Summary of the Results

The main overarching goal of the team was to review existing and develop new approaches to combine global modeling and global datasets in a meaningful way, to extract critical information about magnetospheric dynamics during geomagnetic storms and substorms. State-of-the-art models are often based on assumptions that are difficult to justify, for example ideal MHD models with numerical reconnection. Data mining and artificial neural network models are developing rapidly, providing important insights into global structure and dynamics of the magnetosphere, and constrains for global models. The team reviewed and discussed current approaches to global modeling, including MHD models and “beyond MHD” approaches, data-driven reconstruction of the global magnetic field, data assimilation for ionospheric electrodynamics, global X-ray and energetic neutral atom imaging, and the current state of artificial neural network modeling of the magnetosphere and ionosphere. The team discussed current knowledge gaps, identified future work directions, and started multiple collaborations between scientists from different countries and institutions. The team has published 10 papers with ISSI acknowledgements, including the following works: 1) a study by Stephens et al. (2023) that uses machine learning techniques for data mining of 26 years of magnetometer observations from multiple satellites to reconstruct the overall structure and evolution of reconnection sites in the geomagnetic tail during storms and substorms; 2) an article by Holappa and Buzulukova (2022) that combines 24 years of observations of energetic particle measurements by NOAA POES satellites with simulation results from a global model of the Earth’s magnetosphere to study the effect of the solar wind magnetic field By component on geomagnetic storms; 3) an article by Samsonov et al. (2024) that uses global X-ray imaging of the Earth’s magnetopause to quantify the magnetospheric response to the enhanced solar wind energy input; 4) an article by Smirnov et al. (2023) that presents a new neural network model of electron density in the upper ionosphere, using 19 years of GNSS radio occultation data and consisting of four parameters describing the shape, height and gradient of the F2 peak. The resources provided by the ISSI greatly facilitated teamwork and accelerated progress in multiple research directions, in a relatively short time.