Evaluation of the best important Landslide hazrad zonation methods in Semirum region (in part of Marbor River). (2000-2003)

In this research, Marbor river catchment's (hydrologic unit of Kahredan) with 800 square kilometers area, located in south part of Esfahan province is studied. Distribution map of landslide was obtained by using aerial photos, geological maps and field survey. Moreover by Using, There information sources such as maps, the factors affecting on landslide (8 factors) determined and proper maps in 1:50000 scales was prepared. Order to promoting accuracy, speed and facility, descriptive and spatial data digitized using Ilwis software. A number of 7 conventional methods in Marbor catchment were analyzed. Before analyzing of the methods, using effective parameters in each method, landslide hazard map with appropriate classes was prepared. The using landslide Index equation and landslide precision, different methods were analyze and compared. In Multivariate linear regression method, 17 classes of parameters were used in the equation in 99 percent confidence level in which correlation coefficient, R, is 70 percent. In Multivariate discriminat analysis, regarding of no groups i.e. no landslide, discriminant factor based on 17 classes obtained. The accuracy of stable to non-stable grouping about 64 percent. Moreover chi-square test for discriminant factor was 556 in compare with discriminate coefficient (a) and table is acceptable in 99 percent of confidence level. The using landslide Index equation and landslide precision, different methods were analyze and compared. Using the obtained index and comparing landslide distribution map with landslide hazard zonation map (obtained with each method) show the best methods as below: Multivariate using linear regression analysis, multivariate using discriminant analysis, bivariate method of Information Value, Bivariate method of Density Area, Revised Nilsen method, Revised MoraandVarson method and MoraandVarson method. As a result, statistical multivariate methods consider not only, the role of each factor in landslide but also the effectes of simultaneous dependent factors so the ultimate map is better overlaped to landslide potential of the region. There is no doubt that more accuracy in mapping of effective parameters will affects on output map directly.