Landslide mapping in the coastal area between the Strymonic Gulf and Kavala (Macedonia, Greece) with the aid of remote sensing and geographical information systems

Abstract

The main aim of this study is to photointerpret land cover change along hill slopes in order to detect existing landslides with the aid of methodologies such as false colour composites (FCCs), principal component analysis (PCA) and the normalized difference vegetation index (NDVI). Then, by combining geological data (lithology, vegetation cover), and geomorphologic factors (slope, aspect, distance from the rivers), landslide susceptibility maps were produced. The region where the study took place was the coastal area between the Strymonic Gulf and southwest Kavala prefecture, which was selected because of its intense landslide activity. The identification of locations where landslides occurred was achieved with the use and processing of TERRA/ASTER satellite images, while the data, which were mainly collected from the digitization of contours from topographical maps at 1 : 50 000 scale, were used to construct the final landslide susceptibility maps. The resulting FCC images provided satisfactory information about the locations of landslide sites and the landslide susceptibility maps indicated areas that were more prone to produce such phenomena.