IODP Proceedings    Volume contents     Search
iodp logo

doi:10.2204/iodp.proc.304305.103.2006

Appendix C

Declination bias

The data presented in “Paleomagnetism” (Fig. F240) indicate a bias of normal polarity declinations for both archive halves (toward +x, or an azimuth of 360°) and for discrete samples (toward –x, or an azimuth of 180°). The origin of these directional artifacts apparently lies in the crosstalk between the SQUID sensors (i.e., a dipole moment along one sensor axis generates a signal on the other sensors). The magnitude of the off-axis response may represent a significant fraction (more than half of the axial dipole moment) near the outside of the cylindrical sample region (Parker and Gee, 2002). Most off-axis components would be effectively cancelled for a whole-core measurement. In contrast, significant asymmetries remain for the half-core measurements typically made on the 2G magnetometer onboard the JOIDES Resolution.

To investigate this declination bias, we measured two (nearly) axial dipole standards in the standard measurement position for discrete samples and in the bottom of the sample tray (to estimate the bias when half cores are measured). The standard moment was aligned along the +z-axis of the magnetometer and measured during a series of clockwise rotations about the +z-axis of the magnetometer (Fig. AF1). When the standard is approximately centered in the measurement region (as are discrete samples), the y-magnetic moment describes a sinusoid with approximately zero mean and an amplitude reflecting the small deviation of the standard sample from a purely axial dipole. In contrast, the x-magnetic moment is significantly offset toward negative values. When the same standard is placed in the bottom of the sample region, the resulting measurement bias is exactly reversed (i.e., the x-magnetic moment is significantly offset to positive values).

These results provide a qualitative explanation for the observed bias in normal polarity components both on discrete samples and from the archive-half data. Multiple measurements (e.g., with the moment directed along –z) indicate that, for an axial moment parallel to +z (as is the drilling induced magnetization), a spurious signal along –x is generated for samples in the center of the measurement region and a spurious signal along +x is generated for samples near the bottom of the sample region. Although a small component is generated along the y-coordinate axis, this component is substantially smaller (by up to an order of magnitude) than is the spurious signal along the x-coordinate axis (Fig. AF1). These off-axis components result in a declination bias toward 180° (–x in the core reference frame) for discrete samples and a corresponding bias toward 360° (+x) for archive halves. The bias for discrete samples should be averaged for samples measured in multiple positions but will be present if samples are measured in a single position (the standard measurement protocol).

This bias is most pronounced for samples in which the +z magnetization is much larger than the components in the horizontal (x-y) plane. This is precisely the case for diabase samples from Hole U1309B that acquire a very substantial drilling-induced remanence. The same effect, though less pronounced, is evident in the normal polarity overprints in other lithologies as well. It is important to note that, upon demagnetization, little declination bias should remain when the drilling remanence is no longer substantially larger than the primary magnetization.