In a collaboration with Helmholtz-Zentrum Dresden Rossendorf we will employ artificial intelligence methods to explore the predictability of Earth’s magnetic field changes and the probability for the occurrence of the next reversal.

The geomagnetic field plays an important role in shielding our habitat against space weather influences. The GFZ, with expertise on geomagnetic field observations from ground and space, regularly contributes to the International Geomagnetic Reference Field (IGRF), which is traditionally obtained by non-physical data inversion, and used for a wide variety of scientific and applied purposes, e.g., navigation and exploration. However, methods to predict the geomagnetic core field evolution reliably, even for the next five years as required for the IGRF, are lacking. The HZDR has ample expertise in geodynamo theory and developed the unique DRESDYN experiment. We propose to combine our expertise in a pilot study to apply machine learning methods to investigate the predictability of the geomagnetic field. On the short time-scales, we aim to contribute to IGRF improvements and in particular its five year secular variation forecast. Moreover, we will study longer time-scales in order to better understand the underlying processes in Earth’s core.  Artificial intelligence methods will be employed to infer a realistic physical model of the geodynamo during excursions and/or reversals. We aim to provide an estimate for the probability of the occurrence of the next geomagnetic polarity change within the next centuries to few millennia. This project will provide the basis for broader studies on the mid-term predictability of geomagnetic field evolution.

Project duration

2024 – 2026

Funding

Helmholtz Impulse and Networking Fund

Principal Investigator

• Dr. Monika Korte (GFZ)
• Dr. Frank Stefani (HZDR)