Search

We develop a tomographic inversion method that uses teleseismic P wave polarization data to obtain velocity structure. Polarization inversion has some intrinsic advantages over travel time inversion: It is not influenced by source location and origin time errors; it is not sensitive to deep mantle velocity structure and can be used iteratively to improve the tomographic result. Polarization inversion is more sensitive to near-station velocity structure and to velocity gradient and is complementary to travel time inversion in this sense. The method is applied to California Institute of Technology-U.S. Geological Servey southern California array data. The result is generally consistent with previous work and also reveals that the high-velocity feature beneath the Transverse Ranges is bounded between 40 and 200 km depths and possibly has a second small piece at about 300 km depth. The slow velocity anomaly under the Salt on Trough is limited to shallow depths, less than about 60 km.

Teleseismic Pn waves are modeled as a sum of whispering galley waves, which propagate in a waveguide composed of a simple high velocity mantle lid underlain by a low velocity zone. This model is able to account for those Pn wave propagation properties that are not dominated by scattering, i.e., their apparent velocity and lack of long period energy. Pn apparent velocity data constrain the P wave velocity gradient in the upper 100 km of mantle to be low; dvp/dz

We use physical models to examine the effect of scatterers in the D″ region of the mantle on the decay rate of the Pdiff phase in the core shadow. Scatterers are in the size range 7.5–81 km, and measurements of the decay rate are made in the period range 3–60 s. We find that the presence of scatterers in D″ can change the decay rate of Pdiff by as much as 50%. The heterogeneity does not affect the decay rate for periods greater than about 10–20 s. At shorter periods, the decay rate can be either lower or higher than the laterally homogeneous case, depending on the scatterer geometry. We find that a layer of small (7.5 km) scatterers can decrease the decay rate, while layers of larger scatterers (41–81 km) mainly increase it. We did not identify any major effect of scatterer aspect ratio on the decay rate in the one relevant test that we performed (dikes and sills on the core-mantle boundary).

We derive time-domain apparent-attenuation operators from both laboratory data and a single-scattering theory. The scattering medium consists of a homogeneous elastic block containing parallel cylindrical voids, with volume fractions in the range from 0.66% to 3.75%. The attenuation operators were computed by first measuring (or theoretically predicting) the attenuation of spectral amplitude and then constructing a causal seismogram using the Kramers-Kronig (KK) relation. Theory and experiment broadly agree, with the following exceptions: theory underestimated the overall level of attenuation, compressional wave attenuation is more poorly predicted than shear wave attenuation, and the theoretical attenuation operators tend to be larger in amplitude and less oscillatory than is observed. The KK attenuation operators are compared with attenuation operators derived from a least squares deconvolution (LSD) of the attenuated and unattenuated time series. The KK and LSD attenuation operators are superficially very different, reflecting different limitations in the methods. The LSD attenuation operators tend to be narrower than the KK attenuation operators and should not be used to measure pulse width.

Thermal gradient measurements, consolidation tests, and pore water compositions from the Mariana Trough imply that water is moving through the sediments in areas with less than about 100 m of sediment cover. The maximum advection rates implied by the thermal measurements and consolidation tests may be as high as 10−5 cm s−1 but are most commonly in the range of 1 to 5×10−6 cm s−1. Theoretical calculations of the effect of the highest advection rates upon carbonate dissolution indicate that dissolution may be impeded or enhanced (depending upon the direction of flow) by a factor of 2 to 5 times the rate for diffusion alone. The average percentage of carbonate is consistently higher in two cores from the area with no advection or upward advection than the average percentage of carbonate in three cores from the area with downward advection. This increase in average amount of carbonate in cores with upward moving water or no movement cannot be attributed solely to differences in water depth or in amount of terrigenous dilution. If the sediment column acts as a passive boundary layer, then the water velocities necessary to affect chemical gradients of silica are in the range 10−9 to 10−10 cm s−l. However, if dissolution of silica occurs within the sediment column, then the advection velocities needed to affect chemical gradients are at least 3×10−8 cm s−l and may be as high as 3×10−6 cm s−l. This order of magnitude increase in advection velocities when chemical reactions occur within the sediments is probably applicable to other cations in addition to silica. If so, then the advection velocities needed to affect heat flow ( >10−8 cm s−1) and pore water chemical gradients are much nearer in magnitude than previously assumed.