Prague, Octobre 2, 2026
Our colleague, space physicist Benjamin Grison, has completed a major update and extension of the unique GRMB (Geospace Region and Magnetospheric Boundary) data set. This data set provides a detailed mapping of the four Cluster spacecraft crossings through various geospace regions and magnetospheric boundaries over more than two decades (2001 to 2023). Conducted alongside colleagues from the Institute of Atmospheric Physics of the Czech Academy of Sciences, the research was carried out in collaboration with the European Space Agency (ESA) and international partners. “The dataset is a great tool for any scientist doing large statistical studies,” says Benjamin Grison. The updated methodology provides also a reference framework for training machine learning models and defining orbits for future space missions.
The Cluster mission, developed by the European Space Agency (ESA), has been investigating the interaction between the solar wind and Earth’s magnetosphere for more than two decades. “Processing such a massive volume of data required creating homogenized classification rules—essentially a ‘legend’ to reliably identify which geospace region the spacecraft is passing through,” explains Benjamin Grison.
Updated Methodology Resolves Ambiguities
Cluster (Artist’s view, Credit: ESA)
The updated GRMB methodology addresses previous ambiguities in identifying transition regions. It refines the identification of key regions such as the magnetotail lobes and the plasmasheet, while providing explicit guidance for assigning the simple versus complex tags to distinguish standard plasma signatures from more heterogeneous or fluctuating environments.
When researchers reprocessed the data from the first two years of the mission (2001–2002), they found that the spacecraft spent approximately 70 more days in the lobe regions (IN/LOB) than initially cataloged in the original data set. In addition, the data set was extended to cover the full year 2023. “Despite severe operational constraints, such as the quality degradation of some dataset, the updated methodology demonstrated strong robustness—the proportion of unknown or uncertain regions remained well within a few percent,” says Benjamin Grison.
First 3D Statistical Map of Geospace Regions
Based on the refined Cluster data set, an international team led by the Institute of Atmospheric Physics, in collaboration with researchers from the Royal Belgian Institute for Space Aeronomy, reconstructed a three-dimensional statistical spatial probability map of geospace regions. While reproducing the large-scale organization of the magnetosphere, the map reveals detailed structural characteristics. The cross-tail probability distributions demonstrate that the plasmasheet is not a uniform layer across the magnetotail, but appears significantly thicker on the flanks than near the center. “The resulting map does not merely convey a schematic representation from a textbook, but reflects the real, long-term average structure of the magnetotail, accounting for natural boundary motion and temporal variability,” describes Benjamin Grison.
Machine Learning Training and Future Mission Design
According to the authors, the GRMB data set can also serve as an ideal reference for developing and training automated machine learning models.
“Identifying geospace regions automatically is challenging due to instrumental noise, missing data, or varying instrument availability. Our data set provides a physically consistent, context-aware ‘ground truth’ by enforcing a minimum region duration of 20 minutes to eliminate short-lived fluctuations. Furthermore, this framework is transferable to other magnetospheric missions such as MMS, THEMIS, or future space projects,” adds Benjamin Grison.
Official operations of the mission concluded in September 2024, and the last two satellites underwent controlled re-entry and disintegrated in the Earth’s atmosphere at the turn of August and September 2026.
The complete GRMB data set for Cluster 1–4 covering 2001–2023 is publicly available through the ESA Cluster Science Archive (CSA). Reprocessing for the year 2024 is scheduled for completion in November 2026.
Scientific Paper: Published in Journal of Geophysical Research: Space Physics / AGU: https://doi.org/10.1029/2026JA035382
On 1 September 2026, Tango reentered Earth’s atmosphere at 21:30:31 UTC / 23:30:31 CEST, the last of the four Cluster satellites to do so. Tango had been specially lined up for a targeted reentry safely over the South Pacific Ocean, with the precision so high that a plane could fly out from Tonga and successfully record its demise for reentry science.