Previous   |    Next

doi:10.2204/iodp.proc.330.103.2012

Paleomagnetism

The natural remanent magnetization (NRM) intensity of samples from Hole U1372A spans a very broad range from 3 × 10^–5 to 39 A/m (median = 1.7 A/m), with the lowest values associated with volcaniclastic units. Relatively well defined principal component directions with misfit values ≤ 3.40 were obtained for 1364 intervals from archive-half core measurements (for pieces longer than 9 cm). These directions are generally consistent with stepwise alternating-field (AF) and thermal demagnetization results from 100 discrete samples. Data from the recovered core should provide reliable inclinations for ~20 in situ cooling units. Directions in the volcaniclastic units are more scattered, reflecting the fact that some of the basalt pieces recovered from these intervals are clasts. Nonetheless, some of these basalt intervals may represent in situ lavas that would further increase the number of flow units for determining the paleolatitude at Site U1372.

Archive-half core remanent magnetization data

The remanent magnetization of archive halves from Cores 330-U1372A-4R through 38R was measured at 2 cm intervals using the cryogenic magnetometer. All data acquired within 4.5 cm of either piece end were filtered out prior to further processing, and thus only pieces longer than 9 cm were considered (see “Paleomagnetism” in the “Methods” chapter [Expedition 330 Scientists, 2012]).

The intensity of NRM and magnetic susceptibility vary similarly downhole (Fig. F62B, F62C), with prominent decreases in the volcaniclastic-dominated Units XII and XV. NRM intensity ranges from 3 × 10^–5 to 39 A/m (median = 1.7 A/m). The lower value corresponds to the practical sensitivity level of the cryogenic magnetometer (see “Paleomagnetism” in the “Methods” chapter [Expedition 330 Scientists, 2012]). When volcaniclastic Units XII and XV are excluded, the remaining stratigraphic units show a slight decrease in both NRM intensity and susceptibility downhole. The magnetization values for these 15 stratigraphic units are well approximated by a log-normal distribution (geometric mean = 2.0 A/m ± 0.43 log units). Geometric means for these units decrease from 4.0 A/m in Unit III to 1.4 A/m in Unit XVII (linear regression with R² = 0.4).

Archive halves were progressively AF demagnetized, and representative results are shown in Figure F63. Most intervals display relatively simple behavior during demagnetization, with nearly univectorial behavior after removal of a lower coercivity component, typically by 5–15 mT (Fig. F63A–F63C, F63E, F63F). Although in some instances this lower stability component has a steep inclination consistent with a drilling-induced remanence (Fig. F63C), more commonly this component has a moderate inclination (Fig. F63B, F63E). For a minority of intervals, including both basalt and hyaloclastite breccia lithologies, the highest attainable alternating field (70 mT) was insufficient to significantly demagnetize the NRM (Fig. F63D, F63H), which presumably reflects the dominance of higher coercivity phases such as (titano)hematite (see thermal demagnetization results in “Discrete sample remanent magnetization data”). For some intervals the demagnetization path displays erratic behavior or does not trend toward the origin (Fig. F63G, F63I), likely reflecting the acquisition of a spurious anhysteretic remanent magnetization at the highest demagnetization levels.

Characteristic remanent magnetization (ChRM) directions were calculated automatically using principal component analysis (PCA; Kirschvink, 1980) over a range of demagnetization treatments, and the lowest misfit value was used to identify the most reliable linear segment (see “Paleomagnetism” in the “Methods” chapter [Expedition 330 Scientists, 2012]). The resulting ChRM directions and intensities (3392 intervals) are shown in Figure F62D. More robust PCA picks with misfits ≤ 3.40 (40% of all data; n = 1364) are emphasized with a larger dark red symbol. In intervals that can be most confidently identified as in situ lava flows (see “Igneous petrology and volcanology”), inclinations are generally steep and negative, consistent with Southern Hemisphere normal polarity. There is a distinct increase in the spread of inclinations in the sedimentary sequence (Unit II) comprising breccia and conglomerates and also in volcaniclastic intervals (particularly Unit XV). In volcaniclastic units it is more difficult to determine whether the basalt fragments recovered represent fragments of lava flows, pillow lavas, or possibly larger clasts, particularly when recovery is low. Nonetheless, consistent negative inclinations in several intervals in the volcaniclastic units suggest that some intercalated in situ flow lobes or pillows may be present. Even though most intervals with steep positive inclinations likely represent clasts, the interval of positive inclination at 71.5–72.7 mbsf at the bottom of Unit V may represent a narrow zone of reversed polarity (see “Discussion”). ChRM intensity fluctuations generally parallel those of NRM (Fig. F62B). The similarity of these two values is, in part, attributable to the minimal amount of drilling-induced remanent magnetization, which is consistent with the moderate stability inferred from the median destructive field (MDF′), which is defined as the alternating field required to reduce the vector difference sum to half its initial value (Fig. F62E).

Discrete sample remanent magnetization data

Remanent magnetization of 100 discrete samples (8 cm³ cubes) was measured with the spinner magnetometer, and magnetic susceptibility was also determined for every sample using the Kappabridge. NRM intensities of discrete samples range from 8.4 × 10^–5 A/m (hyaloclastite breccia) to 13.4 A/m (basaltic lava) (Fig. F64), with a geometric mean of 2.1 A/m (Table T13). These values are generally consistent with those determined in the same interval of the archive-half cores. As noted for whole-round core measurements (see “Physical properties”), magnetic susceptibility values also span several orders of magnitude (3.4 × 10^–4 to 0.69 SI). The Königsberger ratio (Qn; the ratio of remanent magnetization to induced magnetization) provides one estimate of the stability of the remanent magnetization (Fig. F64). Qn values in the basalt samples are typically well above 1, indicating that the magnetization is dominated by the remanence. In contrast, Qn values of hyaloclastite breccia samples are as low as ~0.01, which indicates that the dominance of induced magnetization provides a plausible explanation for the unstable measured magnetization and that likely contributes as well to the apparently consistent directions in individual archive-half core pieces.

The majority of the discrete samples (63) were subjected to stepwise AF demagnetization. The other 37 samples were subjected to stepwise thermal demagnetization, and in order to monitor thermal alteration during heating, magnetic susceptibility measurements were made after every heating step.

Most basalt samples exhibited high-quality demagnetization behavior, with similar results obtained from adjacent samples subjected to either AF or thermal demagnetization (Fig. F65A, F65B). In contrast, many hyaloclastite breccia samples exhibited less ideal behavior, with immediately adjacent AF and thermally demagnetized samples often having discrepant results and in at least one case the opposite polarity (Fig. F65C). In hyaloclastite samples with large centimeter-sized clasts, the results were equally variable. Some samples had multicomponent remanence and others univectorial remanence (Fig. F65D, F65E). Note that the highest stability component in both of these samples has a moderate negative inclination. Shore-based studies will be required to assess whether this apparent agreement with the inclinations from (bracketing) lava flows at the site might represent a geologically meaningful remanent magnetization acquired at or shortly after deposition.

Thermal demagnetization also reveals the presence of a range of magnetic minerals (Fig. F65F). Many basalt samples, particularly those from the upper five stratigraphic units, are characterized by dominant unblocking near 575°C (near the 580°C Curie point of pure magnetite) and a generally co-linear component unblocking above 600°C, consistent with remanent magnetization carried by (titano)hematite. Other samples, particularly those from deeper in the hole and characterized by reducing conditions (see “Alteration petrology”), have dominant or substantial unblocking by 200°–300°C. These lower unblocking temperatures are consistent with the presence of high-Ti titanomagnetites (O’Reilly, 1984) that have experienced little deuteric (or high-temperature) oxidation.

Anisotropy of magnetic susceptibility

The anisotropy of magnetic susceptibility was measured on all discrete samples (Table T14). Hyaloclastite samples from Units XII and XV and the vitric-lithic sand of Unit XIII were all statistically isotropic. The remaining samples have dominantly (60%) oblate magnetic fabrics, although shape factors (T, where [prolate] –1 < T < 1 [oblate]; Jelinek, 1981) range from +0.96 to –0.75. With two exceptions, the samples are only weakly anisotropic, with a degree of anisotropy (P′) value of <1.05 (Jelinek, 1981). The two most anisotropic samples (both from Unit XVII Section 330-U1372A-38R-3) have substantially higher P′ values (1.17 and 1.21) and prolate fabrics.

Although scattered, the minimum eigenvectors are generally steeply plunging, and the maximum eigenvectors are subhorizontal (Fig. F66), a combination that might be expected for approximately horizontal lava flows. The eigenvectors in this plot were rotated about a vertical axis by the angle necessary to restore the ChRM declination for each sample to 0°, providing a nominal common orientation for all samples. This reorientation did not result in any pronounced clustering of the maximum eigenvectors. Interestingly, the maximum eigenvectors of the two most anisotropic samples have an average orientation of 214°/10° in this coordinate system, consistent with downslope transport from the center of the guyot.

Discussion

Documentation of paleolatitude requires a sufficient number of flow units to provide a robust estimate of the time-averaged geomagnetic field at Site U1372. In addition, these units must also be in situ or, alternatively, the effects of later tilting/reorientation must be quantifiable from independent information. The presence of volcaniclastic material and the limited core recovery, particularly from the lower part of the hole, make it challenging to determine whether individual recovered basalt fragments are in situ. The petrology group defined 81 lithologic units and provided a qualitative assessment of whether each unit is an in situ cooling unit by determining its in situ confidence index (ISCI; see “Igneous petrology and volcanology”). We used these ISCI confidence scores to provide a graphical representation of confidence in the inclination data, with the darkest (blue) colors reflecting the highest confidence (Fig. F67). The shipboard archive-half data provide the most spatially complete representation of downhole variations in inclination. The inclination-only statistical technique of Arason and Levi (2010) was then used to calculate the mean inclinations and associated 95% confidence angle (α₉₅) for all 2 cm PCA directions from a single lithologic unit and all Fisher piece-average inclinations for individual pieces and for each single unit (Fig. F67C–F67E; Table T15; see “Paleomagnetism” in the “Methods” chapter [Expedition 330 Scientists, 2012]).

Overall, the archive-half core data agree very well with the more robust inclination estimates from demagnetization of discrete samples (Fig. F67). The samples that are most likely to be in situ occur in Units III–VIII and at the base of the hole (Unit XVII). These units all yield consistent steep negative inclinations (ranging from about –45° to –75°), with the exception of a narrow interval in Core 330-U1372A-11R (see below). The intervening volcaniclastic-dominated units show considerable scatter, although a number of steep negative inclinations were documented from basalt in this interval as well. Shore-based studies may provide additional constraints on whether these units in fact may be interpreted as intercalated in situ lava flows or pillows.

Results from Sections 330-U1372A-11R-1 and 11R-2 at the bottom of Unit V (Fig. F68) provide tentative evidence for reversed polarity magnetizations in the dominant normal polarity record of Hole U1372A. The lower boundary of this reversed polarity interval coincides with the boundary between lithologic Units 10 and 11. The upper boundary appears to occur in an unrecovered interval at 65 cm in Section 11R-1 (i.e., between Piece 1 and Piece 2). Discrete sample thermal demagnetization data confirm the normal polarity above and below the reversed polarity zone. A single discrete sample (330-U1372A-11R-1, 84–86 cm; green triangle in Figure F68) from the reversed polarity zone was AF demagnetized but had low coercivity and did not yield a characteristic direction. The bimodal distribution of inclinations in lithologic Unit 10 required calculation of separate normal and reversed polarity averages (Fig. F67; Table T15). This interval will be the subject of detailed shore-based studies.

Summary

The archive-half core data exhibit a wide range of inclinations that are mostly attributable to inclusion of data from conglomerates, breccia, and hyaloclastite units (Fig. F69A). However, the dominant direction is a steeply plunging negative (normal polarity) inclination recurring in multiple in situ lava flow units. When considering only the distinct lava flow units defined with the highest confidence by the igneous petrology group (ISCI of 3 or 2), the range of inclinations becomes focused toward a steeper average inclination (Fig. F69B). Finally, the most confidence can be placed in discrete sample data taken from the most confidently defined in situ flow units. This selection returns the steepest average inclination values of all data sets (Fig. F69C). The arithmetic and inclination-only means of this distribution of inclinations agree within error. The removal of data corresponding to lava flow units with a lower ISCI of 2 has little effect on the calculations. The geocentric axial dipole inclination value expected from the present location of the Louisville hotspot at 51°S is ±68°. This value is similar to the mean inclination obtained from Hole U1372A samples within the uncertainty limits. However, the limited number of in situ lava flow units recovered suggests that caution is warranted in estimating a paleolatitude.

Top of page   |    Previous   |    Next