EMPIRICAL ANALYSIS AND DEVELOPMENT OF LANDSLIDE RUNOUT PREDICTION RELATIONSHIPS FOR THE CEDAR RIVER VALLEY, KING COUNTY, WASHINGTON

This study characterized the landslides and developed empirical relationships to estimate potential future landslide runout on slopes bordering approximately 27 miles (44 km) of the Cedar River Valley (CRV) in western WA. The CRV study area originates in the Cascade Mtns. foothills and is incised into glaciated uplands. There are multiple prior studies relating landslide runout parameters of length (L), height (H), reach angle (α), volume (V) and area (A) using linear/log regressions. Most data sets include varying slide types and sizes, geology, and geographic locations and/or are for large, long runout debris flows, lahars and rock avalanches. The CRV has residential, commercial and infrastructure development and is well suited for a valley-specific empirical study because of available high-quality LiDAR, a large number of landslides, and relatively consistent topography and geology of glacial and interglacial deposits with underlying weathered sedimentary rock on some lower slopes.

Our data set comprises 66 landslides as mapped by King County (2016) and the WA Geological Survey (2019) and 75 additional/modified landslides we mapped using the most recent available LiDAR. It includes debris slides/avalanches (114), debris flows (12) and long runout flow slides (15). The failure H minimum is 34 ft (10 m) and maximum is 955 ft (291 m), with 129 of 141 slope failures ranging in H from 100 to 600 ft (30 to 185 m). Runout lengths range from 127 to 3945 ft (39 to 1203 m). All debris and flow slides have unconstrained or partially constrained flow paths, whereas debris flows have constrained flow paths between the source area and valley floor. Mapped deposit areas and estimated average thicknesses are used to obtain volume values. Analysis results indicate good correlations (R² = 0.62 to 0.95) for regressions of H vs. L and V vs A for the total data and subsets by landslide type; however, unlike some prior studies, CVR relationships using α or tan α as a parameter have very poor correlations. The study successfully met the intended purpose of developing equations for use with the H and V of a potential CRV slide mass, derived from LiDAR-based imagery and topographic data, to estimate the potential runout length and deposit area. Furthermore, uncertainties in the source parameters and the equations can be used to assess a range of probabilities for risk assessment. The methodology could be used to develop similar relationships for other river valleys traversing the glaciated uplands in western WA.