This proposal seeks to address the following issue: Do the seismic waves generated by catastrophic slope failure retain enough information on the forces driving a landslide to allow inference of its dynamics? The dynamics of large, fast-moving landslides are poorly understood for two reasons: (1) the physics of multi-phase sliding and granular flow is complex; (2) empirical constraints on such catastrophic failures are rare, and direct measurements of properties such as acceleration and velocity are almost non-existent. Landslide seismology has the potential to fill the data gap by making remote measurements of the dynamics of mass movement, but the field is in its infancy and fundamental questions need to be answered before major investment in it can be justified. Consequently, this project is focused on a proof of concept of the fundamental methodology behind landslide seismology. The project aims to develop a method for inverting the long-period seismic waves induced during the motion of massive, rapidly accelerating landslides, and evaluate the validity of the forces, accelerations, momenta and energies inferred from it. Emphasis in this pilot study will be on analysis of catastrophic landslides described in the literature for which global broadband seismological data are available.
The aim of this study is to develop the ability to detect catastrophic landslides remotely and in near-real-time. This has obvious benefits to communities engaged in natural hazards mitigation and disaster management, notably in cases where massive slope failures strike in remote areas or under conditions where communications are severely compromised. In addition, application of this tool to the global archive of broadband seismology data will enable the collection of a wealth of new information on catastrophic landslide dynamics, which will serve a wide range of science and engineering communities concerned with slope instability.
This Early-concept Grant for Exploratory Research (EAGER) award addressed the following question: Do the seismic waves generated by catastrophic slope failure retain enough information on the forces driving a landslide to allow inference of its dynamics? The project involved the development of a prototype seismic-analysis technique for collecting, processing, and modeling long-period seismic waves generated by the motion of rapidly accelerating landslides. Such waves are routinely recorded on broad-band seismic stations around the world. The seismic waves are modeled as resulting from the forces acting on the solid Earth in reaction to the momentum changes of the sliding mass. The result of the waveform fitting is a three-dimensional time history of forces, which we call the Landslide Force History (LFH). By Newton's laws, the LFH is the negative of the momentum-change history of the landslide mass, and we use this relationship to reconstruct the motion of the slide directly from the seismic observations. Only the product of mass and displacement can be constrained by determinations of the active forces; we use satellite imagery and other information to estimate the spatial dimension of the slide and can thereby also estimate the landslide mass. The new technique was successfully applied to more than 20 globally distributed landslides. Subsequent to LFH inversion of the seismic data, landslide trajectories were computed and landslide masses were calibrated based on satellite-imagery information. Comparisons of the determined masses and trajectories with field and remote-sensing observations showed that the seismological analysis yields consistent information about the dynamics of the landslide process. Demonstration and validation of the seismological approach to the determination of landslide parameters is the main outcome of this project.
