Runout Number: A New Way to Measure Landslide Runout and the Associated Hazards

Abstract:

While researchers have made great progress toward understanding landslide susceptibility and triggers, far less work has focused on understanding their mobility (i.e. the travel distance and velocity) of landslides following the initiation of movement. This is problematic because landslide runout distance is the key factor that determines the area impacted and therefore the potential hazard. Small landslides with high mobility can have greater impacts than larger landslides with low mobility because they can travel at higher velocities and impact areas farther from their sources. Landslide runout has traditionally been quantified by the height-to-length ratio, H/L, also known as the Travel Distance Angle or Angle of Reach. While this ratio is widely used and easily measured, it is strongly influenced by the slope of the runout path, and simply equal to the slope gradient, regardless of the travel distance, unless the landslide traverses a change in slope angle. Therefore, we propose an additional measurement, the unitless Runout Number, measured as the length of runout divided by the square root of the landslide area. We use a database of 158 landslides of varying runout distance from three locations in northern California, Oregon, and Washington state to compare the two methods of runout measurement and to explore their predictability using parameters than can be measured before any landslide movement. The Runout Number better describes the overall runout for the typical topographies and geometries encountered. However, when compared to predictive parameters shown by prior research to relate to landslide runout, the two methods of runout measurement show different correlations. The Angle of Reach shows correlation to initial slope angle, upslope contributing area, expected landslide area, and grain size distribution (for the parameters percent clay, silt, total fines, and sand). The Runout Number shows correlation to plan curvature, upslope contributing area normalized by expected landslide area, and weak correlation to percent sand. While none of these correlations are strong enough for prediction, they indicate the validity of both methods of runout measurement and the need to include both numbers in a hazard assessment.