Introduction:

The term landslide comprises almost all varieties of mass movements on slopes, including some, such as rock falls, topples, and debris flows, that involve little or no true sliding (Varnes, 1984). Landslides are resulted the continuous spatial-temporal processes including some, hydrologic processes (rainfall, evaporation, ground waters), weight of vegetation, root resistance, soil condition, bedrock, topography, and anthropologies activity (Wu & Sidle, 1995). Since the landslide susceptibility maps dramatically  improves land use planning, it can be efficient way to reduce losses of life and property caused by the landslide, so an important step in the management of  landslide hazard to the safety of human lives, infrastructure development and environmental protection (Neuhauser & Terhorst, 2007:12).  So far for landslide susceptibility mapping using heuristic, deterministic, and statistical methods much has been researched. Heuristic method is a qualitive approach, which is based entirely on initial observations and expert knowledge, thus assigning values and weights are subjective and unrepeatable results (Gorsevski et al, 2006 Wati et al, 2010. ( On the other hand, deterministic methods is based on slope stability analysis and they are applicable only when ground conditions across the study area is relatively homogenous and types of landslides have been identified (Dahal et al, 2008 Gokceoglu & Aksoy, 1996(. The statistical methods are somewhat indirect and based on preliminary observations and expert knowledge and statistical calculations are weight or probability of landslide occurrence (Atkinson & Massari, 1998). Many studies have been done on statistical methods: Ahmadi et al. (1382), mass movements' hazard zonation studied by using multivariate regression analysis and analytical hierarchical process in Germichai Ardebil watershed. Analytical hierarchical process because of the many variables and normative classification compared to multivariate regression method, it preferred due to have greater accuracy and more important. Effective Factors on mentioned landslides included lithology, slope, land use, lineament elements, rainfall, aspect, and altitude. Yalcin (2008) prepared landslide susceptibility mapping of Ardsen area in Turkey, based on using geographical information system and applying analytical hierarchical process and bivariate statistical methods. For producing landslide susceptibility mapping was used from lithology, weathering, slope, aspect, vegetation, drainage density, and distance to drainage and road. It was determined in study area that lithology, weathering, land use, and slope were the most effective factors, respectively. Based on the obtained results, the analytical hierarchy process was introduced as the most appropriate model. Landslide vulnerability analysis was performed in northern Iran by Kelarestaghi and Ahmadi (2009). They applied density area method in weighting of effective parameter maps on landslides. Landslide hazard zonation map was reclassified to very low, low, moderate, high, and very high, respectively. The high and very high area approximately were 26 percent. The purpose of this research is landslide hazard zonation by using bivariate density area, certainty factors, and analytical hierarchy process models, in order to determine landslide hazard zonation and also selection of the most appropriate method zonation.

Methodology

The nine primary factors influencing landslide in the study area, including lithology, land use, slope, aspect, elevation, rainfall, distance to fault, distance to road, and drainage density were identified by interpretation of satellite imagery, aerial photography, and field studies. The used base map in this work including geological map at a scale of 1: 100,000, aerial photographs on a scale of 1: 40,000, topographical maps with a scale of 1: 50,000, ETM +satellite images and precipitation (rain-gauge stations) were prepared by ArcGIS10.2 software. The digital elevation model (DEM) with 30 meter multiplied by 30 meter pixel size was prepared by using topographic map 1:50000. The distance to drainage and road was extracted by drainage and road networks from study area topographic map. The land use map was provided by including unsupervised classification ETM+ image satellite, field survey, and accuracy control. Also geologic map was prepared by digitizing and polygonize of rock units of geologic map 1:100000 and using ArcGIS10.2.

Results and discussion

The Vanak watershed is located in the political realm of Semirom administrate. The area watershed is 168547 hectares. It is located from 51 14 50 to 51 48 15 longitude and from 31 21 05 to 31 52 10 latitude geographic coordinate system. In this work, after preparing landslides inventory map, it was overlaid with effective parameters and were extracted landslide density of effective parameter classes. Then, weight of classes were calculated by certainty factor (Wcf), analytical hierarchy process (WAHP), and area density (WAD) as showed in tables 4 and 5. After calculating of weights and overlaying them, the landslide hazard zonation were provided by using from each operating of models as illustrated in figure 3 until figure 5.

Conclusion

The results of the models assessment showed that area density method by applying quality sum index (QS) is the highest value (0.35), then certainty factor and analytical hierarchy process (AHP) are values of 0.29 and 0.11 in the next category, respectively. So the area density method has a better performance than the other methods in study area. A high value of qualify sum of area density indicates that the priorities of effective parameters were precisely. Results of effective parameters investigation show that landslide occurrence have a positive relationship with slope, altitude classes, rainfall, and drainage density and a negative relationship with distance to road and fault. Finally, after determining of factors weight, desired catchment classified into 5 classes from very low to very high by area density, certainty factor, and analytical hierarchy process models. The area of very high and high classes in area density, certainty factor, and analytical hierarchy process models are 67.89, 70.65, and 11.93 per cents, respectively.