Influence of seismic acceleration on landslide susceptibility maps: a case study from NE Turkey (the Kelkit Valley)


Das H. O., SÖNMEZ H., GÖKÇEOĞLU C., Nefeslioglu H. A.

LANDSLIDES, vol.10, no.4, pp.433-454, 2013 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 10 Issue: 4
  • Publication Date: 2013
  • Doi Number: 10.1007/s10346-012-0342-8
  • Journal Name: LANDSLIDES
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.433-454
  • Hacettepe University Affiliated: Yes

Abstract

Particularly in the last decade, landslide susceptibility and hazard maps have been used for urban planning and site selection of infrastructures. Most of the procedures for preparing of landslide susceptibility maps need high-quality landslide inventory map. Although the rainfall and seismic activities are accepted as triggering factor for landslides, designation of the triggering factor for each landslide in the inventory is almost impossible when well-documented records are unavailable. Therefore, during preparation of landslide susceptibility map, whole landslide records in the inventory map are used together without classifying based on the triggering factors. Although seismic activity is accepted as a triggering factor, possible effect of the use of seismic activity on production of landslide susceptibility map was investigated in this study, and the subject is open to discussion. For this purpose, a series of stability analyses based on circular failure and infinite slope model were performed considering different pseudostatic conditions. The results of analyses show that gentle slopes have higher susceptibility to failure than steeper ones, even if their stability conditions (susceptibilities) are similar for static condition. The seismic forces acting on failure surfaces may not be sufficiently taken into consideration in the conventionally prepared landslide susceptibility maps. Employing the general decreasing trend in stability condition based on slope face angle and the seismic acceleration, a new procedure was introduced for preparing of the landslide susceptibility map for a scenario earthquake. The prediction performance of occurring landslides increased after the procedure was applied to the conventionally prepared landslide susceptibility map. According to the threshold independent spatial performance analyses of the proposed methodology and the produced landslide susceptibility maps, the area under ROC curve values were calculated as 0.801, 0.933, and 0.947 for the maps prepared by considering conventional method and scenario earthquakes having M (w) values of 5.5 and 7.5, respectively.