Magnetic susceptibility is a property that is measured both on core and downhole. Remanent magnetization is measured on core and can be derived from the downhole magnetic field and susceptibility data. Comparison of the log to the core data offers insights into the advantages and limitations of both data sets, and, in particular, an assessment can be made of the sediment magnetization both before and after the drill string overprint has been imparted to the cores.
Differences in depth positioning of the cores and logs are due to ship heave, tides, inaccurate determination of the depth to seafloor, incomplete core recovery, core expansion, and so forth. On the basis of matching polarity intervals, a bulk depth shift (downward) of 6 m has been applied to the Hole 1095B core data and of 3 m to the Hole 1096C core data. There are some small variations in the depth offset downhole, but the bulk depth shifts applied here are typically good enough to bring the different data sets to within 1 m of each other.
Both the Bartington core susceptibility system and the SUMS logging tool work on the same principle. They contain a transmitter coil that creates an alternating magnetic field, which induces a magnetization that is proportional to the susceptibility of the sediment. The frequency of the two systems are similar: 220 Hz for the SUMS and 565 Hz for the Bartington meter. The Bartington meter output unit is ~10-5 SI units, and the SUMS output unit is "ppm SI" equivalent to about 10-6 SI units. Both units are volume normalized. The volume measured by the SUMS is much larger than that measured by the Bartington meter. The vertical resolution of the SUMS is about 40 cm, and the depth of investigation is about 80 cm (Vibert-Charbonnel, 1996).
The overall shape of the two curves is similar (Fig. F9); however, differences in amplitude are significant and are attributed to difficulties in volume normalization. The range of variation is greater for the core susceptibility, which may be due to the variable thickness of core, especially in the extended core barrel cores.
The remanent magnetization (Jr) of the sediment can be calculated from the GHMT remanent anomaly (Bfr) using Equation 3 (see "Relationship of the Field in the Borehole to the Magnetization of the Sediment"). This magnetization has uncertainties due to the possibly incomplete removal of the other components that make up the (measured) total field, to variations in hole diameter, and to uncertainties in the volume of sediment measured.
The shipboard cryogenic magnetometer was used to measure the magnetization of split cores (Barker, Camerlenghi, Acton et al., 1999). The initial core magnetization was measured, followed by the magnetization remaining after 20- and 30-mT alternating-field (AF) demagnetization. The cores acquire a drill string overprint during coring and their trip up the pipe to the ship. The overprint is commonly directed vertically downward (e.g., Roberts et al., 1996). The cores have no induced magnetization because they are measured within shielding that creates a very low field environment.
The magnitude of the magnetization is similar to, or sometimes slightly lower than, the initial core magnetization (Fig. F9). However, whereas the core magnetization is always in the downward direction because of the overprint, the in situ magnetization derived from the GHMT logs shows both normal and reversed polarity. The core magnetization after the 20-mT demagnetization step is of lower magnitude than the in situ magnetization, but the two curves have generally the same shape.
We interpret these observations to mean that the drill string magnetization replaces the low-coercivity remanent magnetization in the cored sediment and does not significantly affect the higher-coercivity components. The drill string effect does not extend as strongly into the sediment surrounding the borehole, leaving the in situ magnetization largely intact. A unidirectional overprint would manifest itself in the GHMT-derived magnetization as a reduction in amplitude without a change in the polarity, because the remanent anomaly derived from the GHMT total field log is set so that there are both positive and negative values. The generally good agreement of the polarity determinations from cores and logs shows that the remanent component of the field has been well isolated.
The Koenigsberger ratio is the ratio of the remanent to the induced magnetization. Its value depends on two main factors: the type of sediment and the paleointensity of the Earth's magnetic field. The dependence on sediment type is caused in large part by the grain size. Larger (multidomain) magnetic grains have a smaller magnetization per volume than smaller (pseudo-single domain) grains and are less able to overcome mechanical resistance from the surrounding matrix and align along the magnetic field. Hence, a sediment containing predominantly large ferrimagnetic grains will have a smaller remanence than one containing smaller grains. Lithic clasts that contain magnetic grains are even more unlikely to be able to align along the magnetic field, a situation that is likely to apply to diamict sediments. Because the susceptibility, and hence induced magnetization, is fairly constant with magnetite grain size (Heider et al., 1996), the Koenigsberger ratio should also be smaller for coarser-grained sediments. The ratio also depends on paleointensity because an increased intensity of the Earth's field leads to better alignment of magnetic grains along it and therefore an increased remanent magnetization. This is demonstrated in independent sediment sections that record similar paleointensity variations through time (e.g., Guyodo and Valet, 1996), although there are still doubts about the completeness of the normalization techniques (e.g., Kok, 1999).
The Koenigsberger ratios derived in this way are only approximate because of the uncertainties in the remanent magnetization and susceptibility. The ratio cannot be calculated absolutely from the core measurements, because of the large overprint and subsequent demagnetization. However, the shapes of the log-based ratio and the ratio of the 20-mT demagnetization step to the core susceptibility are found to be reasonably similar below 400 mbsf in Hole 1096C.