A GIS System Development for Evaluating 3D Slope Stability

When using a 3D deterministic model to evaluate the landslide susceptibility of regional area, processing large amounts of spatial data and identifying the objective for applying the deterministic model would be arduous works. In this research, an enhanced hydrology tool is developed to automatically identify the objectives of research, which are called slope units, from terrain raster data. A new grid-based 3D slope stability model is applied to evaluate the failure probability. Finally, the landslide susceptibility map showing the location and the shape of potential slopefailures is obtained. All the approaches are implemented by an integrated system which has been developed in ArcGIS 9.0 using ArcObjects. This system convert all related input data, including topographic and geological raster data, to a point dataset for calculation. The output will be a polygon shapefile of the critical slip surfaces with the attributes of failure probability and the minimum safety factor. Introduction In Japan, about 70% of the total national territory is composed of mountainous terrain, where catastrophic slope failures occur on an annual basis. The identification of potential slope failure zones has long been an important issue in landslide mitigation in Japan. So far, a large number of methods for slope stability calculation have been developed. Basically, they can be categorized to two types: statistical methods and deterministic models. Considering the statistical method has the shortage of mechanical meaning, the deterministic model, which calculates a safety factor quantitatively, is a more reasonable method for an individual slope. However, applying the deterministic model to a regional slope can be particularly difficult or sometimes even impossible because of the difficulties in obtaining, checking and processing large spatial data sets, defining the applied objective of deterministic model and identifying unknown slide surface. Therefore, a helpful tool is expected for resolving these problems. The Geographic Information System (GIS), with its power and versatility for processing spatial data, has attracted significant attention for the assessment of natural disasters. The GIS provides strong functions both in geostatistical analysis and database processing. In addition, the extension of the analysis to include environmental impact assessment of a slope failure can be easily and effectively performed using GIS. In this paper, a new GIS method is proposed for identifying the applied objectives of deterministic model, which are called slope unit, from the DEM (Digital Elevation Model) data. After dividing the study area to many slope units, a GIS-based 3D slope stability evaluation system is developed to identify a critical slip surface which has the minimum 3D safety factor for each slope unit, and as the result, a map that shows the distribution of probability of slope failure is obtained. As a case study, the probability of slope failure of a wide weathered granite zone along the route No.49 in Gouto area of Japan is evaluated quantitatively. The location and the shape of potential slip surfaces also are predicted effectively. Hydrological Tool for Dividing Slope Units (1) Concept of Slope Unit In the 3D slope stability studies focus on a large mountain area with complicated geometry and soil and water conditions, a very important problem is how to extract appropriate study objects. These study objects are usually called slope units. The slope unit, namely, the portion of land surface that contains maximum internal homogeneity differing from the adjacent units, has relatively similar topographic and geological characteristics respectively. The significance of partition of the slope units for landslide hazard assessment and for other land-related study has been recognized. (2) Method of Dividing Slope Unit Breaks of slope are often identified as significant topographic features, namely dividing lines, indicating the boundaries between adjacent geomorphological units on a map. Since it is virtually impossible to consistently draw dividing lines on topographic maps covering large regions, an automatic computer procedure is required. In this paper, a new method based on GIS spatial analysis and hydrology modeling has been proposed for identifying slope units from DEM (digital elevation model) data automatically. The slope unit can be considered as the left or right side of a sub-basin of any order into which a watershed can be partitioned, therefore, it can be identified by a ridge line and a valley line. By using the hydraulic model tool of ArcGIS 9.0, the watershed polygon and the stream line of a study area can be obtained easily from the DEM data. Topologically, the outline of the watershed polygon can be considered as the ridge line. Considering water always flows downhill along the path of the steepest descent, the stream line can be looked upon as a part of the valley line in a mountain area. After dividing the watershed polygon by the valley line, two slope units can be obtained. However, the stream line which can be created by the readymade GIS hydraulic model tool is not a single line but all lines that consist of cells whose flow accumulation value are greater than a specified threshold value. Because there is not a ready-made application that can extract the valley line from watershed and create slope unit automatically, a GIS-based program has been developed to implement this task. Fig.1 shows the distribution of slope units extracted from a study area by this program. Fig.1 3D view of slope unit distribution 3D Slope Stability Model All slope failures have a three-dimensional (3D) geometry, which varies in space even along a short distance. Therefore, it is rational to use a 3D model to analyze slope stability. Since the middle of the 1970s, increasing attention has been directed toward the development and application of threedimensional stability models. Several threedimensional methods of analysis have been proposed in geomechanical literature. However, these methods are commonly limited on being applied for an individual site due to the difficulties in large spatially distributed data processing. Considering that most of the methods have used the column method and all of slope-related GIS data can be changed to grid-based data, it is possible that these column–based 3D models can be used for the 3D stability calculation by using the GIS grid-based data. By combining the GIS spatial analysis function with an improved Hovland (1977) 3D slope stability analysis model, a new GIS-based 3D deterministic model is applied. (1) GIS-based 3D Model Using the functions of GIS spatial analysis, all original data (such as elevation, underground water, strata etc.) for the 3D safety factor calculation are available with respect to each grid pixel. By inputting these data into a deterministic model of slope stability, a value of the safety factor can be calculated. The global geometry of a potentially sliding mass is illustrated in Fig.2(a). In GIS system, by using the GIS spatial analysis function, the slope-stabilityrelated data of the whole study area can be represented as the GIS vector layers. For each layer, a grid-based layer (see Fig.2(b)) can be obtained by using the GIS spatial analysis function and the grid size can be set with the requisite precision. The stability of the landslide is related to the geological information, the geomorphological aspect, the geomechanical parameters, and the hydraulic condition, by the discretization of the study mass to the small soil column, as shown in Fig.2(c). The slope failure has now been evenly divided into columns and related to grid-based columns. With reference to the gridbased column in Fig.2(d), the spatial data of surface, strata, underground water, fault, and slip surface can be obtained from the grid-based layers. Because there are so many of slope-related data, it is not effective to manage all these grid-based datasets. Therefore, a point dataset is used to store all these grid datasets. In this point dataset, a feature table is used to manage all the slope-related data.