Imag(in)ing the continental lithosphere

This paper is primarily concerned with seismically imaging details in the mantle at an intermediate scale length between the large scales of regional and global tomography and the small scales of reflection profiles and outcrops. This range is roughly 0.1– 1 kmbab10–10 km, where a is the scale. We consider the implications of several models for mantle evolution in a convecting mantle, and possible scales present in the non-convecting tectosphere. Reflection seismic evidence shows that the structures preserved from continental accretion within and at the margins of the Archean cratons are subduction related, and we use subduction as an analog for scales left by past events. In modern orogenic belts we expect to find subduction structures, small scale upper mantle convection structures, and basalt extraction structures. We examine some of the scales that are likely formed by orogenic processes. We also examine the seismic velocity and density contrasts expected between various upper mantle constituents, including fertile upper mantle, depleted upper mantle, normal and eclogitized oceanic crust, and fertile mantle with and without partial melt. This leads directly to predicting the size of seismic signals that can be produced by specular conversion, and scattering from layers and objects with these contrasts. We introduce an imaging scheme that makes use of scattered waves in teleseismic receiver functions to make a depth migrated image of a pseudo-scattering coefficient. Image resolution is theoretically at least an order of magnitude better than traveltime tomography. We apply the imaging scheme to three data sets from 1) the Kaapvaal craton, 2) the Cheyenne Belt, a Paleoproterozoic suture between a protocontinent and an island arc, and 3) the Jemez Lineament, a series of aligned modern volcanic structures at the site of a Proterozoic suture zone. The Kaapvaal image, although not defining a unique base of the tectosphere, shows complicated “layered” events in the region defined as the base of the tectosphere in tomography images. The image of the transition zone discontinuities beneath the Kaapvaal craton is remarkable for clarity. The migrated receiver function image of the upper mantle beneath the Cheyenne belt is complicated, more so than the tomography image, and may indicate limitations in the receiver function imaging system. In contrast the Jemez Lineament image shows large-amplitude negative-polarity layered events beneath the Moho to depths of ∼120 km, that we interpret as melt-containing sills in the upper mantle. These sills presumably feed the Quaternary–Neogene regional basaltic volcanic field. © 2006 Elsevier B.V. All rights reserved.