Department of Earth and Environmental Sciences
My primary research interest is in seismic tomography, the science and
art of making three-dimensional maps of the interiors of solid bodies -- first and foremost, the earth. Seismic waves are mechanical waves that propagate
through solid objects, and they can be used to compute 3D reconstructions of interiors that are otherwise inaccessible. I apply these
tools across all spatial scales: from ultrasonic waves, which sample a
piece of building stone a few centimeters deep, to body
waves from large earthquakes, which sample the whole planet from crust to core. This amounts to remote surveying of
the earth's interior, equivalent to the
non-invasive imaging methods used for medical diagnostics. Progress is largely driven by the growth of
supercomputing capabilities (and prospective graduate students should
be motivated to use these tools).
Seismic tomography of mantle structure beneath western North America. This is a bird's-eye view from south-east. The rainbow-colored 3-D contours outline the Farallon plate, from the surface down to 1800 km depth. The tectonic plate once formed the sea floor of what is now the Pacific Ocean basin, but it has been gradually been sinking back into the mantle over the past 200 million years. In the process, it has thrust up the spectacular mountains that characterize western North America today (surface topography is overlain in translucent blue/green/brown shades). This ancient plate is still detectable because seismic waves travel slightly faster through it than through the ambient mantle. We used waves from over 600 earthquakes, recorded by more than 1000 seismic stations to compute this 3D reconstruction of the deep subsurface. Plot adapted from Sigloch et al. 2008, Nature Geoscience.
On the small spatial scales, in non-destructive material testing and the mapping of the usable subsurface (e.g. acquifers, building sites), seismic tomography mainly tries to detect variations in elastic moduli and in wave attenuation (fluids). On the planetary scale, we essentially map temperature variations in 3D. Thermally driven motions shape the face of our planet and ultimately enable life on Earth. Understanding this "convection machine" is therefore of fundamental interest, and seismic imaging is the most immediate leverage point on the problem.
I pursue two main lines of research:
(1) The push for better methodology in the imaging techniques that should result in much sharper, more detailed pictures. Collectively, such methods are referred to as "waveform inversion", and have in common their extensive use of high-performance computing infrastructure. A variant called "finite-frequency tomography" is particularly suited to imaging on continental and global scales. More recently, I have also become involved in so-called "global search" (Bayesian, Monte Carlo) methods, which can fully assess the uncertainties and "blurriness" of the obtained images. Together with colleagues from applied geophysics, we are developing tomography methods for very small sample scales (centimeters in the case of ultrasonic waves, tens of meters in the case of shallow seismics.)
(2) Understanding tomographic pictures in terms of earth processes and earth history. This big-picture challenge requires making qualitative and quantitative links to the neighboring disciplines: geology, geodynamics, plate reconstructions, mineral physics, and geochemistry. Many structures that we observe in the interior (subducted plates) were located at the surface in the distant geological past, and left their traces on ancient land forms and life, which are independently observed by geologists today. Hence tomography provides a complementary window into the past ~200 million years. Of course the interpretation of tomographic images must be consistent with current surface activity, and they may even give glimpses of the distant future (e.g., sites of future volcanism).