CAREER: From Robust, Reproducible Geophysical Inference to Geological Interpretation: New Perspectives on the Continental Lithosphere

Princeton University, Princeton NJ

Investigators

Abstract

For this CAREER proposal the researcher will study the structure of the lithosphere, from a characterization of its long-term mechanical strength (elastic thickness from gravity and topography inversions) and its short-term elastic response (seismic structure all the way down to the oil exploration scale), to its propensity for deformation in the context of plate tectonics and geodynamics (via the anisotropies detectable in both types of observables). The researcher uses existing data from land campaigns (gravity, magnetics, heat flow), satellites (gravity, magnetics, geodesy) and seismic networks and data centers (USArray, IRIS); some processed, some raw, some for which new processing methods are very much part of the research. He is building on his recent development of mathematical functions and statistical estimation theory. These share the common ground of analyzing noisy and incomplete geophysical observations in a variety of acquisition geometries, and building parsimonious interpretations of them based on recent insights in statistical inference and geophysical inverse theory. In this way the geophysical and geological interpretation that he produces is accompanied by a rigorous analysis of modeling errors, an assessment of the uniqueness in the inversion, and the effective communication of uncertainty to the audience. The tools the researcher develops are put at the service of the community via a series of outreach initiatives aimed at enabling rather than merely informing the public over a range of prior exposure and ability, from middle-school students to post-doctoral scholars. The education and outreach building on these activities are geared towards training young people to do science rather than watching it being done. How strong is the rigid outer layer of the Earth called the "lithosphere", how is this strength best measured, and how does it relate to other geophysical and geological observables such as the depth and frequency of earthquake occurrence, the measurable heat flow through the crust, the large-scale patterns of deformation and mountain building? How do we tackle huge data-rich "inverse problems" in seismology, by which the structure of the Earth at both the largest scale or at the scales of an oil field is mapped via the recording and modeling of earthquake- or explosively generated seismic waveforms? What characterizes the magnetic properties of the lithosphere and how are they best measured from primary satellite observations? The answers we hope to provide are of a geological and geophysical nature and significance, but also impact us in our daily lives. Earthquakes are an obvious example; but perhaps not everyone knows, that oil drilling requires detailed models of the lithospheric magnetic field in order to keep the desired bearing, or that medical imaging techniques such as x-ray tomography and magneto-encephalography work with some of the same principles by which we probe the Earth's interior. With our focus on the development of theory and algorithms that help us extract signal from observations, and facilitate its representation, analysis and characterization, and the subsequent evaluation of the distribution of errors in the modeling, we serve a broader purpose than the specific areas in which these methods are first applied. To ensure this, the researcher also develops lesson materials at the middle-school and high-school level with the student as participant, a hands-on research and writing-heavy Freshman Seminar that combines geology and geophysics to aid with archaeological excavations on Cyprus, the continued publication and documentation of computer software on the web, and the mentoring of undergraduates, graduate students and post-docs on their path to delivering ``reproducible research'' to the society at large.