کاربرد منطق فازی در پهنه بندی خطر زمین‌لغزش در حوضه آبخیز نازلوچای

Introduction
Landslides are one of the most important natural threats in mountainous and sloping areas that are affected by various natural and human factors.  Therefore, knowing the most important factors affecting sloping instability and landslide occurrence will help us to make developmental and infrastructural plans using appropriate methods. Therefore, using the fuzzy model, its evaluation and validation, sensitive zones of the basin should be identified and classified. The quality and quantity of existing data, the work scale, and the selection of appropriate analysis approach and modeling, guarantee the validity of landslide sensitivity maps. The process leading to such plans involves several qualitative and quantitative approaches (Ayalew & Yamagishi, 2005). Judging by experts is the criterion of qualitative approaches that do risk or sensitivity analysis (Aleotti& Chowdhury, 1999).  Qualitative methods can be divided into two types: scatter analysis or inventory and exploration. Mass movement inventory maps are the basis of most methods for landslide hazard zonation (Van Westen & Soeters, 1998). Quantitative methods (quoted by Roostaei and colleagues) are based on the numerical expression of the relationship between controlling factors and landslides (Ayalew & Yamagishi, 2005) and are divided into two approaches: Statistical and algebraic analyzes. The fuzzy logic membership functions range from 0 to 1, with zero absent Statistical and algebraic analyzes. The fuzzy logic membership functions range from 0 to 1, means that 0 is not a member of the fuzzy set and 1 means full membership (Juang et al., 1992). Therefore, in this research, based on the effective factors in the establishment of landslide, the study area is classified by using fuzzy logic.  The study area is located between 44 ° 24 'and 45 ° 53' E longitude and 37 ° 30' to 37 ° 58' N latitudes in the northwest of the country and in the political zone of the West Azerbaijan Province. The area of the catchment is 1518/8 km2.                                                                                                              
Research Methodology
In order to investigate the landslide susceptibility zonation, the provision of a landslide inventory map is the most important part of the work, which can be done by using of geographic information systems with high accuracy. The accuracy of landslide zonation is largely dependent on this stage. So, at first, the existing landslides were identified by using various tools including aerial photos, satellite imagery (Google Earth), existing information, GPS, and in particular field surveys. In the present study, ten factors affecting the occurrence of landslides were considered: elevation, slope, gradient direction, distance from the waterway, distance from the road, distance from the faults, lithology, land use, rainfall and vegetation index. Landslide susceptibility zonation was done after determining fuzzy membership values for each of the thematic layer classes using different fuzzy operator functions (fuzzy OR, fuzzy AND, fuzzy algebraic product, fuzzy algebraic sum and fuzzy Gamma operators).                                                                                                          
Research Findings
For landslide hazard zonation using fuzzy logic model, various factors including elevation, slope, gradient direction, distance from the waterway, distance from the road, distance from fault, lithology, landuse, rainfall and vegetation index were studied. Existence and density of landslides in the western slopes show the role of geological formations, the gradient direction, the distance from the waterway and precipitation in the occurrence of landslide. To evaluate the accuracy of the operators of the fuzzy logic model, the density ratio index and the quality sum index were used. Density ratio index was used to evaluate the model's ability in hazard classes' separation. The more distinction between risk classes is, the model is more capable, and the quality sum index is used to compare the performance of different models. Finally, with respect to the resulting values, the zoning with a 0.7-gamma operator was found to be desirable for the studied area.                                                                
Conclusion
According to the results of zoning (using fuzzy gamma methods), geology, gradient direction and distance from the waterways are the most important factors controlling the landslide occurrence in the studied area. Particularly geologic factors are of great importance. Most of the landslides in the study area occurred on limestone and conglomerate, which are similar to the results of the research by Mr. Tangestani in the southwest of Iran and Mr. Amir Ahmadi who worked for Iran, while these formations do not have much area in the basin. Limestone and a small amount of dolomitic limestone with an occupancy level of 15.5% of the basin, contain more than 30% of landslides. More importantly, limestone is coinciding with north orientation that confirms the role of gradient direction in occurrence of landslides. The impact of the human factor mainly depends on changing environmental conditions, such as road construction, inappropriate plowing, excessive grazing and water diversion for agricultural use. Therefore, by studying the researches in Iran and in different parts of the world, the slipping factors in different basins and regions are not the same and in fact, different slip conditions are present in different regions.