The characteristics of the landslides triggered by the Wenchuan M s 8.0 earthquake from Anxian to Beichuan

Abstract

This paper presents a quantitative analysis of the number and area of the landslides triggered by the Wenchuan M s 8.0 earthquake from Anxian to Beichuan. In this study the controlling factors of landslides are obtained from geological maps, relief amplitude, slope, and topographic roughness calculated from DEM data in ArcGis 9.2. A total of 4134 landslides are identified from digital aero-image of 1 m resolution, covering 19.7 km 2 that accounts for 6.4% of the study area. The number of landslides with areas less than 5000 m 2 is up to 85.22%. The smallest landslide is 6 m 2 and the largest 1515,000 m 2. The number of landslides which have occurred is 1.6 times more prevalent at hanging walls as opposed to foot walls. Also the relationship between the numbers of landslides and the distances to an earthquake rupture at a hanging wall is linear but exponential at a footwall. These results prove that landslide activity is more severe at hanging walls than at foot walls. The outcrop area of geologic unit is proportional to the number of landslides except the Jurassic, Triassic and Quaternary units ( R 2 = 0.952) in the study area, and landslide density is 19. Meanwhile the area of a landslide is linearly proportional to outcrop area of geologic unit except Jurassic, Lower Cambrian and Quaternary unit ( R 2 = 0.8279) in the study area and appear to indicate that the lithology does not play an important role in landslides in the study area. Earthquake-triggered landslides tend to be concentrated in specific zones associated with the event-controlling parameters such as reliefs, slopes and topographic roughness. The landslides occurred particularly in the range of 600–1293 m relief amplitudes, and the number in this range amounts to 90% of total slides in the studied region. Moreover, landslides easily occurred in the curvature 1–15 with slope between 5° and 55°. In addition, landslides also usually occurred in the range of 1–1.5 topographic roughness. Therefore these correlations provide a way to assess regional landslide hazards and are of great significance to regional geohazard planning and prediction.