Previous   |    Next

doi:10.2204/iodp.proc.345.110.2014

Paleomagnetism

Remanence data

Continuous measurements

Remanence measurements were made at 2 cm intervals on all archive-half core pieces longer than ~9 cm. All archive-half cores were subjected to stepwise alternating field (AF) demagnetization at 5 mT steps up to maximum peak fields of 80 mT. Remanence data and corresponding archive-half core point magnetic susceptibility data were filtered to preserve only data corresponding to the intervals where remanence measurements were made and to discard data obtained within 4.5 cm of piece ends.

For the purpose of characterization based on bulk magnetic parameters, the lithologies recovered from Hole U1415J are grouped into two categories. Group 1 consists of gabbroic rock (olivine-bearing gabbro, olivine gabbro, orthopyroxene-bearing olivine gabbro, gabbronorite, olivine-bearing gabbronorite, and oikocrystic gabbro) with a geometric mean natural remanent magnetization (NRM) intensity of 22.6 mA/m (range = 0.9–893 mA/m; n = 66) (Fig. F87). These values are considerably lower than geometric mean values for gabbroic samples (~1 A/m) reported by Gee and Kent (2007), who compiled NRM intensities (reduced to equatorial latitudes) from a range of sites where generally more evolved gabbros were encountered than during Expedition 345 (Varga et al., 2004; Gee et al., 1992, 1997; Kikawa and Pariso, 1991; Dick, Natland, and Miller, 1999; Kelemen, Kiwawa, Miller, et al., 2004; Cannat, Karson, and Miller, 1995; Pariso et al., 1996; Blackman, Ildefonse, John, Ohara, Miller, MacLeod, and the Expedition 304/305 Scientists, 2006). Group 2 consists of troctolitic rock (troctolite, troctolitic olivine gabbro, and clinopyroxene oikocryst-bearing troctolite) with a geometric mean NRM intensity of 1.00 A/m (range = 109 mA/m to 8.140 A/m; n = 36) (Fig. F87). The geometric mean magnetic susceptibilities of Group 1 and 2 samples are 77.6 × 10^–5 and 471.4 × 10^–5 SI, respectively. Higher mean intensities and susceptibilities in the troctolitic rock most likely reflect variable degrees of serpentinization of these more olivine-rich lithologies leading to the production of secondary magnetite (see “Metamorphic petrology”).

Remanent magnetization directions were calculated by principal component analysis (PCA; Kirschvink, 1980) at all measurement points along the core pieces where linear components could be identified on orthogonal vector plots of demagnetization data. Only principal components with a maximum angular deviation <10° were considered acceptable. Figure F88 shows representative examples of AF demagnetization behavior, and Figure F89 shows downhole variations in NRM and PCA pick inclinations, NRM and PCA pick intensities of magnetization, and low-field magnetic susceptibility measured using the Section Half Multisensor Logger (SHMSL). The majority of samples have an initially downward directed remanence, with evidence of inclination steepening caused by acquisition of a drilling-induced magnetization that is at least partially removed by low-field treatments (<15 mT). This is followed by removal of a moderately inclined, downward-directed component, typically by fields of 25–30 mT (Fig. F88). The mean inclination of this component is 37.5° (k = 20.7; α₉₅ = 3.9°; n = 69), calculated using the Arason and Levi (2010) maximum likelihood method. During removal of this component, remanences typically migrate to the upper hemisphere, but no linear components with negative inclination are present. Instead, at demagnetization fields >30 mT remanence directions migrate back to the lower hemisphere and intensities of magnetization increase continuously up to the peak applied field of 80 mT (Fig. F88D, F88F). This behavior is due to acquisition of spurious, laboratory-imparted, anhysteretic remanent magnetization (ARM) along the z-axis of the superconducting rock magnetometer (SRM) system, which has been a characteristic problem of the SRM system observed during several IODP expeditions (e.g., Teagle, Ildefonse, Blum, and the Expedition 335 Scientists, 2012). In rare examples, linear components decaying toward the origin without significant ARM acquisition are successfully isolated (Fig. F88G, F88H), although in some cases this final component is directed along the z-axis of the SRM (Fig. F88I) and must be treated with caution as a potential ARM. Unfortunately, clear examples of demagnetization unaffected by ARM acquisition are limited in number. In the majority of cases, anomalous ARMs in archive-half core samples prevent isolation of sufficient high-coercivity components to allow geological interpretation. The significance of the moderately inclined, downward-directed component with medium coercivity is discussed in “Reliability of linear remanence components in superconducting rock magnetometer data.”

Discrete samples

Shipboard experiments were conducted on 27 discrete minicube samples from Hole U1415J. Two samples were AF demagnetized, one sample was thermally demagnetized, and the remaining 24 samples were subjected to two cycles of low-temperature demagnetization (LTD) followed by thermal demagnetization. In all cases, remanent magnetization directions were calculated by PCA for all demagnetization intervals where linear components could be identified on orthogonal vector plots of demagnetization data. Only principal components with a maximum angular deviation <10° were considered acceptable (Table T10). Two samples displayed erratic demagnetization data. One of these (Sample 345-U1415J-13R-1, 28 cm) came from a piece of white, highly altered troctolite with an exceptionally low NRM intensity (0.5 mA/m) in which intense alteration may have resulted in destruction of remanence-carrying phases. These samples are not discussed further. The remaining samples are divided into three categories according to demagnetization behavior and reliability:

1. Samples with large oikocrysts,

2. Samples with high unblocking temperature single components of magnetization, and

3. Samples with multicomponent remanences.

Samples with large oikocrysts

Five discrete samples were collected in gabbro and troctolite containing large pyroxene oikocrysts. Figure F90 shows orthogonal vector plots and stereographic projections of demagnetization data from two adjacent samples collected from Section 345-U1415J-8R-1. These discrete samples include large oikocrysts in the 8 cm³ minicubes. Sample 8R-1, 80 cm, unblocks across the full range of temperatures (see inset normalized intensity decay curve in Fig. F90) and then displays a high unblocking temperature component that decays to the origin with a declination of 023.2° and inclination of –41.7°. In contrast, Sample 8R-1, 84 cm, has a significant drilling-induced magnetization that is removed by low-temperature demagnetization to leave a steeply inclined component that unblocks close to the magnetite Curie temperature (585°C), with a declination of 206.3° and an inclination of –79.5°. Hence, these two adjacent samples record significantly different magnetization directions, with a solid angle of 59° between PCA components. This is beyond the variability that may be attributed to unresolved secular variation, especially in samples that must have experienced nearly the same cooling history. The source of this variability in this heterogeneous, coarse-grained, oikocryst-bearing rock is uncertain. As a precaution, these data and those from the remaining three oikocrystic discrete samples are excluded from subsequent geological interpretation.

Samples with high unblocking temperature single components of magnetization

Samples from a wide range of lithologies (gabbroic and troctolitic) display well-defined linear remanence components following removal of variably developed drilling-induced magnetizations by low-temperature demagnetization or treatment by fields of <20 mT. A selection of typical examples is shown in the orthogonal vector plots of Figure F91. Some samples display a bend in the thermal demagnetization path (e.g., Samples 345-U1415J-5R-1, 130 cm, and 13R-1, 42 cm) at ~550°C before decaying linearly toward the origin. In such cases, the highest unblocking temperature segment has been picked as the characteristic remanence component.

Unblocking temperature spectra for these samples are remarkably similar to each other (Fig. F92), with ~70% of remanence lost within 40°C of the magnetite Curie temperature. Such discrete unblocking at high temperatures is characteristic of thermoremanent magnetizations carried by fine-grained (single domain and pseudosingle domain) grains of magnetite, and such remanences are unlikely to have been thermally overprinted after thermoremanent magnetization acquisition. Hence, we interpret these components as the primary magnetization in this rock, suitable for subsequent geological interpretation below.

Samples with multicomponent remanences

Three discrete samples in Hole U1415J display complex demagnetization paths that include multiple linear remanence components within each sample and are illustrated in Figure F93. Each exhibits a high-temperature remanence component that unblocks above 500°/520°C (red labels) and an intermediate temperature component of nearly antipodal direction that unblocks between 425°/450°C and 500°/520°C (green labels). In addition, Sample 345-U1415J-23R-1, 22 cm, displays a third component that unblocks between 100° and 350°C (blue label) that is nearly parallel to the highest unblocking temperature component in this sample. These multiple, nearly antipodal components strongly suggest that remanence in these samples was acquired in different geomagnetic polarity periods. However, without independent reorientation of the core piece and recovery of original declinations it is impossible to define the original polarity of each component.

Similar multicomponent, multipolarity remanences have been seen previously in lower crustal rock recovered by drilling in slow-spreading crust along the Mid-Atlantic Ridge. Meurer and Gee (2002) reported three components of different polarities in gabbro from the Mid-Atlantic Ridge Kane Fracture Zone area sampled during ODP Leg 153 and interpreted these as components acquired across the Jaramillo Subchron and Matuyama and Brunhes Chrons during protracted construction of the lower crust by intrusion of thin sills. Morris et al. (2009) reported three component remanences from gabbro recovered from the footwall of Atlantis Massif sampled during IODP Expedition 304/305 and showed that the multicomponent remanences resulted from prolonged cooling of the section across the polarity reversals on either side of the Jaramillo Subchron. In both cases, the data provide constraints on the thermal history of the sampled sections. Alternatively, such remanences might result from successive phases of alteration and acquisition of thermoviscous magnetizations. Clearly, further sampling and detailed thermal demagnetization and rock magnetic experiments are now required in order to understand the distribution, origin, and geological significance of multicomponent remanences within gabbro recovered from Hole U1415J.

Reliability of linear remanence components in superconducting rock magnetometer data

As noted previously, archive-half core data from the SRM show a characteristic, moderately inclined, linear remanence component at AF steps below ~30 mT. This occurs after apparent removal of the drilling-induced remanence and prior to the acquisition of significant anhysteretic remanence at higher fields. In Hole U1415J, this component has a well-defined mean inclination of 37.5° (k = 20.7; α₉₅ = 3.9°; n = 69). It is important, therefore, to determine whether this linear component is a distinct natural component of potential geological significance or whether it represents an artifact of spurious origin.

Figures F94 and F95 provide a comparison of LTD/thermal demagnetization data from discrete samples (measured using the JR6A spinner magnetometer) and the equivalent archive-half core demagnetization data measured using the SRM (within 2 cm of the discrete sample). Only one sample (345-U1415J-8R-3, 112–113 cm; Fig. F94A) shows a good correlation between these data types and is a sample with no significant drilling-induced magnetization or ARM. Such behavior in the archive-half core data is rare (seen at 14 out of 102 measuring points). The remaining samples show a variety of mismatches between archive-half core and discrete sample data. Sample 8R-3, 84 cm, has a single, well-defined remanence component with an upward-directed inclination (Fig. F94B). Equivalent archive-half core data display the common, moderately inclined, downward-directed component, but demagnetization data follow a great circle connecting the NRM direction to the final ARM direction that passes through the discrete sample direction without reaching a stable endpoint. Samples 5R-2, 61 cm (Fig. F94C), and 13R-1, 42 cm (Fig. F94A), have high unblocking temperature components with southeast declinations that are distinctly different from all parts of the equivalent archive-half core demagnetization paths. Sample 18R-1, 144 cm, displays two well-defined components with upward-directed inclinations and northeast declinations (Fig. F95B). The equivalent archive-half core data define a curved path on the lower hemisphere that connects the NRM and ARM directions without passing onto the upper hemisphere. Finally, archive-half core data corresponding to the clearest sample carrying a multicomponent remanence (Sample 23R-1, 22 cm) define a box-shaped demagnetization path on an orthogonal vector plot with no convincing linear segments between the drilling-related and ARM components (Fig. F95C). Hence, in the majority of cases, poor to no correlation exists between well-defined discrete sample magnetization directions and equivalent archive-half core data, which must therefore be treated with extreme caution.

Figure F96A shows a stereographic equal-area projection of the downward-directed linear component (present in the majority of archive-half core demagnetization data) from measurement points in the immediate vicinity of discrete samples. These data come from a range of core pieces that are azimuthally unconstrained in the core reference frame yet show significant clustering toward northwest–northeast declinations. Equivalent discrete sample high unblocking temperature components (Fig. F96B) are widely scattered in declination, which is expected for data from core pieces that are free to rotate in the core barrel with both upward- and downward-directed inclinations. These data further confirm that the downward-directed linear component in the archive-half core samples has no geological significance.

Detailed previous studies of the geometry of drilling-induced remanences in ODP core sections have documented a pronounced radial drilling-induced magnetization component in addition to the dominant subvertical component. In particular, experiments conducted during ODP Leg 206 (Shipboard Scientific Party, 2003) on a suite of subsamples cut from a whole-round basaltic core piece in Hole 1256D demonstrated that drilling-related overprints result in a bias of archive-half core remanences toward northerly declinations in the core reference frame (i.e., toward the center of the core section). The preferred clustering of directions seen in Figure F96A may therefore result from the effects of this drilling-related bias.

Geological interpretation of inclination data from discrete samples

Inclinations of the highest unblocking temperature components in discrete samples from Hole U1415J are illustrated in Figure F97. Samples are presented sequentially rather than by depth scale to allow comparison of data from widely spaced sampling points. Data from oikocrystic rock are excluded. Both upward- and downward-directed inclinations were observed, with inclinations ranging from +55° to –73° (compared to an expected geocentric axial dipole field inclination of ±4.7°). Therefore, these data indicate substantial rotation of the cored section. A clear distinction exists between downward-directed magnetizations in the upper part of Hole U1415J (Cores 345-U1415J-3R through 8R) and upward-directed magnetizations in the lower part (Cores 9R through 23R), although data are insufficient to determine statistically robust mean inclinations. Within these groups, variation in inclination between samples may be due to incomplete averaging of geomagnetic secular variation at the sample level or to the effects of strong anisotropy in sampled lithologies. Sample 345-U1415J-13R-1, 42 cm, is inferred to be anomalous, as structural observations in this section suggest sampling of a block that has fallen in the hole.

With a subhorizontal expected direction of magnetization at Hess Deep, the magnetic polarity of this rock cannot be uniquely determined in the absence of reoriented core samples. However, U-Pb dating of zircons from samples collected in the immediate vicinity of Site U1415 during the JC21 site survey cruise yielded ages of 1.42–1.27 Ma (Rioux et al., 2012). These dates lie in the middle of reversed polarity Chron C1R (Cande and Kent, 1995), suggesting that any primary magnetizations preserved in the sampled rock should be of single, reversed polarity. Hence, coherent tectonic rotation of the sampled section is incompatible with observed upward- and downward-directed remanences. These data strongly suggest, therefore, that at least two displaced blocks with independent rotation histories have been sampled in Hole U1415J. The interval from the top of the hole to ~40 mbsf (Sample 345-U1415J-8R-3, 113 cm) should therefore be treated separately from the interval below 40 mbsf in lithologic and structural syntheses. This grouping is equivalent to lithologic Units II and III. In addition, the anomalously high inclination of Sample 9R-1, 47 cm, and the presence of shallowly inclined, high unblocking temperature components in multicomponent-bearing Samples 23R-1, 22 and 51 cm, potentially indicate that these data also come from blocks that have experienced differential rotation during emplacement to their present orientation (although data are too restricted to confirm this).

Magnetic susceptibility, natural remanence intensity, and Königsberger ratio

In mafic igneous rock, low-field magnetic susceptibility (k) is principally controlled by the volume concentration of magnetite. NRM variability is also controlled by variations in magnetite content but may also be influenced by variability in the magnitude of drilling-induced remanent magnetizations. The relation of NRM intensity and susceptibility is expressed by the Königsberger ratio (Q), which is defined as the ratio of remanent to induced magnetization in a rock (where induced magnetization equals the product of k (SI) and the geomagnetic field strength (A/m). Values of Q > 1 indicate that remanence dominates the magnetization of a rock unit.

Figure F98 shows a log-log plot of NRM intensity against k for archive-half core and discrete samples from Hole U1415J, together with lines of equal Q calculated for a field of 25 A/m. The majority of samples plot above Q = 1 and close to Q = 10. This suggests that parts of the sampled section with higher NRM intensities may contribute a significant fraction to marine magnetic anomalies when in situ. However, caution is required in the interpretation of Q ratios calculated for these samples, as NRM intensities may be artificially increased by drilling-induced magnetization.

Anisotropy of magnetic susceptibility

Anisotropy of magnetic susceptibility (AMS) was measured from 27 discrete samples in Hole U1415J. Each sample was measured three times to check consistency and maintain quality control on any shipboard noise. These three measurements were averaged using a bootstrap approach to determine the average AMS eigenparameters for each discrete sample (Table T11). Eigenvectors and their associated bootstrapped uncertainties are illustrated on the equal-area stereographic projections in Figure F99. Susceptibility tensors are weakly to moderately anisotropic (corrected anisotropy degree [P′] < 1.40; mean = 1.13) (Jelinek, 1981). A majority of ellipsoid shapes are oblate, although all three shapes (triaxial, oblate, and prolate) are also represented in these samples (Fig. F99C). Bulk susceptibilities range from 1.26 × 10^–4 to 5.87 × 10^–2 SI, with an average of 7.35 × 10^–3 SI that indicates predominance of ferromagnetic mineral contributions to the AMS signal in most samples.

Five samples have large (>25°) bootstrapped confidence ellipses around their eigenvectors, suggesting weak signals and inconsistent repeated measurements from the same sample. These five samples (345-U1415J-3R-1, 26 cm; 8R-3, 68 cm; 12R-1, 109 cm; 21R-1, 57 cm; and 21R-1, 120 cm) have an average bulk susceptibility of 2.40 × 10^–4 SI, suggesting that some paramagnetic minerals may contribute to the signal. These samples are not shown in Figures F99A and F99B, and they are also disregarded from further geological interpretation until the AMS source can be better determined.

Structural measurements of magmatic foliation in the upper portion of Hole U1415J show close agreement with the AMS fabrics (minimum eigenvector [k_min] near the magmatic foliation pole) and indicate well-developed fabrics that consistently dip eastward within the core reference frame (Fig. F99A). However, intersample fabric consistency is not evident deeper in Hole U1415J (below ~45.66 mbsf), and a wider range of eigenvector directions is illustrated in Figure F99B. Overall, AMS magnetic foliation is nearly parallel to the magmatic foliation (average solid angle between k_min and magmatic foliation pole <30°), with only one sample (345-U1415J-18R-1, 144 cm) showing poor agreement with the structural fabric (solid angle = 66°).

Top of page   |    Previous   |    Next