Previous   |    Next

doi:10.2204/iodp.proc.314315316.123.2009

Paleomagnetism

Paleomagnetic directions

Archive-half core sections from Holes C0001E, C0001F, and C0001H were measured on the shipboard pass-through cryogenic magnetometer, except for a few sections where sediments were severely disturbed by drilling or flow-in. Natural remanent magnetization (NRM) was measured at 5 cm intervals in each core section, followed by alternating-field (AF) demagnetization at 5, 10, 15, and 20 mT peak fields. Occasionally we opted to reach higher AF fields on short core segments from Hole C0001H that were of particular interest for structural geology. In this case, the goal was to obtain a more reliable magnetization direction based on principal component analysis after analysis of demagnetization (Zijderveld type) diagrams (Zijderveld, 1967). Only a few discrete samples were taken at this stage. Most of the measured cores displayed directions biased toward steep inclinations. After AF demagnetization at 20 mT, magnetic inclination agrees with the expected value (~52°) (Fig. F24). Such steeper NRM inclinations are indicative of drilling-induced overprint, typically observed in ODP/IODP studies. The overprint can be generally removed after AF demagnetization at 5–20 mT, although we observed that cores obtained with the RCB often showed a more resistant overprint and required higher AF fields to overcome. Figure F25 shows the overall trend of both NRM and magnetization after 20 mT, which shows a decrease of quite a large proportion in intensity. For characteristic remanent magnetization (ChRM), in order to establish a magnetic reversal stratigraphy, we used the directions obtained after blanket demagnetization at 20 or 30 mT, following standard ODP/IODP procedures. Note that sometimes the magnetic directions do not reach a stable endpoint at the highest demagnetization value used, suggesting that the primary magnetization has been only partially isolated. A more reliable ChRM by vector analysis is planned with shore-based measurements.

Hole C0001H produced paleomagnetic results more difficult to interpret. We believe that there are basically two issues related to the data set quality. On one hand, a magnetic overprint, typically the type that is drilling induced (i.e., steep and downward), is widespread in cores from Hole C0001H (obtained with the RCB). On the other hand, and because of RCB coring, there are numerous cores that show biscuiting. Measuring nonuniformly magnetized cores (i.e., because of biscuiting or a polarity transition) with the pass-through magnetometer produces steeper magnetic inclinations and very scattered declinations (e.g., Shipboard Scientific Party, 2002). Reliable declination and inclination values can be inferred only when measurements are consistent across an interval longer than the sensing region of the superconducting rock magnetometer (SRM) coils (20 cm). Consequently, in our data interpretation we excluded all cores from Hole C0001H which show biscuits too short for the pass-through magnetometer. Seemingly homogeneous cores have been included, as well as cores for intervals of particular interest, because of the appearance of tectonic microstructures. We took two different approaches. If a section contained biscuits longer than 20 cm, they were measured one at a time in order to avoid a convoluted signal. Alternatively, some sections were measured as usual, but in the analysis we averaged directions that correspond to the center of biscuits longer than 20 cm.

Core orientation

Declinations obtained from archive-half core pieces were fundamental in core orientation corrections and structural analysis (See “Structural geology”). After AF demagnetization of each section, visual inspection of the Zijderveld diagrams allowed us to determine whether the end vector at the highest AF level (typically 20 mT) could be used as the representative paleomagnetic direction for a given interval. If so, paleomagnetic directions were grouped by section and a mean direction was computed with the associated statistical parameters (Table T16). In addition, a number of biscuit fragments were selected for more complete AF demagnetization analysis in order to isolate reliable paleomagnetic directions. On these fragments, the AF peak was 80 mT and the characteristic remanent magnetization component direction was calculated using principal component analysis (program “pca;” Tauxe, 1998), guided by visual inspection of orthogonal demagnetization plots (program “plotdmag;” Tauxe, 1998).

There are three basic assumptions in using paleomagnetic directions to determine the core azimuth. First, the direction of the ambient field in which the remanence was acquired is known. Rocks sampled during Expedition 315 are young enough for Earth’s magnetic field to be modeled satisfactorily by a geomagnetic axial dipole (Quaternary to Neogene). Average inclinations in Holes C0001E and C0001F are 49.7° for positive inclination and –44.2° for negative inclination (Fig. F26), so the expected inclination (52°) for the present latitude of the Nankai Trough is only slightly underestimated. The phenomenon of inclination shallowing is well known in sediments, and it often produces anomalies of up to 15° and occasionally more (King, 1955; Tauxe and Kent, 1984; Garcés et al., 1996). Overall, it seems reasonable to take geographic north as the reference direction. Second, the time interval represented by a sample is sufficiently long for geomagnetic secular variation to be averaged. Third, bedding is horizontal or subhorizontal.

We are aware of the response value of the SRM pickup coils, estimated to be ~20 cm, and know that only measurements taken every 20 cm are truly independent from each other. Hence, for more precise orientation studies, it is recommended to analyze discrete samples based on shore measurements.

Core twisting

We often noticed a change in magnetic declinations along the core, referred to here as core twisting, in Holes C0001E and C0001F (Fig. F27). Core twisting was observed because of the reference lines scribed along the liners before they were placed into the barrel. Some cores (e.g., 315-C0001H-7F) for which liners were visibly twisted after removal from the barrel show extremely rotated declinations, documenting such a phenomenon (Fig. F28). We also documented core twisting in many cores for which liners were undeformed and where sediment was seemingly intact. In these cases, we see conspicuous declination changes with depth, which reveal a progressive rotation of the core; degree of rotation is variable.

The severity of twisting likely depends on several factors, and we did not observe a particular trend. The most extreme example is Section 315-C0001F-7H-4 (Fig. F28), where from top to bottom the core experienced rotation of about two and a half turns (Movie M1). Be that as it may, the striking dispersion of declinations documents core twisting between or even within sections and needs to be further examined in future expeditions, as it is typically assumed to be a constant orientation of the cored material.

It is also interesting to note that NRM directions (i.e., before AF demagnetization) typically show better clustering than directions obtained at 20 mT, which sheds some light on the timing of the twisting. Whereas NRM often shows a cluster, some sections show a wide distribution of declinations at 20 mT (Fig. F29), possibly reflecting core twisting. Such a pattern suggests that rotation or twisting of the core occurs before the acquisition of the secondary magnetization or the drilling-induced overprint. Nevertheless, rotation with an axis parallel to the NRM directions (i.e., vertical along the core axis) would have little or no effect on its distribution. That means that a rotation or twist of the core parallel to the core axis cannot be totally excluded. A rotation of the liner could, for example, occur when it is pulled out of the core barrel on deck. It is important to note that core twisting was observed in cores obtained with the HPCS, which poses an interesting question as to how and when core twisting occurs. Further experiments and observations will be necessary to elucidate the timing and processes controlling such core deformation documented by paleomagnetism but otherwise typically imperceptible. Needless to say, the implications for physical properties that are directional in nature are important, for it is commonly assumed that cored material does not rotate within the liner.

Magnetic reversal stratigraphy

We used magnetic inclination after AF demagnetization at 20 mT to determine the polarity pattern. The magnetic polarity record was then identified using biostratigraphic datums (nannofossils and foraminifers) (see “Biostratigraphy”) and correlated with the geomagnetic polarity timescale (GPTS) of Gradstein et al. (2004). The identified chrons and subchrons are given in Table T17.

In Hole C0001E, below a somewhat noisy uppermost 8 m, magnetic inclination is dominantly positive to 86 m CSF (Fig. F30). This change from normal to reversed polarity (Section 315-C0001E-10H-6) is interpreted as the Brunhes/​Matuyama Chron boundary dated at 0.781 Ma (Gradstein et al., 2004), owing to the identified nannofossil Zone NN19 (see “Biostratigraphy”). The Matuyama Chron (C1r) is characterized by a predominantly reversed polarity, although we noted the presence of a number of short events of normal polarity (e.g., 148 and 259 m CSF). A normal polarity interval between ~127 and ~131 m CSF is interpreted as a part of the Jaramillo Subchron (C1r.1n) (Sections 315-C0001E-3H-2 through 3H-4) (Fig. F31), although the upper and lower limits are absent. The portion of the Matuyama Chron between the base of the Jaramillo Subchron and the top of the Olduvai Subchron possibly shows as many as five short events of normal polarity. Given the constraints of onboard analysis (long-core measurement, partial demagnetization, and response of the pickup coils), with the present data set we cannot unambiguously determine whether the positive inclinations we observe within the Matuyama Chron between ~148 and ~160 m CSF reflect the geomagnetic field or whether they are instead artifacts. More detailed shore-based measurements will be needed to determine the meaning of such short positive inclination intervals within the Matuyama Chron, and our discrete sampling strategy on board was directed toward that goal.

In Hole C0001F (Fig. F31), we could reliably determine the top of the Olduvai Subchron (C2n) at 174.7 m CSF (Table T17). The bottom of the Olduvai Subchron occurs at 193.5 m CSF, although the subsequent older Matuyama Chron is only represented in 3.3 m of sediment. Below, we encountered a thick sandy unit (lithologic Subunit IIB) that is completely disturbed (see “Lithology”), preventing any continuous reliable paleomagnetic analysis. Silty layers, a few centimeters thick, interbedded within the sands might help in obtaining some occasional magnetic polarity readings. The age of the uppermost part of this sandy unit is roughly estimated as ~2 Ma from our magnetostratigraphic correlation (see below). Below the sandy unit, ChRM inclination is dominantly positive in the uppermost ~20 m of the section, interrupted by a short reversed interval at ~218–220 m CSF. The interpretation of such a normal interval as Gauss (C2An) is a possibility, given its proximity to the Olduvai Subchron, but this interpretation is not entirely supported by biostratigraphy data. Rather, nannofossil Zones NN13–NN15 and foraminifer Zone N18 are indicative of Chrons C3n to C2Ar (Fig. F23).

The reliability of the remanence in sediments of the middle and lower part of Hole C0001H needs to be carefully considered. We note that magnetization often does not reach an end vector by 20 mT in the interval from ~324 to 374 m CSF. Within this interval, many samples show traces of great circles upon progressive AF demagnetization (Figs. F32, F33). Such behavior is indicative of a higher coercive component that is not resolved by AF demagnetization applied onboard. Although we applied progressive thermal demagnetization on several discrete samples, this high-coercivity component could not be fully isolated (Fig. F34). The trend of great circles suggests that such a high-coercivity component often has negative inclination (Figs. F32, F33). Therefore, we suspect that the primary magnetization for the interval between 324 and 374 m CSF might indeed be negative.

There are two more issues affecting the overall data quality of Hole C0001H. The first, discussed earlier, is biscuiting. When the core is not uniformly magnetized (e.g., because of differential rotation of segments or biscuiting), data should be treated with caution. Results from Shipboard Scientific Party (2002) documented that across a rotated break between segments of uniform magnetization, x- and y- components, in particular, are underestimated. More detailed demagnetization on shore will allow us to determine a more reliable magnetic reversal stratigraphy. The second issue relates directly to the paleomagnetic behavior of sediments. Very often demagnetization diagrams show large scatter of end vectors at 20 mT in numerous cores. Erratic remanence behavior could be due to the acquisition of spurious magnetization in the magnetometer or perhaps to the presence of ferrimagnetic iron sulfides (greigite or pyrrhotite). The presence of pyrite has been documented in numerous parts of the hole, supporting the possible presence of ferrimagnetic iron sulfides as carriers of stable remanence.

As far as polarity, calcareous nannofossils are indicative of Zones NN15–NN13 (see “Biostratigraphy”), which correlate to Chron C3n or Gilbert Chron (Gradstein et al., 2004) for the interval of positive inclinations between 218 and 220 m CSF. Under this assumption, the polarity change at ~320 m CSF could be tentatively correlated to the bottom of Chron C3n.4n (Fig. F35). Nevertheless, the magnetostratigraphic record below ~220 m CSF is rather noisy, preventing any satisfactory correlation to the GPTS at this time. The correlation suggested here implies that the sandy unit was deposited sometime between ~2 and ~5 Ma (Fig. F36) and that the gross sediment accumulation rate in the upper ~200 m is ~10 cm/k.y.

Top of page   |    Previous   |    Next