GIGJ: a crustal gravity model of the Guangdong Province for predicting the geoneutrino signal at the JUNO experiment

Gravimetric methods are expected to play a decisive role in geophysical modeling of the regional crustal structure applied to geoneutrino studies. GIGJ (GOCE Inversion for Geoneutrinos at JUNO) is a 3D numerical model constituted by ~46 x 10$^{3}$ voxels of 50 x 50 x 0.1 km, built by inverting gravimetric data over the 6{\deg} x 4{\deg} area centered at the Jiangmen Underground Neutrino Observatory (JUNO) experiment, currently under construction in the Guangdong Province (China). The a-priori modeling is based on the adoption of deep seismic sounding profiles, receiver functions, teleseismic P-wave velocity models and Moho depth maps, according to their own accuracy and spatial resolution. The inversion method allowed for integrating GOCE data with the a-priori information and regularization conditions through a Bayesian approach and a stochastic optimization. GIGJ fits the homogeneously distributed GOCE gravity data, characterized by high accuracy, with a ~1 mGal standard deviation of the residuals, compatible with the observation accuracy. Conversely to existing global models, GIGJ provides a site-specific subdivision of the crustal layers masses which uncertainties include estimation errors, associated to the gravimetric solution, and systematic uncertainties, related to the adoption of a fixed sedimentary layer. A consequence of this local rearrangement of the crustal layer thicknesses is a ~21% reduction and a ~24% increase of the middle and lower crust expected geoneutrino signal, respectively. Finally, the geophysical uncertainties of geoneutrino signals at JUNO produced by unitary uranium and thorium abundances distributed in the upper, middle and lower crust are reduced by 77%, 55% and 78%, respectively. The numerical model is available at http://www.fe.infn.it/u/radioactivity/GIGJ

💡 Research Summary

The paper presents GIGJ (GOCE Inversion for Geoneutrinos at JUNO), a high‑resolution three‑dimensional crustal gravity model specifically designed for the Jiangmen Underground Neutrino Observatory (JUNO) in Guangdong Province, China. The authors argue that accurate regional crustal models are essential for interpreting geoneutrino measurements, because the crust contributes roughly three‑quarters of the total geoneutrino signal at a detector site. Existing global models (e.g., CRUST1.0) lack the spatial detail needed to resolve local variations in crustal thickness, composition, and density, leading to significant systematic uncertainties in the inferred mantle contribution.

To overcome these limitations, the authors combine satellite gravity data from the GOCE mission with a comprehensive set of terrestrial geophysical observations: twelve deep seismic sounding (DSS) profiles, receiver‑function analyses, teleseismic P‑wave velocity models, and three independent Moho depth maps. Each data set is assigned a 3σ uncertainty based on the original publications, and the information is integrated into an a‑priori model that defines the topography, sediment‑basement interface, and the upper, middle, and lower crustal boundaries (TUC, TMC, TLC, Moho).

The inversion framework is Bayesian. The study volume (6° × 4° centered on JUNO) is discretized into ~46 000 voxels of 50 m × 50 m × 0.1 km. Each voxel carries a material label (UC, MC, LC, etc.) and a constant density. The likelihood function assumes Gaussian observation noise with a covariance matrix derived from GOCE measurement errors; the prior incorporates the seismic‑derived boundary constraints and a Laplacian smoothness term that enforces spatial continuity. A stochastic optimization based on Markov‑Chain Monte Carlo sampling explores the posterior distribution, allowing the model to balance the high‑precision GOCE data against the less precise but geologically informative seismic constraints.

The resulting GIGJ model reproduces the GOCE gravity field with a standard deviation of residuals of about 1 mGal, which matches the satellite’s intrinsic accuracy. Compared with global models, GIGJ reveals a systematic redistribution of crustal thickness: the middle crust is on average 21 % thinner, while the lower crust is 24 % thicker beneath the study area. These structural adjustments translate directly into geoneutrino flux changes—specifically, a 21 % reduction in the middle‑crust contribution and a 24 % increase in the lower‑crust contribution.

Crucially, the authors quantify how these refinements affect the uncertainties on the geoneutrino signal. By propagating the posterior density uncertainties and the fixed‑sediment‑layer systematic error, they demonstrate that the uncertainty on the uranium and thorium contributions from the upper, middle, and lower crust are reduced by 77 %, 55 %, and 78 %, respectively, relative to estimates based on global crustal models. This substantial reduction enhances JUNO’s ability to isolate the mantle geoneutrino component, thereby improving constraints on the Earth’s radiogenic heat budget and bulk silicate Earth composition models.

The GIGJ model and its associated data products are made publicly available (http://www.fe.infn.it/u/radioactivity/GIGJ), providing a template for similar regional crustal modeling efforts at other neutrino observatories such as SNO+ and Borexino. The study demonstrates that integrating satellite gravimetry with detailed seismic information through a Bayesian inversion yields a robust, site‑specific crustal model that significantly lowers geophysical uncertainties in geoneutrino research.