Sampling of sediment cores using plastic U‐channels has made possible the acquisition of detailed records of paleomagnetic secular variation, geomagnetic polarity, environmental magnetic studies, and relative paleointensity over the past several million years. U‐channel measurements provide the great advantage of rapid measurements of long sediment cores, but the signal resolution is attenuated by the response function of the magnetometer sensors, which therefore restrains the recovery of rapid and large‐amplitude field changes. Here we focus on the suitability of the dynamics of reversals and excursions derived from U‐channel measurements. We compare successive individual paleomagnetic directions of 1.5 cm × 1.5 cm × 1.5 cm cubic discrete samples with those of a 1.5‐m equivalent U‐channel sample train obtained by placing the samples adjacent to each other. We use varying excursion and transition lengths and generate transitional directions that resemble those of the most detailed paleomagnetic records. Excursions with opposite polarity directions recorded over less than 7.5 cm are barely detected in U‐channel measurements. Regarding reversals, U‐channel measurements smooth the signal of low‐resolution records and generate artificial transitional directions. Despite producing misleading similarities with the overall structure of transition records, longer transitional intervals fail also to reproduce the complexity of field changes. Finally, we test the convolution of magnetization by different response functions. The simulation reveals that even small response function changes can generate significant differences in results.

Precambrian paleomagnetic records from dyke swarms provide a unique source of information regarding the Archean geomagnetic field and more specifically the average field strength produced by the early dynamo. We sampled 16 paleomagnetic sites from the Dharwar giant dyke swarm in southern India which was emplaced between 2.365 and 2.368 Ga. Despite taking great care in selecting locations exempt of any geological disturbance, only two of these sites provided primary directions with very steep inclinations and therefore were emplaced in close to a magnetic pole. Paleointensity experiments were conducted on a subset of samples from the dyke margins. The characteristic magnetization is carried by single domain magnetite grains with a very narrow range of unblocking temperatures inferred from the sharp decrease by at least 75% of their remanence above 520 C. The paleointensity results indicate an average low field of 9.2 6 7 mT, consistent with reported values from Canadian dyke swarms for the same period. These results combined with the Thellier-Thellier determinations obtained so far for the Precambrian suggest that a low field period prevailed from circa 2.3 to 1.8 Ga, while the preceding and following time intervals are characterized by significantly stronger paleointensities. Although this suite of episodes is not fully incompatible with previous models for the long-term evolution of the geodynamo, it is tempting to make the link with the recent suggestion of an early dynamo sustained within a conductive magma layer at the base of the mantle from 3.5 to 2.5 Ga which progressively declined until convection became sufficiently efficient to reactivate a strong dynamo process within the Earth's liquid core.

Rapidly deposited layers (RDL) such as turbidites or hyperpycnites are mostly studied for their sedimentological properties, but are carefully avoided in paleomagnetic studies due to the disturbances caused by such sudden and rapid sediment accumulation. Therefore, these layers can also be seen as potential indicators of sediment parameters susceptible of affecting the alignment of magnetic grains and ultimately the acquisition of the natural remanent magnetization (NRM). We have compiled 13 Holocene rapidly deposited layers from core MD99-2222 in the Saguenay Fjord, eastern Canada (St-Onge and al., 2004) with varying thicknesses (from 7.1 cm to 1,510 cm) and 4 Quaternary turbidites of different origins, to document the influence of sedimentary and magnetic parameters on natural remanent magnetization acquisition. We found a logarithmic relationship between rapidly deposited layers thickness on the one hand, and the amplitude of inclination changes and magnetic grain sizes on the other. Inclination and magnetic grain sizes are themselves correlated to each other by a logarithmic law. As there is no relationship between inclination deviation and stratigraphic depth, compaction alone cannot account for such large effects on inclination. Flocculation is grain size sensitive, but it is expected to affect mainly the natural remanent magnetization intensity, rather than its direction. Turbulence that prevails during the rapid deposition of sediments during such events is most likely the dominant factor.