A new approach to use AHP in landslide susceptibility mapping: a case study at Yenice (Karabuk, NW Turkey)


NATURAL HAZARDS, vol.63, no.2, pp.1157-1179, 2012 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 63 Issue: 2
  • Publication Date: 2012
  • Doi Number: 10.1007/s11069-012-0218-1
  • Journal Name: NATURAL HAZARDS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.1157-1179
  • Hacettepe University Affiliated: Yes


This study aimed to investigate the parameter effects in preparing landslide susceptibility maps with a data-driven approach and to adapt this approach to analytical hierarchy process (AHP). For this purpose, at the first stage, landslide inventory of an area located in the Western Black Sea region of Turkey covering approximately 567 km(2) was prepared, and a total of 101 landslides were mapped. In order to assess the landslide susceptibility, a total of 13 parameters were considered as the input parameters: slope, aspect, plan curvature, topographical elevation, vegetation cover index, land use, distance to drainage, distance to roads, distance to structural elements, distance to ridges, stream power index, sediment transport capacity index, and wetness index. AHP was selected as the major assessment methodology since the adapted approach and AHP work in data pairs. Adapted to AHP, a similarity relation-based approach, namely landslide relation indicator (LRI) for parameter selection method, was also proposed. AHP and parametric effect analyses were performed by the proposed approach, and seven landslide susceptibility maps were produced. Among these maps, the best performance was gathered from the landslide susceptibility map produced by 9 parameter combinations using area under curve (AUC) approach. For this map, the AUC value was calculated as 0.797, while the others ranged between 0.686 and 0.771. According to this map, 38.3 % of the study area was classified as having very low, 8.5 % as low, 15.0 % as moderate, 20.3 % as high, and 17.9 % as very high landslide susceptibility, respectively. Based on the overall assessments, the proposed approach in this study was concluded as objective and applicable and yielded reasonable results.