Investigation of Effective Factors on Landslide Occurrence and Susceptibility Zonation Using the Dempster-Shafer Model in the Middle Mazlaghan Chai, Markazi Province

Introduction and Goal
Middle Mazlaghan Chai watershed located in west of Saveh county. The activity of faults and the presence of sensitive rocks have created suitable conditions for the occurrence of small and large landslides. One of the important strategies for reducing losses caused by landslides is avoiding high-risk and very high-risk areas. For this purpose, it is necessary to prepare a relatively accurate landslide susceptibility zoning map from among the existing methods. One of the best methods for landslide zonation is the Dempster-Shafer model. The purpose of this research is determination of the factors affecting the occurrence of landslide, presentation and evaluation of the landslide susceptibility zonation map using the Dempster-Shaffer method and area under curve (AUC), respectively in Middle Mazalaghan Chai.
Materials and Methods
The Middle Mazlaghan Chai watershed, with an area of ​​21,746 ha, is located in Markazi Province. Maximum elevation is 2833 m above sea level in the northern heights of the watershed and minimum elevation is 1399 m at the outlet of the Bivaran River. The average annual temperature of the study area is 13 °C and the average annual precipitation varies from 246 mm in the south to 500 mm in the north. The climate of the region is arid and semi-arid. 52% of the soils in the study area are in the type of entisol and without profile development, and 37% of the vegetation cover consists of moderate and poor rangeland. Initially, a landslide inventory map was prepared in environment of geographic information system using fieldworks, aerial photography and satellite imagery. Then using fieldworks and related research, the most important factors affecting landslides in the study area, including slope, aspect, elevation, distance to fault, distance to road, distance from stream, land-use, lithology and precipitation were investigated and determined. After preparing the information layers and weighting in the GIS, a landslide susceptibility zonation map was prepared and classified using Dempster Shafer method. Finally, the efficiency of the Dempster Shafer method was evaluated using the area under curve (AUC).
Results and Discussion
Fossiliferous limestone units and tuff and lava alternations showed the highest susceptibility to landslides. The results of this research showed the most landslides occurred more than 100 m from streams and at distances greater than 200 m from the roads and faults. Slopes more than 40% are most susceptible to landslides. The northern aspect, elevation more than 2600 m, and precipitation more than 450 mm also showed the highest susceptibility to landslides in the studied area. The zoning results showed that about 22% of the study area is located in high and very high susceptibility zones and approximately 75% of landslides occurred in high and very high susceptibility zones. The area under the curve of the landslide susceptibility zonation map was also obtained 0.849.
Conclusion and Suggestions
Accurate identification of landslide locations using GPS device is one of the results obtained from this research. As a result, the landslide distribution map of 192 landslides was prepared and it was checked and recorded for the first time in the study area. Among the lithological units, fossiliferous limestone units and alternations of tuff and lava showed the highest susceptibility to landslide, and the landslide susceptibility zones are in complete relation to the geological structures of the region. It is suggested that this model be compared with other models and the results of this research will be used as basic information for environmental management and planning.