Thermo-chemical characterization of the Italian crust from seismic data and thermodynamic modeling

Abstract

The lithosphere and in particular its shallowest part, the crust, largely contributes to all natural processes at the surface, including biological ones. Volcanism, CO2 degassing, seismic activity and, ultimately, plate tectonics have deeply shaped our planet, allowing the appearance and the evolution of life in its superb and diversified forms. This profound connection with the shallow subsurface is particularly relevant for our species, whose development and sustainability have been relying on, or have been affected by processes occurring within the crust. The exploitation of natural resources fueling two industrial revolutions, the humanitarian disaster following a M=9 earthquake in a highly populated area are just two, non-exhaustive examples of the two-sided connection between the human civilization and the subsurface. Owing to its importance for human activities and its interaction with the atmosphere and biosphere, the crust has been the target of extensive studies to unravel its structure and chemical composition and to constrain its behavior and role in a variety of disciplines. However, despite its proximity, the crust is largely inaccessible and our knowledge is still limited if compared to its complex heterogeneity, as suggested by geophysical and geological evidence. Lithology, mineralogical associations, preferential mineral arrangement, temperature, pressure, porosity, water content, anisotropy are examples of properties that, with their spatial variations, largely outnumber the data we can collect at surface, resulting in uncertain models of the subsurface or even dissimilar reconstructions of a certain property that still fit the observations equally well. Temperature is one of the key parameters for characterizing the crustal behavior and its relationship with the underlying mantle and neighboring plates. As an example, the depth of the transition from brittle to ductile behavior affects the seismicity of a certain area and, defining the volume of rock involved in the seismic slip, controls the maximum earthquake magnitude in a certain area. While it is relatively straightforward to infer past crustal and mantle temperature through studies on xenoliths and exhumed portion of deep crustal bodies, the understanding of the present-day thermal structure poses numerous challenges and is only poorly known. In fact, current temperatures at depth are extrapolated through heat flow measurements in boreholes, that are limited and sparse thus unable to capture the likely spatial variability of crustal temperature. Moreover, such an approach requires the assumption of a conductive heat transfer, strictly valid only in cold, cratonic regions where heat up-flow by advection is assumed to be absent. Similarly to any material, a varying temperature causes a mineral aggregate to change its stiffness and rigidity, thus varying the velocity of propagation of P-waves and S-waves. Therefore, the anomalies in seismic wave propagation can be exploited for temperature inference. Moreover, since higher temperature causes a faster decay of high-frequency amplitude, the attenuation of seismic waves can be informative on anomalous high temperature at depth. The advantages of using seismic models are rooted in the more homogeneous coverage of the crust and, although the inherent problems of non-uniqueness in the inversion process, the highest resolution images of the subsurface. However, seismic velocities depend on temperature, chemical composition, and pressure through non-linear relationships that must be accurately assessed, as much as the sensitivity of seismic velocities on such variables at crustal conditions. This can be achieved through thermodynamic modeling, by predicting the specific mineralogical aggregate and its seismic features in any conditions of pressure, temperature and chemical composition. The main motivation of this 3-years long Ph.D. project is assessing and applying an interdisciplinary approach to map the temperature distribution at crustal depth through a combination of seismic data and thermodynamics. We investigate the sensitivity of seismic observables (e.g. shear wave velocities, VS) to temperature at depth, in comparison to other variables such as lithology and water content. This work represents the experimental basis for further advance in the estimation of crustal temperature from seismic data. We demonstrated that changes in temperature cause the strongest effect of seismic velocities compared to all other variables. In addition, we carefully analyze the temperature-driven transition of quartz from its alpha to beta form, proving that is seismically relevant in regions characterized by hot local conditions and therefore suitable for the estimation of temperature at depth. We apply a joint-inversion of two different and complementary seismic datasets, receiver functions and Rayleigh wave dispersion curves, for the Italian peninsula. The inversion scheme relies on a Bayesian framework that allows a detailed definition of the model uncertainty and does not require any prior parametrization. Building on our previous relationships, we were able to provide a physically meaningful interpretation of the seismic model(s) we inverted. The analysis of our results allows us to identify intra-crustal jumps in VP/VS ratio that we interpreted due to the alpha- to beta-quartz transition. We also discuss the possible pitfall of using different datasets in complex areas, proving the importance of a Bayesian inversion scheme like the one we use to avoid misinterpretations. Finally, we infer the temperature distributions at the Moho for the Italian peninsula by applying our thermodynamics-bases relationships to absolute shear wave velocities determined by surface waves in the lower crust. The new thermal map proves the deep crustal origin of several shallow geothermal regions. We obtain indeed a very good agreement with the known geological and tectonic framework of the Italian Peninsula and neighboring regions. Also, we find, for the first time, evidence of a deep thermal source beneath NW Italy, in the Piedmont area.