| dc.description.abstract | Reliable palaeointensities (PIs) are necessary to determine the full vector variation of Earth’s magnetic field as a function of geological time and hence to provide constraints for geodynamo models. Knowledge of PIs also has implications for cosmic abundance ratios, as the strength of Earth’s magnetic field influences the amount of radiation entering the atmosphere. Most current palaeointensity determination methods are based on the classic Thellier-Thellier protocol. These protocols are usually fairly laborious and are characterised by rather low success rates.
A new method of determining palaeointensities was proposed by Dekkers and Böhnel in 2006. This ‘multispecimen parallel differential pTRM method’ (MSP-DB) is based on the linearity of pTRM with inducing field, a property that is independent of domain state. It uses multiple specimens to ensure that all specimens have experienced the same magnetic history. However, especially for intermediate grain sizes the MSP-DB protocol seems to overestimate palaeointensities. For this reason, Fabian and Leonhardt (2010) proposed a new, domain-state-corrected protocol (MSP- DSC), which uses three additional steps in comparison to the original DB protocol to isolate the overprinted pTRM, make an estimate of the domain state and detect alteration. Both the MSP-DB and the DSC protocols were tested on historical lavas from La Palma and compared to Thellier-Thellier and microwave results obtained using the Aitken protocol and pTRM checks.
Nine flows were sampled, including the 1949 and 1971 flows, which were sampled at five and three sites respectively. These 20th-century flows are particularly important, as their PI results can be compared to IGRF values. Palaeomagnetic directions determined using thermal and alternating-field demagnetisation for the 20th-century sites were within error of their IGRF values. All samples were within the pseudo-single-domain (PSD) range, but their hysteresis ratios varied widely, from close to the single-domain (SD) range to very near the multi-domain (MD) range. Curie temperatures varied from 80 to 540 oC, showing a lot of variation between sites within one flow and between different flows.
When applying the MSP-DB protocol, the three 20th-century sites with high Curie temperatures (540 !) all yielded PI underestimates within 10% of the IGRF value, whereas the sites with low (135 !) to very low (80 !) Curie temperatures yielded large under- or overestimates (up to 50%), perhaps fortuitously. The overestimates were reduced considerably by the MSP-DSC protocol, but were still quite high. Underestimates were some times corrected upward, usually related to a reduction in scatter. The Thellier-Thellier method yielded technically acceptable plots for ten out of fifteen sites. Two of the 20th-century sites failed, three yielded large underestimates (up to 27%) and three sites yielded results within 10% of the IGRF value. The microwave method yielded acceptable results for nine out of thirteen measured sites. Three of the 20th-century sites failed, the other five yielded results within 12% of the IGRF values.
Comparing results from the three different methods indicates that if two or three methods agree to within a few percent, the obtained PI is close to the actual PI. For sites for which the three methods do not agree, the microwave method yielded the best results. Plotting the obtained PIs against time shows a lower PI (38 μT) 1.09 kyr ago and a decrease from 44 μT in 1585 to 38-41 μT for the 20th-century sites. | |
