Secular variation of the middle and late Miocene geomagnetic field recorded by the Columbia River Basalt Group in Oregon, Idaho and Washington, USA

Abstract

This study of 118 discrete volcanic flows from the Columbia River Basalt Group is aimed to determine their distribution of geomagnetic field directions and virtual geomagnetic poles (VGPs) and to compare the inherent secular variation parameters with those from other studies. The magnetic signature of these rocks is uniformly carried by primary titanomagnetite, indicating that magnetic changes are due to variations in the magnetic field. Although most flows are flat lying, those that are tilted pass the Tauxe and Watson tilt test. Sequential flows with statistically similar site means were grouped, and directions that were considered outliers were evaluated and removed using the Vandamme cut-off method. Three normal-polarity (N-polarity) and three reversed-polarity (R-polarity) intervals are revealed by the stratigraphically ordered flows and have mean directions of N polarity (dec/inc=6.6°/+61.2°, k=29.3, α 95 =4.2°), and R polarity (dec/inc=178.2°/−59.2°, k=16, α 95 =5.5°). Regression analysis indicates that the secular variation analysis has not been affected by regional rotation, and that apparent polar wander is negligible. The VGP distribution is almost perfectly circular and supports the preference of VGP positions for the dispersion analysis. Dispersion parameters with corrections for within-site scatter (S b ) show a range of 14.3°–25.5°, including error limits, and were consistently higher for R-polarity results than for those of N polarity. Published dispersion parameters for extrusives <5 Ma show S b values slightly lower than ours, yielding values of 16°–19°, although the difference is not statistically significant. In contrast, published dispersion parameters from high quality data from the Cretaceous Normal Superchron are lower than those for the Neogene, which suggests that the noisiness of the magnetic field correlates with the frequency of reversals. Our new results allow us to extend the Plio-Pleistocene palaeosecular variation database to the bottom of the middle Miocene. Many Miocene formations on a variety of continents are suitable targets for future analysis. Furthermore, the significant difference between the reversed and N-polarity dispersion parameters is intriguing and needs substantiation.