GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 1.4 243 1. the solutions are filtered by imposing constraints on the values of the indexes, thus filtering 17 solutions among the 10 3 computed; 2. the one which vector of the indexes has the minimum norm has been chosen. From now on, this model is called GIGJ (GOCE Inversion for Geoneutrinos at JUNO). The output of the GIGJ solution is made up of ~46×10 3 voxels, each one assigned with density and label values. Its geometry and density distributions are shown in Fig. 2 and Fig. 3, respectively. As expected, the GIGJ crustal model exhibits a crustal thinning moving from the north west (continental crust) to south-east (oceanic crust) , together with a higher spatial heterogeneity of the UC density with respect to the MC and LC layers. Fig. 3 - Frequency distributions of the density values for each label of the GIGJ model for the Upper Crust, Middle Crust, Lower Crust and Uppermost Mantle. The latter corresponds to the portion of continental lithospheric mantle down to a constant depth of 50 km. With the aim of estimating the geophysical contribution to the geoneutrino signal uncertainty for each crustal layer, the overall mass and volume uncertainty of the GIGJ solution was calculated. It comprises an estimation error component associated to the solution of the inverse gravimetric problem and a systematic error component due to the adoption of a fixed sedimentary layer. Since the joint posterior distribution P ( ρ , L | y o ) of all the voxels cannot be evaluated, the estimation error component of GIGJ was split into a density and a geometry contribution, both estimated by sample statistics on proper marginal distributions of the individual voxels. Expected geoneutrino signal. The GIGJ model was divided into 7 × 10 7 cells of 1 km × 1 km × 0.1 km, each one assigned with crustal layer label, density value and unitary U and Th