Seismic Signals of the 2014 Landslide near Oso, Washington

Abstract:

The 22 March 2014 landslide near Oso, Washington rapidly moved a large volume of material (~8 million m^3), resulting in the efficient generation of seismic waves that were recorded over 350 km away. Analysis of these seismic signals significantly improves our understanding of the dynamics and timing of events. In contrast to the double couple mechanism of earthquakes, at long periods, the equivalent mechanism of a landslide is a single force. Inversion of the long-period waves for the forces exerted on the earth by the landslide yields a time-series that peaks at nearly 10^10 N and lasts ~1.5 minutes. This result, when combined with higher-frequency wave analysis, eyewitness reports, and field observations, implies a complex failure sequence. The earliest force pulses begin before the buildup in high-frequency energy, suggesting the slide began coherently before transitioning within a minute into the highly disrupted and destructive debris-avalanche flow that killed 43 people. This transition may have been due to a collapse of additional material that loaded the material downslope.

Seismically observable “aftershock” landslides continued for weeks. The first and largest occurred a few minutes after the main failure sequence, and was followed by 15 more over the next ~4 hours that were observable at the closest seismic station (11 km away). Three USGS “spiders” equipped with GPS and seismic sensors were deployed by helicopter 10 days later as part of a monitoring effort. Due to their proximity, these seismometers detected signals from even minor collapses, some visually identified by human observers. This augmented network revealed interesting temporal patterns in the post-slide activity, which was dominated by sloughing of material from the headscarp, but also creep of the upper block of the failure mass at a rate of about 1 cm/day. This study shows the value of seismic analysis in landslide investigations to provide timing constraints and help improve our understanding of landslide dynamics.