Data and code for "A Hybrid Parker–Physics-Informed Neural Network Approach for Moho Depth Inversion with Uncertainty Quantification: Application to the Spratly Archipelago and Adjacent Areas, East Vietnam Sea"

This repository contains the input gravity/bathymetry/sediment grids, the derived Moho-depth and uncertainty models, and the Python code implementing the hybrid Parker–physics-informed neural network (PINNs) inversion described in the associated manuscript. It includes the differentiable fifth-order Parker forward operator, the iterative-rescaling (ITRESC) estimator, the PINNs training scripts, the figure-generation code, the digitized OBS Moho control points, and the Parker–Oldenburg baselines, with fixed random seeds and pinned software versions for full reproducibility. Data are released under CC-BY-4.0 and code under the MIT license.