Earthquakes and volcanism are phenomena that directly affect a large number of people in the US and throughout the world. A major driving force for these mechanisms can be found deep down under the surface inside subduction zones -- areas where two tectonic plates move towards each other, one sliding underneath the other. Seismic imaging is a method of estimating rock properties, which in turn reveal physical processes that take place in the subsurface. The quality of seismic images depend on many factors, the most important of which are the location of seismic sources (earthquakes) and recording devices (geophones) as well as signal distortions caused by the complicated nature of the Earth's interior. This project aims at developing a novel seismic signal processing algorithm that removes many of the problems associated with traditional imaging methods. This algorithm is based on seismic interferometry, a technique that allows to numerically turn sources such as earthquakes into virtual geophones capable of recording a signal at a location where no physical device can be placed. It is expected that the proposed algorithm will ultimately result in higher-quality and higher-fidelity images. Those images will provide additional clues about the important physical processes inside the subduction zones and their role in triggering earthquakes, tremor, and volcanic activity.
When choosing the acquisition geometry for imaging purposes, it is ideal to place physical sources and receivers as close to the area of interest as possible. Doing so boosts the signal energy, improves the illumination of the area, increases the effective receiver aperture and greatly reduces the errors in the signal introduced by wave propagation over long distances. While many earthquakes naturally originate inside the subducted crust or mantle, placing receivers inside the subduction zone at the depth of tens of kilometers is technologically impossible. Seismic interferometry provides a method of redatuming (numerically moving) receivers located at the surface to be as if they were located at the source locations. Local earthquakes can then act as physical sources or, when desired, as virtual receivers. Our new technique, which combines classical seismic interferometry and source-receiver wavefield interferometry, uses teleseismic events as well as data produced by large local earthquakes and converts them into virtual data that we would have observed had we actually had physical receivers inside the subducting slab. The great benefit of this redatuming procedure is the construction of a virtual local seismic survey that can be used to produce an image of precisely the area that we are most interested in. We have already developed the method for the acoustic case; with this project we will extend the work to the elastic case and investigate the necessary source and receiver distribution for the method to produce high-quality data from which a detailed image of the subducting slab can be formed.
Subduction zones are regions in which the ocean floor moves underneath a continent. An example of this phenomenon is the Pacific Northwest where the Pacific plate subducts beneath Oregon, Washington and British Columbia. In seismology we try to image the oceanic crust using waves generated by Earthquakes and recorded on the Earth's surface. Seismologists then use this information to understand how subduction works, which improves our understanding of how our planet has evolved and also to better understand the processes that generate Earthquakes. The images constructed in this way are limited by the relatively small number of seismometers that record the Earthquakes as well as by the locations of the Earthquakes that generate the signals. In this project, we developed a method that allows us to estimate the data we would record if we were able to place our seismometers in the subducting slab itself. By placing our virtual seismometers directly into the subducting slab we should enhance the resolution of our images and thus improve our ability to image and understand the subduction process. Our method uses signals from both local and distant Earthquakes and combines them in new ways to estimate the signals that such in-slab seismometers would record. In this project, we began from a very simple version of the method and have expanded it significantly to more realistic situations. In addition, we have identified the first data set (in Greece) to which we will apply the method and have begun extracting the subset of these data that are necessary for the method.
