Optimal slope units partitioning in landslide susceptibility mapping

Abstract

In landslide susceptibility modeling, the selection of the mapping units is a very relevant topic both in terms of geomorphological adequacy and suitability of the models and final maps. In this paper, a test to integrate pixels and slope units is presented. MARS (Multivariate Adaptive Regression Splines) modeling was applied to assess landslide susceptibility based on a 12 predictors and a 1608 cases database. A pixel-based model was prepared and the scores zoned into 10 different types of slope units, obtained by differently combining two half-basin (HB) and four landform classification (LCL) coverages. The predictive performance of the 10 models were then compared to select the best performing one, whose prediction image was finally modified to consider also the propagation stage. The results attest integrating HB with LCL as more performing than using simple HB classification, with a very limited loss in predictive performance with respect to the pixel-based model.

Mapping Susceptibility to Debris Flows Triggered by Tropical Storms: A Case Study of the San Vicente Volcano Area (El Salvador, CA)

“…This technique splits the range of the covariates into intervals, whose extreme values are called "knots", and finds an optimal linear function between two consecutive knots, which is known as "basis function" [28,29,33,36]. This technique has been already successfully used for stochastic modelling of geomorphological phenomena, including soil erosion and landslides [9,29,[33][34][35][36]54]. The MARS model is the result of the weighted sum of terms that include a basis function or a product of two or more basis functions.…”

“…Figure 9 displays the relative importance of the predictor variables assessed by using the nsubsets criterion [9,35,54]. Slope is the most important independent variable, followed by three classes of GEO (i.e., geo.1, geo.2 and geo.7) and LCL (i.e., lcl.3, lcl.4 and lcl.5), and by ELE, which achieve a relative importance higher than 30%.…”

Geomorphological tools for mapping natural hazards

“…In addition to a specific legend developed for floodrelated and anthropogenic elements, mapping in a typical lowland agricultural landscape with regulated rivers is proposed. Martinello et al (2020) deal with landslide susceptibility modelling, focusing on the selection of mapping units; this is an important topic both in terms of geomorphological adequacy and the suitability of the models and final maps. The authors perform a test to integrate pixels and slope units applying Multivariate Adaptive Regression Splines modelling to the Imera Settentrionale river basin (northern Sicily, Italy) to assess landslide susceptibility based on a 12 predictors and a 1608 cases database.…”

Landslide Susceptibility Mapping by Comparing GIS-Based Bivariate Methods: A Focus on the Geomorphological Implication of the Statistical Results

“…Some studies have demonstrated the efficiency of this kind of terrain segmentation, which can even outperform grid-based models [11][12][13]. One of the main strengths of SU segmentation is that the obtained landslide susceptibility maps are more readable and directly linked to the terrain structure [14].…”