IODP Proceedings    Volume contents     Search
iodp logo

doi:10.2204/iodp.proc.304305.103.2006

Paleomagnetism

Basaltic, diabasic, gabbroic, and troctolitic rocks and minor serpentinized peridotite were recovered from Site U1309 (see “Igneous petrology”). Holes U1309B and U1309D penetrated to ~100 and ~1415 mbsf, respectively, and recovery was sufficient that both holes were well suited for continuous measurements of archive halves. Approximately 1700 core pieces yielded well-defined remanence directions, with ~1400 of these directions defined by averages of multiple measurements. In addition, remanence measurements were made on 35 discrete samples from Hole U1309B and 477 discrete samples from Hole U1309D. During Expedition 304, the combined remanence data from archive halves and discrete samples were used to estimate average declinations (for ~100 core pieces from Hole U1309B and 400 pieces from the upper 400 m of Hole U1309D) that may be used for reorientation studies. Because of the high core recovery rate during Expedition 305, a number of errors in the shipboard data could not be corrected on board. Therefore, all data from Expedition 305 were checked for errors and reprocessed during the postcruise meeting to make the data sets from the two expeditions as compatible as possible. The data from Expeditions 304 and 305 are described separately below, and the two data sets are then combined to review the overall distribution of remanence directions and their tectonic significance.

Expedition 304

Samples from both holes are characterized by dominantly negative inclinations that are taken to represent a reversed polarity remanence. However, there are important downhole fluctuations in the inclination from both archive halves and discrete samples. Samples from the uppermost ~200 m have a mean inclination (–48°) that is essentially identical to that expected from a geocentric axial dipole (±49°), suggesting little tectonic tilt has occurred since the remanence was acquired. In contrast, samples from deeper levels in Hole U1309D have significantly shallower reversed polarity inclinations, suggesting a more complicated tectonic history. In addition, multicomponent remanences (with up to four well-defined components) have been documented from thermal demagnetization studies. These results provide compelling evidence for the influence of later normal polarity magnetizations in several intervals of the core. The presence of both polarities is not entirely unexpected, given the location of the site near magnetic Anomaly 2 (Zervas et al., 1995). With further detailed thermal demagnetization studies, the multicomponent magnetizations should provide important information on the thermal history of the site.

Some intervals in the holes, particularly within diabase and oxide-rich gabbroic units, have positive inclinations and, therefore, may be of normal polarity. These positive inclinations are, however, significantly steeper than the present field inclination (48°) or the geocentric axial dipolar inclination at the site. A detailed discussion of these data is presented below and demonstrates that data from intervals with positive inclinations must be treated with caution. No single-component magnetizations of normal polarity of unequivocal geomagnetic origin were observed.

Remanence data

Continuous measurements

Remanence measurements were made at 2 cm intervals on all archive-half pieces longer than ~7 cm. All archive halves were subjected to stepwise alternating-field demagnetization at 5–10 mT steps up to maximum peak fields of 60 mT, with some core pieces also treated at 70 and 80 mT if significant remanence remained. Whole core and archive-half susceptibility data obtained from the MST and the archive MST (AMST; see “Physical properties”) were filtered to preserve only data corresponding to the intervals where remanence measurements were made. Thus, the resulting remanence and susceptibility data sets (Fig. F237) are significantly less affected by artifacts resulting from small core pieces than are the original unfiltered data sets.

The characteristic remanent magnetization (ChRM) directions were calculated by principal component analysis (PCA) (Kirschvink, 1980) at all measurement points along the core pieces where stable vector components could be identified on Zijderveld plots. Initially, this was done by manually picking directions for each 2 cm interval through Hole U1309B and the uppermost 130 m of Hole U1309D. Approximately 1500 vector endpoint diagrams were individually examined in this manner. This procedure allowed recognition of the demagnetization characteristics of different rock types. Based on these characteristics, we developed a semiautomated method of identifying linear segments on the demagnetization plots using the MacPaleomag software for the Macintosh. Such an automated procedure was necessary because, ultimately, demagnetization data from >5500 intervals were analyzed. Principal components were calculated for each of three overlapping coercivity ranges (40–80, 30–60, and 20–50 mT) at each 2 cm measurement point. Only principal components with maximum angular deviation (MAD) angles <5° were considered acceptable. The highest stability linear component found in this way for each measurement point was retained for further analysis (a significant number of points had no linear segment that passed our selection criteria). These data were then averaged at a piece level for plotting and interpretation.

Various basaltic rocks (aphyric, plagioclase-phyric, olivine-phyric, olivine-plagioclase-phyric, and fine-grained basalt), diabasic rocks (diabase and microdiabase), gabbroic rocks (layered gabbro, microgabbro, gabbro, olivine-bearing gabbro, olivine gabbro, troctolitic gabbro), troctolitic rocks (troctolites and olivine-rich troctolites), and serpentinized harzburgite were described from Site U1309 during Expedition 304 (see “Igneous petrology”). For simplicity, we have grouped these various rock types into seven primary lithologies:

  1. Ultramafic rocks
  2. Troctolite
  3. Olivine gabbro and troctolitic olivine gabbro
  4. Oxide and disseminated-oxide gabbro
  5. Gabbro, including olivine-bearing gabbro
  6. Diabase
  7. Basalt

These divisions are sufficiently distinct to be meaningful in terms of their magnetic properties.

The most significant variations in the NRM intensity and susceptibility for archive halves from Site U1309 are well correlated with lithology (Figs. F237, F238; Table T12). Gabbro and troctolite samples are characterized by low susceptibilities and natural remanent intensities. The mean NRM values are considerably lower than geometric mean values for gabbroic samples (~1 A/m) from the MARK area, Hole 735B, or Site 1275 (Gee et al., 1997; Dick, Natland, Miller, et al., 1999; Kelemen, Kikawa, Miller, et al., 2004). Basalt samples from Site U1309 generally have higher susceptibilities (by approximately a factor of 3–5) than those of the gabbros or troctolites. The remanent intensities of basalt samples are higher than those of the gabbros, although both lithologies show considerable variation. In contrast, diabase samples have much higher NRM intensities (up to 20 A/m) and correspondingly higher susceptibilities. Serpentinized peridotites are the most highly magnetic lithology recovered, with a range of susceptibilities suggesting magnetite concentrations of 1%–3% by volume and remanent intensities, on average, three times those observed in the diabases. Although olivine gabbros have a relatively low average NRM intensity (0.7 A/m), this group also includes very high intensities, presumably reflecting variable degrees of serpentinization of the more olivine-rich lithologies (Fig. F238).

Inclinations of NRMs vary widely downhole from –73° to 89°, but generally migrate to lower values during initial demagnetization steps (Fig. F237). Nearly all samples show evidence of a drilling-induced magnetization, evident from the consistently steep positive NRM inclinations. In many lithologies, this drilling overprint is easily removed by alternating-field demagnetization to 15–30 mT, revealing natural components of lower and dominantly negative inclination.

Demagnetization characteristics are strongly dependent on lithology. Most gabbro intervals are dominated by a low-intensity, high-coercivity reversed component (e.g., Fig. F239A). Similar high-stability reversed components are also observed in many basalt samples. Diabase intervals are dominated by normal polarity components and exhibit three types of behavior:

  1. Type A intervals have a low-coercivity, steeply inclined component that is removed by fields of 10–15 mT, leaving a stable, more shallowly inclined normal polarity component that contributes a significant proportion of the natural remanence and that persists to 60 mT or above (e.g., Fig. F239B).
  2. Type B intervals have a low-coercivity, steeply inclined component that dominates the remanence, and no stable endpoint is reached (e.g., Fig. F239C).
  3. Type C intervals have an extremely weak reversed component of high coercivity that is isolated following removal of the strong normal component (e.g., Fig. F239D).

A number of lines of evidence suggest that the normal polarity components observed as the dominant remanence in the diabase intervals (and as low-stability overprints in all lithologies) represent drilling-induced remanences and/or artifacts of the measuring and demagnetizing process. Perhaps the best indication that most normal polarity components recovered from archive halves have no geomagnetic significance comes from comparison of declinations for samples with positive and negative inclinations (Fig. F240). Principal components with negative inclinations (reversed polarity) show an essentially random distribution of declinations, as expected for azimuthally unoriented core pieces (Fig. F240A, F240B). In contrast, positive inclination remanence components from nearly all archive halves are biased toward the +x coordinate axis in the core reference frame (x on stereo plot; Fig. F240C, F240D), suggesting the measured remanence is dominated by some spurious magnetization or artifact. A small number of archive-half pieces have shallower positive inclinations and more random declinations, suggesting they may preserve some record of geomagnetic field behavior. Demagnetization results from discrete samples provide additional information concerning the origin and significance of these apparently normal polarity components, and we will therefore return to this below.

Discrete samples

Stepwise demagnetization data (from either alternating-field or thermal treatment) were acquired for 35 discrete samples from Hole U1309B and 142 discrete samples from Hole U1309D (Table T13). The demagnetization behavior of discrete samples from Site U1309 varies considerably with lithology, as noted for archive halves. Although a low-stability (presumably drilling related) component is evident in nearly all samples, this drilling component is a relatively small fraction of the remanence (<25%) in most gabbroic samples and is readily removed by alternating-field demagnetization at 10–20 mT (Fig. F241A, F241C). The ChRM of gabbroic samples is uniformly of reversed polarity and shows linear decay toward the origin on vector endpoint diagrams over a range of peak alternating-field values extending as high as 120–150 mT. The median destructive fields (MDFs; the alternating-field value necessary to reduce the NRM intensity [expressed as vector difference sum to account for multicomponent remanence] to one-half its original value) for gabbroic samples average 30 mT, indicating the remanence is very stable.

Numerous gabbroic samples were subjected to stepwise thermal demagnetization (Fig. F241B, F241D). The ChRM direction is typically isolated above 500°–550°C, and median destructive temperatures (calculated in an analogous fashion as the MDF) range from 530° to 560°C for all samples except one. Where multiple discrete samples were taken from the same core piece, alternating-field and thermal demagnetization isolate essentially the same characteristic magnetization direction (cf. Fig. F241A, F241B). The maximum unblocking temperature of 580°–590°C, the discrete range (dominantly 500°–580°C) of unblocking temperatures, and high coercivity all suggest that the remanence is carried by fine-grained magnetite.

Negative inclinations are also characteristic in most other lithologies, although this reversed polarity component is, in some cases, obscured by a substantial drilling-related normal polarity magnetization. The substantial drilling overprint in diabase, basalt, and serpentinized harzburgite samples is reflected in the low MDF values (average = ~3 mT) for these lithologies. After removal of the drilling overprint, the small interval of serpentinized harzburgite recovered in Core 304-U1309B-11R showed well-defined ChRM directions of reversed polarity (Fig. F241E). Altered basaltic breccia samples have remanence characteristics very similar to those of gabbroic samples (i.e., low NRM intensities and high stability [average MDF = 39 mT])—nearly univectorial remanence decay (Fig. F241F). Basalt samples generally have substantial low-stability overprints, but reversed polarity ChRM directions were isolated for all discrete samples from this lithology. As noted above for archive-half measurements, discrete samples from diabase units are characterized by substantial low-stability drilling-related overprints that, in many cases, obscure the primary remanence (discussed more fully below).

The characteristic remanence inclinations of discrete samples are in general agreement with the results from archive halves (Fig. F242). The reversed polarity components are distributed along a small circle (Fig. F240A, F240B), as expected from azimuthally unoriented core pieces and similar to the pattern of reversed polarity ChRMs from the archive halves. Unexpectedly, and similar to the observations from the archive halves, the low-stability normal polarity components from the discrete samples are clustered about the –x-coordinate axis in the core reference frame (Fig. F240C, F240D). A possible explanation of this bias is that discrete samples may acquire a small remanence parallel to the sample –x-direction during laboratory drilling of the working half (see also “Appendix C”).

Discrete samples reveal a strong clustering of inclinations near –40°–50°; however, a small number of discrete samples have positive or shallow inclinations. Steep positive inclinations (>65°) were isolated primarily from thermal demagnetization of diabase and some gabbroic samples. The remanence in these samples (red dots in Fig. F240C) initially is very similar to the direction of the steep low-stability component isolated by alternating-field demagnetization (blue dots in Fig. F240C, F240D) and decays univectorially to the origin. Thus, the steep positive inclinations recovered by thermal demagnetization likely reflect samples in which the remanence was essentially entirely reset by the drilling process. A small number of samples have moderate positive inclinations that apparently represent true normal polarity magnetizations (i.e., unrelated to the drilling process).

Downhole inclination variations in archive halves are corroborated by characteristic remanence directions determined from discrete samples. For the shallowest interval (0–180 mbsf, including data from both Holes U1309B and U1309D), discrete samples with negative inclinations have an average value of –48.7° (+3.4°/–1.7°; n = 76; κ = 38.3; calculated using the inclination only method of McFadden and Reid [1982]). The interval from 180 to 260 mbsf shows more scattered inclinations that average –31.5° (+8.5°/–6.2°; n = 25; κ = 14.9). This interval includes subhorizontal remanent inclinations with high magnetic stability (as judged from both alternating-field and thermal demagnetization studies). Discrete samples from 260 to 400 mbsf in Hole U1309D have intermediate inclinations (–38.0°, +4.2°/–2.9°; n = 43; κ = 34.7) similar to those observed in the more abundant archive-half data. Downhole variations in magnetic remanence and their tectonic significance are discussed more fully at the end of this section in conjunction with results from Expedition 305.

Results from discrete sample demagnetization also provide an opportunity to evaluate the accuracy of the archive-half remanence data that are used (in combination with the discrete sample results) to reorient core pieces to a common geographic framework (Table T14). Numerous core pieces have stable remanence directions identified in both the archive-half data and a discrete sample from the same core piece. Comparison of the declinations of these two types of data (Fig. F243) shows that most pairs of data agree within ~10° (note that points in the upper left and lower right of this diagram are not, in fact, discrepant but simply reflect the discontinuity at 360°).

Origin and significance of normal polarity remanence components

Most normal polarity data from Site U1309 (whether derived from the archive halves or discrete samples) have a number of characteristics that invalidate their use in tectonic studies.

Declination bias

Samples with positive inclinations (whether from the archive or working half) show a bias along the x-axis in core coordinates (Fig. F240) that is unexpected for core pieces that are free to rotate in the core barrel. An attempt is made to align the dominant fabric elements between core pieces prior to core splitting. However, this strategy should be most successful in lithologies with strongly developed petrofabrics (e.g., deformed gabbros). The clustering of normal polarity components is predominantly observed in diabase intervals, which have relatively weak petrofabrics. Thus, it is difficult to attribute the clustering of normal polarity components to accurate alignment of petrofabric elements between diabase pieces. A series of experiments conducted to characterize the response functions of the magnetometer (outlined in “Appendix C”) indicate that the X-superconducting quantum interference device (SQUID) (and, to a lesser extent, the Y-SQUID) sensor picks up a signal from the axial (z)- component of magnetization. Thus, samples with strong, steeply inclined normal components (e.g., drilling remanence) generate declinations biased toward the north or south, depending on the sample position in the sensor region.

Low stability of multidomain magnetite

Figure F244 shows a clear relationship between the inclination of magnetization vectors (determined from PCA of demagnetization data from archive halves) and volume susceptibility (determined on the AMST system). Samples with reversed polarities are strongly clustered around low susceptibilities of ~30 × 10–5 to 40 × 10–5 SI and display a range of inclinations consistent with that expected from geomagnetic secular variation. In contrast, steep normal polarities of magnetization predominantly cluster at inclinations greater than the present-day magnetic field and are present in samples with the highest magnetic susceptibilities (two orders of magnitude higher than that of the reversed polarity samples). This correlation presumably reflects the abundance of coarse, multidomain magnetite grains in the diabase and serpentinized, olivine-rich intervals that are more susceptible to acquiring a drilling-induced remanence. This is also indicated by the rapid decrease in remanence intensities of these lithologies during the initial stages of alternating-field demagnetization, with MDF values that average 3 mT.

Comparison of the results of different demagnetization treatments on discrete samples from a single diabase interval provides further insight into the nature of the magnetization of these rocks (Fig. F245). The remanence of these samples is dominated by the steeply inclined drilling-induced component of magnetization. For the sample shown in Figure F245A and F245B, alternating-field demagnetization successfully removes this component and isolates a high-coercivity reversed polarity component. Curvature of the demagnetization path indicates partial overlap of the coercivity spectra of the grains carrying natural and drilling-induced magnetizations. Substantial loss of remanence during low-temperature demagnetization (Fig. F245C, F245D) provides further evidence that the low-coercivity component is carried by unstable multidomain magnetite, with subsequent alternating-field treatment isolating a poorly defined reversed polarity component. Thermal demagnetization indicates distributed unblocking and a maximum unblocking temperature close to the Curie temperature of magnetite. The magnetization decays univectorially to the origin (Fig. F245E, F245F), with a direction identical to the drilling-induced component observed during alternating-field treatment and with no indication of any reversed polarity component. This indicates complete overlap of the unblocking temperature spectra of magnetite grains carrying the natural and drilling-induced magnetizations.

The presence of multidomain magnetite grains (which are particularly susceptible to acquisition of high-intensity drilling-induced remanences) and the partial to complete overlap of demagnetization spectra of natural and artificial components combine to make recovery of natural components of magnetization of geological significance problematic in the diabase intervals, particularly from the more abundant archive-half (continuous) data. These difficulties are further illustrated in Figure F30, which shows demagnetization results from two diabase intervals where both archive-half and discrete sample data are available. Here, alternating-field demagnetization isolates a high-coercivity, reversed polarity component in both discrete samples (Fig. F30A, F30C), in contrast to corresponding points in the archive-half data where either the alternating-field data are too noisy at high field levels to detect stable components (Fig. F30B) or (more rarely) where only a weak reversed component is observed during the last stages of demagnetization (Fig. F30D).

As a result of the remanence characteristics and analytical difficulties described above, there are insufficient discrete samples or archive-half intervals at Site U1309 where high-coercivity, nondrilling-related components of magnetization have been successfully isolated to allow robust directional characterization of these components in diabase, serpentinized peridotite, and some oxide-rich gabbroic intervals. We note, however, that where such components are identified in discrete samples (e.g., Figs. F245B, F245D, F30A, F30C), they are invariably of reversed polarity and may have some utility in reorienting core pieces. Diabase intervals with Type B behavior during archive-half demagnetization are also observed to migrate to lower inclinations and, eventually, to reversed inclinations during demagnetization at higher fields (Fig. F239C). Together, these data suggest that both the diabase and serpentinized peridotite recovered from Site U1309 held reversed polarity magnetizations prior to drilling, in keeping with the reversed polarity magnetizations documented in most other intervals.

Multicomponent remanence in gabbroic rocks from Hole U1309D

Magnetic results from olivine gabbro and troctolite from Section 304-U1309D-22R-2 provide compelling evidence for the reheating of some gabbroic rock from Site U1309 during a normal polarity interval. Archive-half data from Section 304-U1309D-22R-2 (Piece 1) have ChRM directions with negative inclinations but progressively become shallower toward the bottom of the piece (Fig. F246C, F246D). These shallow inclinations continue through the upper 20 cm of Section 304-U1309D-22R-2 (Piece 2), where an abrupt shift in direction occurs. A ~20 cm wide interval with positive inclinations and nearly antipodal declinations (relative to the top and bottom of the piece) is roughly centered on a small pyroxene-rich vein cutting the olivine gabbro (Fig. F246A). This interval is also marked by high stability to alternating-field demagnetization (Fig. F246E) and an increase in MS (Fig. F246B), though the latter feature appears to correlate more closely with a ~10 cm wide troctolite band. It is worth noting that the low susceptibility values throughout this section preclude a significant amount of magnetite (e.g., from serpentinization of olivine). For example, the highest susceptibility values within the troctolite (~24 × 10–5 SI) could be entirely explained by the paramagnetic contribution from iron in silicate minerals if the bulk rock has ~4%–5% FeO (at the low end of values measured for troctolites during Expeditions 304 and 305). Given the susceptibility of magnetite (3 SI), the troctolite would have a maximum of ~80 ppm magnetite even if no iron-bearing silicates were present in the rock.

Two discrete samples (one from Section 304-U1309D-22R-2 [Piece 1] and the second from Piece 2 near the pyroxene-rich vein) were thermally demagnetized to further investigate this apparent normal polarity magnetization. The upper sample (Sample 304-U1309D-22R-2, 8–10 cm) (Fig. F247B) exhibits a high-stability reversed polarity remanence that is unblocked from 520° to 600°C. In addition, this sample shows a well-defined normal polarity magnetization component (inclination = +57°) removed at temperatures from 350° to 520°C. Results from the corresponding interval of the archive half (Fig. F247C) recover a similar high-stability reversed polarity remanence but show little evidence of the nearly antipodal normal polarity overprint isolated by thermal demagnetization.

The lower sample (Sample 304-U1309D-22R-2, 79–81 cm) (Fig. F247D) also shows a high-stability reversed polarity remanence isolated at temperatures >550°C. This sample, however, shows a much larger normal polarity overprint (inclination = +43°), such that the initial NRM direction has a positive inclination. In this case, the archive-half data show a normal polarity ChRM after removal of a steep drilling-related component, although the peak alternating-field of 80 mT is insufficient to demagnetize the sample. Finally, the outer chip from the minicore sample was also alternating-field demagnetized (Fig. F247F). As with the archive-half data, this sample has very high coercivity and a normal polarity ChRM (inclination = +40°).

The normal polarity components in these thermally demagnetized discrete samples are distinct from the steep, low-coercivity component normally associated with a drilling-induced remanence. Indeed, a steep component with positive inclination (removed by ~300°C) that likely represents a drilling-induced component is evident as a distinct magnetization component in both samples. Thus, the moderate-inclination normal polarity component isolated by thermal demagnetization appears to reflect reheating during a normal polarity interval. Although the normal polarity interval in the archive-half data coincides with a small (1 cm wide) pyroxene-rich intrusion, this small intrusion is unlikely to be responsible for the normal polarity overprint throughout the section. Rather, we suggest that a larger intrusion not sampled by the drill core is more likely responsible for this reheating episode.

In addition to the localized example of reheating described above, thermal demagnetization data elsewhere indicate that such complex multicomponent remanences are widespread in Hole U1309D (Fig. F248), although these complexities are not readily detected by alternating-field demagnetization. The directions and temperature stability of these various magnetization components provide valuable information on the thermal and, potentially, the tectonic history of the samples. The sharp breakpoints between the high-temperature components are most compatible with a simple thermoremanence because chemical remanence should be accompanied by a broad range of unblocking temperatures. An initial steeply inclined component present in some samples is removed by temperatures of 100°–350°C (blue arrows in Fig. F248) and is interpreted as a drilling-induced overprint. A normal polarity component (orange arrows in Fig. F248) is commonly observed between temperatures of 350° and 530°C, with a mean inclination of 53.6° (+6.4°/–8.2°; n = 11; κ = 40.3). This moderate-inclination component is, therefore, distinct from the steep, low unblocking temperature (and low coercivity) drilling-induced component. The highest temperature component (red arrows in Fig. F248) is consistently of reversed polarity and unblocks between 530° and 590°C. This component typically has a shallower inclination than the normal polarity overprint, and the two components are not antipodal. These samples are therefore characterized by two components of geomagnetic significance that were acquired over sharply defined blocking temperature intervals. A sample from Section 304-U1309D-71R-2 (Fig. F248C) reveals an additional lower unblocking temperature reversed component between 200° and 350°C, providing evidence of a (presumably later) reversed overprint in at least this section.

Magnetic fabric

The anisotropy of MS (AMS) was determined for most discrete samples from Holes U1309B and U1309D during Expedition 304 (Table T15). The remaining samples had susceptibilities too low to reliably measure onboard. Most susceptibility tensors are moderately anisotropic (P′ < 1.10, where P′ is the corrected degree of anisotropy) (Jelinek, 1981), but significantly higher degrees of anisotropy (P′ up to 1.37) were noted for several very coarse grained gabbroic samples (Fig. F249E). The majority of samples have triaxial susceptibility ellipsoids, with a range of ellipsoid shapes. In the core reference frame, the only discernable pattern is a concentration of minimum principal axes at moderate to steep inclinations (Fig. F249A).

In order to compare the orientations of magnetic fabrics from different samples, some common reference frame is required. Under the assumption that the stable remanent magnetization approximates the time-averaged reversed polarity direction, the magnetic fabric data have been restored to a common reference frame by a simple vertical axis rotation that restores the remanent declination to the presumed reversed polarity direction (180°). After this restoration, the largest concentration of minimum eigenvectors lie in the southwest quadrant, with a corresponding cluster of maximum eigenvectors in the northeast quadrant (Fig. F249B). To a large extent, these concentrations of eigenvectors can be attributed to diabase and basaltic samples (Fig. F249C). Eigenvectors from gabbroic samples are much more variable but also show a concentration of minimum eigenvectors in the southwest quadrant (Fig. F249D). Two samples from the serpentinized harzburgite in Core 304-U1309B-11R have relatively consistent fabrics, with a shallow east-dipping magnetic foliation.

The orientation of the maximum eigenvector of the susceptibility tensor for mafic dikes typically has been interpreted as parallel to the magmatic flow or emplacement direction, with the minimum eigenvector approximating the pole to dike plane (e.g., Knight and Walker, 1988; Tauxe et al., 1998). Although a minority of dike samples have AMS fabrics that do not conform to this simple model, the AMS results from diabase samples from Site U1309 may provide valuable structural information for the site. The principal axis directions from multiple diabase samples taken from the same core piece are well grouped, as emphasized by the small 95% error bounds of three samples from Section 304-U1309B-19R-2 (Piece 4) (Fig. F249C). We note that a sample (Sample 304-U1309D-15R-2, 126–128 cm) from a potentially correlative diabase in Hole U1309D has a very similar magnetic fabric (maximum eigenvector = 359°/21°). To the extent that the interpretive model for AMS in mafic dikes outlined above is applicable, the minimum eigenvectors should approximate the pole to the dike plane (i.e., the diabase units would represent sills dipping gently to the northeast). Although the number of specimens is small, the consistency of the maximum eigenvectors in diabase and basalt samples may indicate a northeast–southwest flow lineation.

Comparison of core magnetic data and logging results

The General Purpose Inclinometry Tool (GPIT) incorporates a three-axis fluxgate magnetometer that is primarily used to provide azimuthal information for the FMS images (see “Downhole measurements”). In principle, these fluxgate sensors also can provide valuable information on the intensity and direction of the magnetization of the borehole wallrocks. The horizontal and vertical magnetic anomalies from the first FMS/GPIT run in Hole U1309B are shown in Figure F250 (note that north and east magnetic anomalies cannot be resolved because there is no independent estimate of the azimuth of the logging tool string). At first glance, this anomaly pattern appears remarkably consistent with the magnetic inclination measured on the recovered core. The vertical anomaly is positive and the horizontal anomaly is negative, as expected for reversely magnetized material, and the average values (approximately +5000 nT for the vertical and –2000 nT for the horizontal) correspond to an inclination value of about –50°.

Comparison of the borehole anomaly data with remanence intensity data from the cores indicates that this correspondence is simply fortuitous. We constructed a simple forward model of the borehole magnetic field based on horizontal, uniformly magnetized layers with intensities based on average values from the core. Average remanence intensities were calculated for seven intervals corresponding to large-scale lithologic changes in the core (Fig. F250). Intensity data from the archive halves after 10 mT demagnetization were used to minimize the influence of drilling-induced remanence (which should not be present in the borehole walls except in the immediate vicinity of the borehole). The calculated anomalies in gabbroic sections of the core are typically less than a few hundred nanotesla, in contrast to the values of several thousand nanotesla measured by the GPIT. This discrepancy is most readily attributed to improper calibration of the fluxgate sensors.

Despite these calibration problems, there is useful information in the borehole magnetometer data. After adjusting the baseline levels and subtracting a model for the magnetic effect of the pipe on the vertical anomaly, the calculated and observed anomalies are similar in character. The observed anomaly data indicate a highly magnetic interval near 73 mbsf in an interval with incomplete recovery. The magnitude of this anomaly suggests that a significant interval of strongly magnetic material (possibly serpentinized peridotite or diabase) was not recovered. In contrast, the near constant and low anomaly amplitudes in other intervals (e.g., 40–57 and 68–94 mbsf) suggests that the recovered materials are representative of the entire drilled interval. Regions such as the interval from 27 to 37 mbsf, where the modeled (based on reversed polarity remanences) and observed anomalies are of opposite sign, may indicate large induced magnetizations or possibly normal polarity of the natural remanence.

Expedition 305

Expedition 305 deepened Hole U1309D to ~1415 mbsf. Shipboard paleomagnetic studies during Expedition 305 consisted of continuous measurements of archive-half sections and progressive demagnetization measurements of discrete samples, in a manner similar to that performed during Expedition 304 (with the exceptions described in the “Methods” chapter). A total of 335 minicore samples were stepwise thermally or alternating-field demagnetized to evaluate the directional stability and coercivity/unblocking temperature spectra of each sample. The Königsberger ratio, which is defined as the ratio of remanent magnetization to the induced magnetization in the Earth’s magnetic field, was also calculated using the volume MS for these samples. AMS was measured on 216 minicore samples with the Kappabridge KLY-2 using the standard 15-position measuring scheme. The remaining samples had susceptibilities that are too low to measure their AMS reliably onboard.

Remanence data

Continuous measurements

Within the recovered rocks, there are considerable variations in magnetic properties and demagnetization behavior among the various lithologies, which are similar to what was observed during Expedition 304 (see “Expedition 304”). The most important observations during Expedition 305 are summarized as follows:

  1. The NRM intensities of the archive halves span more than four orders of magnitude (ranging from 0.001 to >10 A/m) (Fig. F251). Variations in magnetic susceptibility are consistent with the variations in NRM intensity. Susceptibility data from the core sections provide important information on the Fe-Ti oxide content in the rocks recovered from Hole U1309D. Magnetic susceptibility measurements revealed sharp peaks in numerous oxide gabbro core sections (e.g., Sections 305-U1309D-113R-1, 114R-2, 254R-1, 270R-3, 272R-3, 276R-1, and 277R-1), olivine-rich troctolite zones (Sections 305-U1309D-111R-4, 112R-1, 227R-2, 228R-2, 238R-1, 242R-2, 243R-1, and 247R-2), and olivine-bearing gabbro (Sections 305-U1309D-135R-1, 139R-1, 159R-4, and 241R-1). These susceptibility peaks were verified by further measurements of corresponding discrete samples.
  2. A remagnetization imparted by the coring process is commonly encountered, as noted during previous DSDP, ODP, and IODP cruises (e.g., Gee et al., 1989). This remagnetization is characterized by NRM inclinations that are strongly biased toward vertical (toward +90°) in a majority of cores as is largely removed by alternating-field demagnetization to 30 mT (Fig. F251). In some intervals with relatively low magnetic susceptibility (e.g., ~930–970 mbsf), the remagnetization appears to have only affected inclination and little effect on NRM intensity is apparent (Fig. F251). Interestingly, some intervals in the holes (e.g., Core 305-U1309D-101R, ~505 mbsf), particularly within medium- to coarse-grained gabbro sections, show little drilling-induced overprinting both in NRM intensity and inclinations.
  3. Data remanence inclinations obtained in Hole U1309D can be correlated with lithology. As shown in Figure F252 and Table T16, gabbroic samples (Fig. F252A–F252C) are characterized by dominantly negative inclinations (reversed polarity) and low NRM intensity and susceptibility, whereas diabases, troctolites, and olivine-rich troctolites have higher NRM intensity and susceptibility and, apparently, dominantly positive inclinations (normal polarity) (Fig. F252D–F252F). These rocks with positive inclinations typically have one order of magnitude higher NRM intensity and magnetic susceptibility than those of fresh gabboric rocks and commonly have much lower coercivity, as indicated by a rapid decrease in intensity during lower steps of alternating-field demagnetization experiment. In a majority of cases, the positive inclination zones correspond to olivine-rich intervals where the concentration of serpentine is high (e.g., Fig. F251, 1120–1200 mbsf where troctolite and olivine-rich troctolite are recovered). Interestingly, olivine-rich gabbro and troctolite samples in Hole 735B also recorded apparently opposite magnetic polarity than those of surrounding metagabbros (Kikawa and Pariso, 1991). These observations of normal polarity with serpentinization are intriguing and may have rock magnetic (drilling-induced overprint) or geologic implications (hydrothermal alteration in a normal polarity field). Borehole magnetic measurement results and more complete paleomagnetic and rock magnetic studies are needed to further address this problem (see discussion in “Expedition 304”).
  4. Although the timing of remanence acquisition is uncertain, the presence of both polarities is not unexpected because of the age and location of the site near magnetic Anomaly 2 (Olduvai normal epoch; 1.66–1.88 Ma) (Zervas et al., 1995; Gee and Blackman, 2004). However, the inclinations for the recovered rocks during Expedition 305 (depth range = 400–1415 mbsf) are significantly shallower than expected from a geocentric axial dipole (±49°), suggesting some tectonic tilt occurred after the remanence was acquired, as discussed below.
  5. There is an apparent relationship between the inclination (determined from the 30 mT demagnetization data from archive halves) and volume susceptibility (determined with the AMST system). As shown in Figure F253, samples with reversed polarity are strongly clustered around low susceptibilities of ~100 × 10–5 SI. In contrast, no low-susceptibility samples with unambiguous positive inclinations have been noted.

In summary, preliminary pass-through paleomagnetic data revealed important magnetic signatures that await further verification in terms of age and origin. Results from discrete sample demagnetization allow evaluation of the accuracy of the archive-half pass-through data, as described below.

Discrete samples

To investigate the nature of the remanent magnetization of the discrete samples from Hole U1309D rocks, 335 samples were stepwise thermally (266) or alternating-field (69) demagnetized. The demagnetization behavior of discrete samples from Site U1309 varies widely with lithology, as outlined above for archive halves. The ChRM of various gabbroic samples (medium-grained, coarse-grained, amphibole-bearing, olivine-bearing, and oxide gabbros) is mainly of reversed polarity, whereas troctolitic gabbro, olivine-rich troctolite, and diabase samples mainly have normal polarity. The MDF for gabbroic samples averages >30 mT, suggesting the remanence is of high stability. For many troctolite samples, however, the MDF value is often <10 mT. Stepwise thermal demagnetization isolates the ChRM direction typically >500°–550°C, suggesting that the remanence is carried by fine-grained magnetite.

As mentioned above, the vertically directed drilling-induced magnetization is commonly present in Expedition 305 cores. In most cases, this steep downward component of magnetization is not very resistant to alternating-field demagnetization. Examples are shown in Figure F254A; alternating-field demagnetization to 15 mT on Sample 305-U1309D-82R-2, 114–116 cm, effectively removed the drilling overprint to reveal negative inclination of the ChRM. Thermal demagnetization of minicore samples also successfully removed this drilling-induced magnetization component. Figure F254B illustrates that the drilling-induced remagnetization component is removed after 400°C (Sample 305-U1309D-119R-1, 40–42 cm) or 500°C (Sample 126R-1, 27–29 cm) demagnetization, and a characteristic component decaying linearly toward the origin can be identified. Two representative samples of normal polarity (Samples 305-U1309D-143R-2, 72–74 cm, and 160R-3, 104–142 cm) are also shown in Figure F254B. It is interesting to note that in both cases, thermal demagnetization to temperatures >400°C is required to isolate the highest-stability magnetization component.

Rock magnetic characterization

Königsberger ratio

In general, the Königsberger ratio is used as measure of stability to indicate a rock’s capability of maintaining a stable remanence. The International Geomagnetic Reference Field value at Site U1309 (40,918 nT = 32.56 A/m) was used for calculating Q, where, in the equation

Q = NRM [A/m]/(k [SI] × H [A/m]),

  • k = magnetic susceptibility and
  • H = the intensity of the local geomagnetic field.

The variation of the Königsberger ratios is listed in Table T17 and plotted in Figure F255. The Q ratio, in general, resembles that of the NRM. For example, diabase Sample 305-U1309D-287R-1, 8–10 cm, has a higher intensity of remanence than overlying gabbro Sample 305-U1309D-286R-2, 134–136 cm, and, consequently, the Königsberger ratio of the former (68.7) is much higher than that of the latter (7.8) (see Table T17). Similar examples of this correlation are also seen in samples from other rock types. In general, the Königsberger ratios throughout the hole are >1.0 (Fig. F255), indicating that remanent magnetization is greater than the induced magnetization.

Anisotropy of magnetic susceptibility

AMS was measured on 216 minicore samples (Table T18). The remaining samples have susceptibilities that are too low to measure their AMS reliably on board. The degree of anisotropy (P, where P = maximum/minimum eigenvalue of the susceptibility tensor) ranges from 1.034 to 1.895 (Fig. F255). In general, the degree of anisotropy is <1.2. Toward the bottom of the hole (1150–1415 mbsf), where several ultramafic zones were recovered, rock samples have relatively consistent degrees of anisotropy. Because of the high noise levels of these shipboard measurements compared to shore-based laboratory environments, however, the AMS results obtained on the ship are preliminary and need to be verified by postcruise study. The directional data from Expedition 305 shipboard AMS measurements have not been reoriented to a common reference frame and are not discussed further here.

Tectonic significance of remanence data

The deep penetration and high core recovery (particularly of gabbroic rocks containing fine-grained magnetite) at Site U1309 provide a unique opportunity to examine the thermal and tectonic history of lower crust denuded at an oceanic core complex. Although drilling-induced overprints are a significant problem in some lithologies, the high-quality magnetic data obtained from gabbroic (sensu lato) rocks should provide a record of the geomagnetic field during construction and uplift of Atlantis Massif. Two aspects of the remanence data set, in particular, must be accounted for in any valid tectonic model for the massif.

First, several depth intervals with distinctly different mean inclinations can be identified from the shipboard data. Five inclination groups have been identified based on remanence data from archive halves and discrete samples (Fig. F203). The boundaries between these groups generally coincide with structural features (faults, shear zones) and so structural data have been used to define the precise boundaries (see “Structural geology”). The upper 180 m is characterized by a mean discrete sample inclination (~47°) that is nearly identical to the expected dipole inclination at the site (Fig. F204). A simple possible explanation of this result is that the interval has not experienced detectable tectonic tilt since the remanence was acquired, although this inference depends on the rotation axis and history (see discussion in “Reorientation of structure data using paleomagnetic data”). In contrast, the mean inclinations from all inclination groups deeper in the hole are statistically distinct from the expected direction. The shallower inclinations in these sections cannot be attributed to artifacts of the drilling and/or measurement process, as they are corroborated by numerous high-quality discrete sample demagnetization data. An example of a near-horizontal characteristic magnetization identified from thermal demagnetization of a sample from inclination Group II is shown in Figure F248D. The extremely narrow range of unblocking (550°–590°C) in this sample and the corresponding high-coercivity remanence (and identical direction) recovered by alternating-field demagnetization of a companion sample from the same core piece suggest the shallow inclination represents a primary thermoremanence. Inclination Groups III and IV have mean values steeper than those from Groups II and V and closer to the expected dipole inclination at the site. This nonsystematic change in remanence inclination downhole is difficult to reconcile with any simple model of denudation and likely reflects a more complex thermal and unroofing history (relative motions between multiple fault blocks and/or reheating by later intrusions).

Second, well-defined multicomponent remanences noted during Expedition 304 suggest that remanence acquisition spanned multiple polarity intervals. In addition, we note that the highest stability reversed polarity magnetization is typically shallower than the normal polarity overprint and the two components are not antipodal. Although the number of samples with well-defined normal polarity overprints is small (n = 11), these overprints have a mean inclination (53.6° +6.4°/–8.2°; κ = 40.3) that is not statistically distinguishable from the present-day normal polarity inclination at the site. The directional difference between the normal and reversed polarity components may reflect the influence of tectonic tilting after acquisition of the highest stability reversed polarity magnetization. Although less common, such multicomponent magnetizations are also observed in several discrete samples analyzed during Expedition 305.

The origin and significance of remanence throughout the core will clearly require more detailed shore-based studies (particularly thermal demagnetization of discrete samples) in order to more fully understand the thermal and tectonic history at the site. Further rock magnetic and petrographic analyses are required to establish the likely timing of acquisition of magnetization.