temperatures in January are 15.9 °C and 3.2 °C, respectively, and in June–July 37.6 °C and 22.1 °C, respectively (WMO, 2007). The cold winters, cause snowfall on high mountain peaks and leave piles of snow cover at steep slopes that start melting in succeeding summers. The melting of snow reaches at maximum in June and July because of high temperature and monsoonal precipitation. The maximum melting of snow, in July and August coupled with the torrential monsoon rainfall, results in surge in high energy streams and rivers exacerbating erosion, and landsliding. The combination of active seismicity, easily erodible lithologies, rough terrain, monsoonal climate, and accelerated erosion because of construction on steep slopes and deforestation, make the region prone to landslides.

2.1. Tectonics The Kashmir earthquake affected area, is located in one of the most seismically active regions on earth, with the seismically active Himalaya-Karakoram mountain ranges in the north, Hindu Kush in the northwest and Sulaiman-Kirthar in the western part of the region. These mountain ranges delineate the plate boundary between Indian and Eurasian plates, which has been a site of collision tectonics (Himalayan orogeny) since about the last 100 Ma (Coward et al., 1988; Hodges, 2000). Present-day high seismic hazard in Pakistan, and adjacent regions, is a consequence of continued convergence at rates of 31 mm/year (Bettinelli et al., 2006) at this part of the plate boundary. Earthquakes in the Himalaya are often associated with the east–west trending regional thrust faults, related to this plate boundary. From north to south, these are Main Karakoram Thrust (MKT), Main Mantle Thrust (MMT) and Main Boundary Thrust (MBT) (Sayab and Khan, 2010) (Fig. 1a). Northern Pakistan is located in the NW Himalayan Syntaxis, where crustal-scale north-verging fold structures are superimposed on east–west regional thrusts. The Hazara-Kashmir Syntaxis is one of these major north-trending folds, where major regional thrusts are folded around the northern apex of the Syntaxis (Fig. 1b). The Panjal Thrust and the MBT are partially merged with each other while looping around the Syntaxis. A third fault, termed Muzaffarabad Thrust, bounds the Kashmir Himalaya at their SW and traverses obliquely through the core of the HazaraKashmir Syntaxis, before merging with the MBT-Panjal Thrust at the western margin of the syntaxis. A 75 km segment (Avouac et al., 2006) of this fault from Balakot in the NW, through Muzaffarabad to Bagh in the SE, was ruptured in the Kashmir earthquake. 2.2. Environmental settings The geology of the study area comprises of lithologies ranging in age from Precambrian to recent (Hussain et al., 2004; Latif et al., 2008). The constituent lithologies (Table 1) are predominantly sedimentary/meta-sedimentary in origin including limestone, sandstone, siltstone, shale, slate and schist (Hussain et al., 2004; Kamp et al., 2008; Latif et al., 2008). Involvement in multiple faults (Panjal, MBT, Muzaffarabad and Jhelum faults) and fold

3. Earthquake induced landslides in Kashmir The Kashmir earthquake triggered a large number and widely distributed landslides. Due to the remoteness and large spatial extent of the affected area, aggravated by severe disruption to communication networks, satellite-images were the optimal and efficient source to acquire information regarding the intensity and spatial distribution of co-seismic landslides. Analyses of remote-sensing data, at times combined with field observations, led to the co-seismic landslides inventories, subsequently leading to the development of landslide susceptibility map. Following the earthquake, remote sensing imageries including IKONOS, QuickBird and ASTER, were made freely available because of the Charter, for major and severely damaged population centers including Balakot, Muzaffarabad and Bagh. These images were effectively utilized to detect the earthquake induced structural damages and landslides (Shafique et al., 2011b). 3.1. Landslide inventories Following the Kashmir earthquake, remote sensing images of various sensors, resolution and spatial coverage were utilized to develop co-seismic landslide inventories (Table 2). Owen et al. (2008) used ASTER and QuickBird satellite images, to develop a co-seismic landslide inventory in a study area of >750 km2 that was subsequently validated in the field. They have mapped 1293 co-seismic landslides and characterized their distribution, type and nature. Kamp et al. (2008) developed a co-seismic landslide inventory of 2252 landslides in a study area of 2250 km2, using digital classification of ASTER imagery and field observations. They have used only post-earthquake ASTER imagery (27 October 2005) and therefore the developed landslide inventory, also include preearthquake landslides. Sato et al. (2007) acquired a stereoscopic view of the area using post-earthquake SPOT-5 images (Table 2) and field observations to map 2424 co-seismic landslides in a study area of 2805 km2. They have observed that 79% of the earthquakeinduced landslides are small (<0.5 ha) and 9% were large (P1 ha in area). Like Kamp et al. (2008), the inventory developed by Sato

72

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

Table 2 Studies using remote sensing data for developing inventories of Kashmir earthquake induced landslides. Study

Remote sensing data and spatial resolution

Image acquisition date

Spatial extent of the area (km2)

Number of landslides mapped

Sato et al. (2007) Owen et al. (2008) Kamp et al. (2008) Ray et al. (2009)

SPOT-5 (2.5 m) ASTER (15 m) and QuickBird (2.5 m) ASTER (15 m) Cartosat-1 (2.5 m), Resourcesat-1 (5.8 m), Landsat-TM (30 ml) and ASTER (15 m) QuickBird (1 m), IKONOS (1 m), SPOT (2.5 m) and WorldView1 (0.5 m)

October 2005 ASTER (27 October 2005) QuickBird (October 2005) 27 October 2005 Cartosat (9 October 2005), ASTER (27 October 2005)

2805 >750 2250 54.5

2424 1293 2252 776

36

158

Saba et al. (2010)

Chini et al. (2011)

QuickBird (0.6 m)

Lodhi (2011) Basharat et al. (2014)

ASTER (15 m) SPOT-5 (2.5 m)

QuickBird (13 August 2004), IKONOS (12 October 2005), SPOT (13 November 2006 and 19 October 2007), WorldView1 (20 September 2008) For Muzaffarabad = 13 August 2004 and 22 October 2005, For Balakot = 11 August 2004 and 19 October 2005 27 October 2005 October 2005

et al. (2007) also includes the pre-earthquake landslides. Moreover, Sato et al. (2007) have mapped landslides as points rather than as polygons and thereby ignored the information on spatial extent of the landslides. Basharat et al. (2014) mapped 1460 co-seismic landslides in a study area of 1299 km2 using visual classification of post-earthquake SPOT-5 imagery and field investigation. However, similar to Sato et al. (2007), Basharat et al. (2014) have also mapped landslides as points and therefore the spatial extent of the mapped landslides is missing. Moreover, they have used the post-earthquake image and hence the developed landslide inventory, also includes the pre-earthquake landslides. Ray et al. (2009) have used pre and post-earthquake Cartosat-1, Resourcesat-1, Landsat-TM and ASTER images covering an area of 54.5 km2 and mapped 776 co-seismic landslides with the help of field observations. They have separated the landslides existed before the earthquake, reactivated and new landslides. However, only landslides with an area of =>45 m2 were mapped for the inventory and therefore many small landslides remained unmapped. Chini et al. (2011) used fine resolution images (QuickBird) of pre- and post-earthquake, to map the co-seismic landslides in Muzaffarabad and Balakot towns using digital classification technique. They have compared the pre and post-earthquake images, and demarcated landslides, based on their bright and white slopes in post-earthquake images that can be attributed to the removal of vegetation and exposed fresh rock surface due to landslides. The proposed technique is efficient in mapping the landslides, however, acquiring these fine resolution images, for a regional scale study, is economically very expensive. Lodhi (2011) have applied the image classification techniques of IHS, NDVI and PCA on post-earthquake ASTER image to map the coseismic landslides. The derived landslide inventory was subsequently verified using the IKONOS image for a subset of the study area and field observations. Saba et al. (2010) used pre and postearthquake fine resolution satellite images (IKONOS, QuickBird, SPOT and WorldView1) and mapped 158 co-seismic landslides supported with field observation in a study area of 36 km2. However, the selected study area is only 0.48% of the total earthquake affected area and hence might not reflect the realistic magnitude of landslide hazard in the region. Similar to Saba et al. (2010), Fujiwara et al. (2006) also used the pre and post-earthquake IKONOS images, only for the northern parts of Muzaffarabad, and mapped 100 co-seismic landslides. Kamp et al. (2010) used preearthquake (2001) and post-earthquake (2005) ASTER images, to map the landslides triggered, exclusively by the earthquake. They have observed that from 2001–2005 there is a sixfold increase in landsliding (from 369 in 2002 to 2252 in 2005) and almost eightfold increase in landslide area (from 8.2 km2 in 2002 to 61.1 km2 in 2005).

60

1299

1460

The variation in the number of landslides mapped by different authors, can be attributed to the variety of utilized remote sensing data with varying resolution and different spatial extent of the respective study areas (Table 2). Spatial resolution of the remote sensing data mainly determines the size of the recognizable landslide and is of primary importance for developing a landslide inventory. Kamp et al. (2008), Lodhi (2011) and Owen et al. (2008) have used moderate resolution ASTER images, which are suitable for regional scale studies and to map landslide of considerable size. However, they cannot detect the landslides of small spatial extent, because of smoothening effect (Shafique et al., 2011a). Chini et al. (2011) and Saba et al. (2010), have efficiently used fine resolution images to map co-seismic landslides; however, their high costs, narrow swath size (11 km for IKONOS and 16 km for Quickbird, respectively) and long revisit intervals, limit their use for the entire earthquake affected area (approximately 7500 km2). The majority of the studies are relying only on the spectral information, from the remote sensing images, to detect landslides. However, the spectral information (bright color of the eroded landslides) can easily be confused with landuse characteristics such as agricultural lands, logging sites or eroded slopes (Sato et al., 2007). Therefore, comparing the post and the pre-earthquake images supported with field observations, can minimize such interpretation errors. Moreover, the combined use of spectral, spatial, shape and contextual information in object based image classification can be used to detect a landslide and shall produce a more realistic landslide inventory (Martha et al., 2010). Moreover, the earthquake-induced landslides that occurred in glaciated affected area are difficult to map using only the remote sensing. Eventually, there is still a need to develop a complete and comprehensive inventory of the Kashmir earthquake-induced landslides using remote sensing data following the criteria suggested by Harp et al. (2011). 3.2. Type of landslides Owen et al. (2008) have classified the earthquake induced landslides according to Varnes (1978) classification. They have observed that 71% of the earthquake-induced landslides were rock fall and 18.8% were debris fall, ranging in spatial extent from a few m2 to 1000 m2. According to Owen et al. (2008), the earthquakeinduced landslides were concentrated in six distinct geomorphicgeologic-anthropogenic settings: (i) rock falls in highly fractured carbonate rocks; (ii) rock falls and rock slides in Tertiary siliciclastic rocks along antecedent drainages; (iii) natural failures in fluvial incised steep slopes comprising of Precambrian and Lower Paleozoic rocks; (iv) small debris falls of fluvial undercut

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

73

Fig. 2. Different types of the earthquake induced landslides (a) Rock fall (b) Debris fall (c) Debris slides (d) Rock slide (e) Scree (f) road triggered landslides.

Fig. 3. Panorama of area surrounding the Hattain Bala landslide. Red line highlight lateral spreading, fissuring and creeping as an indications of new landslides. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Quaternary valley fills; (v) many small rock falls and shallow rock slides on ridges and spur crests; and (vi) failures in locations associated with road construction that traverses steep slopes. Saba et al. (2010), have classified 26.58% of the mapped landslides as rock fall and 58% as the translational and rotational landslides that together reached to the 84.58%, and hence is nearby to the observations of Owen et al. (2008). Basharat et al. (2014) and Sato et al. (2007) also observed that majority of earthquake-induced landslides are rock falls and debris falls, however, did not provided the specific classification of the mapped landslides. Basharat et al. (2014), stated that the majority of co-seismic landslides are shallow with >1 m thickness and are mainly responsible for structural damages to the surrounding houses and road network. The deep seated landslides predominantly occurred on the pre-earthquake landslides or unstable slopes, such as the Hattian Bala landslide (Fig. 3). The minor discrepancy between the different studies in mapping landslide types can be attributed to the differences in the spatial extent of the study area, employed classification and types, merging of landslides types in a group and sources for mapping landslides. Remote sensing images are efficient in mapping the spatial extent of landslides, however, field observations are crucial for assigning them to a landslide types. A recent (March 2015) visit by the authors to the earthquake affected area, shows that majority of the landslide are shallow, with top few meters of surface material. Rock fall in the region are mainly occurred in the thinly bedded and highly fractured

dolomites of Muzaffarabad formation that forms steep valleys, many of them are >50° (Fig. 2a and e). Debris slides are more common on the tightly folded and fractured shales, mudstone and clay of the Murree Formation (Fig. 2c). Debris fall are predominantly observed on lower edges of Quaternary deposit undercut by the rivers and streams (Fig. 2b). Fissuring of the landslide is also observed in the region that are the indication of future landslides in the region, mainly on steep slopes associated with Muzaffarabad and Murree formation. Most of the landslide in the region are triggered by the road construction, destabilizing the slopes and often disrupting the traffic flow (Fig. 2f). Translational landslides were frequently observed that can be attributed to the highly fractured, brittle and jointed features of the bedrock, also observed by Saba et al. (2010) and Basharat et al. (2014). 3.3. Hattian Bala landslide The Hattian Bala landslide was the largest landslide triggered by the Kashmir earthquake. It buried several small settlements on the valley slope and causing 1000 human casualties. The volume of the landslide is estimated to be around 6.8  107 m3 (Dunning et al., 2007; Konagai and Sattar, 2012) with a scar of >1 km long, >200 m wide, and >20 m deep (Owen et al., 2008). Basharat et al. (2012) estimated volume for the landslide as 9.8  107 m3. The apex of the landslide is at a distance of 3.3 km from the Muzaffarabad fault rupture, with almost the same elevation and hence

74

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

can be associated to the approaching fault rupture (Ray et al., 2009). Nevertheless, the apex of the landslide starts from the crest of the ridge that is prone to topography induced amplified seismic shaking (Shafique et al., 2008) and hence the landslide can also be attributed to amplified shaking that caused slope failure and eventually a landslide (Lee et al., 2010; Sepulveda et al., 2005). The preearthquake remote sensing data, shows that the area had clusters of many pre-earthquake landslides (Dunning et al., 2007; Petley et al., 2006) that were collectively triggered by the earthquakeinduced ground shaking. The lithologies of the landslide are sandstone, siltstone and shale of the Murree Formation which is hosting most of the landslides in the region, (Basharat et al., 2014; Kamp et al., 2008; Sato et al., 2007). The landslide blocked a tributary of the Jhelum valley to the depths of 130 m and resulted in a landslide-dammed lake that remains a potential risk of flooding in downstream areas (Parvaiz et al., 2011; Sattar et al., 2011). Weather controlled wetting and drying cycles leads to increased deformation in the debris (Kiyota et al., 2011) and ultimately the intensive rains in 2010, have breached the landslide induced dam, resulting in draining of lake water, reducing the water level and flooding risk (Konagai and Sattar, 2012). However, the weathering from the exposed bare walls of the channel can clog the water channel that might lead to an increase in the volume of the lake in the coming years. Moreover, an active landslide on right bank of the lake can also lead to disaster (Konagai and Sattar, 2012; Schneider, 2009) and hence the consistent monitoring of the lake and surrounding unstable slopes is proposed to avoid any devastating situation. The recent visit to the Hattian Bala landslides shows the intense lateral spreading, fissuring and creeping on the areas surrounding the main slide (Fig. 3). Many small landslides have recently occurred in the surrounding areas of the Hattian Bala landslide with devastating impacts. The local people mentioned the increasing cracks in their houses floors and walls, due to the subsurface lateral spreading. Moreover, the rate of spreading accelerated subsequent to the rainfall that can be attributed to the increased sheer stress, due to increased moisture in the soil surface. All these are the alarming indications that areas surrounding the Hattian Bala landslide is prone to more landslides and therefore mitigation measures are required to minimize further devastations. 3.4. Impact of causative factors on landslide distribution and density The inventories of the Kashmir earthquake induced landslides, are in consensus that the spatial distribution of the Kashmir earthquake-induced landslides are concentrated on specific seismic, physical and anthropogenic zones which is in agreement with the findings of Keefer (1994), Rodriguez et al. (1999) and Jibson et al. (2004). Evaluating the influence of earthquake causative fault, lithology, geomorphology, topography, precipitation and anthropogenic factors on spatial distribution and intensity of the earthquake-induced landslides, is crucial for understanding the operating mechanisms and developing a landslide susceptibility map. 3.4.1. Tectonics The Muzaffarabad fault, triggering the Kashmir earthquake, has a strong influence on the spatial distribution of the landslides. The fault has a mean slip of 5.1 m (Bendick et al., 2007) based on GPS measurements. Slip rupture derived from ENVISAT SAR images shows the peak slip of 9.6 m (±1.1 m) (Pathier et al., 2006). Fujiwara et al. (2006) have mapped the 90 km long strip of earthquake induced ground deformation by comparing pre and post-earthquake InSAR and SAR images and calculated a maximum vertical displacement of 9 m in the north of Muzaffarabad. Avouac et al. (2006) have applied sub-pixel correlation of multi-temporal

ASTER images, and mapped continuous surface rupture, over a distance of 75 km, with an average offset of 4 m. The maximum offset of 7 m is found in the northwest of Muzaffarabad (Avouac et al., 2006). Due to the nature of thrust fault, the majority of the earthquake induced landslides were concentrated on the uplifted, hanging wall of the fault (Abers et al., 2013; Basharat et al., 2014; Chini et al., 2011; Peduzzi, 2010; Petley et al., 2006; Ray et al., 2009; Sato et al., 2007; Tatard and Grasso, 2013) that can be mainly attributed to the stronger ground shaking, on the hanging wall of a thrust fault (Abrahamson and Somerville, 1996; Fujiwara et al., 2006). Kamp et al. (2008) found that 26% of their co-seismic landslides were concentrated in a 300 m buffer zone of the fault line, a feature characteristic of reverse-fault ruptures (Ju-Jiang, 2000; Sayab and Khan, 2010). Ray et al. (2009) observed that 52% of their coseismic landslides were located within a 5 km distance of the causative fault, among which 42% are on the hanging wall side and the remaining 10% are on the foot wall side of the fault. Basharat et al. (2014), have observed that concentration of their earthquake induced landslides, are gradually decreasing with moving away from the fault line to a maximum of 11 km, and attributed it to the decrease in ground shaking, with increase in distance from the fault. The concentration of the landslides, on the hanging wall of a thrust fault, is also observed in other earthquakes, such as the 2008 Wenchuan earthquake in China, the 1994 Northridge earthquake in US and the 1999 Chi-Chi earthquake in Taiwan (Chen et al., 2012; Gorum et al., 2011; Harp and Jibson, 1996; Wang et al., 2003). The optical and radar satellite images are frequently and effectively used in a pre-earthquake scenario, to monitor the tectonic movements along active faults and in the post-earthquake scenario to detect the rupture/displacement induced by the earthquake (Avouac et al., 2006; Fielding et al., 2004). The Kashmir earthquake is occurred on the pre-existing active faults, and therefore monitoring of such active faults using SAR images can assist in prediction and hazard assessment of earthquakes (Fujiwara et al., 2006). Moreover, in post-earthquake situations, prompt mapping of the earthquake induced rupture/displacements using remote sensing images, can be effectively used to map the co-seismic landslides and the associated structural damages. With increasing availability of SAR images and diverse characteristics, there is a greater opportunity of utilizing these images for earthquake hazard management. 3.4.2. Geology Geology of the earthquake affected area is predominantly of sedimentary nature, with easily erodible limestone, sandstone, siltstone, mudstone and shales (Table 1), accompanied with active tectonics, rough terrain, monsoon weather and anthropogenic activities on steep slopes and hence providing a favorable conditions for the landslides. The earthquake affected region is witnessed to the devastating landslides, even prior to the Kashmir earthquake (Kamp et al., 2010). Subsequent to the earthquake, the widespread distribution and intensity of the co-seismic landslides are mainly controlled by the underlying geological formations. The Murree formation covering the 52% of the region, is comprised of Miocene interbedded sandstone, siltstones, claystone, red1 clay, shales and host majority of the earthquake-induced landslides (50% – Sato et al., 2007; 63% – Kamp et al., 2008; 65% – Ray et al., 2009; 57% – Peduzzi, 2010 and 67.4% – Basharat et al., 2014) (Fig. 2c). Prior to the earthquake, majority of the landslide activity in the region, was also observed in the Murree Formation (Kamp

1 For interpretation of color in Fig. 2, the reader is referred to the web version of this article.

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

et al., 2010). The largest earthquake triggering landslide i.e. Hattian Bala landslides is also located in the Murree Formation. The high concentration of landslides on Murree Formation, can partly be also attributed to the anthropogenic activities (roads), larger spatial coverage, host of the earthquake triggering Muzaffarabad fault and mostly located on the hanging wall of the fault. The area underlain by the Murree Formation is also declared and observed as the most susceptible to the future landsliding (Guzzetti et al., 1999). The Tanawal formation, composed of Quartzite, Quartzose and Schist, has the highest density of landslides (>3/km2). The Muzaffarabad formation, comprised of the highly fractured Precambrian dolomites, limestone and siliclastics, has the most extensive fissuring and is located along the western margins of the syntaxis between Muzaffarabad and Balakot. The Muzaffarabad formation is prone to rock fall (Fig. 2a), shallow failure (Fig. 2d) and has the highest percentage of the area covered by the landslides (4.3%) (Basharat et al., 2014; Kamp et al., 2008). Moreover, the area covered by the Muzaffarabad formation is demarcated as the most susceptible to future landsliding (Kamp et al., 2008). The Salkhala formation (slates, schists) accounts 14% of the study area and have received 15% of the earthquake-induced landslides (Kamp et al., 2008). Fewer landslides occurred in the Hazara, Panjal, Kamlial, Mansehra, and Muzaffarabad Formations. The authors, evaluating the impact of geological formations on Kashmir earthquake induced landslides, are considering study areas with varying spatial extents, sources and scale of the utilized geological maps and hence differences in the observations and derived results. Moreover, the available geological map for the region is at the formation level, which group various lithologies in a formation and therefore also contribute to differences in findings and observations. The development of a lithological map of the region, shall assist to better understand the mechanism of the landslides and accordingly develop strategies for their mitigation. Remote sensing images are frequently and efficiently used to develop lithological maps (Ninomiya et al., 2005; Rowan and Mars, 2003; Salati et al., 2011), however, the extensive forest, shrub and grass cover (87%, Kamp et al., 2008) in the region, hide the underlying lithology from the satellite images, and therefore, hampering the effective applications of remote sensing for lithological mapping in the study area. 3.4.3. Topography 3.4.3.1. Terrain elevation: Based on the ASTER DEM, the elevation of the area ranges from 614 m to 3789 m asl with a mean elevation of 1770 m. About 61.5% of the area has an elevation between 500– 2000 m asl and most of the landslides have occurred in this range of terrain elevation. Kamp et al. (2008) found that about 90% of the landslides are located below 2000 m asl and half (50%) of the landslides have occurred between 1000 and 1500 m asl. Only 10% of the landslides were observed between 2000–3000 m asl, with no landsliding above 3500 m asl. Similarly, Basharat et al. (2014) observed that 74% of landslides occurred below 2000 m asl and the highest concentration between 1000–1500 m asl. Ray et al. (2009) observed 65% of the landslides are in elevation range of 850– 1750 m asl. Since 97% of the fault rupture triggering the Kashmir earthquake, is exposed at elevations of 668–2000 m asl, therefore concentration of landslides at these corresponding elevations can also be partly attributed to the presence of rupture. Moreover, majority (>90%) of the anthropogenic activities contributing to the destabilization of slopes are also concentrated at 600– 1900 m asl and therefore contributing the concentration of landslides in the elevation range of <2000 m asl. Therefore, it can be concluded that the combination of different factors results in concentration of landslides on elevation range of 500–2000 m asl. Moreover, majority of the studies on the area, have used the freely available ASTER DEM to evaluate the impact of topography on the

75

distribution of landslides, however, the intrinsic errors of 13.7 m in the ASTER DEM (Shafique and van der Meijde, 2014) might also add uncertainty in evaluating the impact of topography on landslides distribution. Nevertheless, the spatial resolution (30 m) and vertical accuracy (13.7 m) of ASTER DEM is still superior than the freely available SRTM DEM with 90 m spatial resolution and 23.71 m vertical accuracy (Shafique and van der Meijde, 2014). 3.4.3.2. Terrain gradient: Terrain gradient has the primary role in occurrence of landslides. Landslides are common on terrains, steeper than the angle of repose of the substrata accompanied with minimum cohesion among the slope material. The 32% of the earthquake affected area has a terrain slope of <20° and 88% is <40° and thus provide a favorable situation for the landslides. The distribution and intensity of the Kashmir earthquake triggered landslides are significantly influenced by the terrain slope of the area. Kamp et al. (2008), used ASTER DEM derived terrain slope and observed that 41% of the landslides, occurred between terrain slopes of 25°–35°; that accounts for 31% of the area, which is also in agreement with Sato et al. (2007), Peduzzi (2010) and Basharat et al. (2014). Ray et al. (2009) observed 35% of the landslides occurred in the terrain slope range of 30°–40°. Moreover, 22.4% of the landslides were observed on terrain slopes of <25° that cover 47% of the earthquake affected area. The typical angle of repose of the unconsolidated material is 25°–40° (Kamp et al., 2008), and therefore, partly also explains the concentration of earthquake induced landslide between this slope range. Moreover, large landslides also occurred in this slope range of 25°–40°. The fewer and small landslides on steep slopes (40°–60°) are because of the absence of unconsolidated materials on these steep slopes. Hence, it can be concluded that the majority of the landslides in the region were on moderately steep terrain slopes, also observed for the 2008 Wenchuan earthquake (Gorum et al., 2011). 3.4.3.3. Terrain aspect: Meunier et al. (2008) highlighted that terrain slopes facing away from the earthquake source, are prone to amplified seismic shaking, due to trapping of energy, that subsequently leads to the slope failure and landslides. For Kashmir, this means that most of the landslides would be expected on the southern to western directions (Basharat et al., 2014; Shafique et al., 2008), also in agreement with the field observations. Sato et al. (2007) found that 73% of the Kashmir earthquake induces landslides, are located on slopes facing south and south-westerly directions. Similarly, Kamp et al. (2008) observed that most of the landslides (>70%) are located on slopes facing southerly directions (south-east, south, south-west). Basharat et al. (2014) and Das et al. (2007) also found high concentration of landslides in southerly directions and attributed this to the active Himalayan tectonic transport from the northeast to the southwest. Similarly, Ray et al. (2009) estimated that 65.2% of the mapped landslides are facing south, south-east and south-west. 3.4.4. Landuse and landcover Land cover has a strong impact on the distribution of landslides, often, barren areas are more prone to landslides than the forested areas. Kamp et al. (2008) used an ASTER derived land cover map, to evaluate the impact of land cover on the earthquake-induced landslides. The region is mostly forested with 45% of area, and 42% is covered with grassland/shrubs. Majority (>70%) of the landslides occurred on shrub land/grassland, about 20% on agricultural land and only 2% on forested land. The relatively lower values for agricultural land, are at first sight maybe surprising since it is widely accepted that agricultural practices, increase the hazard of landslides. When looking into more detail, it is seen that the agricultural fields, are predominantly located on the less steep slopes and close the valley bottoms, whereas it was previously shown that most

76

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

landslides occurred on the intermediately steep slopes, and often originate towards the top of the mountains. The fewer landslides on forest, support the fact that forested land has a strong resistance to landslides, which is in agreement with Peduzzi (2010). 3.4.5. Anthropogenic factors The earthquake affected area is characterized by rugged terrain and therefore most of the road network, is carved out on slopes, leading to slope instability and eventually landsliding (Fig. 2a, c and f). The major portion of road network in the region is at risk, because of the landslide hazard. Triggering factors such as earthquake, torrential monsoon rains and floods greatly aggravate slope instability mainly along the road network. Moreover, human terracing, excavation for buildings material and construction, roads, overgrazing and deforestation are contributing to the slope instability (Rahman et al., 2014; Sudmeier-Rieux et al., 2011). The Kashmir earthquake, triggered majority of the landslides, along the road network in the area. Owen et al. (2008) observed that >53% and Kamp et al. (2008) estimated 26% (within the 50 m buffer zone, as suggested by Van Westen et al. (2003), of the co-seismic landslides were related to the terracing and road construction. Kamp et al. (2010) showed that even before the earthquake, the steep slopes along the road network in the region were affected by landslides. Kamp et al. (2010) compared the 2001 and 2005 landslides along the roads and estimated 11-fold increase in landslides within the 50 m buffer along the road network. The association of landslides with road network and other anthropogenic activities in rugged and tectonically active terrains is also observed by Keefer (1994) and Barnard et al. (2001). 3.4.6. Impact of succeeding monsoons rains Besides triggering extensive landslides, the earthquake-induced ground shaking have also resulted in loss of cohesion in earth stratum, making them vulnerable for landslides in future. Due to the monsoonal climate and active seismicity in the region, it was predicted that these fissured and fractured destabilized slopes, might lead to widespread landslides in the forthcoming monsoon season, or if struck by another low magnitude earthquake (Aydan et al., 2009; Bulmer et al., 2007; Owen et al., 2008; Petley et al., 2006; Schneider, 2009). To evaluate the impact of the subsequent monsoon rainfall on the Kashmir earthquake triggered landslides, Saba et al. (2010), studied the temporal distribution of landslides, using the fine resolution temporal images of QuickBird, IKONOS, Worldview and SPOT-5. They have found that there were greater numbers of landslides during the monsoon seasons of years 2006 and 2007. However, the year 2008, showed relatively lesser number of landslides, despite heavy monsoon rainfall, reflecting a trend towards stabilization of slopes. They also observed that in 2008, only a few new landslides were occurred and majority of the mass movements were restricted to pre-existing earthquake-induced landslide. Khattak et al. (2010) selected 68 earthquake-induced landslides and repeatedly photographed them in years 2005, 2006, 2007 and 2008 to monitor any geomorphic changes. They observed that among the monitored landslides, the majority (80%) shows no geomorphic changes, 11% shows stabilization and only 9% of the landslides show an increase in spatial extent of the landslides. Similarly, Khan et al. (2013) used the same approach as Khattak et al. (2010) for the selected earthquakeinduced landslides and concluded that only 10% of the landslides show significant increase and the remaining show slight or no increases in the landslide area. The studies of Saba et al. (2010), Khattak et al. (2010) and Khan et al. (2013) are in agreement that there is no significant increase in the landslides, caused by the post-earthquake monsoons. However, Saba et al. (2010) considered only 36 km2 study area (0.48% of the total earthquake affected area i.e. 7500 km2), Khattak et al. (2010) and Khan et al. (2013) moni-

tored only selected landslides and hence these studies might underestimate the prevailing landslide hazard in the area. There is still a need to compare the earthquake-induced landslides with the current status over the entire earthquake affected area using fine resolution satellite images to highlight the prevailing landslide hazard in the region. The list of causative factors considered to evaluate the spatial distribution of the Kashmir earthquake induced landslides are mostly similar as considered in similar other earthquake induced landslides such as the 2008 Wenchuan, 1999 Chi Chi, and 2010 Haiti earthquakes (Gorum et al., 2011, 2013; Ju-Jiang, 2000; Weissel and Stark, 2001). In all these earthquakes i.e. Wenchuan, Chi Chi and Haiti, the fault rupture, geology and topography have significant control on distribution of co-seismic landslides (Gorum et al., 2011, 2013; Weissel and Stark, 2001), as is the case for the 2005 Kashmir earthquake. However, contrary to the other inventories we observe that for the 2005 Kashmir earthquake the anthropogenic factors are the most controlling factor for the distribution of landslides, particularly the road network (Owen et al., 2008). Even prior to the earthquake, the majority of the landslides in the region were associated with the road network (Kamp et al., 2010).

4. Landslide susceptibility mapping A landslide susceptibility map, presents zones of varying susceptibility that is mainly used for developing and implementing landslide mitigation strategies. Preparation of a landslide susceptibility map requires input from different parameters including lithology, tectonics, topography, anthropogenic factors, distance from roads and streams and the presence of landslide triggers such as earthquakes and heavy rainfall (Faraji Sabokbar et al., 2014; Guzzetti et al., 1999a; Poiraud, 2014; Vanacker et al., 2003). The landslide susceptibility map can either be qualitative or quantitative. Many studies have evaluated the Kashmir earthquake induced landslide, however, only Kamp et al. (2008), Sudmeier-Rieux et al. (2007) and Kamp et al. (2010) have developed the landslides susceptibility maps for the earthquake affected area. Kamp et al. (2008) used the Barnard Multi-Criteria Evaluation (MCE) employing Analytical Hierarchy Process (AHP) in GIS environment to develop a landslide susceptibility map (Fig. 4) with an accuracy of 67%. MCE evaluates and assigns weights to the event-controlling parameters. The MCE in this case assigned greater weights to the geology considering its significant control on the distribution of landslides and hence the level of susceptibility also suddenly changes with change in underlying geological formation (Fig. 4). The geology is followed by terrain slope and distance from the fault. The landcover, distance to roads and river were assigned moderate weights and least weight was assigned to elevation and aspect. The susceptibility map has demarcated around 1/3 of the area as ‘‘high” to ‘‘very high” susceptibility to landslides and 2/3 of the area with ‘‘moderate” to ‘‘low” susceptibility. The majority of the region underlain by Muzaffarabad formation is demarcated as highly susceptible for landslides followed by the Murree formation. Moreover, the area located in close vicinity of the faults is also demarcated as highly susceptible for landsliding. The landslide susceptibility map for the year 2001 developed by Kamp et al. (2010), is comparable with the landslides induced by the Kashmir earthquake, where 75% of the earthquake-induced landslides, have occurred in the ‘‘high” and ‘‘very high” zones of the 2001 susceptibility map. The landslide susceptibility map developed by Sudmeier-Rieux et al. (2007) using statistical regression shows that the forested areas are less susceptible for landslides than the less vegetated areas, if the terrain slope and distance from fault are similar. The majority of the landslides triggered by the earthquake are stabilized (Khan et al., 2013; Khattak et al., 2010; Saba et al.,

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

77

Fig. 4. MCE derived landslide susceptibility map adopted from Kamp et al. (2008). The Alluvium (A) and Kamlial Formation (K) shows low susceptibility, Hazara Formation (H) has low to moderate landslide susceptibility. Murree Formation (MR) has moderate to high landslide susceptibility. Muzaffarabad Formation (MZ) with very high landslide susceptibility.

5. Conclusions and recommendations

rounding unstable slopes are strongly recommended to avoid any further disaster. Landslides are still a major and active hazard in the area and thousands of people and major infrastructure is vulnerable to the devastating impact of landslides. Hence, to better understand the mechanism of intensity and spatial distribution of the earthquake-induced landslides and accordingly plan for their mitigation, the following recommendations are proposed for future research.

Remote sensing data has been effectively used at the various stages of landslide management during the 2005 Kashmir earthquake. The remote sensing data derived landslide inventories provide an overview of the spatial distribution and intensity of coseismic landslides. ASTER imagery was used to develop landslide inventories at a regional scale; however, because of the image resolution, it cannot detect the small landslides. Fine resolution images, such as SPOT-5, IKONOS and QuickBird are capable of detecting the small landslides; however, acquiring and analyzing these images for regional scale study may not be economically feasible. Visual and digital image classification techniques were used to develop inventories of the earthquake-induced landslide, however, their field verification shall improve the accuracy and completeness of the inventory. Spatial analysis shows that the distribution of co-seismic landslides are strongly influenced by the earthquake causative fault line, geological formations and topographic attributes. However, the variation in the evaluated impact of these attributes on the landslides distribution is because of the variation in the utilized remote sensing data, spatial extent of the study area and the applied techniques. Temporal remote sensing imagery and repeated photography of landslides shows that the earthquake-induced landslides are mostly stabilized. However, the regional temporal comparison of the landslides using optical or radar remote sensing data can provide a more realistic insights into the temporal behavior of earthquake-induced landslides. The active monitoring of Hattian Bala landslides and sur-

1. A comprehensive and complete earthquake landslide inventory should be developed, separating the landslides prior to the event and those triggered specifically by the earthquake. Remote sensing data that can be effectively used to develop the inventory and mapping criteria should follow the suggestions of Harp et al. (2011). The comprehensive and complete landslide inventory of earthquake should assist to develop a realistic landslide susceptibility map that can be used by the concerned agencies to mitigate the impacts of future landslides. 2. It is observed by many authors that the geology of the region has significant control on the intensity and spatial distribution of the earthquake induced landslides. However, the available geological map is only at the formation level, where, several lithological units are combined in a formation and hence it is difficult to evaluate the impact of a specific lithology on spatial distribution of landslides. Therefore it recommended to develop a lithological map of the region and subsequently used to evaluate their influence on the spatial distribution of landslides. 3. Real-time monitoring of active landslides in the region, is crucial to minimize their devastating impacts. The real-time monitoring of landslides can detect indications of any major activity that might lead to sliding. The two big and active landslides i.e. the Manser Camp and Lohar Gali landslides are posing a continuous threat to the surrounding population. Both landslides are located on the river banks and any major sliding might lead to river blocking and damming of water. Real time monitoring

2010), and therefore the landslide susceptibility map developed by Kamp et al. (2008) based on the co-seismic landslides might not be a realistic with the current situation. Therefore, there is a need to develop a landslide susceptibility map considering the current situations.

78

4.

5.

6.

7.

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

of these landslides, should assist to forecast sliding and subsequently inform the concern authorities for effective response measures to minimize the damages. The amplification of seismic induced ground shaking on hill ridges is often responsible for landslide. Impact of topography on amplification of earthquake shaking on ridge crests, often leading to slope failure, is well observed and numerically proven (Lee et al., 2009). Numerical studies on a regional scale, covering a whole earthquake affected area, requires massive computer power and recent studies have used remote sensing based proxies for amplification assessment (Shafique et al., 2012). It is suggested to initiate studies to evaluate the role of topographic induced seismic amplification and propagation of seismic waves, on the intensity and distribution of the earthquake-induced landslides (e.g. Meunier et al., 2008). Subsequently, the findings of such studies can be incorporated in the hazard assessment of earthquake-induced landslides. To facilitate consistent monitoring of the Kashmir earthquake affected unstable slopes, it is recommended to develop landslide inventories on a regular basis. Developing regular landslide inventories is important to understand the role of various causative factors for landslides that should be considered for any slope stabilization measures. In case of the Kashmir earthquake, the studies by Saba et al. (2010), Khattak et al. (2010) and Khan et al. (2013) monitor the earthquake-induced landslides and unstable slopes, however, for a limited area, or for selected landslides only. It is recommended to monitor earthquake-induced landslides on a regional scale using fine resolution remote sensing images, to detect the active landslides and investigate their causes. The optical images can be only used to detect and spatial changes in the landslide extent, however, the vertical changes in the landslide material or surface cannot be identified using optical images. Therefore, the SAR or InSAR images should be utilized in combination with the optical images to also detect the vertical variation in the landslides over a period of time. This combined information should assist in updating the landslide susceptibility map for the region. The actions to minimize the hazard induced damages mainly depend on the knowledge of intensity and recurrence interval of the hazard, information on the physical, social, economic, environmental vulnerabilities, and the prevailing risk in the area (UNISDR, 2005). However, in case of the Kashmir earthquake, there is little information available on the landslides hazard intensity, recurrence interval, vulnerability to the society, directly or indirectly, and the magnitude of the landslide induced risk. Availability of such information, should assist the decision makers to develop and implement landslide management strategies. Moreover, this information can also be used for landuse planning or any developmental projects. Semi-automated image classification techniques such as Object Based Image Classification (OBIC) can be effectively used to promptly develop inventories of earthquake induced landslides on a regional scale.

Acknowledgment The authors would like to thank Higher Education Commission of Pakistan for supporting this study.

References Abers, G.A., Nakajima, J., van Keken, P.E., Kita, S., Hacker, B.R., 2013. Thermal– petrological controls on the location of earthquakes within subducting plates. Earth Planet. Sci. Lett. 369–370, 178–187.

Abrahamson, N.A., Somerville, P.G., 1996. Effects of the hanging wall and footwall on ground motions recorded during the Northridge earthquake. Bull. Seismol. Soc. Am. 86, S93–S99. ADB, WB, 2005. Preliminary Damage and Needs Assessment. Asian Development Bank and World Bank, Islamabad, Pakistan, p. 124. Avouac, J.-P., Ayoub, F., Leprince, S., Konca, O., Helmberger, D.V., 2006. The 2005, Mw 7.6 Kashmir earthquake: sub-pixel correlation of ASTER images and seismic waveforms analysis. Earth Planet. Sci. Lett. 249, 514–528. Aydan, Ö., Ohta, Y., Hamada, M., 2009. Geotechnical evaluation of slope and ground failures during the 8 October 2005 Muzaffarabad earthquake, Pakistan. J. Seismol. 13, 399–413. Barnard, P.L., Owen, L.A., Sharma, M.C., Finkel, R.C., 2001. Natural and humaninduced landsliding in the Garhwal Himalaya of northern India. Geomorphology 40, 21–35. Basharat, M., Rohn, J., Ehret, D., Baig, M., 2012. Lithological and structural control of Hattian Bala rock avalanche triggered by the Kashmir earthquake 2005, subHimalayas, northern Pakistan. J. Earth Sci. 23, 213–224. Basharat, M., Rohn, J., Baig, M.S., Khan, M.R., 2014. Spatial distribution analysis of mass movements triggered by the 2005 Kashmir earthquake in the Northeast Himalayas of Pakistan. Geomorphology 206, 203–214. Bendick, R., Bilham, R., Khan, M.A., Khan, S.F., 2007. Slip on an active wedge thrust from geodetic observations of the 8 October 2005 Kashmir earthquake. Geology 35, 267–270. Bettinelli, P., Avouac, J.-P., Flouzat, M., Jouanne, F., Bollinger, L., Willis, Pascal, Chitrakar, G.R., 2006. Plate motion of India and interseismic strain in the Nepal Himalaya from GPS and DORIS measurements. J. Geodesy 80, 567–589. Bozzano, F., Lenti, L., Martino, S., Paciello, A., Scarascia Mugnozza, G., 2008. Selfexcitation process due to local seismic amplification responsible for the reactivation of the Salcito landslide (Italy) on 31 October 2002. J. Geophys. Res.: Solid Earth 113, B10312. Bulmer, M., Farquhar, T., Roshan, M., Akhtar, S.S., Wahla, S.K., 2007. Landslide hazard after the 2005 Kashmir earthquake. EOS Trans. 88, 53–55. Chen, X.L., Ran, H.L., Yang, W.T., 2012. Evaluation of factors controlling large earthquake-induced landslides by the Wenchuan earthquake. Nat. Hazards Earth Syst. Sci. 12, 3645–3657. Chen, Q., Cheng, H., Yang, Y., Liu, G., Liu, L., 2014. Quantification of mass wasting volume associated with the giant landslide Daguangbao induced by the 2008 Wenchuan earthquake from persistent scatterer InSAR. Remote Sens. Environ. 152, 125–135. Chigira, M., Wu, X., Inokuchi, T., Wang, G., 2010. Landslides induced by the 2008 Wenchuan earthquake, Sichuan, China. Geomorphology 118, 225–238. Chini, M., Cinti, F.R., Stramondo, S., 2011. Co-seismic surface effects from very high resolution panchromatic images: the case of the 2005 Kashmir (Pakistan) earthquake. Nat. Hazards Earth Syst. 11, 931–943. Coe, J.A., Glancy, P.A., Whitney, J.W., 1997. Volumetric analysis and hydrologic characterization of a modern debris flow near Yucca Mountain, Nevada. Geomorphology 20, 11–28. Coward, M.P., Butler, R.W.H., Chambers, A.F., Graham, R.H., Izatt, C.N., Khan, M.A., Knipe, R.J., Prior, D.J., Treloar, P.J., Williams, M.P., Barnicoat, A., 1988. Folding and imbrication of the Indian Crust during Himalayan Collision [and Discussion]. Philos. Trans. R. Soc. London. Series A, Math. Phys. Sci. 326, 89–116. Dai, F., Lee, C., Ngai, Y., 2002. Landslide risk assessment and management: an overview. Eng. Geol. 64, 65–87. Das, J.D., Saraf, A.K., Panda, S., 2007. Satellite data in a rapid analysis of Kashmir earthquake (October 2005) triggered landslide pattern and river water turbidity in and around the epicentral region. Int. J. Remote Sens. 28, 1835–1842. David, K.K., 1994. The importance of earthquake-induced landslides to long-term slope erosion and slope-failure hazards in seismically active regions. Geomorphology 10, 265–284. DCR, 1998. District Census Report (1998), Population Census Organization, Government of Pakistan, Islamabad, Pakistan. Dunning, S.A., Mitchella, W.A., Rossera, N.J., Petleya, D.N., 2007. The Hattian Bala rock avalanche and associated landslides triggered by the Kashmir Earthquake of 8 October 2005. Eng. Geol. 93, 130–144. Faraji Sabokbar, H., Shadman Roodposhti, M., Tazik, E., 2014. Landslide susceptibility mapping using geographically-weighted principal component analysis. Geomorphology 226, 15–24. Fielding, E.J., Wright, T.J., Muller, J., Parsons, B.E., Walker, R., 2004. Aseismic deformation of a fold-and-thrust belt imaged by synthetic aperture radar interferometry near Shahdad, southeast Iran. Geology 32, 577–580. Fujiwara, S., Tobita, M., Sato, H.P., Ozawa, S., Une, H., Koarai, M., Nakai, H., Fujiwara, M., Yarai, H., Nishimura, T., Hayashi, F., 2006. Satellite data gives snapshot of the 2005 Pakistan earthquake. Eos, Trans. Am. Geophys. Union 87, 73–77. Gorum, T., Fan, X., van Westen, C., Huang, R., Xu, Q., Tang, C., Wang, G., 2011. Distribution pattern of earthquake-induced landslides triggered by the 12 May 2008 Wenchuan earthquake. Geomorphology 133, 152–167. Gorum, T., van Westen, C.J., Korup, O., van der Meijde, M., Fan, X., van der Meer, F.D., 2013. Complex rupture mechanism and topography control symmetry of masswasting pattern, 2010 Haiti earthquake. Geomorphology 184, 127–138. Gorum, T., Korup, O., van Westen, C.J., van der Meijde, M., Xu, C., van der Meer, F.D., 2014. Why so few? Landslides triggered by the 2002 Denali earthquake, Alaska. Quatern. Sci. Rev. 95, 80–94. Guzzetti, F., Carrara, A., Cardinali, M., Reichenbach, P., 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy. Geomorphology 31, 181–216.

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80 Guzzetti, F., Ardizzone, F., Cardinali, M., Rossi, M., Valigi, D., 2009. Landslide volumes and landslide mobilization rates in Umbria, central Italy. Earth Planet. Sci. Lett. 279, 222–229. Guzzetti, F., Mondini, A.C., Cardinali, M., Fiorucci, F., Santangelo, M., Chang, K.-T., 2012. Landslide inventory maps: new tools for an old problem. Earth Sci. Rev. 112, 42–66. Harp, E.L., Jibson, R.W., 1996. Landslides triggered by the 1994 Northridge, California, earthquake. Bull. Seismol. Soc. Am. 86, S319–S332. Harp, E.L., Jibson, R.W., 2002. Anomalous concentrations of seismically triggered rock falls in Pacoima canyon: are they caused by highly susceptible slopes or local amplification of seismic shaking? Bull. Seismol. Soc. Am. 92, 3180–3189. Harp, E.L., Keefer, D.K., Sato, H.P., Yagi, H., 2011. Landslide inventories: the essential part of seismic landslide hazard analyses. Eng. Geol. 122, 9–21. Hodges, K.V., 2000. Tectonics of the Himalaya and southern Tibet from two perspectives. Geol. Soc. Am. Bull. 112, 324–350. Hovius, N., Meunier, P., Lin, C.-W., Chen, H., Chen, Y.-G., Dadson, S., Horng, M.-J., Lines, M., 2011. Prolonged seismically induced erosion and the mass balance of a large earthquake. Earth Planet. Sci. Lett. 304, 347–355. Hussain, A., Mughal, N., Haq, I., Latif, A., 2004. Geological Map of the Gari Habib Ullah Area, District Mansehra and Parts of Muzaffarabad District, AJK, Geological Map Series. Geological Survey of Pakistan, Islamabad-Pakistan. Hussain, A., Yeats, R.S., MonaLisa, 2009. Geological setting of the 8 October 2005 Kashmir earthquake. J. Seismolog. 13, 315–325. Jibson, R.W., Harp, E.L., Schulz, W., Keefer, D.K., 2004. Landslides triggered by the 2002 Denali Fault, Alaska, earthquake and the inferred nature of the strong shaking. Earthq. Spectra 20, 669–691. Ju-Jiang, H., 2000. Chi Chi earthquake induced landslides in Taiwan. Earthq. Eng. Eng. Seismol. 2, 25–33. Kamp, U., Growley, B.J., Khattak, G.A., Owen, L.A., 2008. GIS-based landslide susceptibility mapping for the 2005 Kashmir earthquake region. Geomorphology 101, 631–642. Kamp, U., Owen, L., Growley, B., Khattak, G., 2010. Back analysis of landslide susceptibility zonation mapping for the 2005 Kashmir earthquake: an assessment of the reliability of susceptibility zoning maps. Nat. Hazards 54, 1–25. Kazmi, Z.A., Konagai, K., Ikeda, T., 2013. Field measurements and numerical simulation of debris flows from dolomite slopes destabilized during the 2005 Kashmir earthquake, Pakistan. J. Earthq. Eng. 18, 364–388. Kazuo, K., Jorgen, J., Shigeki, T., Takaaki, I., 2009. Huge landslides caused by massive earthquakes and long-lasting geotechnical risks. In: Sassa, K., Canuti, P. (Eds.), Landslides – Disaster Risk Reduction. Springer, Berlin, Heidelberg, pp. 159–176. Keefer, D.K., 1994. Landslides caused by earthquakes. Geol. Soc. Am. Bull. 95, 406– 421. Khan, S.F., Kamp, U., Owen, L.A., 2013. Documenting five years of landsliding after the 2005 Kashmir earthquake, using repeat photography. Geomorphology 197, 45–55. Khattak, G.A., Owen, L.A., Kamp, U., Harp, E.L., 2010. Evolution of earthquaketriggered landslides in the Kashmir Himalaya, northern Pakistan. Geomorphology 115, 102–108. Kiyota, T., Sattar, A., Konagai, K., Kazmi, Z.A., Okuno, D., Ikeda, T., 2011. Breaching failure of a huge landslide dam formed by the 2005 Kashmir earthquake. Soils Found. 51, 1179–1190. Konagai, K., Sattar, A., 2012. Partial breaching of Hattian Bala Landslide Dam formed in the 8th October 2005 Kashmir Earthquake, Pakistan. Landslides 9, 1–11. Korup, O., 2010. Earthquake-Triggered Landslides – Spatial Patterns and Impacts. Swiss Federal Research Institutes WSL/SLF, Davos. Latif, A., Afridi, A.G.K., Majid, A.N., 2008. Geological Map of the Balakot Quadrangle (Earthquake Affected Area), District Mansehra, NWFP, Pakistan, Geological Map Series. Geological Survey of Pakistan, Islamabad-Pakistan. Lee, S.-J., Komatitsch, D., Huang, B.-S., Tromp, J., 2009. Effects of topography on seismic-wave propagation: an example from northern Taiwan. Bull. Seismol. Soc. Am. 99, 314–325. Lee, S.T., Yu, T.T., Peng, W.F., Wang, C.L., 2010. Incorporating the effects of topographic amplification in the analysis of earthquake-induced landslide hazards using logistic regression. Nat. Hazards Earth Syst. 10, 2475–2488. Lodhi, M.A., 2011. Earthquake-induced landslide mapping in the western Himalayas using medium resolution ASTER imagery. Int. J. Remote Sens. 32 (19), 5331– 5346. Martha, T.R., Kerle, N., Jetten, V., van Westen, C.J., Kumar, K.V., 2010. Characterising spectral, spatial and morphometric properties of landslides for semi-automatic detection using object-oriented methods. Geomorphology 116, 24–36. Meunier, P., Hovius, N., Haines, J.A., 2008. Topographic site effects and the location of earthquake induced landslides. Earth Planet. Sci. Lett. 275, 221–232. Meunier, P., Uchida, T., Hovius, N., 2013. Landslide patterns reveal the sources of large earthquakes. Earth Planet. Sci. Lett. 363, 27–33. Nefeslioglu, H.A., San, B.T., Gokceoglu, C., Duman, T.Y., 2012. An assessment on the use of Terra ASTER L3A data in landslide susceptibility mapping. Int. J. Appl. Earth Obs. Geoinf. 14, 40–60. Nichol, J.E., Shaker, A., Wong, M.-S., 2006. Application of high-resolution stereo satellite images to detailed landslide hazard assessment. Geomorphology 76, 68–75. Ninomiya, Y., Fu, B., Cudahy, T.J., 2005. Detecting lithology with Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) multispectral thermal infrared ‘‘radiance-at-sensor” data. Remote Sens. Environ. 99, 127–139.

79

Ouzounov, D., Freund, F., 2004. Mid-infrared emission prior to strong earthquakes analyzed by remote sensing data. Adv. Space Res. 33, 268–273. Owen, L.A., Kamp, U., Khattak, G.A., Harp, E.L., Keefer, D.K., Bauer, M.A., 2008. Landslides triggered by the 8 October 2005 Kashmir earthquake. Geomorphology 94, 1–9. Parkash, S., 2013. Earthquake related landslides in the Indian Himalaya: experiences from the past and implications for the future. In: Margottini, C., Canuti, P., Sassa, K. (Eds.), Landslide Science and Practice. Springer, Berlin, Heidelberg, pp. 327–334. Parker, R.N., Densmore, A.L., Rosser, N.J., de Michele, M., Li, Y., Huang, R., Whadcoat, S., Petley, D.N., 2011. Mass wasting triggered by the 2008 Wenchuan earthquake is greater than orogenic growth. Nat. Geosci. 4, 449–452. Parvaiz, I., Champatiray, P.K., Bhat, F.A., Dadhwal, V.K., 2011. Earthquake-induced landslide dam in the Kashmir Himalayas. Int. J. Remote Sens. 33, 655–660. Pathier, E., Fielding, E.J., Wright, T.J., Walker, R., Parsons, B.E., Hensley, S., 2006. Displacement field and slip distribution of the 2005 Kashmir earthquake from SAR imagery. Geophys. Res. Lett. 33, 1–5. Peduzzi, P., 2010. Landslides and vegetation cover in the 2005 North Pakistan earthquake: a GIS and statistical quantitative approach. Nat. Hazards Earth Syst. Sci. 10, 623–640. Petley, D., Dunning, S., Rosser, N., Kausar, A.B., 2006. Incipient landslides in the Jhelum Valley, Pakistan following the 8th October 2005 earthquake. In: Marui, H. (Ed.), Disaster Mitigation of Debris Flows, Slope Failures and Landslides, Frontiers of Science Series. Universal Academy, Tokyo, Japan. Poiraud, A., 2014. Landslide susceptibility–certainty mapping by a multi-method approach: a case study in the Tertiary basin of Puy-en-Velay (Massif central, France). Geomorphology 216, 208–224. Rahman, A.-U., Khan, A., Collins, A., 2014. Analysis of landslide causes and associated damages in the Kashmir Himalayas of Pakistan. Nat. Hazards 71, 803–821. Ray, P.K.C., Parvaiz, I., Jayangondaperumal, R., Thakur, V.C., Dadhwal, V.K., Bhat, F.A., 2009. Analysis of seismicity-induced landslides due to the 8 October 2005 earthquake in Kashmir Himalaya. Curr. Sci. 97, 1742–1751. Rodriguez, C.E., Bommer, J.J., Chandler, R.J., 1999. Earthquake induced landslides: 1980–1997. Soil Dyn. Earthq. Eng. 18, 325–346. Rowan, L.C., Mars, J.C., 2003. Lithologic mapping in the Mountain Pass, California area using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data. Remote Sens. Environ. 84, 350–366. Saba, S.B., van der Meijde, M., van der Werff, H., 2010. Spatiotemporal landslide detection for the 2005 Kashmir earthquake region. Geomorphology 124, 17–25. Salati, S., van Ruitenbeek, F.J.A., van der Meer, F.D., Tangestani, M.H., van der Werff, H., 2011. Lithological mapping and fuzzy set theory: automated extraction of lithological boundary from ASTER imagery by template matching and spatial accuracy assessment. Int. J. Appl. Earth Obs. Geoinf. 13, 753–765. Sato, H.P., Hasegawa, H., Fujiwara, S., Tobita, M., Koarai, M., Une, H., Iwahashi, J., 2007. Interpretation of landslide distribution triggered by the 2005 Northern Pakistan earthquake using SPOT 5 imagery. Landslides 4, 113–122. Sattar, A., Konagai, K., Kiyota, T., Ikeda, T., Johansson, J., 2011. Measurement of debris mass changes and assessment of the dam-break flood potential of earthquake-triggered Hattian landslide dam. Landslides 8, 171–182. Sayab, M., Khan, M.A., 2010. Temporal evolution of surface rupture deduced from coseismic multi-mode secondary fractures: insights from the October 8, 2005 (Mw 7.6) Kashmir earthquake, NW Himalaya. Tectonophysics 493, 58–73. Schneider, J., 2009. Seismically reactivated Hattian slide in Kashmir, Northern Pakistan. J. Seismol. 13, 387–398. Searle, M.P., Khan, M.A., 1996. Geological and Tectonic map of northern Pakistan and adjacent areas of northern Ladakh and Western Tibet (coloured, 1: 650,000). Sponsored Publication of British Petroleum, Shell & AMOCO. Sepulveda, S.A., Murphy, W., Jibson, R.W., Petley, D.N., 2005. Seismically induced rock slope failures resulting from topographic amplification of strong ground motions: the case of Pacoima Canyon, California. Eng. Geol. 80, 336–348. Shafique, M., van der Meijde, M., 2014. Impact of uncertainty in remote sensing DEMs on topographic amplification of seismic response and Vs 30. Arab. J. Geosci., 1–9 Shafique, M., van der Meijde, M., Kerle, N., van der Meer, F., Khan, M.A., 2008. Predicting topographic aggravation of seismic ground shaking by applying geospatial tools. J. Himalayan Earth Sci. 41, 33–43. Shafique, M., van der Meijde, M., Kerle, N., van der Meer, F., 2011a. Impact of DEM source and resolution on topographic seismic amplification. Int. J. Appl. Earth Obs. Geoinf. 13, 420–427. Shafique, M., van der Meijde, M., Ullah, S., 2011b. Regolith modeling and its relation to earthquake induced building damage: a remote sensing approach. J. Asian Earth Sci. 42, 65–75. Shafique, M., van der Meijde, M., van der Werff, H.M.A., 2012. Evaluation of remote sensing-based seismic site characterization using earthquake damage data. Terra Nova 24, 123–129. Singhroy, V., 2009. Satellite remote sensing applications for landslide detection and monitoring. In: Sassa, K., Canuti, P. (Eds.), Landslides – Disaster Risk Reduction. Springer, Berlin, Heidelberg, pp. 143–158. Singhroy, V., Molch, K., 2004. Characterizing and monitoring rockslides from SAR techniques. Adv. Space Res. 33, 290–295. Siyahghalati, S., Saraf, A.k., Pradhan, B., Jebur, M.n., tehrany, M.s., 2014. Rule-based semi-automated approach for the detection of landslides induced by 18 September 2011 Sikkim, Himalaya, earthquake using IRS LISS3 satellite images. Geomatics, Nat. Hazards Risk, 1–19.

80

M. Shafique et al. / Journal of Asian Earth Sciences 118 (2016) 68–80

Sudmeier-Rieux, Qureshi, R.A., Peduzzi, P., Jaboyedoff, M.J., Breguet, A., Dubois, J., Jaubert, R., Cheema, M.A., 2007. An interdisciplinary approach to understanding landslides and risk management: a case study from earthquake-affected Kashmir. Mountain Forum, Mountain GIS e-Conference, January 14–25, 2008, p. 16. Sudmeier-Rieux, K., Jaboyedoff, M., Breguet, A., Dubois, J., 2011. The 2005 Pakistan earthquake revisited: methods for integrated landslide assessment. Mountain Res. Dev. 31, 112–121. Tang, C., Zhu, J., Qi, X., Ding, J., 2011. Landslides induced by the Wenchuan earthquake and the subsequent strong rainfall event: a case study in the Beichuan area of China. Eng. Geol. 122, 22–33. Tatard, L., Grasso, J.R., 2013. Controls of earthquake faulting style on near field landslide triggering: the role of coseismic slip. J. Geophys. Res.: Solid Earth 118, 2953–2964. UNISDR, 2005. Hyogo framework for action 2005–2015: building the resilience of nations and communities to disasters. In: Reduction, U.N.I.S.f.D. (Ed.), World Conference on Disaster Reduction, 18–22 January 2005, Kobe, Hyogo, Japan. Van Westen, C.J., Getahun, F., 2003. Analyzing the evolution of the Tessina landslide using aerial photographs and digital elevation models. Geomorphology 54, 77– 89. van Westen, C.J., Rengers, N., Soeters, R., 2003. Use of geomorphological information in indirect landslide susceptibility assessment. Nat. Hazards 30, 399–419. Vanacker, V., Vanderschaeghe, M., Govers, G., Willems, E., Poesen, J., Deckers, J., De Bievre, B., 2003. Linking hydrological, infinite slope stability and land-use change models through GIS for assessing the impact of deforestation on slope stability in high Andean watersheds. Geomorphology 52, 299–315.

Varnes, D.J., 1978. Slope movement types and processes. In: Schuster, R.L., Krizek, R. J.E. (Eds.), Landslides-Analysis and Control: National Academy of Sciences Transportation Research Board Special Report, Washington, DC, pp. 12–33. Varnes, D.J., 1984. Landslide Hazard Zonation: A Review of Principles and Practice/ David J. Varnes and the International Association of Engineering Geology, Commission on Landslides and Other Mass Movements on Slopes. UNESCO, Paris. Wang, W.-N., Wu, H.-L., Nakamura, H., Wu, S.-C., Ouyang, S., Yu, M.-F., 2003. Mass movements caused by recent tectonic activity: the 1999 Chi-chi earthquake in central Taiwan. Island Arc 12, 325–334. Wasowski, J., Keefer, D.K., Lee, C.-T., 2011. Toward the next generation of research on earthquake-induced landslides: current issues and future challenges. Eng. Geol. 122, 1–8. Weissel, J.K., Stark, C.P., 2001. Landslides triggered by the 1999 Mw7.6 Chi Chi earthquake in Taiwan and their relationship to topography. In: Geoscience and Remote Sensing Symposium, 2001. IGARSS’01, vol. 752. IEEE 2001 International, pp. 759–761. WMO, 2007. World Meteorological Organization. (accessed 31 August 2007). Xu, C., 2014. Preparation of earthquake-triggered landslide inventory maps using remote sensing and GIS technologies: principles and case studies. Geosci. Front. Zhang, S., Zhang, L.M., Glade, T., 2014. Characteristics of earthquake- and raininduced landslides near the epicenter of Wenchuan earthquake. Eng. Geol. Zhao, C., Lu, Z., Zhang, Q., de la Fuente, J., 2012. Large-area landslide detection and monitoring with ALOS/PALSAR imagery data over Northern California and Southern Oregon, USA. Remote Sens. Environ. 124, 348–359.