Landslide potential zoning in Qatourchai watershed of Khoy city using ANP method

Abstract
Introduction
Landslides are among the geomorphic processes affecting the landscape evolution of mountainous regions, which occur under the influence of various local geomorphological, hydrological and geological conditions (Roering et al, 2005). In a simple definition, the mass movement of earth constituents from a slope (slope) to the bottom of the earth is called landslide. This downward movement of surface materials occurs due to the effect of the earth's gravity and the mobility and displacement of these materials increases with the presence of water in the sediments (Pareta et al, 2012). Landslides cause heavy casualties and financial and life losses every year, which in some cases cannot be compensated or require a lot of time and money. Therefore, planning to prevent these damages is of particular importance. In this regard, landslide risk zoning maps make it possible for relevant executive bodies to identify sensitive and vulnerable areas to landslides and make decisions about the desired programs ).
Jafari et al. (2018) analyzed landslide risk zoning in Badranlu catchment area using Analytical Network Process (ANP) method. In this study, 69 landslides were identified in the first step for landslide risk zoning in the study area. In the next step, the factors involved in the occurrence of landslides, including lithology, land use, land slope, slope direction, average annual rainfall, distance from the river, distance from the fault, and distance from residential centers are evaluated spatially and as thematic layers in the environment. GIS were prepared. These factors were prioritized using expert opinion and ANP network analysis process and were combined accordingly. Finally, the study area was classified into three classes in terms of sensitivity to the risk of landslides. The results show that 21 landslides with an area of 75,600 square meters are in the medium-risk landslide class, 26 landslides with an area of 91,800 square meters are in the high-risk landslide class, and 22 landslides with an area of 76,500 square meters are in the very high-risk landslide class. According to this research, the most important factors affecting the occurrence of landslides in the Badranlo basin are slope, lithology, distance from the river and annual rainfall. Ayaliyo and Yamagishi (2005) used the logistic regression approach in the framework of geographic information system in order to assess and zonate the risk of landslide occurrence in Kakodayahiko mountains located in the center of Japan. In this research, the spatial distribution of landslides was considered as a dependent variable and a set of environmental and human variables such as slope, vegetation, communication networks, etc. were considered as independent variables. Due to the spatial nature of landslide hazard, statistical analysis was done in the framework of Geographical Information System (GIS). The results indicate the high efficiency of the multivariate statistical approach (logistic regression) for modeling landslide risk at the level of the study area. Also, the results show the great importance of variables of distance from roads and slope in the occurrence of landslides in the region. This shows the vulnerability of the domains in the study area to human interference.
Qaturchai Khoi watershed is one of the basins in the northwest of the country, where the instability of the range is considered one of the important processes in the evolution of the region's landscape. Such instabilities in some cases threaten the existing settlements and infrastructures in the region. In the current research, the risk of landslides in this basin is zoned. In order to zone the risk of landslides in the Qoturchai catchment area, 10 variables influencing the occurrence of landslides were used.These variables include: height, slope, slope direction, slope length, lithology, faults, precipitation, distance from waterways, vegetation and distance from the road were used, and their effects on the occurrence of landslides have been briefly discussed.


Methodology
In this research, in order to zone the risk of landslides in the Qoturchai catchment area, the technique of superimposing different variables affecting the occurrence of landslides was used. For this, it is necessary to combine the geographic information system (GIS) with other models. In this context, multi-criteria decision-making models are widely used. One of these efficient and common models is the Analytical Network Process (ANP) model. In the current research, the mentioned model was used to extract the weight or the influence factor of the effective criteria on the occurrence of landslides in the Qotourchai catchment area due to its high flexibility and the ability to establish internal and external communication. In the following, explanations about the ANP model and its implementation method are provided.
In order to assess and zonate the risk of landslides in the Qoturchai catchment area, ArcGIS and Super Decisions software were used as the most important processing and analytical tools of the research.ArcGIS software was used to prepare thematic layers affecting landslide occurrence and spatial analysis.Super Decisions software was also used to perform calculations related to the Analytical Network Process (ANP) model.

Conclusion
In the present study, the Qoturchai catchment area was evaluated in terms of the risk of landslides. The study area is located in the southwest of Khoy city, affiliated to West Azarbaijan province. This basin is one of the basins of the country that has a high potential for landslides. Therefore, in the present study, the spatial assessment and zoning of the landslide risk of this basin were investigated. In order to zone the risk of landslides in Qatourchai catchment area, 10 variables influencing the occurrence of landslides were used. These variables are: height, slope, slope direction, slope length, lithology, faults, precipitation, distance from waterways, vegetation and distance from the road were usesd.The comparison of the effective criteria on the occurrence of landslides using the Analytical Network Process (ANP) model indicates that the three variables of slope, lithology and precipitation are respectively the most important and effective variables affecting the occurrence of landslides in the Qotourchai catchment area. These variables have a total effect of more than 60% on the occurrence of landslides in the study area. In fact, a significant part of the landslides in the region is the result of the separate or combined involvement of the aforementioned variables. On the opposite point, the variables of distance from the road, slope length, slope direction and vegetation have the least importance or influence in the occurrence of landslides in the region and their role is more local.e
Keywords: Landslide Qoturchai Basin; ANP model; Zoning