Multifactor empirical mapping of the protective function of forests against landslide occurrence: statistical approaches and a case study

Abstract

Forests are increasingly valued for services beyond timber and non-timber products including land protection with respect to events such as landslides, soil erosion, floods and avalanches. The most important properties of a protective forest are its ecological and mechanical stability. Planning and implementing multifunctional forest management in protective forests is challenging because of the trade-offs and synergies among the many functions of the forest. In this study, a multifactor empirical method is presented for assessing the protective role of forests on a stand scale with respect to landslide occurrence. Multifactor methodologies typically estimate landslide susceptibility exploiting the relationship between past landslide patterns and site characteristics. Two statistical approaches were here applied to assess the probability of landslide occurrence: the weight-of-evidence technique and the logistic regression technique. Statistical analysis was performed on the basis of landslide detachment zone only. The question of how to estimate protective forest function was answered through the comparison of models established with different sets of predicting factors. This study ultimately aims to provide a decision-support tool focused on mapping the potential role of forests in landslide-prone areas. A case study from the Italian Alps was considered. The density of landslide detachment outside forest areas proves to be more than twice than that within forest areas.

Keywords

• Forest Protective Function
• Landslide Susceptibility
• Logistic Regression
• Weight of Evidence
• GIS
• Alps