Item

Abstract

Obtaining reliable geomagnetic palaeointensity (PI) estimates is still a difficult objective. Most common techniques in PI studies are based on various modifications of the Thellier method, but high failure rates and large uncertainties are common. Furthermore, only rocks which obey stringent criteria are assumed to yield faithful PI estimates. In general, magnetic particles must be of single domain size, the magnetization must be a thermoremanent magnetization and chemical alteration of magnetic carriers must be avoided during the laboratory experiment. Recently, a new method for obtaining PIs was proposed, which, on the grounds of results from two historic lavas and three simulated experiments, is assumed to be independent of magnetic domain state. This method uses multiple samples subjected to only one heating step at moderate temperatures. Here we present a test of the multispecimen parallel differential pTRM method on 11 historical lavas from Iceland and Mexico, from which the actual field intensity is known, either from magnetic observatory data, or from magnetic field models. Our results show, that the majority of PI estimates after the multispecimen parallel differential pTRM method yielded results that are very close or indistinguishable within the range of error from the expected intensity and thus largely confirm the findings from the initial study, however, there is in general a small overestimate, which we show is associated with multidomain material. When compared to a Thellier-type PI experiment, carried out on sister samples from five Icelandic lava flows, the average multispecimen PI is closer to the actual field intensity. Samples from one lava, characterized by low Curie temperatures, failed the Thellier approach due to laboratory-induced alteration, but here the multispecimen method returned a reasonable PI. Even though there is evidence that the method is not entirely independent of magnetic domain state, as previously proposed, we regard this method as a valuable alternative in future PI studies.