Journal of Spatial Analysis

Evaluating and comparing the performance of Frequency ratio coefficient models and network analysis in Rock fall zoning
(A case study of Zanjan-Taham-Tarom Road)

Extended Abstract
One of the natural hazards of the collapse of rocks from the foothills of the mountains, causing great financial losses and loss of life. Especially when it comes to the path of communication. The rock fall is a rapid movement of a mass without cohesion in the powder or a mixture of soil and rock, so that the initial construction is not discernible, the level that occurs along that rupture it is often unclear. The falling stones of a mountain depend on several factors, which have the natural origin or origin of human origin. Natural factors influencing the fall can be rock factors, slopes and altitudes, geological structure, fault and slope of the geological layers, rainfall and temperature changes distance from the river, etc. human factors can also be referred to as road, land use and mining, destruction of vegetation, etc. in Iran, the collapse of rock parts on mountain roads causes massive loss of life and financial damage. Therefore, it is necessary to identify and classify the roads in terms of risk of suitable methods. In the north west of Iran it has mountainous topography and due to the state of tectonic and its seismic and climatic conditions, suitable conditions for landslide are provided in some domains. So, due to the fact that the area studied in the mountainous north - west region and the possession of all the crumbling conditions are very prone to collapse.
The research method is applicable in terms of practical purpose and the process of doing work on a combination of library and field methods. In this study, it has been used to determine the prone areas of collapse and zoning of anp models and frequency ratio. Two models that differ in terms of process and mechanism. In order to organize the research framework, first, a field study of the study area has been studied and the mathematical position of falling points is recorded with gps. Then, in order to model the mentioned models, the layers of GIs for the shape of the Georeferenced and digital were prepared. to provide the zoning layers of geology , slope and Aspect , elevation levels , land use and vegetation , fault , and land cover maps , annual temperature and precipitation , distance from the road , distance from the stream were used . The 20 m x 20 m contour line were originally prepared using the topography map of 1: 25,000 in the ArcGIS environment. Then, the contour line and Dem of the area were constructed. The slope and Aspect maps, elevation levels, Isothermal and isohyet map frost and stream network were created via Dem and meteorological data. Geological map and fault map were created using digital map 1: 100,000 Zanjan and map and vegetation map and road distance using Landsat 8 - 2017 OLI and ArcGIS images. To produce linear layers, the Distance function was used.
Using the statistical method, the frequency ratio and the network analysis method are using the landslide hazard zonation using the statistical method, the frequency ratio and analysis of network analysis to zoning the risk of falling by combination and sum of maps in class were low-risk to very high. From the tangible results of this study, the relationship between slope maps, elevation levels, rock material, Isothermal and isohyet is done. So that each side of the road had operated on the five factors that had happened. With regard to the output of the maps, the risk zones were high to very high for ANP models 14/17, 35/27 and FR 02/6, 35/14 percent. Ranges from high to very high with slopes between 40 and 80 percent and Sedimentary formations such as sandstone, siltstone with tuff layers, elevation levels 1,500 – 2300, Southern and Eastern slopes, Distance between 0 and 500 faults, Road and stream have adaptations . The changes in the percentage of area in both models show despite the difference in the size of the risk zones, Follow a similar process. To assess the zoning accuracy of these methods, two sets of quality and accuracy index (experimental probability) were used. The evaluation of the models showed that in the network analysis model, the indexes were 0.76 and 0.88, respectively, that the relation of frequency ratio coefficient model had optimal quality and accuracy.
In this research, various factors influencing the occurrence of rock falls on Zanjan-Taham-Tarom road were investigated. From there, mass movements such as rock fall on the roads act as a system, as a result, all factors play a role in the occurrence of such phenomena. But some elements have a more vibrant role. In the studied area, among the factors affecting lithology, slop, elevation levels, precipitation, temperature changes, number of freezing days and distance from the road and land use are more than other factors in the occurrence of rock fall. Assessing the quality and accuracy of zoning maps while confirming zoning accuracy showed that the network analysis method has better performance. The risk of collapse on Zanjan – taham- tarom road is always exists. Therefore, we need to use sustainable methods to reduce the risks. Domain stabilization methods are generally done in the form of mechanical, biological and bio-mechanical which, according to the long course of the road and the duration of the road, Mechanical methods such as unloading , embankment , drainage , use of separation walls as well as the use of  net Grid are suggested .
 
Keywords: Frequency ratio, Network analysis, Zanjan Road - Taham - Tarom
 
.