Mehdi Bashiri; Seyedeh-Maedeh Kavousi Davoudi; Ali Afzali
Volume 5, Issue 14 , June 2018, , Pages 157-178
Abstract
Introduction
The mass movement of materials on steep slopes under the influence of gravity and motivation factors such as earthquakes, floods and torrential rains called landslide. Landslide similar the other natural phenomena is an important natural disaster that occur every year in the mountainous ...
Read More
Introduction
The mass movement of materials on steep slopes under the influence of gravity and motivation factors such as earthquakes, floods and torrential rains called landslide. Landslide similar the other natural phenomena is an important natural disaster that occur every year in the mountainous and highland areas of our country. By risk assessment of landslide occurrence, the sensitive areas with a high potential risk can be identified. Also the effect of different environmental factors on the pattern of high-risk areas can be used in risk management, practically.
Methodology
One of the basic measures for achieving methods for mass movements control and management is identifying the effective factors in occurrence of this phenomenon. Also, in countries involved in the landslide, there is an increasing tendency for risk and damage assessment of this phenomenon. Therefore, at the present study, using fractal geometry, the effects of land use, geology and geomorphology on the landslide patterns were evaluated in the Tooye-Darvar watershed, Semnan province. Because the fractal theory studies and recent reviews in the earth science indicate that some geological processes such as mineralization, sedimentation, deposition, volcano, morphology and etc have self-similar characteristics. First, using aerial photo interpretation, field surveying and recording the position of landslides using GPS, the landslide distribution map prepared. Then, information layers for each slope, aspect, elevation, geology, geomorphic units, soil erosion class, distance from fault, distance from road, distance from stream and land-use factors prepared.in the ArcGIS software environment using digital elevation model (DEM) of the region with a pixel accuracy of 20 meters. Also the landslides information layer prepared according to the field studies and rasterized. Then, different layers overlapped and table of combined properties for merged layer includes the information of each pixel, extracted and entered into the Excel environment. In Excel, the relative importance (or frequency ratio) calculated for each different class of information layers.
Statistical analysis
After collecting and recording data and creating the database, the SPSS v.23 used for data analysis. In the first stage, the normalization checked using Kolmogorov-Smirnov test at the 5% confidence level. Then, the effect and significance of each measured variable investigated in landslide occurrence.
Calculating the fractal dimension
The landslide areas available in each unit extracted as a polygon and the resulting image transmitted to the Fractalyse software and its fractal dimension calculated using box-counting method. Then, the fractal dimension of landslides placed in work units transferred to the SPSS environment and statistical comparison performed with the aim of investigating the geometric or morphologic differences of sliding zones in different land uses and geologic and geomorphic units. Then to compare the different landslide hazard classes, the density or compression ratio of landslide used in each hazard class. The density ratio calculated by dividing the landslide density in a particular hazard class to the average density of landslides based on the area density or the number of landslides.
landslide susceptibility zonation
In order to zonate the landslide susceptibilityusing bivariate statistical methods, the information value and the area density, each of the factors affecting landslide occurrence include slope and elevation maps, slope aspect, soil erosion class, geomorphic type, geological unit, land use, distance from the stream, distance from the road and distance from fault in GIS environment digitized and classified. Then, based on the two above mentioned methods, the weight of each factor and its related classes determined. The weighted maps of effective factors combined and using natural breaks method, the obtained maps classified in very low, low, moderate, high and very high hazard classes. In order to evaluate the implemented model in the region , 2/3 of landslide points and 2/3 of landslide areas used for modelling and the remaining 1/3 of each one used to evaluate the model.
Discussion
The results of fractal dimensions study in 146 landslide areas using box counting method showed an average of 1.987. Study of the spatial features in landslide areas include landuse, slope aspect, soil erosion class, geological unit, geomorphic type, height and slope class, distance from road, distance from fault and distance from stream showed that only the effect of geomorphic types on geometric dimension of landslide areas is significant and this significance is caused by high difference (sig=0/000)between mountains and plateaus and upperterraces types. Finally, the density ratio for landslide areas and points in each class of spatial characteristics for landslide occurrence, calculated and the effects of these variables on landslide occurrence severity, presented and analyzed. Also, in the landslide susceptibility zonation about 1/2 of landslides located in the high and very high risk classes that indicates the high potential for landslide in this region.
Conclusion
The results showed that the surface erosion has no significant effect on occurrence of large landslides but it has affected the landslide points. Also, the presence of marl and lime in study area that is a geological unit susceptible to dissolution can be effective in landslide occurrence. Geomorphologically, the mountain and hill types have been effective, which could be due to the high slope of these types. In the case of landslides occurrence in the vicinity of roads, faults and streams, it can be concluded that the small landslides has been affected by road, but it has no significant effect on the occurrence of large landslides. But the existence of fault in the area and proximity to the fault led to the occurrence of extensive landslides. Finally, the presence of stream in the area has also been effective in occurrence of large landslides, but the landslide points, has not been affected by existence of stream in their buffers. Also, high potential for landslide in the study area represents the being endangered for regional installations, agricultural lands, engineering structures and buildings.
