PROBABILITY OF LANDSLIDE OCCURENCE MAPPING USING PROBABILITY DENSITY FUNCTION: A CASE STUDY OF THE MAE THA GROUP IN NAM LI WATERSHED, THAILAND

摘要

Due to the spatial complication of mixing kinds of rocks in a geological mapping unit and its structure, any quantitative engineering property related to the landslide is not a single value. It exists in a range. This leads to existence in range of the landslide susceptibility indexes (LSIs) calculated from varying properties of a rock unit. The objective of the study is to analyze the probability of landslide occurrence of the Mae Tha Group in Nam Li watershed, Thailand using Probability Density Function (PDF). Referring to Tanang et al. (2010), by using GIS analysis, the range of LSIs in this rock unit (0.63-1.603) associated with landslide scars could be extracted. In this study, these LSIs were sliced into 11 ranges captured by 12 values (0.63, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, and 1.603) which were candidates to be a critical LSI. Performing on each TIN polygon of each candidate, slope of the polygon, the mean and standard deviation of LSIs calculated from the possible composite properties of the rock unit were input to PDF and resulted in the Probability of Failure Indexes (PFIs). For each candidate, a raster map layer of PFIs was generated and modified to be PFI-1 layer. Any cell with PFI equal to 1 was kept as 1 or as 0 otherwise. Then, overlay analysis of a PFI-1 raster layer of each candidate and landslide scar layer was operated to obtain the fitting ratio. Any PFI-1 layer depicting the most fitting ratio was identified as the layer of critical LSI, which herein was 1.2. The PFIs raster layer of the critical LSI was then classified into 5 classes - very low, low, moderate, high, and very high, corresponding to the probability of landslide occurrence in the rock unit.