Item

Abstract

Despite the many advances recently made in palaeointensity (PI) studies, more high quality PI data are required for a better understanding of the full-vector evolution of the ancient geomagnetic field. In this study, we applied the recently-proposed multispecimen parallel differential pTRM (MS) method for absolute PI determination to rocks older than those studied hitherto with this method. Fifty-one lava flows from within the Trans-Mexican volcanic belt, which are generally younger than 1 Ma, returned 32 new PI estimates. Rock magnetic investigations revealed that in most cases, the remanence was carried by Ti-poor titanomagnetites of pseudo-single domain magnetic structure although titanomagnetites with higher Ti-contents and additional contributions of (titano-) haematite were also occasionally present. Comparisons with reconstructions of the global dipole moment reveal that the PI estimates from this study were, on average, around 30 per cent higher than expected. Other, mainly Thellier-type, PI data from Mexico were also observed to be high relative to global records, which could indicate that persistent non-dipole features might be responsible for the higher than expected results. However, the paucity of available data obscures the significance of this observation and the balance of evidence rather suggests an artificial biasing of most measurements towards high values. Our results seem to corroborate results from previous studies on historical lava flows and synthetic samples in which domain state effects were found to cause overestimates of the PI by up to 30 per cent in the MS method. We expect that the degree of the overestimate in the majority of these new MS results is no larger than what might be expected from Thellier experiments performed on samples with a similar given degree of multidomain behaviour. However, seven of the lava flows which were studied here using the MS method were subjected to a preceding demagnetisation step of 200 °C in order to erase viscous remanences and these may have produced results which were further biased towards higher intensities. This additional bias is expected from theoretical considerations and may represent a challenge to future application of the MS method to older rock units. It could, however, potentially be avoided in future studies by performing measurements of the sample moment at temperatures higher than the blocking temperature of any overprint.