Structure and Evolution of the North American Upper Mantle: Insight from Integrative Modeling of Gravity, Topography and Seismic Tomography Data

Abstract:

A limitation on the application of geophysical methods for the study of the upper mantle is the effect of lateral variations in the structure of the overlying crust that obscure the signal from the mantle. However, the North American upper mantle is particularly well-suited for geophysical study because crustal corrections can be made based on the results from numerous active- and passive-source seismic investigations that have determined lateral variations in crustal properties, including crustal thickness, P- and S-wave velocities, and crustal density estimated from empirical velocity-density relations. We exploit this knowledge of the crust of North America to construct an integrated 3D model of variations in density, temperature and composition within the upper mantle to a depth of 250 km. Our model is based on a joint analysis of topography, gravity, and seismic tomography data, coupled with mineral physics constraints. In the first step we remove the effect of the laterally-varying crust from the observed gravity field and topography (assuming Airy isostasy) using our crustal model NACr2014 (Tesauro et al., submitted). In the second step the residual mantle gravity field and residual topography (obtained in the first step) are inverted to obtain a 3D density model of the upper mantle. Thermal effects dominate this initial density model. To compensate for the thermal effects we invert for mantle temperatures based on the S-wave velocities determined by two seismic tomography models (S40RTS and NA2007). After removing the thermal effect from the mantle gravity anomalies we are left with the upper mantle density variations that are due to compositional variations. We recover two long-wavelength (5°-10°) features in the upper mantle compositional density model that are not evident in seismic tomography models: (1) a strong (+200 mgal) positive compositional anomaly beneath the Gulf of Mexico, perhaps due to eclogite in the uppermost mantle, and (2) a NE-SW trending positive (+140 mgal) anomaly beneath the Grenville-Appalachian suture that may indicate a remnant slab fragment in the uppermost mantle that dates to the Appalachian orogeny. These and other features in the mantle density model allow us to construct an evolutionary model for the development of the North American upper mantle.