Application of fuzzy logic in the preparation of hazard maps of landslides triggered by the twin Ahar-Varzeghan earthquakes (2012)

Abstract

The twin Ahar-Varzeghan earthquakes (Mw = 6.4 and Mw 6.2, August 12, 2012) are among the most severe and destructive seismic events to have occurred in the northwest part of Iran in the last century. The main shock of this event was felt within a 300 km radius in most of the northwest provinces of Iran, including West Azerbaijan, Ardebil, Gilan, Zanjan, Alborz, Hamedan, and Kurdistan, as well as in neighboring countries such as Republic of Azerbaijan, Armenia, and Turkey. The quake caused some large landslides, which led to remarkable economic losses in the region. Landslide susceptibility mapping is one of the useful tools that can be applied in disaster management and planning development activities in mountainous areas. In this research, a geographic information system (GIS)-based multi-criteria decision analysis model (fuzzy logic) is used to evaluate landslide susceptibility within the area stricken by the twin Ahar-Varzeghan earthquakes (2012). Thus, a rigorous field-based investigation was conducted during several days of fieldwork to prepare a database of landslides triggered in the earthquake-stricken area. The extended fieldwork was carried out to scrutinize the basic map of the slope instabilities plotted immediately after the earthquake. During the fieldwork, 47 fall and topple zones, nine soil slides, 13 rock slides, two areas of lateral spreading, and one rapid soil flow were detected. The ground strength class, slope angle, normalized difference moisture index (NDMI), normalized difference vegetation index (NDVI), distance from the rivers and roads, and shake intensity were selected as the input layers for fuzzy logic analysis in a GIS environment. Next, the performance of various fuzzy operators in landslide susceptibility mapping was empirically compared by applying fuzzy operators [intersection (AND), union (OR), algebraic sum (SUM), multiplication (PRODUCT)] and different fuzzy gamma values of fuzzy overlay. The results showed that the majority of the landslides fall in the “high” and “very high” susceptibility classes. We found that there is a satisfactory consistency between the landslide susceptibility map prepared using the fuzzy union (OR) operator and the landslide distribution map.