GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 1.2 137 values of heat flow in these areas. To overcome the limitations of heat flow measurements, a more accurate and spatially representative method is desirable. In this regard, an approach based on seismic velocities can provide better coverage and less uncertainty. Data and method. In our work, we integrate two types of seismological data and infer the temperatures at depth by two independent methods: (i) detecting the seismic evidence of the temperature-driven transition of the quartz from the α to β form and (ii) using thermodynamics for translating seismic velocities into temperature. Specifically, we jointly invert dispersion curves of Rayleigh surface waves (from Molinari et al., 2015) and Ps receiver functions (RFs) using a reversible-jump Markov Chain Monte Carlo algorithm. The inversion method does not require strong prior information on the geometry of the subsurface structure and treats the number of model parameters as an unknown. The target of our study is the Italian peninsula, an area of recent tectonic activity and associated seismic and magmatic activities that are fundamental for understanding the broader contest of the central western Mediterranean. Dispersion curves and RFs are thus inverted at 50 stations across the Italian peninsula, providing a model of the subsurface in terms of V S and V P / V S . Thermodynamic modeling is used as a tool to predict absolute temperatures at depth from seismic velocities. We employ the code Perple_X (Connolly, 2009), based on Gibbs free energy minimization for the calculation of (i) the stable mineral association at any P-T, given a certain chemical composition and (ii) the density and seismic velocities of the resulting mineralogical aggregate. The employed thermodynamic database accounts for the the peculiar behavior of quartz at the α to β transition (T~ 580°C). We assume the crustal composition of Rudnick and Gao (2003), parameterized into the upper, middle, and lower continental crust. To this composition, we introduce 1% of H 2 O, thus allowing the modeling of hydrated minerals (e.g., amphibole) and the formation of melt for a more realistic representation of the actual crust at high temperature. Results. We present the temperature inferred from thermodynamics using the model from Molinari et al . (2015), in terms of temperature at the Moho (Fig. 1). We observe the highest Fig. 1 - Left: contour plot of the Moho depths obtained from the shear-wave velocity model of Molinari et al. (2015). Right: temperature at the Moho, obtained from the shear-wave velocities and thermodynamic modelling, assuming the chemical composition of Rudnick & Gao (2003). The highest temperatures (exceeding 900°C) correspond to thickened portion of the crust along the Alpine and Apennine belts. On the top, the scatter plot of temperature and depth of the Moho shows that the majority of the data are correlated and lie along the 21°C km −1 geotherm. Data points with low correlation lie along hotter geotherms and depict an anomalously hot Moho.