GNGTS 2017 - 36° Convegno Nazionale

GNGTS 2017 S essione 3.3 715 and is a major component of the surface heat flow even when superimposed with concurring near-surface disturbances (Freymark et al. , 2015). Estimating the distribution of these elements occurring throughout the continental crust is not a trivial task, since direct and indirect observations (outcrops, xenoliths, and tomographies, e.g. Rudnick et al. , 1998; Jaupart and Mareschal, 2011; Huang et al. , 2013) indicate that any simple relationship between crustal thickness and heat production (e.g. ������������ ����� �� Lachenbruch, 1970) is complicated by the large intra-crustal compositional variability. For such reasons stochastic approaches are commonly employed, either exact solutions (Srivastava and Singh, 1998) or random modelling (Jokinen and Kukkonen, 1999), and the results are commonly described with their probability density function. Apart from parametric uncertainty, the entity and predictability of the relationship between crustal thickness and total heat production is difficult to evaluate on itself, due to aforementioned superposition of effects in the observed surface heat flow. Method. Applying a scaling relationship between the undulation of a gravimetric CMB, in- verted from a global gravity model, and bulk heat production is a straightforward operation that can already provide a useful estimate, albeit characterised by large uncertainties: up to 30 mW/ m 2 of interquartile range for a 45 km thick crust (Pastorutti and Braitenberg, 2017). From this approach, we can get insights on the entity of the crustal component of the surface heat flow. It is useful in partitioning different thermal regimes, thus helping the interpolation of existing heat flow measures and their downward continuation, attenuating the effect of incorrect exten- sion of local contributions at large distances (i.e. separating the component due to upper crustal emplacements from the signal due to the variation in heat flow from the mantle). We analyse the heat flow prediction with a set of synthetic tests, for which we developed a versatile framework for joint gravity and temperature modelling. The thermal forward modelling part is based on a 3D finite-difference forward modelling solver, on rectangular domains, with non-homogeneous heat production and conductivity, which was written for this purpose. It solves the steady state diffusion equation in the form, where k is the thermal conductivity, A is the heat production per unit of volume, and x is the position vector. The temperature and pressure dependence of thermal conductivity is taken into account, iteratively, using the simple relationships of Chapman (1986) and Schatz and Simmons (1972) for the crystalline crust and lithospheric mantle, respectively. The gravity forward modelling is done with a prism based algorithm, while the inverse modelling relies on an iterative constrained inversion routine (Braitenberg et al., 2007). The domain-box is designed to represent a portion of the lithosphere, under a flat-Earth approximation, with a flat top and bottom boundary. The top is fixed at T(0) , the surface temperature; the bottom is a flat surface in the upper asthenosphere, it can be alternatively set as a temperature or heat flow boundary condition, which can be iteratively varied to obtain the required LAB. The sediment thickness from the top boundary to the crystalline basement is considered to be known and its gravimetric effect is forward modelled and stripped. Results. We devised the above configuration –which allows the fast prototyping of models with any parameter distribution– to evaluate the joint temperature-gravity effect of different layered geometries and of disturbing bodies in a reference lithosphere; to quantify the required instrumental sensitivity (i.e. the detectability in the measured gravity gradient at orbital al- titude); and to test the suitability and effect of the fundamental assumptions. The issue we enquired here regards this last aspect: we assume a relationship between a gravimetric crustal thickness and crustal heat production –to what extent is this adequate to predict the surface heat flow? We revisit the traditional linear model (Jaupart, 1983; Nielsen, 1987) by including the uncertainty due to inverting the effect of the crustal inhomogeneities to a single CMB. Our analysis shows how concurring effects result in complex phenomena, even in these simplified synthetic conditions. Given a certain a constant lithospheric thickness, an increase in the crustal (radiogenic) heat flow component (due to a thicker or more enriched crust)