Rapid geomagnetic changes inferred from Earth observations and numerical simulations

Earth's magnetic field's large-scale secular variation is well-documented by global models based on recent observations and historical data. Features like hemispheric activity differences and high-latitude rapid changes have been linked to core-mantle interactions and core jets, respectively, providing insights into the geodynamo. Paleomagnetic studies, however, have revealed longer-term changes in field intensity, with rates significantly exceeding those of the modern field, sparking debate about their origin. While the 'intensity spike' around 1000 BCE is controversial, recent estimates align with core flow kinematic bounds. Furthermore, rapid changes in field direction have also been observed, notably in sediments from central Italy, showing angular changes in the Virtual Geomagnetic Pole (VGP) position reaching 1° yr⁻¹, about ten times faster than current changes. The reliability of these paleomagnetic observations and their link to dynamo processes are actively debated. Previous studies using simplified models have shown rapid directional changes, but the rates' geographical preference and relation to core dynamics remain unknown. This study aims to determine the maximum possible rate of geomagnetic field direction change and the locations where these occur, investigating the underlying core surface processes.

The large-scale secular variation of Earth's magnetic field is reasonably well-established through global models using observations from the past two decades and historical records. These models highlight features such as differing activity levels between the Atlantic and Pacific hemispheres, and rapid changes at high latitudes, linked to thermal interactions between the core and mantle and accelerating jets in the core, respectively. Paleomagnetic studies have revealed intensity changes significantly faster than those observed in the modern field, notably the intensity spike around 1000 BCE in the Levantine region. Although controversial, this spike's rate of change is compatible with core flow kinematics. Studies in China and Texas, along with numerical simulations, support the existence of such intensity spikes, often linked to the migration of normal-polarity flux patches across the core surface. Rapid directional changes have also been studied, historically linked to lava flows at Steens Mountain, but now considered untenable. The fastest changes are currently reported in Italian sediments, showing VGP changes of approximately 1° yr⁻¹, significantly exceeding modern rates. However, the reliability of this data is contested. Simple kinematic models have also shown rapid changes, but their relation to core dynamics remains poorly understood.

This study uses two approaches to investigate rapid geomagnetic directional changes: a time-varying paleomagnetic field model (GGF100k) spanning 0–100 ka, and a suite of 16 high-resolution geodynamo simulations. GGF100k offers a long timescale perspective essential for capturing rare, rapid changes. The simulations cover a range of dynamical behavior and reproduce various geomagnetic field features. The analysis focuses on two field properties: the local field vector (**B**) and its transformation to the equivalent Virtual Dipole vector (**P_v**). The rate of change (∂**P**/∂*t* and *dB*/ *dt*) is calculated using differences between values at time *t* and *t* + Δ*t* on a 2° by 2° latitude-longitude grid. The largest changes are labeled 'extremal events'. The study also employs a simplified model of rapid directional changes involving a moving flux patch superimposed on a static axial dipole field to investigate the influence of isolated normal and reversed polarity patches on surface directional changes. Finally, a method for locally reducing the field strength in regions around extreme events was developed to assess the individual contributions of specific CMB field features to surface directional changes.

The analysis of GGF100k and the geodynamo simulations revealed rapid directional changes, with rates exceeding 1° yr⁻¹ in both. Remarkably, simulations show extremal directional changes reaching ~10° yr⁻¹, significantly faster than previously reported paleomagnetic estimates. These rapid changes typically occur during periods of decreasing field strength, but not necessarily at local minima. Most simulated extremal events are distant from reversals, a feature also observed in GGF100k. The latitude of maximum directional change is generally less than 40° in all simulations. Statistical analysis of GGF100k shows a log-normal distribution of *dB*/ *dt* values, with a higher probability of observing faster changes at lower latitudes, a pattern confirmed by the simulations. A simplified model demonstrated that reversed flux patches on the CMB produce faster directional changes than normal patches. Analysis of the simulations suggests that the movement of reversed flux patches towards the observation point is the primary cause of rapid directional changes, with the greatest contributions often coming from equatorial regions.

The consistent findings from both the geodynamo simulations and the GGF100k model strongly suggest that the geodynamo can produce much faster directional changes than previously thought possible. While such events are rare, their occurrence, particularly at lower latitudes where the field is weaker, has implications for future paleomagnetic data acquisition. The association of these rapid changes with times near known excursions and decreased field strength highlights the role of non-dipole field activity. The simplified model and detailed simulation analysis strongly link these extreme directional changes to the migration of reversed flux patches at the core-mantle boundary. Comparison with previous work on intensity spikes indicates that intensity and directional variations reflect different physical processes at the core's top. Future research using enhanced paleomagnetic models may reveal more detailed insights into regional excursions.

This study demonstrates, through a combination of geodynamo simulations and paleomagnetic field models, that the Earth's geodynamo is capable of generating much faster geomagnetic directional changes than previously considered. These rapid changes, often exceeding 10° yr⁻¹ in simulations and reaching several degrees per year in high-resolution models, are primarily linked to the movement of reversed flux patches at the core-mantle boundary, particularly at lower latitudes. Future research should focus on refining paleomagnetic data acquisition strategies to better capture these rare events and further investigate the complex dynamics of the geodynamo.

The geodynamo simulations, while reproducing many features of Earth's magnetic field, do not fully capture the low Ekman and magnetic Prandtl numbers of the Earth's core. The GGF100k model's resolution is limited by the uneven temporal and spatial distribution of paleomagnetic data, potentially underestimating actual field changes. The simplified model of a moving flux patch, while insightful, omits complexities such as multiple interacting patches and time-dependent changes in patch strength and direction. The masking procedure used to isolate the contributions of specific CMB field features involves simplification and potential for aliasing, though efforts were made to minimize these effects.