Evaluating the Effect of Root Cohesion on Shallow Landslides for Physically Based Modeling

• Published in: Sensors and materials Vol. 33; no. 11; p. 3847
• Main Authors: Jang, Su-Jin, Kim, Suk Woo, Kim, Minseok, Chun, Kun-Woo
• Format: Journal Article
• Language: English
• Published: Tokyo MYU Scientific Publishing Division 01-01-2021
• Subjects: Accuracy; Cohesion; Evaluation; Interpolation; Landslides; Landslides & mudslides; Modelling; Rainfall; Safety factors; Slope stability; Soil analysis; Soil properties; Soils; Spatial analysis
• ISSN: 0914-4935; 2435-0869
• DOI: 10.18494/SAM.2021.3606

In physically based landslide modeling, to improve the accuracy of source area prediction, it is important to not only use high-resolution terrain information but also apply field-based soil properties. The objective of this study was to use physically based models to evaluate the effect of root cohesion, which may have influenced slope instability and the occurrence of shallow landslides in a forested basin located in Saam-ri, Chuncheon, South Korea. The shallow landsliding stability (SHALSTAB) model used in this study was applied to three scenarios with different soil and root cohesion. In scenarios 1 (1 kPa) and 2 (2 kPa), 15.0 and 4.7% of the entire basin were classified as unstable (factor of safety (FS) <1.0 for the critical rainfall of 151 mm day-1), whereas in scenario 3 (2–20 kPa), where soil and root cohesion input was interpolated and rasterized spatially, only 4.3% of the entire basin was classified as unstable. The accuracy and precision of each scenario were evaluated using receiver operating characteristic (ROC) analysis. The area under the curve (AUC), accuracy, and precision in scenario 3 were 0.858, 96.4%, and 32.1%, respectively, which were higher than those in scenarios 1 and 2. These results demonstrate the incorporation of spatial analysis of soil and root cohesion in determining the effect of root reinforcement in the SHALSTAB model. Thus, field-surveyed cohesion data and their interpolation can be applied to improve the accuracy and precision of the predictive simulation of shallow landslides.