Previous   |    Next

doi:10.2204/iodp.proc.330.105.2012

Paleomagnetism

The deep penetration and high core recovery from Hole U1374A provided a detailed record of magnetization variations in Rigil Guyot. The highest remanent magnetization intensities are associated with lava flows, lava lobes, and intrusive sheets, but the volumetrically dominant volcanic breccia has only slightly lower intensities. The uppermost ~45 m of Hole U1374A is characterized by reversed polarity, whereas predominantly steep negative inclinations (normal polarity) occur in the interval between ~45 and 522 mbsf. Remarkably consistent normal polarity magnetizations were observed for intervals containing a wide range of lithologies, including lava flows, dikes, and volcanic breccia. Both archive-half and discrete sample data reveal steep inclinations, similar to or greater than that expected from a geocentric axial dipolar field at the current location of the Louisville hotspot at 51°S.

Archive-half core remanent magnetization data

The remanent magnetization of archive halves from Sections 330-U1374A-1R-6 through 73R-1 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.

The natural remanent magnetization (NRM) intensity varies by more than four orders of magnitude (Fig. F73B) from a maximum of ~19 A/m (associated with intrusive sheets in Unit XVIII) to a minimum of ~10^–3 A/m (in Unit II sediments). Several lithologic units and some stratigraphic units are broadly associated with changes in NRM intensity or magnetic susceptibility (Fig. F73B, F73C). For example, many of the isolated intrusive sheets and lava flows in the dominantly volcaniclastic sequence are accompanied by an increase in both NRM intensity and magnetic susceptibility (e.g., the lava of lithologic Unit 2 at ~20–30 mbsf and the dikes of lithologic Unit 103 at ~328 mbsf). The enhanced NRM intensity of stratigraphic Unit XV relative to that of surrounding units and the increase in magnetization toward its base are apparently responsible for a distinctive magnetic anomaly signature in the borehole at this depth (see “Downhole logging”).

The deep penetration and exceptionally high recovery at Hole U1374A resulted in a substantial number of archive-half measurements that, in turn, allow some general assessment of the NRM intensity for many of the recovered lithologies (Fig. F73B). Units identified as intrusive sheets have the highest NRM intensities (geometric mean = 5.5 A/m ± 0.3 log units). Lava flows and lava lobes have geometric mean intensities of 2.1 A/m ± 0.5 log units and 3.3 A/m ± 0.3 log units, respectively. Slightly lower geometric mean magnetization values characterize volcanic/volcaniclastic breccia (0.7 A/m ± 0.5 log units) and sedimentary breccia/conglomerates (0.8 A/m ± 0.5 log units).

Best-fit principal component directions were calculated for each 2 cm interval of the archive halves using an automated routine that maximizes the percentage of remanence incorporated and minimizes the scatter about the best-fit direction and the deviation of this vector from the origin (see “Paleomagnetism” in the “Methods” chapter [Expedition 330 Scientists, 2012a]). On the basis of the misfit value distribution, only directions with misfit values of <3.42 were considered; this represents ~40% of the total number of 2 cm interval principal component directions. Representative demagnetization results from archive-half core intervals with misfit values below and near the cutoff value of 3.42 are shown in Figure F74. The resulting inclinations, intensities, and stability (median destructive field of the vector difference sum [MDF′]) are shown in Figure F73D (dark red circles), F73B (black circles), and F73E.

Inclinations from Hole U1374A are dominantly negative and steep, indicative of Southern Hemisphere normal polarity. This consistent normal polarity signal extends from ~45 mbsf to the base of the hole and is interrupted by intervals with increased scatter. In some cases, this increased scatter is associated with sedimentary breccia and conglomerates (e.g., stratigraphic Units IX and XI). Scattered inclinations, however, are not diagnostic of breccia units because most volcanic breccia also has consistent normal polarity (Fig. F73D). Such consistent inclinations are present in volcanic breccia composed of pillow fragments, clasts with lobate margins, and in some cases clasts with sharp margins (e.g., Sections 330-U1374A-46R-3, 57R-3, and 67R-3, respectively; see “Igneous petrology and volcanology”). A large fraction of directions that were excluded from further consideration on the basis of their calculated misfit values nonetheless have inclinations that are consistent with more reliable (lower misfit) data (cf. Fig. F74D, F74E, F74F, F74H). More moderate negative inclinations were observed in the lowermost 20 m of the hole, corresponding to lithologic Units 143–148. The uppermost ~45 m of the hole, which includes intercalated sediments, lavas, and breccia (Fig. F74A–F74C), is characterized by positive inclinations, indicating reversed polarity.

Discrete sample remanent magnetization data

NRM and magnetic susceptibility were measured for all discrete samples (N = 236; Table T15). NRM intensities for lavas, intrusive sheets, and basalt clasts in volcanic breccia/conglomerate range widely from 0.19 to 23.5 A/m (arithmetic mean = 6.29 A/m), whereas volcanic breccia samples have generally lower values (mean = 1.56 A/m; Fig. F75). Volcanic sandstones from stratigraphic Units II and VII have distinctly lower magnetizations, averaging 0.52 A/m. Discrete sample NRM intensities are generally consistent with the ranges measured for the same lithologies on the archive halves (Fig. F73). The Königsberger ratio (Qn; calculated for a field intensity of 36.6 A/m) for most lavas and intrusive sheets is >10, indicating that the overall magnetization is dominated by remanent rather than induced magnetization. Other lithologies have relatively lower Qn values, and the lowest values are associated with sediments (volcanic sandstone) from the upper portion of the hole (Fig. F75C). Although induced magnetization might complicate interpretation of demagnetization data at sea, more reliable remanent magnetization data can be acquired in the low-field environment of shore-based laboratories.

Results from alternating-field (AF; N = 140) and thermal (N = 96) demagnetization of discrete samples generally reveal relatively simple characteristic remanent magnetization (ChRM) directions (Fig. F76). Both normal and reversed polarity ChRM directions were observed, with reversed polarity generally confined to the uppermost ~45 m of the hole, as noted in the archive-half core data. Most samples show nearly univectorial behavior during demagnetization, with a small to negligible lower stability component (Fig. F76A–F76D, F76F). There is little evidence of any drilling-related remanence, which is consistent with the high MDF′ values for the discrete samples, as shown above for the archive-half core data (Fig. F73E). AF and thermal demagnetization generally recover the same magnetization component. This correspondence also extends to some samples where a small, nearly antipodal lower stability component is present (Fig. F76E).

For a small number of sample pairs (particularly those from the upper portion of the hole), the lower temperature magnetization component agrees well with the ChRM direction determined by AF demagnetization (Fig. F76B), but at higher temperature steps (>400°C) an additional positive inclination component is present. The emergence of this higher temperature component coincides with an order-of-magnitude increase in magnetic susceptibility (monitored after each heating step), and its origin is uncertain. This component is typically poorly defined (maximum angular deviation > 5°) and does not pass our reliability criteria. Shore-based studies will further examine this discrepancy.

Anisotropy of magnetic susceptibility

Anisotropy of magnetic susceptibility was measured for all discrete samples (Table T16). Discrete samples were collected from 10 intervals identified as intrusive sheets, with two adjacent samples taken in most of these intervals. The eigenvectors associated with the minimum eigenvalues for these intrusive sheets are dominantly subhorizontal (Fig. F77A), compatible with flow in steeply inclined dikes (Knight and Walker, 1988). Eigenvectors from adjacent sample pairs in these intrusive sheets typically agree within 5°–10°. For lava flows highly likely to be in situ (i.e., in situ confidence index [ISCI] = 3; see “Igneous petrology and volcanology” in the “Methods” chapter [Expedition 330 Scientists, 2012a]), the eigenvectors associated with the minimum eigenvalues are subvertical, whereas those associated with the maxima are subhorizontal, as might be expected for lava flows (Fig. F77B). Many samples with lower ISCI values also exhibit the same pattern, although with more dispersion.

Discussion

The uppermost portion of Hole U1374A is characterized by reversed polarity, whereas remanent magnetizations with consistent normal polarity were observed for a fairly long stratigraphic interval between ~45 mbsf and the base of the hole at 522 mbsf (Fig. F78). This consistency is especially pronounced in the lower portion of the hole (Cores 330-U1374A-56R through 68R), where numerous volcaniclastic units as well as intercalated lava flows and lobes display less scattered inclinations. In this interval, most individual 2 cm principal component analysis directions as well as piece-average inclinations and discrete sample inclinations are between –60° and –90° (Fig. F78B, F78D, F78F). Most of the intrusive sheets intruding these volcaniclastic materials also have similar steep negative remanence vectors. The lowermost ~20 m (Cores 330-U1374A-68R through 73R) have somewhat shallower negative inclinations, with most values ranging between –45° and –70°.

A small number of positive inclination (reversed polarity) zones are also present in the dominantly normal polarity interval from ~45 mbsf to the base of the hole. Some positive inclinations, particularly where inclinations are scattered, undoubtedly represent clasts with random orientations. The origin of more narrow intervals of positive inclinations (e.g., ~95, ~109, and ~160 mbsf) is uncertain. Some may be related to isolated core pieces that were inverted during curation or measurement (several such inverted core pieces were noted and corrected, but it is possible others were not detected). Others may reflect unstable remanent magnetization that was reset by drilling. Drilling-related overprints are typically removed by low to moderate AF demagnetization (5–20 mT), and therefore the moderate to very high coercivity evident in both the archive halves and many discrete samples suggests that the observed steep inclinations are likely not related to drilling. The use of nonmagnetic core barrels appears to have significantly reduced the drilling-related remanent magnetization in the Expedition 330 cores. However, some positive inclination (reversed polarity) intervals may be geologically significant. For example, a 50 cm interval in the archive-half data for Section 330-U1374A-63R-3 (~445 mbsf) has essentially vertical, positive inclination. A single thermally demagnetized sample in this interval has a comparable final magnetization component (inclination = +89°) with a less stable normal polarity overprint. This and other apparent short polarity intervals will be the subject of postexpedition research.

The archive-half core and discrete sample data from Hole U1374A provide a consistent inclination record, whether the data are treated individually or grouped by lithologic unit. The most reliable (misfit ≤ 3.42; n = 5496) archive-half core data from Hole U1374A (Fig. F78B) are strongly focused toward steep negative inclinations (Fig. F79A). The average inclination remains steep when inclination-only averages are considered (Fig. F78C; Table T17) as calculated for the most likely in situ lithologic units (ISCI = 2 or 3). The resulting inclination-only mean is –73.5° ± 4.3° (α₉₅; n = 19; with the single positive inclination result that corresponds to lithologic Unit 4 inverted) (Fig. F79B). If we consider the Fisher piece-average inclinations for each piece occurring in lithologic units assigned an ISCI value of 2 or 3 (Fig. F78D), then a bimodal distribution is observed (Fig. F79C). Inclination-only statistics (Arason and Levi, 2010) for lithologic units using these piece averages show the same bimodal distribution but with smaller n (Figs. F78E, F79D), resulting in an inclination-only mean of –79.5° ± 7.5° (α₉₅; n = 23). Using the same data set, lithologic Units 1, 2, and 4, which record reversed magnetic polarity, have an inclination-only mean of 57.9° ± 11.4° (α₉₅) that is distinct from that of the overall data set (–79.9° ± 7.5° [α₉₅; n = 23]).

Discrete sample demagnetization data and the manual principal component analysis directions picked for this data subset provide the most reliable inclination estimates (Fig. F79E). Two distinct polarity groups are also seen in these data, and the arithmetic mean inclination (–68.7° ± 13°; n = 50) is equal to that expected for the present-day location of the Louisville hotspot. However, the inclination-only maximum likelihood estimate is –80.2° ± 7.2° (α₉₅), which corresponds to a paleolatitude of 71°. For either conclusion to be validated, further shore-based analyses are required to refine the inclination for each lithologic unit and to test whether secular variation has been adequately averaged. Shore-based research will also include an assessment of whether the data from this site can be combined with paleomagnetic data from Hole U1373A, also drilled on Rigil Guyot.

Top of page   |    Previous   |    Next