Regional and global seismic travel time tomography studies show that the Earth's upper mantle has a high degree of lateral heterogeneity, with fluctuations in seismic velocities up to +/- 6% in Vp and +/- 8-9% in Vs. These anomalies are often found in bodies whose size is at the edge of resolution with teleseismic data using ray-theoretical methods. Recent work to incorporate aspects of wave character (i.e., fat rays or banana-donuts) in the tomographic inversions has illuminated smaller scale, previously hard to image anomalies, such as those from low velocity zones in the deeper parts of the mantle. The larger scale fluctuations in the mantle are indicative of both chemical variations and physical state differences (e.g. melt phases) as well as temperature fluctuations. Determining the geometry and amplitude of the seismic velocity distribution is an important means of understanding the geodynamics of the Earth's surface and and upper mantle. Despite the great successes in imaging the Earth's interior using travel time tomography, the relative coarseness of tomography images has led to a situation in which details of mantle circulation and mantle hetereogeneity derived from large and small scale convection are still obscure. The goal of this research project is develop and apply iterative elastic waveform tomography to teleseismic data. This is a means of seismic image formation by iterative data fitting, i.e., an optimization formulation of the seismic inverse problem which requires a method for simulating the propagation of seismic waves, and a method to update some or all parameters in the (an)elastic wave equation so that the simulated wavefield matches or fits a collection of observed data. Waveform tomography has the possibility of improving the resolution of images of the Earth's upper mantle by more than an order of magnitude.
High resolution images made with waveform tomography will more thoroughly utilize the data from the USArray component of Earthscope and permit better geological interpretations of the structure of the Earth's crust and upper mantle beneath North America. This is particularly important in the western U.S., where both earthquakes and volcanoes are scientifically poorly understood natural hazards. The development of waveform tomography generally, and of extended variants incorporating velocity analysis in particular, also can significantly affect the development of industrial seismic technology used in the exploration for petroleum reserves over the coming decade. Many of the issues addressed in our work are also pressing in the industrial context; advances in algorithm design and computer performance offer similar opportunities for improved imaging. A significant fraction of our students enter the oil industry upon graduation, so this project should provide an excellent and unusually forward-looking training resource. The theoretical and practical aspects of the research undertaken in this project will be disseminated to the scientific community, particularly graduate students and post-doctoral researchers, in part through a series of workshops on imaging science that has been funded by the NSF Earthscope program in the coming 3 years, and being organized by one of the PIs of this grant.
