Partial fluid saturation affects absorption and dispersion in sandstones. The proposed theoretical model describes acoustic relaxation due to local fluid flow. Previously proposed models of local flow were based on microgeometries not representative of sedimentary rocks; they were unable to describe the behavior of partially saturated sandstones. The new model is based upon observed microstructures in sandstones. A fraction of the grain contacts in sandstones are permeated by sheet‐like gaps. The incomplete solid‐solid contact allows an interconnected fluid film to exist between the grain surfaces. The model consists of a narrow gap connected to a finite annular pore. An acoustic stress wave drives the film out of the narrow contact region and into the adjacent pore. The viscous flow results in a dissipation of energy. The model predicts the real and imaginary parts of the complex frame moduli as a function of frequency and fluid saturation. The predictions agree well with experimental results.

Seismic wave attenuation in rocks was studied experimentally, with particular attention focused on frictional sliding and fluid flow mechanisms. Sandstone bars were resonated at frequencies from 500 to 9000 Hz, and the effects of confining pressure, pore pressure, degree of saturation, strain amplitude, and frequency were studied. Observed changes in attenuation and velocity with strain amplitude are interpreted as evidence for frictional sliding at grain contacts. Since this amplitude dependence disappears at strains and confining pressures typical of seismic wave propagation in the earth, we infer that frictional sliding is not a significant source of seismic attenuation in situ. Partial water saturation significantly increases the attenuation of both compressional (P) and shear (S) waves relative to that in dry rock, resulting in greater P‐wave than S‐wave attenuation. Complete saturation maximizes S‐wave attenuation but causes a reduction in P‐wave attenuation. These effects can be interpreted in terms of wave induced pore fluid flow. The ratio of compressional to shear attenuation is found to be a more sensitive and reliable indicator of partial gas saturation than is the corresponding velocity ratio. Potential applications may exist in exploration for natural gas and geothermal steam reservoirs.

Seismic attenuation and velocities were measured in resonating bars of Massilon sandstone at various degrees of saturation. Whereas shear energy loss simply increases with degree of saturation, bulk compressional energy loss increases to ∼95% saturation and then rapidly decreases as total saturation is achieved. This behavior is analogous to the behavior of shear and compressional velocities, but the effect on attenuation is larger by an order of magnitude. Our observations are in excellent agreement with the predictions of several models of energy loss involving partial or total saturation. Pore fluid attenuation mechanisms are expected to be dominant at least in the shallow crust.

Ultrasonic velocity and attenuation measurements were made on dry, brine‐ and oil‐saturated Berea sandstone and fused glass beads. The results for fused glass beads are consistent with the predictions of Biot theory. They indicate that as predicted, the Biot absorption/dispersion mechanism shifts to higher frequencies as the fluid viscosity increases. Similar data for Berea sandstone are not consistent with Biot theory, since observed velocities are generally higher than predicted. Using the Biot theory, we calculate low‐ and high‐frequency velocities for the liquid‐saturated samples. “Biot dispersion” is then defined as the percent difference between the low‐ and high‐frequency limits. “Apparent dispersion” is defined as the percent difference between the measured ultrasonic velocity and the low‐frequency Biot limit. Comparison of these two measures of dispersion gives insight into the presence of a non‐Biot absorption/dispersion mechanism. Whenever the apparent dispersion is larger than the Biot dispersion, the extra dispersion is interpreted as being caused by a local flow relaxation. To be consistent with attenuation data, this relaxation must be distributed over at least five to six decades in frequency.

No abstract