Determining the structure of the Earth's mantle is a fundamental problem of geophysical research. This information, especially the variation in density, provides important insight into the composition and dynamics of the Earth's deep interior. Seismological data have been used to construct models of the mantle, but the data conventionally used for such purpose (body and surface waves) are not sensitive to density variation. A global density model is typically obtained by making a strong assumption that seismic wave speed variations are linearly related to density variation. This assumption loosely translates to three-dimensional variations arising from a purely thermal origin. Results from recent studies, however, suggest that the density heterogeneity within the mantle is poorly correlated with shear-wave speed anomalies, especially in the lower mantle. In this project, we combine multidisciplinary data types to better-constrain properties of the Earth's mantle with a focus on density.
The modern, dense, global GPS sites will be used for the determination of Earth-tide amplitudes across the spectrum of tidal frequencies to yield site-dependent corrections to the (three-dimensional) Love numbers. Previous studies using other ground-based space geodetic systems indicate that the GPS network has great sensitivity to the Love numbers, in part due to the unique temporal spectrum of the tides. These Love numbers are usually calculated using a spherically symmetric Earth model.
The proposed research would explore an approach for combining global geodetic and seismic data to yield a new model for the mechanical structure of the Earth. Seismological data (normal mode splitting information) from previous studies, three-dimensional solid-Earth tidal amplitudes from a new geodetic solution using a global network of Global Positioning System (GPS) sites, and the Earth's gravity field from satellite geodesy will be jointly inverted to obtain a three-dimensional model for the elastic parameters and density of the mantle. Each of these data types has different sensitivities and inherent resolution, and the combination will lead to a model with superior accuracy and resolution relative to the individual techniques.
