LANDSLIDE SUSCEPTIBILITY ZONATION BASED ON FREQUENCY RATIO (FR), FUZZY LOGIC (FL), AND FR – FL COMBINATION METHODS IN KARAHA BODAS GEOTHERMAL FIELD, WEST JAVA

• Main Author: Havirgus Zahara, Giva
• Format: Theses
• Language: Indonesia
• Subjects: Geologi, hidrologi & meteorologi
• Online Access: https://digilib.itb.ac.id/gdl/view/61492
• Institution: Institut Teknologi Bandung

Indonesia has the potential for geothermal energy which is quite unique because of the location of geothermal sources in mountainous areas with quite steep slopes and quite complex geological structures. In general, geothermal fields have high temperatures located in volcanic areas with high relief, steep morphology, thermal alteration rocks that produce thick soil, and high rainfall. This is a threat in geothermal development which causes landslides. This research aims to evaluate the parameters that have a strong influence on landslides in this research area, map the landslide susceptibility in Karaha Bodas geothermal area with Frequency Ratio (FR) method, Fuzzy Logic (FL) method and FR – FL combination method, and to knowing more precise statistical methods to determine landslide susceptibility zoning in the study area. Based on the inventory of landslide incidents obtained from PVMBG, local BPBD, and remote sensing, there were 538 landslide incidents in this research area. Landslide incidents data are divided into two groups randomly, are training data as much as 70% (377 landslide incidents) and test data as much as 30% (161 landslide incidents). There are sixteen parameters used to determine their effect on landslides, including elevation, slope, slope aspect, curvature, lithology, flow direction, distance to the road, lineament density, distance to lineament, drainage density, distance to drainage, stream power index. (SPI), topographic wetted index (TWI), rainfall, normalized differential vegetation index (NDVI), and land use and land cover (LULC). Based on the value of the area under curve (AUC), the parameters that greatly influence the occurrence of landslides in the study area using the FR method are slope, LULC, drainage density, elevation, distance to road, lithology, distance to drainage, rainfall, TWI, NDVI, and curvature. The parameters that greatly influence the occurrence of landslides in the study area using the FL method are slope, LULC, drainage density, distance to road, elevation, lithology, rainfall, and distance to drainage. The parameters that greatly influence the occurrence of landslides in the study area using the FR - FL combination method are slope, LULC, drainage density, elevation, distance to road, lithology, and distance to drainage. Based on the validation using the ROC curve, it shows that the FR method has an AUC value of 0.786 for the success rate and 0.793 better for the prediction rate, which is higher than using the FL method (AUC value 0.642 for the success rate and 0.651 for the prediction rate) and the FR – FL combination method (AUC value is 0.721 for the success rate and 0.702 for the prediction rate), so it can be said that the FR method based on the ROC curve has a better accuracy rate than other methods. Based on Seed Cell Area Index (SCAI) analysist, the landslide susceptibility zoning map produced by the FR method has better results in very low, medium, and high with highest and lowest the SCAI value. The landslide susceptibility zoning map produced by the FL method has better results in low with high the SCAI value. The landslide susceptibility zoning map produced by the FR – FL combination method has better results in high with lowest the SCAI value. Based on spatial domain analysis, FR method produce a higher level of correct pixels than FL method and FR – FL combination method. Based on this research result, it was concluded that FR method has better accuracy than FL method and FR - FL combination method for landslide susceptibility zonation in Karaha Bodas geothermal area and its surroundings.