地理学与自然灾害杂志
开放获取

抽象的

A Novel Methodology for Large-Scale Landslide Susceptibility Mapping Based on Landslide Domain; An Effective Tool for Local Landslide Management and Planning: A Case Study from Darjeeling Himalaya, India

Sumit Kumar1*, Gargi Singh1, Rabisankar Karmakar1, Akshay Kumar Mishra1, Mathiyalagan Rajkumar1, Abhisek Kumar1, Pankaj Jaiswal2, Saibal Ghosh1

A Mesoscale scale landslide susceptibility methodology has been derived with an aim to produce a susceptibility map that can define the susceptibility of a slope to a specific process of landsliding so that it can be used by stakeholders for detailed planning in landslide management and infrastructure developement. The methodology is based on intensive field and laboratory input. The mapping in such a large (1:10.000 or larger) scale helped in defining the domains based on slope material and processes. Domain specific geo factors were derived using the knowledge of landslides in the terrain. To define the actual slope mass character, geotechnical map was prepared using RMR_basic and slope for rocky areas; C and Φ values as well as the slope, in debris and soil covered areas. The modelling was carried out in ArcGIS 10.3. Various geo-factors for landsliding like geotechnical properties of slope forming material, structure, Stream Power Index (SPI), landform, landuse-landcover, relative relief, regolith thickness etc. has been considered for the study. DEM derived slope and aspect were used for preparing the geotechnical map. Structure, slope, aspect and RMR_basic were used to prepare the Kinematic failure map.

In each domain, the responsible geofactors were evaluated and landslide susceptibility was calculated using Multi class Index Overlay Method (McIOM). A combined susceptibility map was prepared for the study area using the susceptibility condition derived for each domain. However, the susceptibility map thus prepared only indicated landslide initiation susceptibility of the area to consider the run-out susceptibility also, the impact probability map for debris flow was derived using a conceptual model (r.randomwalk) wherein the boundary conditions were defined by using the debris flow inventory. Both the initiation and runout susceptibility map was classified into three classes using natural break values into high, moderate and low.

The debris slide, earth slide, rock slide, rock fall and cut-slope domain is calculated to have 38%, 7.5%, 28.81%, 11.9% and 44% area under high susceptibility respectively.

The methodology developed is highly effective in defining the various spatial proneness of the slopes in the area to specific types of landslide and also defines the impact probability of debris flow along its path of propagation. This output provides immense information essential for local land use planning, managing landslide prone areas, prioritizing developmental activities in different sectors and developing safe communication corridors.