GNGTS 2019 - Atti del 38° Convegno Nazionale

240 GNGTS 2019 S essione 1.4 In the following we analyse the provided geological and geophysical information, the inversion algorithm and the obtained results. Study area and available geophysical information. The study area includes the northern margin of the South China Sea, the Guangdong region and the south-eastern part of Guangxi region. It is a part of the South China Block that has a complex tectonic history (John et al. , 1990), as well as poorly understood crustal composition and thickness (Zheng and Zhang, 2007). As our final model assumes a layered crust and a one-layer uppermost mantle, we define the following surfaces: the topography/bathymetry, the bottom of sediments (i.e. the Top of the Upper Crust, TUC), the Top of the Middle Crust (TMC), the Top of the Lower Crust (TLC), the Moho Discontinuity (MD). The constraints for the definition of the crustal model are obtained from published studies, including Deep Seismic Sounding profiles (DSS), Receiver Functions (RF), teleseismic P-wave velocity models and Moho depth maps, as shown in Fig. 1. Each of this data is taken according to its accuracy when introduced as prior probability in the Bayesian inversion algorithm. Bayesian gravity inversion. We introduce an algorithm based on a Bayesian approach (Rossi et al. , 2016) and able to invert the gravity field by integrating some a-priori data on the crustal structure coming from geological and geophysical data. This approach allows obtaining a 3D voxel-wise crustal model beneath the detector. The investigated volume is split into voxels, V i , with index i = 1, 2 ..., N . Each voxel is a regular prism with a fixed size and described by two parameters: a label L i , denoting the material Fig. 1 - Geophysical input data used for the construction of the 3D model in the 6° × 4° area centred at the location of JUNO detector (Guangdong, South China) Data from deep seismic sounding profiles, P-wave velocity profiles and seismograph stations are used as input to build the a priori model for the inversion of gravimetric data. The raw observations of gravimetric disturbances (δg) are represented as a continuous grid with 5 km × 5 km resolution.