South Atlantic Anomaly The hole in the magnetic shield, its two lobes and their future

Colours: total intensity of the field at sea level (magenta = weak, blue = strong); white line: the 32,000 nT limit. Choose a year: 2050* is the extrapolated field (“extrapolated field” method), not a measurement. Light blue: the South American minimum since 1900, orange its forecast, green the African lobe and its forecast. Hover a point for its year and position.

Intensity and extent change steadily: their forecasts are tight. The position moves slowly westwards, about 20 km a year, with errors growing beyond 20 years.

Since 2005 a second minimum exists south-west of Africa (the splitting described by Finlay et al., 2020). In 2025 it is at 23,462 nT against 22,071 nT for the South American lobe, but it weakens by about 87 nT a year against 30: if this goes on, it will become the weakest point of the Earth’s field.

The African lobe has existed for only 20 years: none of these methods could be judged on it. The spread between methods is the only measure of uncertainty, and it is large.

The dipole, the part of the field that looks like a bar magnet, weakened from 32,176 to 29,733 nT between 1900 and 2025 (-7.6 %). To measure its role, the dipole coefficients of both years are swapped, keeping the rest of the field: with the dipole of 2025 and the rest of 1900, the minimum deepens by 65 % of what it actually did; with the dipole of 1900 and the rest of 2025, by 49 %. Both effects combine (intensity is not additive). The westward drift comes entirely from the rest of the field: with the dipole of 2025 alone, the minimum would stay nearly where it was in 1900 (see the table). The anomaly is the surface trace of reversed-flux patches at the core boundary under the South Atlantic (Terra-Nova et al., 2017), which grow and move. See also the dipole moment on the pole forecast.

Dipole and poles. The anomaly grows while the dipole moment falls and the north magnetic pole runs towards Siberia (pole forecast): three aspects of one core field, not three independent causes; a correlation between these curves, all monotonic for a century, would prove nothing more. The dipole swap above gives its physical share.

Reversal? A weakening dipole and a spreading anomaly are sometimes presented as signs of a coming reversal (Pavón-Carrasco and De Santis, 2016); nothing in these data can date a reversal, which takes millennia (pole reversal).

Radiation belts. Above the anomaly the inner belt (protons) comes down to 200–800 km: that is where low-orbit satellites suffer most of their upsets (Heirtzler, 2002). Our belt measurements come from the GOES satellites, in geostationary orbit (36,000 km), where the anomaly has no effect: the link cannot be tested here, and we do not present it as measured.