The interpretation of seismic data is dependent on laboratory investigations of the elastic and anelastic properties of rocks. Important constraints on the composition of the continental lithosphere have been provided by comparing laboratory and field determined seismic velocities. Much less is known about the nature of attenuation in the crust and upper mantle. Furthermore, laboratory attenuations have not been studied as extensively as velocities. Nevertheless, the utilization of laboratory attenuation measurements to tie seismic data to the anelastic properties of rocks is promising. Laboratory velocity measurements are usually obtained with a pulse transmission technique, whereas measurements of seismic attenuation in the laboratory are determined by resonance techniques, ultrasonic pulse propagation, stress-strain hysteresis loop analysis, and torsion-pendulum oscillations. Velocities in rocks increase with increasing pressure, whereas attenuations decrease. The greatest changes, which occur over the first 100 MPa, are attributed to the closure of microcracks. Velocities decrease with increasing temperature. At temperatures below the boiling point of a rock’s volatiles, attenuation appears to be temperature-independent, and above this, attenuation decreases. Increasing pore pressure, which lowers compressional and shear-wave velocities in sedimentary and crystalline rocks, produces a marked increase in Poisson’s ratio. The influence of pore pressure on attenuation in crystalline rocks is at present poorly understood. At high pressures, velocities are primarily a function of mineralogy, whereas mineralogy may only be of secondary importance for attenuation. Anisotropy, which is common in velocities of many crustal and upper-mantle rocks, may also be an important property of rock attenuation. Attenuation may vary significantly with frequency for saturated rocks, but appears to be frequency-independent for dry rocks. The effect of frequency on velocity is minimal.

At present there are many ambiguities involved in arriving at reasonable models for crustal and upper-mantle compositions. We review several approaches previously employed by geologists and geophysicists to estimate bulk chemical and mineralogic compositions. Recent studies focusing on the petrology of xenolith suites and geologic mapping of high-grade metamorphic massifs, such as the Ivrea zone, support the thesis of heterogeneity in crustal composition. Even though crustal composition varies laterally, there is strong evidence pointing to the importance of metamorphic grade, which generally increases with depth, in controlling crustal petrology. Seismic refraction and reflection methods show the most promise for understanding the lateral variability in petrology. Laboratory studies of the seismic properties of rocks at crustal and upper-mantle pressure and temperature conditions show that seismic data can provide valuable constraints on crustal and upper-mantle composition. Seismic anisotropy is likely an important property of the continental lithosphere at all levels. Field and laboratory experiments carefully designed to investigate this directional dependence of seismic velocities will provide valuable constraints on the fabric and composition of the continental crust and upper mantle. Within the upper crust, physical properties are likely to be strongly influenced by the presence of fractures containing fluids at high pore pressures. A model for the continental crust and upper mantle, emphasizing probable extreme lateral variability, is constructed from information available on exposed deep crustal sections, xenoliths, and laboratory and field seismic studies.

The seismic velocity structure of the Earth’s crust is examined in relation to the role played by high pore pressures. Compressional and shear-wave velocities measured at carefully controlled confining and pore pressures show significant decreases with increasing pore pressure. This behavior is important in sedimentary rocks and in crystalline rocks likely to occur at deep crustal levels. Velocities are shown to be a function of effective pressure rather than differential pressure. Similar results are found for crystalline rocks in which gas or water is the pore fluid. Thus, pervasive CO 2 at elevated pore pressures will lower crustal velocities. Low-velocity regions originating from high pore pressures have high Poisson’s ratios. Within the crust, dehydration accompanying progressive metamorphism can produce high pore pressures and regions of low velocities if pore fluids are contained. A conceptual model is presented for the continental crust in which midcrustal discontinuities such as the Conrad, where present, separate an over-pressured upper region of igneous and metamorphic rocks from underlying dry rocks. Fluids released by lower crustal dehydration are trapped above the Conrad discontinuity. In midcrustal regions, where pore pressures are low due to loss of fluids, progressive metamorphism with depth will be detected as a gradual increase in velocity rather than a sharp discontinuity.