Introduction

Knowledge of the thermal and compositional structure of the upper mantle is essential to understanding the evolution of our planet. Direct constraints on shallow upper mantle chemical composition and temperatures are given by mineralogy and the geochemical signature of outcropping rocks. Overall, these data give a good idea of the ranges of temperatures and compositions expected in the first 200km. In general, there is a consensus that the average composition should not be far from the less depleted peridotites we sample, and therefore close to the "pyrolite" composition proposed by Ringwood (Ringwood, 1975). Shallow upper mantle T are consistent with those expected by a mantle adiabat with a potential T at surface of 1300$^oC$ ($\pm 50^oC$). Also, melting conditions of peridotite rocks with depth (see, for example, the KLB1N solid curve by Hirschmann, 2000) are never fully reached, as indicated by seismic observations. This limits the maximum T we expect in the upper mantle. Below 200km, the best constraints on temperatures and compositions come from interpretations of seismic observations based on mineral physics. Recently, interdisciplinary studies which combine current knowledge of material properties at high P and T and geophysical observations are providing important insights into the nature of the upper mantle.

In a recent paper, (Cammarano and Romanowicz, 2007), we found that long period seismic waveforms require a high gradient of shear velocity between 250 and 350 km. Although some variations between continent and oceans are still seen until $\sim$ 300km, the feature we observed is clearly global and raises questions about the average thermal or compositional structure of the upper mantle. In that paper, we discussed the thermal and compositional feature based on the family of PREF models and applying a pressure and temperature-dependent Q model, developed by the previous work of one of us (Cammarano et al., 2003). Also, we showed that the $V_S$ gradient with depth would require a value of G' (i.e. the pressure derivative of shear modulus) for olivine equal to 2.5 (experimental olivine G' are between 1.2 to 1.6) to make possible to explain such a feature without modifying the thermal or compositional structure of the upper mantle and we argued that an enrichment in garnet component with depth is the most plausible explanation.

Here, we refine our interpretation by testing two thermoelastic models (PEPI03, Cammarano et al., 2003 and LARS07, updated by Stixrude and Lithgow-Bertelloni, 2005) and the Faul and Jackson Q model (Faul and Jackson, 2005), discussed also in the preview report. We use the same 3-D $V_S$ model obtained by inverting with respect to PREF, but we invert for T using the two average pyrolitic models coupled with the Faul and Jackson Q model at a given grain size and reference period at 150s. We test effects of grain size and extremely different activation volumes (V*) on the interpretation. Resulting 3-D structures are averaged and the $<$$Q_S$$>$ profile is compared with observations (the same of the previous report). Effects of variations in dry composition are tested using the LARS07 model. We model composition from harzburgite (1) to MORB (0). Pyrolite is given by a $\sim$ 17% of MORB component (0.17). In this report, we show uniquely average properties of the 3-D models.

Figure 2.55: Comparison of thermal interpretation using PEPI03 and LARS07 pyrolitic models plus P-T and GS dependent Q model. Solid lines refer to GS=1cm, dashed for GS=1mm and dotted $V*=0.6\times 10^{-5}$. Color scheme is given in the legend. Reference period is 150s. Starting $V_S$ and thermal (adiabatic) model are in black
\begin{figure*}\begin{center}
\epsfig{file=fabio07_2_1.eps, width=16cm}\end{center}\end{figure*}

Figure 2.56: Comparison of thermal interpretation for 4 different compositions using LARS07 model plus P-T and GS dependent Q model at constant GS of 1cm and reference period of 150s. Color scheme is given in the legend. Starting $V_S$ and thermal (adiabatic) model are in black
\begin{figure*}\begin{center}
\epsfig{file=fabio07_2_2.eps, width=16cm}\end{center}\end{figure*}

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2007, The Regents of the University of California