Landslide manual and automated inventories, and susceptibility mapping using LIDAR in the forested mountains of Guerrero, Mexico

Abstract

Landslides are a pervasive natural disaster, resulting in severe social, environmental and economic impacts worldwide. The tropical, mountainous landscape in SouthWest Mexico is predisposed to landslides because of frequent hurricanes and earthquakes. The main goal of this study is to compare landslide susceptibility maps in Guerrero derived using high-resolution LIDAR (light detection and ranging) data from both a manual landslide event inventory and an automated landslide inventorying algorithm. The paper also highlights the importance of applying LIDAR data in landslide inventorying and susceptibility mapping. We mapped landslides based on two approaches: (1) manual mapping using satellite images and (2) automatic identification of landslide morphology employing the Contour Connection Method (CCM). We produced a landslide susceptibility map by computing the probability of landslide occurrence from statistical relationships of inventoried landslides detected with LIDAR digital terrain models (DTMs) and derived landslidecausing factors using the logistic regression method. Our results suggest that the automated inventory derived through the CCM algorithm with LIDAR DTMs effectively minimizes the timeconsuming and subjective manual inventorying process. The high overall prediction accuracy (up to 0.83) from logistic regression demonstrates the validity and applicability deriving reliable landslide susceptibility maps from an automated inventory; however, LIDAR data are required.

On the Importance of Train–Test Split Ratio of Datasets in Automatic Landslide Detection by Supervised Classification

“…In CCM, contours and nodes are applied to mapping and vectors are used to connect the nodes to evaluate gradients and associated landslide features based on criteria defined by the users. Another study that continued the application of this method was presented by Gaidzik et al [72]. The authors mapped landslides based on two approaches: (1) manual mapping using satellite images and (2) automatic landslide morphology detection by employing the CCM.…”

A Simplified, Object-Based Framework for Efficient Landslide Inventorying Using LIDAR Digital Elevation Model Derivatives

“…Slope was computed using the average maximum technique [35] and profile curvature was computed using directional derivatives of slope [36]. This new raster, called the 'mixture raster' herein, or the scarp index [37], has large negative values for steep and convex terrain, and large positive values for steep and concave terrain ( Figure 3, Step 1). The large positive mixture values, located at the base of scarps and in similar non-scarp terrain, were then extracted to facilitate mapping of scarp terrain.…”

Tectonic control on slow-moving Andean landslides in the Colca Valley, Peru

“…(e.g. Guzzetti et al 2008;Wu and Qiao 2009;Lacroix et al 2013Lacroix et al , 2014Lacroix et al , 2015Lacroix et al , 2019Zerathe et al 2016;Gaidzik et al 2017;Ramírez-Herrera and Gaidzik 2017;Samia et al 2017). Tectonic processes produce lines of weakness and failure surfaces (passive role) affecting the slopes, as well as the topographic growth and over-steepening of the slopes (active role) that promote the occurrence of large landslides (Carlini et al 2016).…”