Study on geographical features of deep‐seated catastrophic landslides using high resolution geographical data

Yasutaka TANAKA, Gengo YOSHIMURA, Hiroaki SUGAWARA, Kazuya FUNAKOSHI,　Norihisa SOMEYA and Kazuyuki OKANO

Abstract

This study specifies the geographical features of slopes in the basin of the Totsu River in Nara Prefecture, Japan, where many deep‐seated catastrophic landslides (DCLs) occurred due to the Typhoon Talas in 2011. Based on LiDAR data acquired before the DCLs, we calculated slope gradients and eigenvalue ratios across entire slope areas and their DCL portions, and compared the DCL‐to‐slope ratios of these indexes in a 2x2m grid. In comparison to the entire slope area, DCL portions had a relatively large area with a minimum slope gradient of 25° and eigenvalue ratio of < 5.5.Using these values as thresholds, we classified landforms into four major categories, and quantified microtopographic features highly related to DCLs. Our numerical analyses reveals that, 1) steep and smooth slopes where valley landforms had not been formed are widely distributed;2) slopes with shallow landslides, gullies and other erosional landforms are in the lower part; and 3) gentle slopes with depressions are in the upper part. The study results indicate, it is considered that the use of slope gradients and eigenvalue ratios for numerical analysis of microtopographic characteristics is an effective method for predicting potential location and scales of DCLs.

Key words

LiDAR, deep‐seated catastrophic landslides, eigenvalue ratio, slope gradient, microtopography