Dynamically consistent interpretation of the seismic structure at the base of the mantle

Citation

Sidorin, Igor A. (1999) Dynamically consistent interpretation of the seismic structure at the base of the mantle. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/v5ch-c986. https://resolver.caltech.edu/CaltechETD:etd-11042005-150916

Abstract

There is increasing evidence of large degrees of heterogeneity in the seismic structure of the lowermost 200-300 km of the mantle constituting the D" layer. The region is believed to play an important role in the dynamics of the mantle and the Earth as a whole. This is the place where hot plumes, reaching the Earth's surface may originate. This is also the region where most of the lithosphere subducted from the surface may ultimately settle. The interface between the two largest structural domains of the Earth, the core-mantle boundary, is also a zone of active chemical reactions. One of the most diagnostic seismologically observed features in the D" region is an apparent seismic velocity discontinuity 200-300 km above the core-mantle boundary, generally referred to as the D" discontinuity. The primary evidence for the discontinuity comes from the observed seismic triplication with phases Scd or Pcd arriving between the direct arrival, S or P, and the core-reflected, ScS or PcP, in the 65°-85° distance range. The cause of this abrupt velocity increase is unknown and various explanations have been advanced, including sharp thermal gradients, a chemical interface, or a solid-solid phase transition. However, neither seismology nor geodynamics alone can distinguish between the alternatives. In addition, no satisfactory explanation has yet been given to the apparent intermittance of the D" discontinuity, as the triplication is strong in some regions (such as Alaska or Central America) but weak or missing in other regions (such as Central Pacific). We use a combination of dynamic and seismic waveform modeling to provide tighter constraints on this structural feature of D" and reduce the tradeoffs that exist in both seismological studies and dynamic modeling. The dynamic models are based on the adiabatic model that is computed in Chapter 2 by integrating available mineral physics data. The temperature field, chemical heterogeneity, and the distribution of phases computed from convection models are mapped to seismic velocities which are then used to compute synthetic seismic waveforms. By comparing these waveforms with data, we rule out some classes of dynamic models in favor of others. In particular, in Chapter 3 we demonstrate that a model with a chemical layer at the base of the mantle does not provide a consistent explanation for the seismological observations of the D" discontinuity. We propose that the strength of the triplication is conditioned by both the abrupt velocity increase at the D" discontinuity and the local velocity structure accompanying the discontinuity. Variations of the local structure strongly modulate the strength of the observed triplication and provide a natural explanation for the apparent intermittance. We also show that purely thermal gradients computed from convection models do not produce a sufficiently strong Scd phase. In Chapter 4 we suggest that the observed regional patterns in the strength of the D" triplication are most compatible with a phase change model of the D" discontinuity. In Chapter 5 a variety of convection models with a basal phase transition are tested to obtain the characteristics of the phase transition most compatible with observations. We find that the best value for the ambient elevation above the core-mantle boundary is about 150 km and the best value for the Clapeyron slope is about 6 MPa/K. In Chapter 6 this model is further tested by placing a discontinuity in context of the global shear velocity structure recovered by Grand's [1994] tomography model. We find that such a synthetic velocity model with a phase change characterized by a shear velocity contrast of 1.5%, ambient elevation ~ 200 km and Clapeyron slope ~ 6 MPa/K predicts the observed differential travel times patterns for the D" triplication beneath Alaska, Eurasia and Central America. The model also provides an explanation for the apparent intermittance of the D" discontinuity by predicting very weak triplication for Central Pacific and north-eastern Caribbean where convincing evidence for the D" triplication is lacking.

Thesis Files