G-DRIFT (Geodynamics Data Reformatting and Integration Facilitation Toolkit) is the data backbone of the G-ADOPT ecosystem. It provides a unified, FAIR-compliant Python interface for loading and working with the geodynamic and seismological datasets that underpin mantle convection studies — things like 1D reference Earth models (PREM), solidus profiles, 3D seismic tomography models (25 published models, including S40RTS, SAVANI, and SEMUCB-WM1), thermodynamic lookup tables for mineral physics properties, and anelasticity correction models.
G-DRIFT handles automatic downloading, caching, and cryptographic verification of datasets from a central repository, and provides physics-aware interpolation (1D spline, 3D KD-tree with multiple kernels) for mapping between different model parameterisations. It solves a real problem: the datasets used in geodynamic inversions are typically scattered, poorly documented, and difficult to reproduce. G-DRIFT centralises them in one well-curated package.
Source: github.com/g-adopt/g-drift