Expected sensitivity of Neutrino Oscillation tomography to the structure and composition of Large Low Velocity Provinces in the lower mantle

Continued advancements in the treatment of oscillation parameter uncertainties, atmospheric neutrino flux modeling, and detector response characterization are not only advancing precision in neutrino physics but are also enabling a novel application of neutrinos to geophysics: Neutrino Oscillation Tomography. This technique probes the Earth's interior by exploiting the sensitivity of Earth-crossing atmospheric neutrino oscillation probabilities to variations in the electron number density (defined as the product of matter density and proton-to-nucleus ratio) along the neutrino’s path. As such, it offers a novel method complementary to conventional geophysical techniques for inferring the density or chemical composition of deep Earth regions, such as the mantle and core. A potential target of study for this method is the Large Low-Velocity Provinces (LLVPs), two seismically anomalous regions (~1000 km in height) located atop the core–mantle boundary (CMB) and collectively covering approximately 20% of its surface. Depending upon their physical characteristics, LLVPs may play a critical role in large-scale geodynamic processes, such as mantle convection, plate tectonics, and the geodynamo. However, their density, composition, and origin remain poorly constrained. Conflicting geophysical interpretations propose that LLVPs may be either denser or less dense than the surrounding mantle, or may exhibit depth-dependent density gradients. Their mineralogical composition is also uncertain, with no direct observational method currently available for verification. In this study, we evaluate the sensitivity of neutrino oscillation tomography to variations in the electron number density and spatial geometry of LLVP-like structures. Using three-dimensional Earth models and the OscProb solver to compute oscillation probabilities, we simulate atmospheric neutrino event distributions incorporating a full von Mises–Fisher distribution to describe the detector's angular resolution in both zenith and azimuth. Two Earth models are considered: (1) a reference model adopting the mantle structure of the Preliminary Reference Earth Model (PREM), and (2) a perturbed model that includes an LLVP-like volume composed of three overlapping slab segments, each approximately 300 km thick, characterized by a specified electron number density contrast, relative to the ambient mantle. Sensitivity to LLVP parameters is quantified using a log-likelihood ratio (LLR) test, allowing assessment of constraints on the electron number density of these enigmatic structures.