International Journal of Science and Research (IJSR) ISSN (Online): 2319-7064 Index Copernicus Value (2013): 6.14 | Impact Factor (2014): 5.611 Volume 4 Issue 10, October 2015 www.ijsr.net Licensed Under Creative Commons Attribution CC BY Weights of Evidence Method for Landslide Susceptibility Mapping in Tandikek and Damar Bancah, West Sumatra, Indonesia Sumaryono 1 , Dicky Muslim 2 , Nana Sulaksana 3 , Yunara DasaTriana 4 1 Geological Agency of Indonesia and Ph.D student of Padjadjaran University, Bandung, Indonesia 2, 3 Faculty of Geological Engineering, Padjadjaran University, Jalan Raya Bandung-Sumedang Km.21, Jatinangor, Indonesia 4 Geological Agency of Indonesia, Diponegoro No 57, Bandung, Indonesia Abstract: The September 30, 2009, Padang earthquake has induced huge landslide in Padang Pariaman district, West Sumatera, Indonesia. The goal of this paper is to assess the landslide susceptibility of Pariaman, using the weight of evidence statistical method. The method is based on the decision of which state is more likely to occur grounded on the presence or absence of a predictive variable and the occurrence of an event (e.g., landslide) within a pixel. This method can be used without requiring geotechnical, groundwater or failure depth data. However, the other factors to influence landslide occurrence, such as slope, curvature, peak ground acceleration, geological condition, lineaments/faults distance of landslide to the rivers and, land use can be known. This is quite impressive, based on AUC The highest factor controlling the landslide is lineament, than following with lithology, peak ground acceleration, slope and land use. Weights calculated individually for those themes were added to produce a probability estimate of the area. The predictive of the map was tested on the basis of 35% sample of landslides that were not used in the modeling process (test data). Analytical result verified by using test data of landslide shows AUC prediction rate is 0.849 and AUC success rate using all landslide data is 0.894 with difference 0.045. This condition to the allowed tolerance of 15%. This is showing good model of landslide susceptibility. The obtained landslide susceptibility map and landslide inventory data base can be used for landslide hazard prevention and mitigation, and proper planning for land use in the future. Keywords: Weights of evidence, landslide susceptibility map, area under curve. 1. Introduction Landslides are a complex natural phenomenon that constitutes a serious natural hazard in many countries (Hutchinson, 1988; Cruden and Varnes, 1996). Nakano (2015) founds that the landslide at Padang Pariaman strongly controlled by geological factors as weathering of mixed layer at the base of pumice fall deposit, slope paralel to the pumice fall deposits, which had been cut at the foot slope and undercutting by river incision Landslide susceptibility maps can be obtained through different methods: by empirical or heuristic, by qualitative or semi-quantitative combination of thematic layers, which are estimated as correlated with landslide occurrences (Rupke et al. 1988), or using deterministic and statistical techniques (descriptive, univariate, bivariate, or multivariate statistical analysis, Lee et al. 2002). In spatial analysis, WOE was proposed by Agterberg and developed in the mineral and mining fields in late 1980s. WOE modeling for landslide susceptibility mapping, using the log-linear form of the Bayesian probability model, has been recently applied in geomorphologic risk assessments (van Westen 1993; van Westen 2002 ; van Westen et al. 2003, 2006; Lee et al.2002; Ranjan et al. 2008; Barbieri and Cambuli 2009). Bivariate statistical analysis methods such as WoE is one of the methods used to conduct landslide susceptibility mapping. This method utilizing historical data events to gain patterns geofactor or parameters that controlling and influence of the landslide occurrence (Sumaryono. 2013). GIS-based statistical methods have become very popular in landslide susceptibility assessment (van Westen et al. 2006), because of effective data management, simultaneous use, graphic and attribute crossing of these digital layers, and providing accurate output data and superior image quality. Advantages of this method are the accuracy that can be accounted and can be done quickly. Several statistical approaches that have been discussed and used in Van Westen, 2003, and is: "Information Value", "Frequency Ratio", "and Weight of Evidence". 2. Study Area Pariaman regency located in West Sumatra Province, Indonesia (Figure 1). Landslide in this area triggered by West Sumatra Earthquake, 30 September 2009. Based on GIS analysis, West Sumatra Earthquake trigger 154 landslides. Extensive landslides occurred in Agam and Padang Pariaman Regency, West Sumatra Province, Indonesia, causing more than 250 peoples lost their live, and burying some villages. The landslides occurred during rainfall, and originated on mountains mantled with loose pumice. Translational slides and flowside are spread over the entire area on hornblende hypersthene pumiceous tuff. A lithological unit affected by translational slides is hornblende hypersthene pumiceous tuff underlying clay layer from the thepra alteration. From a geological point of view, shows the landslide occurred dominantly on hornblende hypersthene Paper ID: SUB158816 1283