Modeling the contribution of trees to shallow landslide development in a steep, forested watershed

• Revised landslide model can improve model performance.
• Rainfall interception affects the time when shallow landslide initiates.
• Root reinforcement and tree surcharge affects shallow landslide development.

The objective of this study was to identify the contribution of trees to shallow landslide development in a steep forested watershed using a deterministic modeling approach. Rainfall interception, tree root reinforcement, and tree surcharge were considered the main factors. A revised version of the Transient Rainfall Infiltration and Grid-based Regional Slope-stability (TRIGRS) model was employed in the approach. Hydrological modifications included adding the processes of rainfall interception using an application of the Rutter model. The revised infinite slope stability model was also used to consider tree root reinforcement and tree surcharge. A comparative analysis was conducted with the results simulated by TRIGRS and the revised model to quantify the contribution of trees to landslide development. The Bonghwa site in South Korea, which was damaged by an extreme storm with 228 mm of rainfall on July 24–25, 2008, was selected as the study site. Data related to the local topography, soil, and forest properties were measured in the field for use in the model simulations, although some data were taken from the literature or assumed by the authors on the basis of the site characteristics. The results showed the rainfall interception did not significantly affect the amount of rainfall reaching the soil surface, but it changed the temporal distribution of the rainfall intensity. Additionally, the rainfall interception was found to have little influence on infiltration from the simulation results of pore water pressure. The results of the simulated factor of safety indicated that root reinforcement and tree surcharge made significant contributions to the enhancement of slope stability. The simulation results were compared to the results from locations in which landslides occurred, indicating that the revised model estimated the landslide susceptibility over the entire study site well, while TRIGRS appeared to overestimate the risk of shallow landslides. In conclusion, trees appeared to make a significant mechanical contribution to shallow landslide development during a severe storm event in steep, forested watersheds. Efforts to revise the existing model improved its performance to assess the shallow landslide susceptibility of mountainous watersheds despite some limitations of the current study.