PALEOMAGNETISM

Pass-through cryogenic magnetometer measurements were taken on all split-core archive sections and a few discrete samples from working sections. In addition, magnetic susceptibility (MS) was measured for whole cores, all archive-half core sections, and discrete samples. Coherent basalt pieces that could be oriented unambiguously with respect to the top were measured at 1-cm intervals. In order to isolate the characteristic remanent magnetization (ChRM), archive-half core sections were subjected to alternating-field (AF) cleaning. The number of AF demagnetization steps and peak field intensity varied, depending on lithology and the natural remanent magnetization (NRM) intensity. On average, the basalt half cores were treated using a minimum of eight AF steps. The maximum applied field ranged between 40 and 70 mT. A total of 26 minicores (~10.5 cm³ in volume) were taken from the working halves for AF treatment. Of these, five samples were each cut into two subminicores (5.2 cm³ in volume) before AF treatment for the purpose of comparison between AF and thermal treatments. One set of the subdiscrete samples was stepwise thermally demagnetized to 580°C, to determine the directional stability during stepwise demagnetization, to verify the pass-through measurement from the cryogenic magnetometer, and to help identify the magnetic mineralogy. All the other samples were progressively AF demagnetized up to a peak field of 70 mT. We analyzed all results from long-core and discrete samples in Zijderveld and stereoplot diagrams and calculated the ChRM direction using principal component analysis (Kirschvink, 1980). To investigate rock magnetic properties, we also conducted isothermal remanent magnetization (IRM) experiments on some of the previously AF-treated discrete samples followed by thermal demagnetization of the IRM samples.

One of the major experimental requirements in paleomagnetic research is to isolate the ChRM by selective removal of secondary magnetization. In Figure F27, we show the NRM intensities and inclinations before and after 25-mT AF demagnetization for all the recovered cores. All split-core and discrete samples exhibited strong signs of drilling-induced remagnetization (overprint), as shown by the NRM inclinations that are strongly biased toward relatively high positive values (with a mean of +40°), which are significantly higher than that expected for the actual geocentric axial dipole inclination at Site 1243 (~+10°). The effect of removing this overprint by AF cleaning is a significant decrease in intensity, from a mean value of 3.3 A/m (all measurements without cleaning) to 1.7 A/m (all measurements after 25 mT cleaning) and a shift toward shallower or even negative inclination (Fig. F27). Although the amount and coercivity of overprinting varied, most of it seems to be removed with 25- to 35-mT AF demagnetization for the majority of samples, allowing us to isolate the ChRM direction using the principal component analysis method on higher field demagnetization steps. For rocks from Cores 203-1243B-6R through 7R, the drilling overprint is more resistant to demagnetization, requiring AF treatment up to 70 mT to remove most of the overprint. In some cases the ChRM is not resolved, suggesting that additional AF and thermal demagnetization is needed on these cores at a shore-based laboratory.

After the drilling overprint was removed, the intervals measured generally display linear vector paths that trend toward the origin of the vector plots. In Table T5, we list the ChRM direction, NRM intensity, and the maximum angular deviation angle for all coherent basalt pieces longer than 15 cm. A downhole plot of ChRM inclinations is also shown in Figure F28. The 15-cm threshold was considered to be a reasonable size to allow reliable and consistent estimations of directions and magnetization within a sample, as the first few centimeters of both section ends are affected by edge effects. The edge effects translate into an incorrect assumed volume in the ODP database, as determined by the shipboard software used for NRM calculations and bias in directions detected by the 2-G sensor because of the changing geometry of the sample. For coherent pieces longer than 50 cm, we list data roughly every 5 cm and for smaller pieces every 4 or 2 cm, excluding the edges. The ChRM inclinations obtained from different parts of long coherent pieces generally agree within a few degrees as shown in the detailed results from Section 203-1243B-9R-2 represented on Figure F29.

Because many core sections recovered in Hole 1243B contain small and unoriented pieces, long, continuous, and vertically oriented core sections are limited in number. Consequently, oriented samples for shipboard studies had to be shared between research groups. For this reason, we were able to perform AF demagnetization on only 26 discrete samples and thermal demagnetization on 5 of the subsamples, as mentioned previously. Experiments revealed that samples have two magnetic components, a low-coercivity drilling overprint, which is removed after 20- to 50-mT or 400°C demagnetization, and the ChRM. Several representative examples are shown in Figure F30. It appears that thermal demagnetization is more effective in isolating the ChRM. Judging from the unblocking temperatures, the ChRM is carried by a fine-grained magnetite or titanomagnetite. This is further confirmed by IRM acquisition and thermal demagnetization of IRM experiments (Fig. F31).

The ChRM directions from discrete samples agree roughly with those obtained from split-core measurements (note the 180° change in declination resulting from the change from archive- to working-half cores), but the inclinations are shallower in almost all the cases (see blue dots in Fig. F29). Particularly interesting is the variation in demagnetization behavior within a pillow lava of Unit 6, where a pair of closely spaced samples were targeted for a comparison study. As shown in Figure F32, AF demagnetization of 25 mT on Sample 203-1243B-12R-1, 85-87 cm, removes the drilling-induced overprint and reveals a well-defined ChRM component. The drilling-induced remagnetization component is more resistant to AF demagnetization for Sample 203-1243B-12R-1, 88-90 cm, which is only 1 cm distant from Sample 12R-1, 85-87 cm (see inserted core photo in Fig. F32). The ChRM in this sample (Sample 203-1243B-12R-1, 88-90 cm) was not revealed until after 50-mT demagnetization. We propose that as this sample is farther away from the chilled margin it may contain a higher concentration of multidomain titanomagnetite grains more sensitive to the drilling overprint. This explanation is consistent with observations made on magnetic properties of young MORB (e.g., Carlut and Kent, 2002).

A reliable magnetostratigraphy at Site 1243 is hampered by several factors. First, recovered cores at Site 1243 are azimuthally unoriented; therefore, the declination does not provide direct evidence for magnetic polarity. Second, although it is a common practice to use the magnetic inclination derived from the ChRM as the indicator of magnetic polarity where positive inclinations indicate normal polarity and negative inclinations indicate reversed polarity (for Northern Hemisphere sites), positive and negative inclinations at Site 1243 are not reliable indicators for polarity changes. Because Site 1243 was located only 3°N of the equator 11 m.y. ago (Pisias et al., 1995), the corresponding mean magnetic inclination would change only ~12° from a normal state to a reversed state (i.e., from ~+6° or ~-6°). Yet, the magnitude of ordinary secular variations during a period of stable polarity at that latitude would encompass >12° (see McFadden et al., 1991). Third, there are no independent age constraints available for the recovered basalt; thus, it is impossible to correlate polarity intervals, if any, with the geomagnetic polarity timescale.

Nevertheless, we tried to use a "statistical" approach to identify the magnetic polarity in this study. The 26 AF-demagnetized discrete samples provide the most reliable estimate of ChRM and, when averaged, give a null mean inclination. This tends to demonstrate that the section encompasses periods of reversed and normal polarity (otherwise, given the sample-set size, a mean close to +6° or -6° should be found). Indeed, variations in inclination, with respect to depth, indicate that drilling in Hole 1243B may have penetrated through a period of magnetic field reversal. The magnetic inclinations for cores in aphyric basalt from Unit 3 show normal polarity, with a mean positive inclination of ~20°. In the plagioclase olivine phyric basalt from Units 4, 5, and 6, the ChRM inclinations of discrete samples range from 0° to -20°, suggesting a reversed polarity zone. The fresh to slightly altered aphyric basalt from Unit 7 once again shows normal magnetization with a mean inclination of ~25°. Two observations need further mention:

1. We note that the negative or near-zero ChRM inclinations are not randomly distributed throughout the sequence but are confined to Cores 203-1243B-7R through 13R, within which most long continuous basalt pieces are contained (Fig. F29). Thus, the negative inclinations in these long continuous core sections suggest that these pieces were homogeneously magnetized and lend support to the notion that the ChRM in these core sections may indicate a reversed polarity.
2. As shown in Figure F28, there is a discrepancy between mean inclinations calculated using the discrete samples and long-core measurements. The steeper inclination values observed in the long-core measurements may be due to an overprint that has not been completely removed, to paleosecular variation, or to tectonic rotation. For the last possibility, however, we note that the sediment-basement seismic reflection is horizontal and there is no evidence for local tectonic rotation of the basalts at Site 1243.

Shore-based studies of discrete samples are required to confirm the polarity results and explain the steep inclination in the long-core measurements. One implication of this polarity result, if confirmed, is that the ChRM data probably give a reasonable indication of the time-averaged direction, and the thin igneous section we cored may have erupted in several phases separated by at least several tens of thousands of years.

MS was measured independently at 2-cm intervals with the Bartington meter along whole-core sections of all cores recovered and every 1 cm on the point susceptibility meter. The two MS data sets compare very well with each other (Fig. F33). Susceptibility peaks in lithologic units commonly correlate with lithologic changes. For example, the peak at 166.4 mbsf (Fig. F33) corresponds to a pillow top at Section 203-1243B-13R-1 [Piece 8]. The higher susceptibility value may be explained by a finer grain size of the pillow lava top. Another signal of particular interest is the sharp peak in MS between 128.60 and 130.49 mbsf, corresponding to intervals in Sections 203-1243B-6R-1 and 2. This peak was verified by further individual sample susceptibility measurements from the minicores that were used for demagnetization experiments. Thus, it is not due to contamination from core liners. We note that the susceptibility maximum also coincides with the stronger NRM intensities in the pass-through magnetic measurements. The origin of this spike is unknown at present. Shore-based rock magnetic studies on discrete samples will be used to investigate the cause of the MS anomaly. Interestingly, this may have implications about whether these strongly magnetized rocks may have remagnetized the rocks in the uppermost part of Unit 4 (Fig. F28).

In summary, the basaltic cores recovered from Site 1243 appear to record a stable component of magnetization with possibly both normal and reversed inclinations after removal of the pervasive drilling-induced remagnetization. Shipboard AF and thermal demagnetization studies and comparison between continuous and discrete samples indicate that some of the drilling-induced magnetization may remain on a part of the long-core measurements even at the highest AF demagnetization level (up to 70 mT) and that thermal demagnetization is probably more effective for removing this component. Preliminary data from IRM acquisition experiments, unblocking temperatures, and coercivity determinations suggest that magnetite and titanomagnetite, with varying titanium content, are the most likely magnetic carriers in these cores. The lava sequence recovered at Site 1243 may have recorded a reversal sequence (normal-reversed-normal). These hypotheses will be tested in subsequent shore-based investigations.