Estimates of three-dimensional rupture surface geometry of deep-seated landslides using landslide inventories and high-resolution topographic data

Abstract

Deep-seated landslides cause significant damage to the built environment. Characterizing the shear surface and volume of these features is often limited to intensive site-specific monitoring, or at a larger scale, generalized empirical relationships developed from change analysis. However, there is limited means of generalizing regional-scale rupture surface information for deep-seated landslides where deposits still overlay much of the failure surface. Herein, a method is proposed that uses high-resolution bare earth digital elevation models with an inventory of mapped landslide extents to gather first order estimates of three-dimensional rupture surface geometries and volumes for rotational and compound landslide features at a regional scale. This approach is first calibrated to a suite of well-characterized landslides and then applied to over eight hundred deep-seated landslides in three different geologic regimes. Estimates of landslide area-volume relationships are proposed that are consistent with previous empirical assessments. The aforementioned relationships describe mean landslide thickness; added relationships relating landslide area to maximum landslide thickness are proposed. The presented results suggest limits on magnitudes of depth with landslide classification between geologic settings. Consistent with prior research, it is observed that landslide area exhibits a first-order control on maximum and mean landslide thickness, but further insight is provided towards the influence of a limited set of landslide classification, which exhibit controls on the range of magnitudes observed for maximum landslide depth associated with failure. Lastly, the implications and limitations of the developed three-dimensional mapping approach is discussed.