RAINFALL INDUCED LANDSLIDE - DETECTION AND DEVELOPMENT OF EARLY WARNING SYSTEM

Abstract

Landslides are a prominent natural disaster that impacts numerous mountainous areas worldwide, like the Himalayas, Andes, and Alps. It can cause injuries, loss of life, and widespread destruction of land and property, leading to substantial economic losses. Mountainous areas heavily rely on road transport, and when landslides obstruct these routes, it brings severe difficulties to the affected communities. Around 12.6% of India's total land area, equivalent to approximately 0.42 million square kilometers, is susceptible to landslide occurrences, with the north region being particularly susceptible. The Himalayan region is characterized by substantial rainfall, an extended monsoon season, heightened seismic activity, a relatively young geological profile, anthropogenic activities and formidable mountain terrain. These factors collectively result in the Himalayas being responsible for more than 70% of fatal landslides worldwide, with a significant share of these destructive events occurring within India. Despite the increasing information and awareness about landslides, the harm and destruction keep on rising with a massive increase in landslide occurrence throughout the rainy season. Therefore, conducting a well-targeted research study is of utmost importance to understand the mechanism of rainfall-induced landslides and their potential threat to human life and property. Due to the unpredictable nature of landslides, understanding the complex mechanisms behind such events is crucial for effective monitoring and mitigation strategies. In this study, physical and numerical modelling methods have been utilized to effectively study the mechanism and determine the relative factor of safety under the given rainfall condition. Physical modelling offers a valuable approach to simulate and analyse the processes that trigger and control landslide occurrences under varying rainfall conditions. Hydro-mechanical parameters have been calculated, and a semi-similar material physical model test has been conducted to study the sliding mechanisms. In order to simulate the desired rainfall, a self-developed artificial rainfall generator is used. Furthermore, numerical modelling has been employed to determine the safety factor under dry and rainfall conditions. The study further affirms the validation of numerical simulations when compared to physical modelling. This validation is particularly valuable given the inherent complexities associated with physical modelling, making numerical modelling a more feasible alternative. v In this study, two landslide sites, Jhakri (N31º29’08”, E77º41’43”) in Shimla district and Kotrupi (N31º54‟37.60, E76º53”26.30) in the Mandi district of Himachal Pradesh, lying in the northern region of the Indian subcontinent have been selected. Laboratory investigation has been performed to determine the geotechnical parameters of soil, which have further been utilised in physical and numerical modelling. The physical modelling performed for the Jhakri landslide revealed that a rainfall depth of 80mm and an intensity of 30mm/hr led to debris type of slope failure. The numerical analysis confirmed the slope's stability with a safety factor of 1.23 pre-rainfall and its subsequent instability with a safety factor of 0.626 post-rainfall, highlighting the primary role of rainfall in triggering landslides. Although numerical and physical techniques are frequently used, their limitations in dealing with unpredictable rainfall-induced landslides highlight the importance of sensor-based monitoring. This study introduces an inventive and cost-effective slope monitoring system that incorporates micro-electromechanical system (MEMS) based tilt and moisture sensors. It allows the collection of real-time data on tilt deformation and moisture content. The effectiveness of this monitoring system has been verified using a custom direct shear-based testing setup and physical slope modelling. The developed low-cost monitoring system proved its efficiency in detecting both gradual and sudden movements during rainfall-induced landslides, with precise tilt angle measurements and moisture content readings contributing to its accuracy and precision. The tilt sensor can record even the slightest changes in the slope angle with a precision of 0.01 degree, enabling early detection of slope movement. Additionally, the volumetric water content sensor can detect variations with a precision of 1 percent, aiding in the identification of critical conditions that could lead to landslides. The developed early warning system, designed for identifying impending slope failures, utilized a combination of tilt angle and moisture content variations. Through continuous monitoring, the system observed a gradual shift in the tilt angle of the slope over a two-hour period, displaying a variation ranging from approximately 0.5 degrees to 1.5 degrees. This specific range can be served as a predefined warning threshold. At the crucial second-hour mark, coinciding with the slope failure, there was a sudden and notable deviation 3 degrees to 3.5 degrees in the tilt angle. This deviation acted as a key indicator, marking the system's ability to accurately detect the onset of failure conditions Furthermore, the soil moisture sensor integrated into the system exhibited substantial variations of approximately 40% during periods of rainfall. These variations signalled a notable rise in soil saturation, reaching up to 95%, suggesting that elevated moisture levels may serve as a potential triggering factor for slope failure. The physical and numerical analysis performed on the Jhakri landslide revealed a decrease in the safety factor from 1.045 pre-rainfall to 0.670 post-rainfall conditions, affirming the role of rainfall as the primary trigger for slope failure. The results of the physical and numerical modelling very well establish the initiation of rainfall induced landslides. The factor of safety is one of the crucial parameters in assessing the stability of slopes. The obtained safety factor values highlight the role of hydrology (rainfall) in activating and triggering the failure of slope mass. The findings, thus, corroborate the recent increase in landslide occurrences in the monsoon season. This study further demonstrated the greater significance of hydrological conditions, and recommends Bureau of Indian Standards for emphasizing the importance of assigning higher weightage than other contributing factors. This study also proves the suitability and feasibility of numerical modelling to analyse different slopes, providing scientific guidance for monitoring and early warning so that preventive measures can be taken to reduce its effect. The proposed low-cost monitoring system for rainfall induced landslides is effective and accurate and holds potential for wide-scale implementation in monitoring precarious slopes in hilly terrains.