Paleointensities with Arai Signature Corrections for Multi-Domain Lavas from Galapagos

We developed a technique for correcting concave-up Arai diagrams for multi-domains (MD) to determine unbiased paleointensities for a set of 24 basalt samples from sites GA78, 79, 84 and 85 from Floreana in the Galapagos Islands [Rochette et al., 1997; Kent et al., 2010] that have mutually consistent directions deviating ~40° from the reverse polarity time-averaged field. Rock magnetic experiments (loop, FORC and Js-T) show the main magnetization carriers are fine-grain low-titanium MD magnetite particles with grain size variations. Repeating of the loops and FORCs after thermal treatments to 600°C indicate the samples suffer very little from alteration. We used a comprehensive back-zero-forward (BZF) heating technique by adding an additional zero-field heating step in between the original Thellier two opposite in-field heating steps. This triple heating BZF experiment allows us to calculate natural remanence (NRM) loss and thermal remanence (TRM) gain in different ways, in order to estimate paleointensities with various protocols to provide internal self-consistency checks. Lab-applied field of 15 μT and 14 temperature steps up to 575°C were used. Partial TRM (pTRM) checks were inserted by adding forward-field heating steps to previous lower temperatures after the zero-field heating steps. After the first BZF experiment, we gave the samples total TRMs by cooling from 575°C in the same lab field. We then repeated the BZF described above, with the lab-applied total TRMs as synthetic NRMs, using the same lab field and target temperatures. Any systematic departures from linearity in the resulting synthetic Arai diagrams should therefore only represent the MD magnetization recording signatures of samples. From the first BZF, we estimate the uncorrected paleointensities by using the original NRM loss and pTRM gain ratios from a fixed temperature segment of 400°C to 500°C chosen to avoid viscous remanences (VRM) from lower temperatures and alterations from higher ones. From the repeated BZF, we calculate the synthetic NRM loss and pTRM gain ratios from the same temperature segment as the Arai signature correction slope (ASCS), which is the recording bias from the MD effects for each sample. The corrected paleointensity for each sample is then determined by dividing its original paleointensity by its ASCS. The ASCS correction technique is designed to minimize the inevitable paleointensity estimation biases caused by the concave-up Arai magnetization recording properties of MD particles, where use of lower/higher temperature steps tends to over/underestimate paleointensities. Supported by internal self-consistence provided by the BZF, a high success rate (14/24), and good within-site agreements, the ASCS correction technique provides what we believe is an accurate paleointensity for the collective site of 3.7±0.6 μT, a low value that is consistent with the excursional direction (D=212.8°; I=-26.5°). These samples thus provide a severe test because of their weak paleointensity relative to potential VRM acquisition in the present day field of about 30 μT for this locale. We will use these techniques to determine a time-averaged equatorial paleointensity for Plio-Pleistocene from all sampled Galapagos lavas (>50).