Earth’s magnetic field is probably not reversing

Abstract

The geomagnetic field has been decaying at a rate of ∼5% per century from at least 1840, with indirect observations suggesting a decay since 1600 or even earlier. This has led to the assertion that the geomagnetic field may be undergoing a reversal or an excursion. We have derived a model of the geomagnetic field spanning 30-50 ka, constructed to study the behavior of the two most recent excursions: the Laschamp and Mono Lake, centered at 41 and 34 ka, respectively. Here, we show that neither excursion demonstrates field evolution similar to current changes in the geomagnetic field. At earlier times, centered at 49 and 46 ka, the field is comparable to today's field, with an intensity structure similar to today's South Atlantic Anomaly (SAA); however, neither of these SAA-like fields develop into an excursion or reversal. This suggests that the current weakened field will also recover without an extreme event such as an excursion or reversal. The SAA-like field structure at 46 ka appears to be coeval with published increases in geomagnetically modulated beryllium and chlorine nuclide production, despite the global dipole field not weakening significantly in our model during this time. This agreement suggests a greater complexity in the relationship between cosmogenic nuclide production and the geomagnetic field than is commonly assumed.

One Hundred Thousand Years of Geomagnetic Field Evolution

“…Figure confirms the hypothesis by Brown et al (), (Korte, Brown, Panovska, & Wardinski, ) that excursions occur when dipole power drops to nondipole levels at the CMB in field models with sufficient resolution to around SH degree 5: D and ND have similar values around 100–93 and 60 ka, ND clearly dominates during the Laschamp, the strongest of the excursions, the levels remain similar after the Laschamp from 36 to about 28 ka, and again about 21–18 ka. There is a clear difference between the Laschamp, where dipole power drops far below ND power for ∼5 kyr, with the axial contribution getting close to 0 and the other excursions.…”

“…Summary diagram of geomagnetic field models, paleointensity stacks, and dipole moment reconstructions covering timescales from the Holocene to 5 Ma. References: OH1992 (Ohno & Hamano, ), OH1993 (Ohno & Hamano, ), CALS10k.1b (Korte et al, ), CALS10k.2 (Constable et al, ), HFMx, HFM.OL1 (Panovska et al, ), HFM.OL1.A1 (Constable et al, ), YOS2000 (Yang et al, ), SHA.DIF.14k (Pavón‐Carrasco et al, ), MS1982 (McElhinny & Senanayake, ), GEOMAGIA VADM (Knudsen et al, ), C2018‐Overall stack (Channell et al, ), NAPIS‐75 (Laj et al, ), IMOLE (Leonhardt et al, ), LSMOD.1 (Brown et al, ), LSMOD.2 (Korte, Brown, Panovska, & Wardinski, ), GLOPIS‐75 (Laj et al, ), SAPIS (Stoner et al, ), GGF100k (Panovska, Constable, & Korte, ), sint‐200 (Guyodo & Valet, ), IMIBE (Lanci et al, ), NOPAPIS‐250 (Yamamoto et al, ), SAS‐300 (Hofmann & Fabian, ), SASC‐300 (Hofmann & Fabian, ), RADM (Ziegler & Constable, ), S1999 (Shao et al, ), IMMAB4 (Leonhardt & Fabian, ), IT2008 (Ingham & Turner, ), sint‐800 (Guyodo & Valet, ), PISO‐1500 (Channell et al, ), HINAPIS‐1500 (Xuan et al, ), M1995 (Mazaud, ), sint‐2000 (Valet et al, ), PADM2M (Ziegler et al, ), EPAPIS‐3Ma (Yamazaki & Oda, ), NARPI‐2200 (Channell et al, ), SK1990 (Schneider & Kent, ), GK1993 (Gubbins & Kelly, ), MM1977 (Merrill & McElhinny, ), LN1, LR1 (Johnson & Constable, ), MMM1996 (McElhinny et al, ), LSN1, LSR1 (Johnson & Constable, ), KG1997 (Kelly & Gubbins, ), CC1998 (Carlut & Courtillot, ), MF‐1, MF‐2, MF‐3 (Shao et al, ), LN3, LN3‐SC (Cromwell et al, ).…”

“…A model (LSMOD.1) for the time interval 50–30 ka, including both the Laschamp and Mono Lake (∼34 ka) excursions, has been presented by Brown et al () and recently slightly updated by correction to one of the underlying data sets to version LSMOD.2 (Korte, Brown, Panovska, & Wardinski, ). All available sediment and volcanic data were checked carefully.…”

“…Conversely, if a concurrent variation in different locations is assigned different ages the difference might be described incorrectly by the model as rapidly changing regional secular variation. Prior age alignments of records as frequently done for stacks (e.g., Channell et al, ; Laj et al, ; Stoner et al, ) or applied regionally by Brown et al (), or alignment by iterative modeling (Leonhardt et al, ) can mitigate the effect and can also be prone to reflect potentially wrong assumptions about contemporaneity of magnetic field variations at different locations. Several attempts at tackling age uncertainties have been made in paleomagnetic models, for example, by adjusting the sediment record chronologies by randomly stretching and compressing the individual timescales while preserving the stratigraphic relation (Nilsson et al, ), using a probabilistic approach (Hellio et al, ; Lanos, ) or bootstrap resampling to create an ensemble of models (Korte et al, ).…”

“…Although the global from GGF100k only exceeds the 0.5 threshold for excursions and reversals suggested by Panovska and Constable () during the Laschamp excursion, Figure shows mildly elevated values at ∼28 ka (close to Mono Lake/Auckland), ∼62 ka (Norwegian‐Greenland Sea) and from ∼95 ka back to the beginning of the model (Post‐Blake), suggesting that the model can inform us about global excursion characteristics. Specialized models exist for the Laschamp (IMOLEe, 44–36 ka, Leonhardt et al ()) and the Laschamp and Mono Lake/Auckland (LSMOD.1, 50–30 ka, Brown et al () excursions. Korte, Brown, Panovska, & Wardinski, () have compared them, analyzed the robustness of their excursion features, and updated the 50–30 ka model to LSMOD.2.…”