Constraints on the acquisition of remanent magnetization in fine-grained sediments imposed by redeposition experiments

Abstract

The magnetization of sediments is acquired through complex processes involving a large number of physical, mineralogical and magnetic parameters. Despite many attempts, the degree to which these processes distort the record of the geomagnetic field as it is archived as a natural remanent magnetization (NRM) remains poorly documented. Among many other parameters, it is important to evaluate the amount of smoothing inherent to the signal, its relation with the field intensity and its variability in the sediment column. In order to address these problems, we performed new redeposition experiments using carbonate-rich, Ocean Drilling Program (ODP) Site 851, and clay-rich, ODP Site 854, sediments. We used a dilute solution of gelatin, which gels below 20 °C, thereby allowing mechanical blocking of the magnetic grains. We observed two critical results: (1) The efficiency of detrital remanent magnetization (DRM) decreases with increasing sediment concentration for a given slurry. Sediment concentration is defined as: c = m_s / (m_s + m_H2O), where m_s and m_H2O are the sediment and water mass, respectively. Higher c would then reflect greater compaction, lower water content and, presumably, greater depth in the sediment column. This effect reduces DRM efficiency nearly to zero for c > ∼ 50%. (2) Post-depositional remanent magnetization (pDRM) is important for c < ∼ 50%. pDRM is carried by grains covering the entire coercivity spectrum. By comparing the mean value of NRM divided by anhysteretic remanent magnetization from the previous magnetostratigraphic study at Site 851 with the relevant ratio derived from our redeposition experiments, we were able to estimate that pDRM was significant within the depth interval where ∼ 44% < c < ∼ 56%. If the sediment concentration profile for the uppermost sediment was known at Site 851, we could define the transfer function for the deconvolution of the field variations. Finally, the dependence of DRM efficiency on c suggests that changes in the thickness of the surface mixed layer would change DRM efficiency. Thus, fluctuations in maximum bioturbation depth could possibly cause DRM intensity changes, regardless of changes in earth's field.

Introduction

For some 50 yr, the remanent magnetization of sediments has been widely used in a large variety of applications. These studies have improved our knowledge of the earth's magnetic field over time scales covering almost the entire range of field variability (from hundreds to millions of years). The data have also been applied to problems concerning tectonic deformation, crustal block rotations and environmental studies. More broadly, these data provide important constraints on the geodynamics of the deep earth and are important to problems ranging from the growth of the inner core to mantle convection. Paleomagnetic constraints on these problems are unique in comparison with others in that they can resolve these problems in terms of geological history.

Considering the importance of sedimentary paleomagnetic records, our understanding of the physical processes that control detrital remanent magnetization (DRM) is woefully underdeveloped. (For the sake of clarity, we define DRM as a magnetization produced by settling or the physical agitation of the sediment and post-depositional remanent magnetization [pDRM] as a magnetization produced in settled sediments by a field [H] of similar magnitude to the earth's by magnetic grains that are free to mechanically rotate parallel to H without physical agitation of the sediment.) As we pursue increasingly detailed paleomagnetic records, it becomes critical to understand how these records filter the signal of the earth's magnetic field. The problem is a simple convolution, m = f ⁎ s, where m is the paleomagnetic record, s is the true variation of the earth's magnetic field and f is the filter, or transfer function that controls any lag or smoothing that may be present in the imperfect recording of the field.

The most recent published redeposition experiments were conducted by Katari et al. [1] who concluded that f causes no smoothing of s and that, at most, causes an unspecified amount of lag. Their conclusion is uncertain because they did observe evidence for smoothing, but claimed that it was due to a thermally activated viscous remanent magnetization (VRM). In spite of this claim, Katari et al. [1] could not completely rule out the possibility that they had observed a true pDRM. The existence of pDRM has been defended in previous laboratory experiments [2], [3], [4], [5], [6] and also on the basis of paleomagnetic observations from sediment cores [7]. Katari et al. [1] pointed out that the observed pDRM involved samples that were dried in order to produce a consolidated sample. They then argued that the drying out process could be an important factor in producing the pDRM. For most sediments, drying does not play a role in consolidation and, thus, these experiments may not be applicable. In summary, one cannot say whether or not pDRM smoothes s based on previous laboratory experiments.

Technical difficulties involved in simulating DRM in the laboratory have substantially hindered the development of our understanding of DRM. Simple gravitational settling of fine-grained material is often not sufficient to consolidate the sediment [8] and fine sediment can stabilize at high water concentrations. The sediments used in the present experiments stabilize at ∼ 80% water by mass (all percents given are mass based). Two techniques have previously been employed to address this major difficulty. Both approaches generate critical problems. Some experimentalists [5], [9], [10] removed water through drying and/or filtering but drying might disturb the fabric of sediment and produce a “drying remanence”. Others [1], [11], [12], [13] used a cryogenic magnetometer to measure unconsolidated sediment without disturbing its fragile fabric. However, such measurements have been made in zero field; therefore magnetic grains that are not “locked in” can rotate away from their magnetized position when removed from the field in which the deposition took place. Grains that would contribute to a pDRM would be in such a state. Measurements performed on unconsolidated sediments are, thus, particularly maladapted to detect pDRM.