| Rainfall-induced landslides and debris-flows are calamitous natural hazards that are difficult to predict. Point measurement surveys of subsurface hydro-mechanical properties are often used alongside slope stability assessments to predict landslide occurrence. However, such surveys can be tedious and costly in the field scale, and invasive in the laboratory scale. Presented here are newly applied remote sensing techniques intended to improve slope stability characterization methods at a variety of scales. Electrical resistivity tomography (ERT) was used to quickly estimate soil thickness over a steep (33--40°) zero-order basin in the Oregon Coast Range. After characterizing the hydroelectrical properties at the study site, Depth-to-bedrock was interpreted from the geophysical dataset with a root mean squared error of 27 cm compared to point measurements. In the laboratory scale, a particle image velocimetry (PIV) tool was used to observe shear plane development and strain localization in a tabletop vertical cut slope simulator prior to slope failure. A vertical sliding trap door was gradually removed until the slope failed abruptly, and digital images were taken concurrently for the PIV analyses. Areas of maximum strain localization were found to coincide with the location of the eventual failure plane, showing the PIV technique can be used to detect developing shear planes in the soil. Furthermore, Culmann's Method, a commonly used two-dimensional critical height analysis, was extended to three dimensions and for use in unsaturated soils. Experimental failure heights agreed with the extended theory (within 14.3% relative error) for a range of soil moisture content and cut slope widths, compared to an 88.5% error without the three-dimensional correction. Using the extended theory, a theoretical threshold was also proposed and tested for sidewall width influence on laboratory cut slope failures. |
