Landslide Inventory Mapping and Dating using LiDAR-Based Imagery and Statistical Comparison Techniques in Milo McIver State Park, Clackamas County, Oregon

Abstract

A landslide inventory was conducted for the Redland and Estacada Quadrangles of western Oregon using LiDAR DEMs. Many of these landslides were field verified. In total, 957 landslides were mapped using LiDAR whereas previously, only 228 landslides were believed to exist in the study area based on SLIDO information. In Milo McIver State Park, 41 landslides were mapped using LiDAR. SLIDO indicated only three landslides present within the park. A sequence of seven terraces of the Clackamas River is mapped in Milo McIver State Park. Landslides in the park predominantly occur between these terraces. Soils studied from representative areas within landslide complexes and terrace surfaces help to formulate a soil chronosequence for the study area. The youngest soils, Entisols, develop in less than 1,600 years, Inceptisols between 1,600-10,000 years, and the oldest soils, Alfisols, develop in at least 10,000 years. Classifications of soil profiles netted ten Alfisols (mainly on upper terraces), 49 Inceptisols, and 20 Entisols (reactivated slides in the complexes). The soils are predominantly ML soils and have Loam and Silt Loam textures. Results of spectral analysis, carried out on the LiDAR DEMs, indicate that the spectral character of landslides changes with age. However, applying statistical tools such as the Kolmogorov-Smirnov test (K-S test) and cluster analysis suggest that it is not possible to use spectral analysis to determine the relative age of failed surfaces. The K-S test showed that the spectral character among landslides varies widely. Cluster analysis resulted groupings not based on age or terrain type. The result of the cluster analysis illustrates that it may not be realistic to use a single cutoff, which separates failed terrain from unfailed, in the spectral distributions to analyze an entire region. In all, the results of the spectral analysis were not conclusive. Individual landslides, not complexes, should be used in future studies, since complexes have slides that are continually reactivating. The landslides were also too young to display very much differentiation in age based on soils and spectral analysis. Essentially, a similar study should be conducted using individual landslides with a large age range for more conclusive results.