Magnetization of axial lavas from the southern East Pacific Rise (14°–23°S): Geochemical controls on magnetic properties

Although the spatial association of iron‐rich lavas and high‐amplitude magnetic anomalies is well documented, a causal link between enhanced iron content and high remanent magnetization has been difficult to establish. Here we report magnetic data from approximately 250 samples, with 8–16% FeO* (total iron as FeO), from the southern East Pacific Rise (EPR) that provide strong support for the presumed geochemical dependence of remanent intensity. The limited age range (0–6 ka) of axial lavas from this ultrafast spreading ridge (∼150 mm/yr full rate) effectively minimizes variations resulting from time dependent changes in geomagnetic intensity or low‐temperature alteration. Systematic sampling relative to the chilled margin illustrates that substantial grain size‐related variations in magnetic properties occur on a centimeter scale. Both microprobe data and Curie temperatures suggest that the average groundmass titanomagnetite composition in the southern EPR samples is approximately constant (modal modified ulvöspinel content = 0.67) over a wide range of lava compositions. Saturation magnetization and saturation remanence are highly correlated with FeO* ( R = 0.73 and 0.83, respectively), indicating that more iron‐rich lavas have higher abundances of otherwise similar titanomagnetite. We show that there is a good correlation between natural remanent magnetization (NRM) and FeO*, provided that sufficient specimens are used to determine the average NRM of a sample ( R =0.63). Because the range of iron contents in mid‐ocean ridge basalts is limited, the best fit slope (4.44 A/m per %FeO* in an ambient field of 0.030 mT) should provide reasonable bounds on the equatorial magnetization of submarine lavas (∼10 A/m at 8.5% FeO* and ∼50 A/m at ∼16% FeO*). Finally, we demonstrate that along‐axis variations in NRM closely parallel geochemical changes along the southern EPR. Where magnetization values deviate significantly from those predicted from the range of measured FeO* contents, these discrepancies may reflect additional unrecognized geochemical variability.