POTENTIAL INITIATION OF THE CARMEL KNOLL AND DEVILS PUNCH BOWL LANDSLIDES OF LINCOLN COUNTY, OREGON

We investigated the mechanisms potentially responsible for the formation of the Carmel Knoll and Devils Punch Bowl landslides, located on the central coast of Oregon. Although much is known about reactivation of the landslides, there is a lack of understanding of factors that could result in formation of future, similar landslides. The studied slides are typical of nearby deep-seated landslides in the region that pose hazards to property and the main transportation route, U.S. Hwy. 101. Our characterization involved literature review, engineering geologic mapping, subsurface exploration, soil sampling, laboratory testing of residual strength parameters, nearly continuous monitoring of displacement and meteorological and groundwater conditions, and slope stability analyses. The landslides occur in weak Tertiary sedimentary rock that dips 19 to 28° toward the ocean and are mainly translational in movement with a small rotational component in the toe. No evidence was observed of sliding along any particular beds or bedding planes, but seemed to cut across the beds. The slides are 100 m long by 150 to 200 m wide and occur along a 10 to 20 m high, 30° to 50° coastal bluff. Groundwater generally occurs 5 to 15 m below the ground surface. Based on reconstructions from offsets of bedding and manmade structures and current displacement rates, the slides have an approximate age of 200 to 700 yrs. The climate of the region has remained fairly constant for the last 600 years with an average rainfall of 170 cm/yr. The coastal area of Oregon is also tectonically active and has been subject to seismic accelerations during the past with the last earthquake dated in 1700. The estimated potential peak seismic acceleration for the sites is 0.35 g. To evaluate the potential triggering mechanisms, we used RocSciences Slide to perform two-dimensional slope stability analyses. We estimated the pre-failure topography and peak strengths and modeled three different scenarios: high groundwater, strain softening and seismic acceleration. Modeling results suggest that the slides occurred due to seismic shaking. Unrealistically high groundwater levels and strain softening did not result in failure during modeling. Our findings may be useful for predicting formation of future, similar landslides.