Earth’s magnetic field is dynamic and has repeatedly flipped polarity throughout geological history, events known as geomagnetic reversals. During these episodes, the planet’s magnetic north and south poles switch places, leaving behind magnetic signatures preserved in volcanic rocks, marine sediments, and oceanic crust. These records form the basis of the Geomagnetic Polarity Time Scale (GPTS), a critical tool used by geoscientists to date geological formations, reconstruct past plate movements, and understand changes in Earth’s interior. However, scientists have long suspected that the geological record is incomplete and that some geomagnetic reversals may remain undiscovered due to limitations in observational resolution.
To investigate this possibility, researchers led by the National Institute of Polar Research analyzed the most recent global dataset of reversal timings, GPTS2020. Their aim was to determine whether hidden or unresolved polarity flips might exist within the past 155 million years. Instead of relying on traditional analytical methods, the team employed a statistical technique called adaptive kernel density estimation (AKDE). This method evaluates how frequently events occur across a timeline by assigning probabilities to each recorded data point and generating a smooth estimate of event density through time. In essence, it allows researchers to identify periods when events cluster together or become unusually sparse.
Using this improved statistical framework, the scientists examined how the frequency of geomagnetic reversals has varied through geological time. Previous studies had suggested a long-term trend: reversal frequency declined gradually from around 155 million years ago until the onset of the Cretaceous Normal Superchron—an extended period between roughly 121 and 83 million years ago when the magnetic field did not reverse at all. After the superchron ended, reversal frequency appeared to increase steadily toward the present day. This pattern was generally interpreted as a smooth long-term change driven by processes deep inside the Earth.
The new analysis refined this understanding by applying a more robust method for selecting the statistical parameters used in the AKDE model. Specifically, the researchers used cross-validation to determine the optimal initial bandwidth—the parameter that defines the temporal resolution of the analysis. Earlier studies had relied on empirical assumptions when setting this value, but the new approach allowed the model to capture finer variations in reversal frequency.
With this improved resolution, the researchers discovered four distinct dips in the estimated reversal frequency after the Cretaceous Normal Superchron. These dips may represent periods when short-lived polarity flips occurred but were not captured in the existing geological record. Because brief geomagnetic reversals can be difficult to detect—especially when rock layers or sediment records lack sufficient temporal resolution—they may simply be missing from the GPTS rather than absent from Earth’s magnetic history.
The team further tested their model by incorporating newly identified polarity events known as the Lima–Limo reversals, dated to roughly 31 million years ago and discovered through high-precision studies of Ethiopian flood basalts. When these events were included, the dip in reversal frequency around 32 million years ago became smoother. This result suggests that the model is capable of identifying where missing events may exist and supports the idea that the Earth’s magnetic field likely changed more continuously than previously documented.
Understanding the timing and frequency of geomagnetic reversals is important not only for geophysics but also for broader Earth science. Periods with frequent reversals provide many magnetic markers that help scientists precisely date rocks, track plate tectonic movements, and reconstruct environmental changes. Conversely, long intervals with few reversals—such as superchrons—make geological dating more difficult but offer valuable clues about processes occurring deep within the planet.
Ultimately, the study identifies four time intervals that are promising targets for future research. By examining deep-sea magnetic anomalies, volcanic sequences, and sediment cores with higher resolution, scientists may be able to uncover previously undetected geomagnetic reversals, improving our understanding of the geodynamo and the evolving dynamics of Earth’s interior.
https://phys.org/news/2026-02-geomagnetic-reversals-earth-incomplete.html