Some Problems in Mantle Structure and Dynamics. Part 1. Inversion for Depth Variation of Spectra of Mantle Compressional and Shear Velocity Heterogeneity. Part 2. Physical Model of Source Region of Subduction Zone Volcanism

Citation

Davies, John Huw (1990) Some Problems in Mantle Structure and Dynamics. Part 1. Inversion for Depth Variation of Spectra of Mantle Compressional and Shear Velocity Heterogeneity. Part 2. Physical Model of Source Region of Subduction Zone Volcanism. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/1pkp-t397. https://resolver.caltech.edu/CaltechTHESIS:09242010-084738829

Abstract

Part 1. The scatter in ISC P- and S-wave travel-time residuals was inverted to give a measure of the data-set's incoherent noise and the depth variation of the spectra of the Earth's seismic heterogeneity. The P- and S-wave models are similar in pattern with most of their power shallower than 400km. The power generally decreases with depth and is lowest around 1500-2600km depth. Both models show a slight increase locally below the 670km discontinuity. The long-wavelength half-width (l<50) is around 500km through the upper mantle increasing to around 1200km in the lowermost mantle. The variance in the travel-time residuals requires that (δ1nVₛ/ δ1n Vₚ) ≈ 2, if they are correlated. Our results suggest values as high as 5 from 60-1400km; these could be correct but our preferred explanation is that it's a result of poor depth resolution of the shallowest layer and a difference in the spectral resolution of the two studies.

Part 2. Thermal modeling of a generic subduction zone suggests that the oceanic crust does not melt extensively to be the source of subduction zone magmas. The slab dehydrates and the water is transported laterally into the wedge by a mechanism involving transport fixed in amphiboles and vertical porous flow when free. This water generates melting at the amphibole saturated solidus. Melting reaches a maximum at the hottest geotherm, which also caps the source region. Melts depart the source region in cracks whose direction of propagation is perpendicular to the least compressive stress. For a corner flow regime this leads to focusing of melt towards the wedge corner. The model correctly predicts the location of, the volcanic front. The melt and residue provide buoyancy that leads to local flow reversal and modulates the volcanism with a period of ≈ 1Ma. Estimates suggest more water is subducted than reappears in extrusive volcanics. We suggest the excess water is stored in melts trapped deeper in the section that later become the precursors of granitic and tonalitic plutons.

Thesis Files