SEISMICALLY INDUCED LANDSLIDE HAZARD ANALYSES FOR LOWER INDIAN HIMALAYA

Abstract

The entire Himalayan arc is recognized as a global hotspot for landslide and seismic events; which may be ascribed to the orogeny processes that had formed the Himalaya. Every year landslides and related natural disaster events claim many lives and destroy property, infrastructure, and the environment of the Himalaya. It is estimated that Himalayan landslides kill 1 person/100 sq. km per year and average losses due to Himalayan landslides is more than USD 1 million/year. Given the great relief, high seismicity, active tectonism, high volume of precipitation, and wide variety of rock and sediment types; landslides seem ubiquitous in the Himalaya and is perhaps the major present-day process shaping the landscape. With ingress of roads and other heavy constructions like dams and hydro power plants in this fragile mountain chain, the overall risk of landslide hazard increases manifold. This necessitates an accurate and updated landslide hazard zonation (LHZ) for the Himalayan belt, based on which future landuse pattern can be envisaged. LHZ is a scientific practice of predicting the spatial distribution of landslides over a region which is determined as a function of landslide occurrence and various landslide related factors. Considering the high incidences of landslide disasters and their long term socio-economic impact, national guidelines are drafted to guide the activities envisaged for mitigating landslide risk through Landslide Hazard Zonation (LHZ) mapping. The classical approach of LHZ mapping is based on examination of various static landslide causative factors with occasional inclusion of triggering factors like rainfall and earthquakes. As the positive correlation between seismicity and landslide occurrence had become more and more prominent, the classical approaches got changed; and a paradigm shift has been observed in LHZ studies of-late. More emphasis is now given on comprehending the occurrence and mechanism of seismically induced landslides due to the complexity and enormity of such events. There have been several earthquakes in the Himalayan region viz. Chamoli earthquake (Mw-6.8), Kashmir earthquake (Mw-7.6), Sikkim earthquake (Mw-6.9), Nepal earthquake (Mw-7.8), which caused widespread landslide events. In fact, in many cases, losses due to seismically induced landslides have been more than those caused directly due to shaking. Out of the all earthquake related casualties, which are not caused directly by ground shaking, approximately 70% may be attributed to landslides. In this context, LHZ studies considering earthquakes as main triggering factor is a time bound priority for the Himalayan region. However, a critical review of the existing literature reveals that there is a paucity of macro-scale, regional level studies quantifying the role of seismicity in LHZ mapping for the Himalayan arc in general, and the lower Himalayan belt in particular. An endeavour has been iv made in this research work to carry out LHZ mapping under both static and seismic conditions for a part of lower Indian Himalaya. The research work has been carried out in three phases: (a) in the first phase, the study area's present scenario of landslide susceptibility under static conditions is assessed using statistical method of LHZ mapping; (b) in the second phase, suitable method for carrying out seismically induced LHZ mapping is formulated; and (c) in the third phase, LHZ maps of the study area are prepared under different seismic conditions to quantify the role of seismicity in landslide occurrence and spatial distribution in the study area. The study area encompasses approximately 12,350 sq. km.; with estimated population of more than 15 lakhs as per the 2011 India census. Several important and thickly populated cities of Uttarakhand and Himachal Pradesh are located in the study area. Geologically, the study area exhibits a complex and heterogeneous amalgamation of fifteen formations from different ages'. The study area falls in Zone IV as per IS 1893(Part I): 2016, indicating that the whole study area is seismically very active. The area caters three major thrusting systems of the Himalayan arc: Main Frontal Thrust (MFT), Main Boundary Thrust (MBT) and a portion of Main Central Thrust (MCT), along with numerous transverse lineaments. In the first phase, eight static landslide causative parameters are identified for the study area. A comprehensive landslide inventory has been prepared, which is the primary step in LHZ mapping and data has been extracted from various sources. The prepared landslide inventory is used for proximity analyses to establish correlation between landslide activities and various causative parameter. Information Value method, one of the widely used statistical methods of LHZ mapping, has been applied to prepare the initial LHZ map of the study are under static causative parameters. The prepared LHZ map has identified almost 37% of the total study area as the zones of high to very high landslide susceptibility. Different statistical methods, which are widely used for landslide susceptibility assessment, generally lack in incorporating seismic indicators. This may be attributed to the paucity of sufficient earthquake induced landslide inventories, which is attributed to the rarity of an extreme earthquake event. Moreover, the conventional studies correlating earthquake magnitude and landslide distribution, types and coverage area drew criticism from researchers due to limitations of the dataset used and the regional and characteristic biasness associated with earthquake events. Such scenarios become exaggerated for the Himalayan region, where not until recently, much attention have been paid to seismically induced landslide hazard zonation. Most of the LHZ studies carried out for the Himalayan belt considered static landslide causative factors only; and the few studies that did consider earthquake scenarios, are concentrated around the Chamoli earthquake, Sikkim earthquake and Nepal earthquake. All v these earthquakes, in-spite-of having originated in the Himalaya only, differ significantly from one another in terms of their characteristics. Thus, it is understood that any statistical method derived from earthquake induced landslide inventory developed for a particular earthquake event may not be adequate enough for a different tectonic set up. Alternatively, Map combination method of LHZ mapping has been used in this research work where, probabilistically generated peak ground acceleration (PGA) is considered as landslide triggering seismic factor. The biggest advantage of this method is that various landslide causative parameters (static as well as triggering) can be incorporated as thematic layers, which are assigned a weight depending upon their perceived control on landslide occurrence. The weights of various thematic layers are numerically integrated to generate the LHZ map. However, the subjectivity in weight assignment procedure is the main limitation in this method. To address this issue, a landslide susceptibility scale is developed for the study area statistically, which is used to assign the weights of various thematic classes. Information Value method, Frequency Ratio method and Fuzzy Cosine Amplitude Methods are correlated to develop the susceptibility scale, which is further used for multi-hazard integration. The LSZ map prepared using the developed scale, is compared with other LSZ maps prepared using statistical methods for performance evaluation of the developed susceptibility scale. The developed susceptibility scale has produced better results for the study area. Use of probabilistic PGA values as landslide triggering factor in LHZ mapping has a distinct advantage: it eliminated the regional and characteristic biasness associated with an single earthquake, which increases the applicability of the method. The predicted PGAs are not from a single event, but rather represents the stress deformation expected in the region. Moreover application of PSHA in LHZ allows incorporation of seismotectonic environment (in terms of faults and lineaments) of a bigger area (R~300 km) which would likely to produce earthquakes in the study area, and recorded past seismicity. A detailed PSHA study has been carried out for the study area. The results of PSHA is discussed in terms of expected PGA for five scenario earthquakes with return periods of 10, 50, 100, 225 and 475 years. Consideration of the entire range of earthquake sizes quantifies the impact and implications of seismicity in landslide hazard comprehensively. Assignment of weights to different earthquake scenarios is a difficult task in LHZ mapping. There is no statistical correlation available to quantify the size of a scenario earthquake with landslide spatial distribution. Therefore a new method has been implemented in this research work to assign the weights objectively. The method, which is based on the normalized PGA values of different scenario earthquakes, could portray the relative importance vi of different earthquake size on LHZ mapping effectively. Five LHZ maps of the study area are prepared under seismic conditions to understand the role and impact of seismicity in landslide occurrence and their spatial distribution in the lower Himalaya. It is observed that for an earthquake scenario with 475 years return period, almost 51% of the total area falls under very high landslide hazard. This is a significant outcome of the study, which highlights the consideration of seismicity in LHZ mapping for the Himalayan arc. The results of the research work shows that in case of moderate to great earthquakes, there is paradigm shift of hazard zones from very low towards very high. Based on the results, the present study concludes that inclusion of earthquake scenarios will enhance the understanding of landslide hazard with a more pragmatic vision, especially for seismically active mountainous belts like the Himalaya. The LHZ maps prepared for the scenario earthquakes with 225 years and 475 years return period will be of practical use for implementing frame works for risk mitigation and disaster response.