Author(s)

A. Grozavu, M. C. Mărgărint & C. V. Patriche

Abstract

In Romania, landslides are a frequent reality in the geomorphologic landscape, particularly in the plateau units, with favourable conditions for the occurrence of such phenomena. This paper focuses on the applicability of landslide susceptibility assessment, by using a logistic regression approach, to a sector of Iaşi Cuesta in the Moldavian Plateau, where landslides happen to a significant extent. Ortophotoimages, topographical and geological maps were used as input data. The landslide inventory, multicriteria qualitative and quantitative analysis and GIS application have facilitated data processing and spatial visualization results. A database of landslide and landslide causative factors was constructed, consisting of several environmental data layers: land use as a qualitative variable and geomorphometrical parameters as quantitative variables. The latter were obtained from a DEM, at a resolution of 20x20 m, derived from 1:25,000 topographic maps. Application of this model at the scale of one small administrative unit (commune) shows that landslide occurrence is best explained by terrain slope, land use and altitude. The landslide susceptibility map reveals that terrains displaying high and very high susceptibility represent about 29% of the study area. The logistic regression model seems to be an adequate tool for the assessment of landslide susceptibility in an objective and quantitative manner, providing also parameters allowing the evaluation of the output quality. Starting from primary data sources at fine scales (1:5000), our model may be extended to other similar regions in the Moldavian Plateau. Keywords: landslide susceptibility assessment, GIS application, logistic regression, geomorphometry, Moldavian Plateau, Romania.

Keywords

landslide susceptibility assessment, GIS application, logisticregression, geomorphometry, Moldavian Plateau, Romania