WHEN LANDSLIDES DO AND DO NOT RESPOND TO CLIMATE, SEISMICITY, AND SLOPE GEOMETRY CHANGES OVER SEASONAL TO HOLOCENE TIME SCALES

Deep-seated landslides may react to climatic and seismic events or changes to their slope geometry, resulting in significant hazards and rapid landscape change. Despite a general understanding of these landslide triggering mechanisms, predicting if, when, and how rapidly a given landslide will move from data on those triggers remains challenging. Here we highlight results from landslide-prone regions in the northwestern United States to show the diverse mechanisms by which landslides respond or do not respond to changes in forcing mechanisms over both Holocene and seasonal time scales. Two important techniques that capitalize on high resolution topographic data (lidar and structure-from-motion) are essential to answer this question. First, landslides roughen topography when they occur, and then their deposits smooth with time, which can be quantified with spectral analysis to effectively date large numbers of prehistoric landslides. Second, repeat topographic data can be used to derive detailed 3D surface displacements to determine if and how landslides respond to different forcings. Over Holocene timescales, we find that deep-seated landslides in western Washington and northwestern Oregon most frequently occur in ‘hotspots’ where fluvial erosion and geologic structures prime slopes for failure. Landslide frequency more subtly increases in conjunction with wetter climate and increasing frequency of local earthquakes, but these trends are not as pronounced. Case studies of recent landslides highlights why this may be the case. At the scale of an individual landslide, subtle changes to slope geometry, which affect driving and resisting stresses, can dominate movement patterns even during periods of extreme changes in climatic forcing. For example, at the Silt Creek Landslide in Oregon, a localized episode of head scarp retreat caused a major movement episode during a period of drought. Similarly, at the Mill Gulch Earthflow in northern California, its heterogeneous displacement field adjusted to subtle changes in the surface slope and toe erosion in addition to seasonal precipitation variation. These modern examples highlight why landslides in the geologic record may not be precisely tied to climatic or seismic events, and instead integrate a variety of causal mechanisms into their long term behavior.