Landslide Susceptibility Mapping by Using Geospatial Technique: Reference from Hofu City, Yamaguchi Prefecture, Japan

Abstract

Hofu City in the southern part of Honshu Island, Japan, was hit by a severe landslide on July 21, 2009. This chapter aims to determine the sites where the landslides are most likely to happen by creating a susceptibility map within the case study sites. We used remote sensing and geographic information systems (GIS) datasets that include ALOS AVNIR-2 satellite imagery, the digital elevation models (DEM), geology records, the local rain gauge data, and geographical representation of the history of landslides. Seven parameters, including land cover, elevation, slope, aspect, geology, and boundary extraction, were integrated using logistic regression with an isohyet map (i.e., from the rainfall dataset) to model a landslide susceptibility map (LSM). Following the research's findings, landslides were more likely to occur at elevations between 50 and 350 m, slope angles between 5 and 50 degrees, slope directions northeast and north, all of the land cover types and lithological types of granodiorite, fan deposits, and middle terrace. Among those statics' parameters, elevation, land cover, and slope were the most significant in determining the landslide susceptibility model. Further, almost half of Hofu City was categorized as high and very high susceptible areas. Moreover, out of the 928 inventory landslide dataset, 916 were placed at the very high susceptibility areas of the LSM.