New Paleomagnetic Data from South America and Implications for the Southern Hemisphere Geomagnetic Field Evolution

The Earth’s magnetic field (EMF), generated by geodynamo processes, varies across multiple timescales ranging from years to billions of years. Integrated analyses of paleodirectional and paleointensity data over the past 10 million years reveal persistent non-dipolar features in the Southern Hemisphere, linked to the South Atlantic Magnetic Anomaly (SAA). However, the spatial and temporal coverage of high-quality paleomagnetic data remains uneven and is particularly scarce in South America. To address this gap, we provide new paleodirections and paleointensities from different volcanic units in Argentina, Colombia and Brazil. New paleodirectional data were obtained from 23 sites of the Auca Mahuida Volcano (0.19-1.53 My; Argentina) by thermal and alternating field demagnetisation protocols. Paleointensity data were obtained from 41 sites in Colombia, Brazil, and Argentina, ranging in age between 0.005-1.95 My using multiple experimental methods, including Triaxe, Wilson, Thermal Thellier, Microwave Thellier, and the Double Heating Technique of Shaw (DHT-Shaw). The magnetic mineralogy of the studied volcanic rocks comprises dominantly low-Ti titanomagnetite, as indicated by thermomagnetic curves with susceptibility drop between 550 °C and 580 °C and IRM curves with saturation fields below 300 mT. FORC analyses and Day diagrams reveal the dominance of the Pseudo Single Domain (PSD) state of the magnetic grains. Directional data for the Auca Mahuida Volcano yielded a mean direction (D =356.4°, I = -52.04° and α95 = 6.4°; N = 14; K>50; Vandamme cutoff) that is statistically indistinguishable from the expected direction for a geocentric axial dipole (GAD) field (IGAD = −57.0°) in that location, within the 95% confidence limits. The corresponding VGP dispersion Sb =12.799.3°15.9° agrees well with other studies carried out in the Southern Hemisphere and Paleosecular Variation (PSV) models.Out of 332 specimens analysed in paleointensity experiments, approximately 20% met our selection quality criteria (f>0.3, n>4, 𝛽