Previous   |    Next

doi:10.2204/iodp.proc.335.103.2012

Paleomagnetism

The primary goal of paleomagnetic studies is to quantify the structure, direction, and intensity of natural remanent magnetizations (NRMs) in the different rock types that make up the lower oceanic crust and to use these data to provide insights into crustal accretion, deformation, and the source of marine magnetic anomalies. In addition, magnetic fabric analysis can provide valuable information on the degree and direction of weak preferred alignments of mineral phases, leading to improved understanding of magmatic and tectonic processes. Because of low recovery during Expedition 335, the majority of data discussed here were acquired from analyses conducted on samples taken from Expedition 312 section halves (predominantly from Gabbro 1 and Gabbro 2 lithostratigraphic units). Discrete samples prepared from these legacy core section halves during Expedition 335 are curated with the prefix 335(312).

Remanence data

Continuous measurements

The lack of significant recovery of archive section halves resulted in an exceptionally limited number of continuous remanence measurements during Expedition 335.

In order to assess the effectiveness of the 2G superconducting rock magnetometer (SRM) inline demagnetizing coils in recovering characteristic remanent magnetizations (ChRMs) from core sections affected by drilling-induced magnetization, prior to analysis of Expedition 335 cores a number of archive section halves from Expedition 312 were subjected to stepwise alternating field (AF) demagnetization from 40 mT up to a maximum peak field of 80 mT. These sections had previously been subjected to stepwise AF treatment up to 40 mT during Expedition 312, but no stable endpoint component was reached by this level of demagnetization. A typical result of this two-stage demagnetization process is illustrated in Figure F77A. The initial treatment to 40 mT was effective in removing most of the steeply inclined drilling-induced remanent magnetization (DIRM) in the sample, resulting in a great circle path trending toward shallower inclinations. However, treatment at higher fields reverses this trend and the remanence migrates to a near-vertical direction by 80 mT. To confirm that this was not the result of incorrect processing and amalgamation of data collected during different expeditions by different operators, the corresponding (previously undemagnetized) working section half was subjected to stepwise AF treatment from 0 to 80 mT. Results are shown in Figure F77B and again exhibit a hairpin-shaped trajectory that migrates to vertical inclinations at high fields.

These experiments demonstrate that the SRM inline AF demagnetizing coils generate a spurious, strong anhysteretic remanent magnetization along the z-axis of the system at fields >40–50 mT (which is often insufficient to isolate non-DIRM components in hard rock samples). This may result from the presence of a residual magnetic field directed along the z-axis in the region of the coils as a result of the less effective magnetic shielding along this direction. However, the measured field at the position of the z-coil was only ~30 nT during these experiments. Alternatively, it may indicate a problem with the AF waveform produced by this coil. Until this instrumental problem is resolved, demagnetization data from the SRM system must be treated with caution.

Because only section half pieces >10 cm in length may be reliably measured using the SRM system because of artifacts arising from edge effects within 5 cm of the ends of core pieces (see “Paleomagnetism” in the “Methods” chapter [Expedition 335 Scientists, 2012b]), only one archive section half piece recovered during Expedition 335 was suitable for analysis. Results are shown in Figure F78 (note that demagnetization data above 60 mT are omitted, as they are contaminated by instrument-induced anhysteretic remanence). This sample of granoblastic dikes from Dike Screen 2 near the current base of Hole 1256D does not exhibit the near-vertical DIRM component ubiquitous in previous recovered samples (see “Discrete samples” below), suggesting that a change in operating procedure since Expedition 312 (most likely the use of nonmagnetic core barrels) may have reduced DIRM acquisition. The ChRM component of this sample has an inclination of +30°. Although data from a single sample cannot be interpreted with any confidence, we note that this inclination is indistinguishable from the overall mean inclination calculated from discrete samples downhole from the top of Gabbro 1 (see “Discrete samples,” below).

Discrete samples

Given the very limited recovery of oriented core pieces during Expedition 335, shipboard experiments on discrete samples were conducted principally on a suite of 11 minicube samples cut from Expedition 312 working section halves within Gabbro 1 and Gabbro 2. These samples were subject to AF or thermal demagnetization, with six samples being pretreated with two runs of low-temperature demagnetization (LTD) in order to remove a significant proportion of the ubiquitous DIRM encountered in all previous drilling phases at the site. These shipboard data were augmented by unpublished shore-based data provided by R. Anma and D. Wilson derived from demagnetization of discrete samples performed subsequent to Expedition 312. A total of 53 samples measured by R. Anma were treated with standard AF demagnetization, with anisotropy of magnetic susceptibility determined for a subset of 22 samples. A total of 19 samples measured by D. Wilson were split into 2 subsamples: one half minicube was subject to AF demagnetization, the other to thermal demagnetization (following initial AF demagnetization to a maximum field of 10 mT to reduce DIRM). Data from the AF demagnetized half minicubes show evidence of laboratory-induced anhysteretic remanence prior to isolation of the characteristic remanence and are not discussed further. Principal component analysis of data from the combined suite of 83 samples was conducted during Expedition 335 to provide consistency and insights into remanence structure and directions through Gabbro 1, Dike Screen 1, and Gabbro 2. However, only demagnetization data from the 11 specimens measured during Expedition 335 are presented in detail here (Figs. F79, F80; Tables T8, T9), although results of principal component analysis of the combined sample set are represented graphically in summary figures.

Orthogonal vector plots of demagnetization data (Figs. F79, F80) reveal a range of magnetic components in samples from Gabbro 1 and Gabbro 2. A low-stability, steeply inclined component is evident in all samples and is interpreted as a DIRM. This component dominates the vector difference sum (VDS; i.e., the sum of the vectors removed at each demagnetization step, representing the total magnetization present in a sample; Tauxe, 2010; Gee et al., 1993), contributing on average 66% of the VDS (with a range of 18%–91%). The dominance of the DIRM component results in a distribution of NRM directions for all samples that is clustered around steep inclinations (Figs. F81, F82), whereas principal component analysis demonstrates that the DIRM is tightly clustered around near-vertical inclinations.

Six samples were subjected to two cycles of LTD (Merrill, 1970; Dunlop, 2003; Yu et al., 2003) to determine whether the DIRM could be successfully reduced by this treatment prior to subsequent AF or thermal demagnetization. The first cycle of LTD was found to remove (on average) 52% of the NRM, with the component removed having inclinations ranging from +78° to +86° (Table T8). Hence, a single cycle of LTD is demonstrably successful in removing a significant proportion of the drilling-induced magnetization in these rocks. The second LTD cycle only removed an additional 5% of the NRM (on average). Data from adjacent samples in Section 335(312)-1256D-222R-2W in igneous Unit 89 in Gabbro 1 (Fig. F80A) illustrate improvements in the resolution of low unblocking temperature components resulting from LTD. The sample at 44 cm in this section was not pretreated with LTD cycles and exhibits a steep but nonlinear demagnetization path until a ChRM component of moderate inclination is isolated between 560° and 600°C. In contrast, the sample at 47 cm exhibits a large, vertical DIRM component that is removed by two LTD cycles and thermal demagnetization at 100°C. Magnetization then jumps to a shallower inclination by 200°C and thereafter exhibits a ChRM that is linear between 450° and 600°C. LTD treatment has therefore improved separation of the DIRM and ChRM in these samples. The jump in magnetization direction between 100° and 200°C in Sample 335(312)-1256D-222R-2W, 47 cm, suggests that a non-DIRM-related, low-unblocking temperature component may be present in these rocks (possibly representing a thermoviscous component acquired in the present-day field at elevated temperatures). Further experiments with more closely spaced temperature steps are required to better define this component.

Following removal of the DIRM component, principal component analysis identifies linear components (with maximum angular deviations < 10°) trending toward the origin in 63 of 83 samples analyzed. These components are typically isolated above 35 mT or 540°C and are considered to represent the ChRM of the samples. Maximum unblocking temperatures of 580°–600°C are consistent with remanence being carried by magnetite. The dispersion in ChRM declinations (Fig. F81C) reflects the lack of azimuthal control on the orientation of IODP cores. ChRMs have moderate inclinations (Figs. F81C, F82) with a mean of +30.7° (α₉₅ = 3.3°, k = 30.0, and n = 63; Arason and Levi, 2010). There is no significant difference between the mean inclinations for Gabbro 1 and Gabbro 2 (+31.6° and 29.9°, respectively), although sample dispersion is slightly higher in Gabbro 1 (k = 17.9, α₉₅ = 6.5°, and n = 29 and k = 59.3, α₉₅ = 3.7°, and n = 26, respectively). ChRM inclinations for the granoblastic rocks of Dike Screen 1 are indistinguishable from the overall mean of the section, but the number of samples in this lithostratigraphic unit marked by low core recovery is too small to allow calculation of a separate, statistically meaningful mean inclination.

The overall mean inclination of the section (Fig. F82) is significantly steeper than that expected for the paleoposition of Site 1256, which restores to an equatorial paleolatitude in the Miocene (Wilson, Teagle, Acton, et al., 2003). Similar anomalous inclinations were observed in shipboard analyses during Expedition 312 (Teagle, Alt, Umino, Miyashita, Banerjee, Wilson, and the Expedition 309/312 Scientists, 2006). Potential explanations for the origin of the moderately inclined observed inclinations are discussed below.

Multicomponent remanence in gabbroic rocks from Hole 1256D

In addition to DIRM and ChRM components, thermal demagnetization of Sample 335(312)-1256D-231R-1W, 115 cm, provides evidence for acquisition of remanence during at least two geomagnetic polarity chrons. This sample exhibits a high-stability remanence component with a positive inclination that is unblocked from 500° to 600°C and forms a linear ChRM above 550°C (Fig. F80B). In addition, this sample shows a well-defined negatively inclined component removed at temperatures from 200° to 450°C. These components are nearly antipodal, strongly suggesting that remanence 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. In addition, a significant jump in remanence direction between 100° and 200°C may indicate the presence of a third low-unblocking temperature positive inclination component, although more closely spaced demagnetization steps would be required to test this inference. Similar low-unblocking temperature components may also be present in Samples 335(312)-1256D-222R-2W, 47 cm, and 231R-3W, 114 cm (Fig. F80).

Sample 335(312)-1256D-231R-1W, 115 cm, represents the first multicomponent (excluding DIRM overprints) gabbroic sample reported in Hole 1256D. Similar multicomponent, multipolarity remanences have been seen previously in lower crustal rocks recovered by drilling in slow spreading rate crust along the Mid-Atlantic Ridge. Meurer and Gee (2002) reported three components of different polarities in gabbros from the Mid-Atlantic Ridge Kane Fracture Zone (MARK) area sampled during ODP Leg 153 and interpreted these as components acquired across the Jaramillo Subchron and the 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 gabbros recovered from the footwall of Atlantis Massif sampled during IODP Expedition 304/305 and showed that these resulted from prolonged cooling of the section across the polarity reversals either side of the Jaramillo Subchron. In both cases, data provide constraints on the thermal history of the sampled sections. 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 Hole 1256D gabbros.

Origin of moderately inclined remanence components

Characteristic components of magnetization throughout the studied interval of Hole 1256D have inclinations that are significantly steeper than the expected subhorizontal direction, with an overall mean inclination of +30.7°. The cause of these anomalous inclinations is difficult to establish with certainty, but a number of potential contributing factors are discussed and assessed below.

Drilling-induced artifacts

As noted previously, the majority of samples analyzed have a significant, near-vertical DIRM component. Principal component analysis identifies a clear linear component decaying to the origin in most samples once this DIRM is removed by low-field/low-temperature demagnetization. However, these final ChRM components potentially could be biased toward positive inclinations if a residual DIRM persists to higher treatment levels. This explanation has previously been invoked to explain anomalously steep inclinations in Hole 1256D (Teagle, Alt, Umino, Miyashita, Banerjee, Wilson, and the Expedition 309/312 Scientists, 2006). Simple forward modeling of superimposed natural and drilling-induced remanence components indicates that significant steepening of linear ChRM components requires near-complete overlap of the high treatment level parts of the coercivity/unblocking temperature spectra of the natural and drilling-induced components. The degree of bias also depends on the relative intensities of DIRM and ChRM. Assuming that the DIRM is carried by multidomain magnetite grains and the stable ChRM by finer, (pseudo)single domain grains (Allerton et al., 1995), overlap sufficient to produce linear ChRMs biased by +30° is unlikely.

Several additional aspects of the paleomagnetic data set analyzed during Expedition 335 also suggest that contamination by DIRM cannot fully explain the observed anomalous inclinations:

• There is no relationship between the strength of DIRM and ChRM inclination, as illustrated by plotting the ratio of the intensity of the DIRM and the vector difference sum of the demagnetization path (DIRM/VDS) against the inclination of the ChRM for discrete samples from core sections from the top of Gabbro 1 to the bottom of Hole 1256D (Fig. F83). High values of DIRM/VDS indicate a remanence dominated by the drilling-induced component. No significant correlation is seen, and indeed the five lowest ChRM inclinations occur in samples where the DIRM forms >65% of the VDS. There is a similar lack of correlation of ChRM inclination with low-field magnetic susceptibility, a parameter that directly reflects the concentration of coarse, multidomain magnetite grains that are more susceptible to acquiring a DIRM.

• Nearly antipodal components of magnetizations are observed in Sample 335(312)-1256D-231R-1W, 115 cm. Significant contamination of the remanence in this sample by a persistent downward-directed DIRM overprint should steepen the downward-directed natural component and pull the upward-directed component to a shallower inclination. The difference in inclinations between these components (Table T8) does not match this scenario and cannot be explained by a DIRM-related bias.

• The single result obtained from an archive section half recovered during Expedition 335 using a nonmagnetic core barrel does not exhibit a subvertical DIRM component (Fig. F78) and yet has a ChRM of 30° (i.e., indistinguishable from the overall mean inclination calculated from discrete samples).

• Detailed previous studies of the geometry of DIRMs in ODP core sections have documented a pronounced radial DIRM component in addition to the dominant subvertical component. In particular, experiments conducted during Leg 206 on a suite of samples cut from a whole-round basaltic core piece from Hole 1256D demonstrated that DIRM results in a bias of working section half remanences toward southerly declinations in the core reference frame (i.e., toward the center of the core section). This bias is not observed in the distribution of ChRM components in discrete samples cut from working section halves (Fig. F81C), where (if anything) there is a predominance of ChRMs in the northern hemisphere.

In conclusion, contamination of ChRM components by residual DIRM extending to high demagnetization levels may partially contribute to anomalously steep inclinations but seems unlikely to fully account for the total apparent steepening of 30° inferred from comparison of observed inclinations with a geocentric axial dipolar reference inclination.

Present-day overprint

The International Geomagnetic Reference Field (IGRF) at Site 1256 (calculated using the IGRF.py program of Tauxe, 2010) is as follows: declination = 004°, inclination = 30.9°, and intensity = 33,666 nT. The IGRF inclination is close to the mean inclination of the section. However, the presence of multipolarity remanence in Sample 335(312)-1256D-231R-1W, 115 cm, is difficult to reconcile with a complete overprint by a present-day field component in these rocks. In addition, the estimated ambient temperature of the section prior to drilling (even if held constant for 15 m.y.) would result in maximum laboratory unblocking temperatures of only ~250°–300°C for ideal magnetite-hosted remanence (Tauxe, 2010), compared to maximum observed unblocking temperatures of >540°.

Nondipole fields/inadequate sampling of paleosecular variation

Inclinations are anomalous compared to an expected direction calculated from a geocentric axial dipolar (GAD) field. Modeling of geomagnetic field data covering the last 5 m.y. demonstrates that nondipolar field contributions may be persistent on geological timescales (e.g., Gubbins and Kelly, 1993; Kelly and Gubbins, 1997). Additional information on nondipole components may be derived from analysis of the skewness of marine magnetic anomalies (Gee and Kent, 2007). Results for the Galapagos Ridge region for the last 1 m.y. (Schneider, 1988) suggest an inclination anomaly of –3.4° (deviation from a GAD value). It is not known whether an anomalous geometry of the geomagnetic field persisted over the equatorial Pacific region at 15 Ma, but in any case it would be unlikely to result in more than a few degrees deviation in inclination from the GAD value.

An alternative source of anomalous inclination would be inadequate sampling of paleosecular variation (PSV) of the geomagnetic field by the suite of discrete samples. Statistical field models allow prediction of the variability of field directions due to PSV at Site 1256. Figure F84 shows the distribution of vector endpoints calculated from 1000 realizations of the statistical field model TK03.GAD of Tauxe and Kent (2004), calculated for a normal polarity field only for clarity. At equatorial latitudes there is a pronounced north–south elongation in the distribution of field directions, most evident when data are presented after rotation of the GAD direction to the vertical. Interpretation of observed paleomagnetic data by comparison with GAD reference directions relies on sufficient sampling of PSV, assuming that the long-term geomagnetic field has a GAD geometry. This is typically achieved by combining data from rocks that acquired their remanence over time periods on the order of 10⁵ y or more. It is generally assumed that cooling rates in lower crustal gabbros are sufficiently slow to adequately average PSV (Gee and Kent, 2007), and this is clearly the case in intervals containing multipolarity remanences. Hence, anomalous inclinations are unlikely to result from systematic undersampling of PSV.

Tectonic rotation

Tectonic tilting of the section after remanence acquisition is capable of producing significant changes in magnetization direction, but it is critically dependent on the orientation of the tilt axis relative to the initial remanence direction. Ridge-parallel tilt axes are likely to dominate at a spreading axis. Reconstruction of the geometry of the East Pacific Rise in the Miocene suggests a ridge orientation of 340°, with Site 1256 located very close to the Equator (Wilson, Teagle, Acton, et al., 2003). The effect of rotation around this axis on originally horizontal remanences of normal and reversed polarity is illustrated in Figure F85. This geometry would require unrealistic amounts of tilting to account for substantial changes in inclination but may contribute part of the observed inclination steepening. The orientation of sheeted dikes in Hole 1256D has been determined independently by Tominaga et al. (2009). The average dip and dip direction of inferred dike contacts observed on FMS images is 79° ± 8° and 053° ± 23°, suggesting a maximum permissible tilt of ~20° (down to the southwest), assuming that dikes were emplaced vertically. Tilting of this magnitude around a ridge-parallel axis would rotate initially horizontal reversed/normal polarity remanences to inclinations of ±5°–10°, respectively. Rotation around alternative axes at higher angles to the initial remanence declination would be more effective at producing the observed moderate inclinations but are unlikely in this tectonic setting.

Remanence deflection due to significant anisotropy

Pronounced magnetic anisotropy in a rock (resulting from a preferred orientation of minerals) can produce a deflection of remanences away from the ambient geomagnetic field direction at the time of magnetization (Stephenson et al., 1986; Potter, 2004). Anisotropy of magnetic susceptibility (AMS) fabrics in the sampled rocks are quite weak (1%–3% anisotropy; see “Magnetic fabric”). However, AMS includes contributions from both remanence-carrying and nonremanence-carrying minerals. A given AMS fabric may not, therefore, accurately reflect the orientation distribution of the remanence-carrying phases in the rock and hence provides a poor measure of the potential amount of remanence deflection. This can only be effectively quantified using the anisotropy of remanence (usually by determination of the anisotropy of anhysteretic remanence [AARM]; Potter, 2004). The degree of remanence anisotropy is usually much more pronounced than that of AMS (Stephenson et al., 1986). AARM fabrics in lower crustal gabbros from Atlantis Massif have been shown to potentially result in up to ~15° of remanence deflection in some samples (J.S. Gee, pers. comm., 2007).

Reoriented AMS ellipsoids in gabbros in Hole 1256D have maximum principal axes that broadly trend north–south (see “Magnetic fabric”). Hence, if remanence anisotropy and AMS ellipsoids are similarly oriented, significant deflection of remanence away from the ambient geomagnetic field direction is potentially possible and may partially account for observed anomalously steep inclinations. AARM analysis will therefore form a focus of postcruise research on these rocks.

Magnetic susceptibility, NRM intensity, and Königsberger ratio

In mafic igneous rocks, low-field magnetic susceptibility (k) is principally controlled by the volume concentration of magnetite. Discrete samples from Gabbro 1, Dike Screen 1, and Gabbro 2 have a mean susceptibility of 39.7 × 10^–3 SI (range = 5.8 × 10^–3 to 116 × 10^–3 SI) and a mean NRM intensity of 5.7 A/m (range = 4.4–15.5 A/m). NRM variability is also controlled by variations in magnetite content but may be influenced by variability in the magnitude of DIRM. 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 in A/m). Values of Q > 1 indicate that remanence dominates the magnetization of a rock unit.

Figure F86 shows a log-log plot of NRM intensity against k, together with lines of equal Q calculated for a field of 27 A/m. The majority of samples plot close to Q = 10. This, combined with high NRM intensities, suggests that this section may contribute a significant fraction to observed magnetic anomalies (as noted previously by Teagle, Alt, Umino, Miyashita, Banerjee, Wilson, and the Expedition 309/312 Scientists, 2006). 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. A full discussion of modeling of marine magnetic anomalies in the region around Site 1256 is provided by Wilson (1996).

Magnetic fabric

AMS was determined for all discrete samples prepared during Expedition 335, and results are summarized in Table T9. Additional data were provided by R. Anma from analysis of discrete samples requested after Expedition 312. In both cases, measurements were conducted using an AGICO Kappabridge, ensuring consistency within the data. Combined results for Gabbro 1 and Gabbro 2 are summarized in Figure F87. Most susceptibility tensors are weakly to moderately anisotropic (P′ < 1.09 [mean = 1.03], where P′ is the corrected anisotropy degree) (Jelinek, 1981). There is a range of ellipsoid shapes from strongly oblate to strongly prolate. In the core reference frame there is no coherent arrangement of maximum and minimum principal axes, reflecting the lack of primary azimuthal control on the orientation of core pieces. In order to compare the orientations of magnetic fabrics from different samples, some common reference frame is required. Under the assumption that the stable ChRM for each sample approximates the time-averaged geomagnetic field direction at the site at the time of accretion, magnetic fabric data have been restored to a common reference frame by a simple vertical axis rotation that restores the ChRM declination to 000°. This correction is not dependent on the indeterminate polarity of the ChRM, as AMS principal axes are bidirectional, but does ignore the natural variability in remanence directions that may arise from secular variation. After correction, AMS principal axes become more coherently organized, with maximum axes forming two north–south oriented clusters and minimum axes forming a girdle distribution that trends east–west. Separating samples with prolate and oblate fabrics results in improved clustering of maximum/minimum axes for prolate/oblate ellipsoids, respectively.

AMS maximum axes represent the magnetic lineation, and their preferred orientation is normally interpreted as a proxy for magmatic flow/emplacement directions in igneous rocks (Tarling and Hrouda, 1993). The north–south alignment of maximum axes in gabbros of Hole 1256D is broadly aligned with the presumed orientation of the East Pacific Rise and suggests a component of ridge-parallel fabric development during crystallization of these rocks. Similar ridge-parallel magnetic lineations have been reported in lower crustal gabbros of the Troodos ophiolite by Abelson et al. (2001), where it is inferred to reflect redistribution of melt toward a spreading segment end.

Top of page   |    Previous   |    Next