Evaluation of Landslide Susceptibility Zonation applying Fuzzy Gamma Operators in Taleghanroud Watershed of Qazvin Province

Landslide is one of the most destructive types of erosion on slopes, which causes a lot of financial and human losses. Since it is difficult to predict the occurrence of landslides, it is very important to identify landslide-sensitive areas and the zoning of these areas based on the potential risk of this phenomenon. Evaluation of landslide susceptibility is one of the basic tools for managing and reducing potential damages. The present study has attempted to assess the efficiency of various fuzzy gamma operators for landslide susceptibility zonation in Taleghanroud watershed of Qazvin province. Therefore, the landslide distribution map and also 11 effective factor were first prepared which include layers including slope, slope direction, altitude, land use, lithology, distance to road, distance to stream, distance to fault, earthquake acceleration, precipitation, and maximum daily precipitation. A total of 15 landslides were identified, 70% of which were used to model and 30% of which were used to evaluate the results of the models. Then, after determining the values of Frequency Ratio and fuzzy membership for different classes of effective factors, landslide susceptibility maps were produced using fuzzy gamma operators (for gamma values equal to 0, 0.1, 0.2, 0.3, 0.4, 0.5 , 0.6, 0.7, 0.8, 0.9 and 1). The evaluation process using Density Ratio and Sum of Quality indices showed that the gamma of 0.7 has higher accuracy than other gamma values in the study area. The landslide hazard zoning map of the superior model will be useful in land use planning and reducing the landslide risk of the region. Keywords: Fuzzy membership values, Frequency Ratio, Hazard zoning, Density Ratio, Sum of Quality. Extended Abstract

Landslides are one of the most destructive types of erosion on slopes, which causes sediment, muddy floods, filling dam reservoirs, and also lots of damage to engineering structures, residential areas, and agricultural lands. Due to landslide damage, it is necessary to prepare a landslide susceptibility zoning map using appropriate methods, especially in areas that are prone to landslides. These types of maps are among the basic and essential tools for managing and reducing possible damages of this phenomenon. The method of gamma fuzzy operators is one of the relatively conventional and new methods for landslide susceptibility zoning, which, due to the use of fuzzy logic, has no limitations of algebraic addition or multiplication of layers. The present study aims to evaluate the efficiency of various fuzzy gamma operators for landslide susceptibility zonation in Taleghanroud watershed of Qazvin province.

In this study, the landslide distribution map and also 11 effective factor were first prepared which include layers including layers including slope, slope direction, altitude, land use, lithology, distance to road, distance to stream, distance to fault, earthquake acceleration, precipitation, and maximum daily precipitation. A total of 15 landslides were identified, 70% of which were used to model and 30% of which were used to evaluate the results of the models. All factor layers were crossed with the landslide distribution map to determine the importance of each class of the factor layers. The area of the factor classes and also the area that covered by the landslide in each class were determined to calculate the importance of each factor class via frequency ratio relationship. Then, landslide susceptibility maps were produced using fuzzy gamma operators (for gamma values equal to 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1). Density Ratio (Dr) and Sum of Quality (Qs) indices were applied to evaluate the validity of the used models.

A total of 15 landslides with the minimum, maximum and total areas of respectively 3027, of 534779 1669377 m2 were recorded in this watershed. The results of the frequency ratios and fuzzy memberships of factors showed that the classes including slope 35°- 45°, northeast slope direction, altitude 1050-1400, precipitation 280-380, maximum daily precipitation 51 -55 mm, 0.246-2242 earthquake acceleration, 1000-2000 m distance from fault, 100-200 m distance from waterway, distance of more than 400 m from road, rainfed agriculture, and class 4 lithology units (high sensitivity) have had the highest values. Therefore, the have the most important role in the occurrence of landslides of the study area. The different landslide zonation maps showed that the percentage of area under more susceptible classes has increased steadily by increasing gamma values from zero (fuzzy multiplication) to one (fuzzy sum). Therefore fuzzy multiplication operator has resulted in most of the surface area with very low landslide hazard, and the fuzzy sum operator has resulted in most of the surface area with very high landslide hazard. This is due to the decreasing nature of the fuzzy multiplication operator and the increasing nature of the fuzzy sum operator. Evaluation of hazard zonation maps using Dr and Qs indices showed that the Qs index values for different fuzzy integration models range from 0 (fuzzy sum) to 93.3 (Gamma = 0.7) which indicates that the fuzzy combination method with gamma equal to 0.7 has provided the best zoning map. In addition, the values of the Dr Index in fuzzy integration model with gamma 0.7, have an increasing trend for hazard classes from one (very low) to 5 (very high) which indicates that the zonation map of the superior model is classified correctly.

The Fuzzy Gamma operators are among the conventional and relatively new methods for landslide susceptibility zoning. These methods have no limitations of algebraic addition operators or multiplication of layers due to the use of fuzzy logic. The landslide susceptibility map obtained from this study provides proper information for designers, managers, policymakers, and engineers who can develop various measures to reduce landslide risk in the region. However, the conditions and degree of instability of areas under high and very high hazard classes should be studied more accurately by experts before development plans of the region. Landslide is one of the most destructive types of erosion on slopes, which causes a lot of financial and human losses. Since it is difficult to predict the occurrence of landslides, it is very important to identify landslide-sensitive areas and the zoning of these areas based on the potential risk of this phenomenon. Evaluation of landslide susceptibility is one of the basic tools for managing and reducing potential damages. The present study has attempted to assess the efficiency of various fuzzy gamma operators for landslide susceptibility zonation in Taleghanroud watershed of Qazvin province. Therefore, the landslide distribution map and also 11 effective factor were first prepared which include layers including slope, slope direction, altitude, land use, lithology, distance to road, distance to stream, distance to fault, earthquake acceleration, precipitation, and maximum daily precipitation. A total of 15 landslides were identified, 70% of which were used to model and 30% of which were used to evaluate the results of the models. Then, after determining the values of Frequency Ratio and fuzzy membership for different classes of effective factors, landslide susceptibility maps were produced using fuzzy gamma operators (for gamma values equal to 0, 0.1, 0.2, 0.3, 0.4, 0.5 , 0.6, 0.7, 0.8, 0.9 and 1). The evaluation process using Density Ratio and Sum of Quality indices showed that the gamma of 0.7 has higher accuracy than other gamma values in the study area. The landslide hazard zoning map of the superior model will be useful in land use planning and reducing the landslide risk of the region. Keywords: Fuzzy membership values, Frequency Ratio, Hazard zoning, Density Ratio, Sum of Quality. Extended Abstract

Landslides are one of the most destructive types of erosion on slopes, which causes sediment, muddy floods, filling dam reservoirs, and also lots of damage to engineering structures, residential areas, and agricultural lands. Due to landslide damage, it is necessary to prepare a landslide susceptibility zoning map using appropriate methods, especially in areas that are prone to landslides. These types of maps are among the basic and essential tools for managing and reducing possible damages of this phenomenon. The method of gamma fuzzy operators is one of the relatively conventional and new methods for landslide susceptibility zoning, which, due to the use of fuzzy logic, has no limitations of algebraic addition or multiplication of layers. The present study aims to evaluate the efficiency of various fuzzy gamma operators for landslide susceptibility zonation in Taleghanroud watershed of Qazvin province.

In this study, the landslide distribution map and also 11 effective factor were first prepared which include layers including layers including slope, slope direction, altitude, land use, lithology, distance to road, distance to stream, distance to fault, earthquake acceleration, precipitation, and maximum daily precipitation. A total of 15 landslides were identified, 70% of which were used to model and 30% of which were used to evaluate the results of the models. All factor layers were crossed with the landslide distribution map to determine the importance of each class of the factor layers. The area of the factor classes and also the area that covered by the landslide in each class were determined to calculate the importance of each factor class via frequency ratio relationship. Then, landslide susceptibility maps were produced using fuzzy gamma operators (for gamma values equal to 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1). Density Ratio (Dr) and Sum of Quality (Qs) indices were applied to evaluate the validity of the used models.

A total of 15 landslides with the minimum, maximum and total areas of respectively 3027, of 534779 1669377 m2 were recorded in this watershed. The results of the frequency ratios and fuzzy memberships of factors showed that the classes including slope 35°- 45°, northeast slope direction, altitude 1050-1400, precipitation 280-380, maximum daily precipitation 51 -55 mm, 0.246-2242 earthquake acceleration, 1000-2000 m distance from fault, 100-200 m distance from waterway, distance of more than 400 m from road, rainfed agriculture, and class 4 lithology units (high sensitivity) have had the highest values. Therefore, the have the most important role in the occurrence of landslides of the study area. The different landslide zonation maps showed that the percentage of area under more susceptible classes has increased steadily by increasing gamma values from zero (fuzzy multiplication) to one (fuzzy sum). Therefore fuzzy multiplication operator has resulted in most of the surface area with very low landslide hazard, and the fuzzy sum operator has resulted in most of the surface area with very high landslide hazard. This is due to the decreasing nature of the fuzzy multiplication operator and the increasing nature of the fuzzy sum operator. Evaluation of hazard zonation maps using Dr and Qs indices showed that the Qs index values for different fuzzy integration models range from 0 (fuzzy sum) to 93.3 (Gamma = 0.7) which indicates that the fuzzy combination method with gamma equal to 0.7 has provided the best zoning map. In addition, the values of the Dr Index in fuzzy integration model with gamma 0.7, have an increasing trend for hazard classes from one (very low) to 5 (very high) which indicates that the zonation map of the superior model is classified correctly.

The Fuzzy Gamma operators are among the conventional and relatively new methods for landslide susceptibility zoning. These methods have no limitations of algebraic addition operators or multiplication of layers due to the use of fuzzy logic. The landslide susceptibility map obtained from this study provides proper information for designers, managers, policymakers, and engineers who can develop various measures to reduce landslide risk in the region. However, the conditions and degree of instability of areas under high and very high hazard classes should be studied more accurately by experts before development plans of the region.