Global Empirical Modeling of Magnetospheric Electrons for MIT Research

Simulations of the coupled magnetosphere–ionosphere–thermosphere (MIT) system require grounding in data to be credible. This can be established through data assimilation, cross-validation with empirical models, or the use of realistic initial and boundary conditions. In this context, MIT research would benefit from a global, pitch-angle-resolved empirical model of magnetospheric electrons. We introduce GENET, a data-driven digital twin of the near-Earth electron environment that reconstructs pitch-angle distributions of 0.1–100 keV electron fluxes at distances within 20 RE. Trained on two decades of Cluster observations, GENET accurately reproduces canonical magnetospheric structures and their large-scale dynamics during various space weather conditions. The model can serve as an observational reference for MIT simulations, provide initial and boundary conditions to numerical codes, and enable multiphysics coupling with other machine learning models. To support global MIT dynamics research, we welcome collaborations on model coupling and cross-validation.