SOIL ENGINEERING PROPERTIES OF LANDSLIDE AND DEBRIS FLOW INITIATION SITES, COAST RANGE, OREGON

Hundreds of landslides occurred in the Oregon Coast Range as a result of heavy rainfall during December 2005, and January 2006. Shallow translational slides that did not mobilize into debris flows occurred adjacent to slides that produced debris flows in geologically homogenous terrain. This field setting provides an opportunity to identify physical differences between hillside materials that mobilized into debris flows and those that did not. Previous experimental work shows that the propensity of landslides to mobilize into debris flows depends on the initial porosity and grain size distribution of the material (e.g. Iverson et al., 2000 and Wang and Sassa, 2003). Initial porosity influences the likelihood of debris-flow mobilization. Upon shearing, materials more dense than a critical porosity dilate reducing pore-water pressure and thus reduce the propensity for flow. Materials less dense than a critical porosity contract upon shearing, increasing pore-water pressure, and increasing the likelihood of debris-flow mobilization. The effect of the fine-grained fraction is to generally decrease the permeability of the material which retards the dissipation of excess pore water pressures, allowing the material to partially liquefy and mobilize into a debris flow. We present field-based test results that generally support these findings.

Undisturbed samples were collected using a modified California sampler at 35 shallow landslide source areas to determine porosity and grain-size distribution. Twenty-one of the landslides mobilized as debris flows and 14 did not. The soils from all the source areas were classified as gravels. However, at a 95% confidence level, materials from debris-flow source areas contain a statistically greater fine-grained fraction of silt and clay (7.0% ± 5.0%) than the materials from source areas of landslides that did not mobilize into debris flows (3.7% ± 2.6%). Materials from debris-flow source areas have a statistically greater mean porosity (0.55 ± 0.04) than those from source areas that did not produce debris flows (0.48 ± 0.03) at a 95% confidence level. These results suggest that the spatial variability of bulk density of landslide-prone materials has important implications for debris-flow hazard assessment. Critical porosity will be determined in the future via direct shear.