Previous   |   Next

doi:10.2204/iodp.proc.301.106.2005

Paleomagnetism

There were several objectives for paleomagnetic measurements at Site U1301. For the sedimentary section, the goal was simply to characterize magnetism within the cores for the initial reports. In the igneous section, one goal was to measure the variation in magnetic directions, which may result from several sources, including geomagnetic secular variation, magnetic reversals, and tectonic processes in the crust. Another goal was to look at variations in magnetic characteristics, which may result from differences in basalt composition or alteration.

Sedimentary section

There are two limitations on the usefulness of paleomagnetic measurements from the sedimentary section of Site U1301: lithology and age. The section did not promise good paleomagnetic results because the sediments consist of a mix of hemipelagic mud, sand turbidites, silt turbidites, and sandy debris flow deposits with mud clasts (Shipboard Scientific Party, 1997) (see "Lithostratigraphy"). All but the hemipelagic mud are poor candidates for producing good paleomagnetic data. In part, this is because coarse-grained sediments are poor paleomagnetic recorders; also, coarse-grained sediments are difficult to core, frequently resulting in less than complete cores, often with liquefaction and redeposition during coring and recovery. Another problem is that the section was not continuously cored, which breaks the continuity of the record and makes interpretation more difficult. Furthermore, most of the section is <1 Ma in age, so no significant change in paleomagnetic inclination owing to plate motion should have occurred and only one magnetic reversal, at the base of the Brunhes Chron, is to be expected. Finally, the cores were not oriented using the Tensor tool because of time constraints. The actual measurements lived up (or down, as the case may be) to expectations.

Measurements were made with the pass-through cryogenic magnetometer at 5 cm intervals on all archive-half core sections that appeared to have intact stratigraphy. Many sand cores appeared to have liquefied and settled inside the core liner; these were not measured. In addition to the natural remanent magnetization (NRM), the cores were also measured after alternating-field (AF) demagnetization at 10, 20, 30, and 40 mT levels.

Initial NRM measurements typically showed steep inclination values, approaching 90°, with consistent declinations near 0°. This situation is typical of cores that are overprinted with a drill string induced isothermal remanent magnetization and sidewall core deformation (Acton et al., 2002). AF demagnetization produced data that appear less affected. With paleomagnetic inclinations plotted versus depth (Fig. F42), it can be seen that most measurements give positive inclinations around 40°–80°, with considerable scatter. This result is consistent with the expectation that the entire sedimentary section was deposited during the Brunhes Chron. Scattered negative inclinations and other inclinations near zero are likely spurious data points caused by internal core deformation, rather than actual reversed or transitional field directions. However, near the bottom of the hole, in Core 301-U1301C-17H, is a zone of consistent negative inclinations that may indicate reversed polarity at the top of the Matuyama Chron. This finding is consistent with ages of sediments determined from prior drilling, which suggest that the Brunhes/Matuyama boundary should occur around 200 mbsf (Shipboard Scientific Party, 1997). Given the poor stratigraphic control, high paleomagnetic data scatter, and poor continuity, it is impossible to know whether this apparent reversed section is a true record of the paleomagnetic field.

NRM intensities of Hole U1301C sediments are quite high for sediments, with values as high as 1.7 A/m (Fig. F43). The lowest reliable NRM intensity is 2.1 × 10^–2 A/m, so the intensity range is relatively small, covering only about two orders of magnitude. The strong intensities are likely a result of the sediment sources, which include the Columbia River basalts and Cascade volcanics, and are therefore rich in magnetic minerals. The high scatter in intensities plotted versus depth are probably the result of the sediment transport mechanisms, which are episodic, high energy, and unlikely to produce uniform sediments.

Igneous section

Because the igneous section was rotary cored, producing generally small, often unoriented pieces, whole igneous core sections were not measured with the pass-through cryogenic magnetometer. This produces a situation where the magnetic fields of several small, variously oriented rock fragments can combine within the 30 cm sensing region of the pass-through magnetometer and result in spurious measurements. Instead, measurements were made on 286 discrete samples (Table T9). Most samples were 6–10 cm³ cubes cut from working-half core rock pieces with straight sides that indicate vertical orientation. In all, 158 cube samples were measured, including 23 samples taken by physical properties scientists, which were used to increase the number of paleomagnetic data points. To keep from affecting the physical properties of these samples, magnetic cleaning was restricted to AF demagnetization. Additional data were derived by measuring oriented pieces borrowed from the archive-half cores (128 total samples). Archive-half samples were irregular in volume and shape because cutting or modifying the archive rock pieces was not permitted. All such samples had straight vertical sides, indicating vertical orientation, and the sample volumes were ~50–200 cm³. Such irregular pieces are not often used in paleomagnetic studies because of concern that the irregular shapes may affect the accuracy of measurements rendered by the magnetometer. However, these measurements were considered useful in the current study because the recovery was often low and measurements were possible on many pieces that were either sampled by other investigators in the working half or were unsuitable for sampling because of fractures and cracks. Because these rock pieces are part of an archive, heating was not permitted, so AF demagnetization was used.

Most AF demagnetization was done using the cryogenic magnetometer's inline AF coils. The usual routine was to use 5 mT steps from 10 to 50 mT, with an additional step at 60 mT. For thermal demagnetization, 50°C steps were used from 100° or 150°C up to 350°–375°C, with 25°C steps up to 500°–550°C. The top end of the thermal demagnetization as well as the beginning of 25°C steps was changed slightly during the course of the study, depending on prior results. Characteristic remanent magnetization (ChRM) directions (i.e., the magnetization believed to be the thermal remanent magnetization [TRM] acquired initially upon cooling of the igneous rocks) was determined for each sample by examining orthogonal vector plots of demagnetization steps and looking for a consistent direction showing univectorial decay toward the plot origin. If such a vector was found, its mean direction was calculated using a least-squares line fit (principal component analysis [PCA]) of Kirschvink (1980).

NRM values for the igneous samples range from 1.33 to 60.81 A/m (Table T9), values consistent with other young ocean crustal basalts (e.g., Johnson et al., 1996). Demagnetization results showed that many samples have the ubiquitous steep, downward-directed overprint that is usually attributed to an isothermal remanent magnetization imparted by the drill string (Acton et al., 2002). This remagnetization does not significantly affect all samples but probably inflates NRM values for those with low-coercivity magnetic grains.

Orthogonal vector plots show that Hole U1301B basalts have a complex set of demagnetization behaviors (Fig. F44). Some samples have simple, single magnetizations, consisting of the ChRM only, which shows univectorial decay to the orthogonal vector plot origin (Fig. F44B). Some others have similar ChRM with a variable amount of steep, downward-directed overprint, likely a result of exposure to high magnetic fields in the drill string and coring equipment (Fig. F44A, F44C, F44E, F44F). With many samples, however, the magnetization appears to have several components and complex demagnetization curves are common. Some samples appear to carry a high-coercivity magnetization that is not easily removed by AF demagnetization. Indeed, many such samples show low demagnetization step vectors converging on the origin, only to veer away at the highest steps (Fig. F44D). Furthermore, with thermal demagnetization, many samples do not stabilize on a stable ChRM direction until high temperatures, in excess of 450°–500°C (Fig. F44E).

Thermal demagnetization curves suggest that the complex demagnetization data may result from the samples having several different types of magnetic grains. In Figure F45, curve A shows a sample magnetization that is not reduced significantly until high temperature is applied. The trajectory of this curve and its rapid decay to near zero at 550°C is consistent with relatively pure magnetite. Curve C shows a sample whose magnetization is nearly gone by 300°C, consistent with impure magnetite (i.e., containing titanium in its crystal lattice) or another mineral with a lower Curie temperature. Curve B has two plateaus, implying two different populations of magnetic grains with different Curie temperatures. Curve D shows an increase in magnetization between 200° and 250°C, which is a characteristic of the mineral pyrrhotite, a metastable iron sulfide that becomes more magnetic upon heating in that range, but whose magnetization is destroyed upon heating to 300°C (Thompson and Oldfield, 1986). In sum, the magnetizations of Hole U1301B basalt samples are variable and appear to be the result of several types of magnetic grains. This should come as little surprise because the basalts show extensive evidence of variable hydrothermal alteration (see "Igneous and metamorphic petrology").

An initial plot of ChRM inclination values versus depth (Table T9) showed considerable scatter, thought to result from including samples with poorly defined magnetizations. To distinguish between more and less reliable sample magnetizations, the samples were grouped by their demagnetization behavior. The most reliable samples are those that show consistent, univectorial decay toward the origin of an orthogonal vector plot. These were labeled A1 for the most consistent samples and A2 for those with fewer consistent steps or higher scatter between demagnetization steps (see Table T9). Samples that displayed univectorial decay, but with ChRM vectors that did not converge directly on the origin, were labeled class B. Those samples that were close to converging on the orthogonal vector plot origin were labeled B1, but those that missed significantly were labeled B2 (Table T9). Samples whose magnetizations veered away from the orthogonal vector plot origin at high demagnetization steps were labeled class C, with those that passed close to the origin labeled C1 and those that missed by a wide margin or trended away labeled C2. Samples with irregular, inconsistent demagnetization steps were labeled class D. Classes A1 and A2, considered the most reliable, compose ~57% of the samples. Classes B1 and C1, those that were close to ideal univectorial behavior, are another 14%. Samples in classes B2, C2, and D were considered too unreliable for inclusion in magnetic field inclination interpretations.

Even when unreliable samples are removed, the plot of inclination versus depth still shows considerable scatter, especially in the lower part of the cored section (Fig. F46). In the section of the hole covered by Cores 301-U1301B-1R to 15R (~350 to ~450 mbsf), inclination values cluster, with most samples giving values between 30° and 70° (Fig. F46). From Core 301-U1301B-16R and downward, many samples give similar inclination values but many others yield lower inclinations. In addition, many samples produced negative inclinations. Below 450 mbsf, 29 samples have negative inclinations, whereas only 5 occur above that depth (Fig. F46). Many of the recovered core pieces are small, so it is likely that a few negative inclinations reflect accidental inversion of core pieces during handling. However, the large number below 450 mbsf is too many to be explained simply by such mishaps. Furthermore, the direction of the low-coercivity, downward-pointing overprint, probably imparted by the drill string, should point upward if a sample were inverted. Though inconclusive for some samples without much overprint, virtually all of those with steep overprints and negative inclinations have a downward-pointing overprint. Thus, the likeliest conclusion is that most of the negative inclinations reflect the true magnetization direction of the rocks.

How could the rocks become magnetized with a negative inclination? Magnetic reversals are an obvious possibility; however, this is an untenable explanation because the negative-inclination samples are interspersed with positive-inclination samples in many places. One would have to postulate a multitude of short periods of magnetic field reversal or an uncharacteristically long time of basalt emplacement that would include millions of years of reversal record. Neither fits with current knowledge of the recent geomagnetic reversal record or crustal formation. Whatever causes these samples to have negative inclinations must be specific to the rocks themselves.

Given that constraint, two other possibilities are remagnetization or self-reversal. Self-reversal is a process through which basalts of certain compositions record a magnetization direction that is opposite to that of the prevailing geomagnetic field (Verhoogen, 1956). Although self-reversal is thought to be relatively rare, occurring in a restricted range of maghemite compositions, recent research suggests it may be more prevalent than normally regarded (Doubrovine and Tarduno, 2004). Without further tests, this possibility cannot be ruled out. Curiously, of nine sets of matched samples from the same rock pieces, three give negative magnetizations through thermal demagnetization and positive inclinations through AF demagnetization (Fig. F44A, F44E; Samples 301-U1301B-30R-1, 98–100 cm, and 30R-1, 99–101 cm). All of these are comparisons of working-half core cube samples with archive-half irregular samples, so the possibility of accidental sample inversion during handling cannot be ruled out.

Remagnetization can occur through heating or chemical changes within a rock that changes the magnetic minerals. Although alteration changes the magnetic minerals in most crustal basalts (e.g., Johnson et al., 1996), usually the new magnetization has the same direction as the old as long as the remagnetization occurs rapidly. However, a reversed magnetization could occur if the alteration and magnetic mineral replacement occurs during a period of time with an opposite magnetic polarity. This seems the most likely explanation for Hole U1301B samples because geologic observations indicate extensive hydrothermal alteration (see "Igneous and metamorphic petrology") and because shipboard paleomagnetic studies point to multiple magnetization components as well as the presence of pyrrhotite in some samples. Pyrrhotite is a mineral that is a common byproduct of the dissolution of magnetic minerals, such as magnetite, and the conversion of the iron into iron sulfide minerals (Thompson and Oldfield, 1986).

If the remagnetization interpretation is correct, then the negative inclinations show zones where greater alteration has occurred. Looking at Figure F46, it appears that most samples with negative inclinations occur in two depth bands. The upper one spans 460 to 500 mbsf (Fig. F46), and the lower spans from ~515 mbsf to the bottom core at ~577 mbsf. It is difficult to determine whether the apparent gap between the two zones is real or an artifact of core recovery. Interestingly, the negative inclinations in the upper band have lower absolute values (10°–40°), whereas those in the lower band are generally higher (up to ~80°). This difference could be explained by the two sections having been remagnetized at different times, when the magnetic field was reversed, but at different inclinations owing to magnetic field secular variation. Another indicator that supports two bands of more intense alteration is the stratigraphic positions of unreliable samples (Fig. F46). Although poor-quality samples are scattered throughout the entire igneous section, there are clusters of them around 470–510 mbsf and 550–560 mbsf. The clustering of these samples partly results from sampling bias, but they also reflect zones where samples with poor quality occur in greater abundance. The overall implication of these arguments is that the upper part of the section has experienced less magnetic alteration, whereas the lower has been more affected, with perhaps two zones in the lower part of the section in which the greatest alteration has occurred.

Interestingly, the inclinations that are from apparently reliable samples give inclinations that tend to be shallower than expected for the latitude of the site (dashed lines, Fig. F46). This is true even when the expected inclination is corrected to account for the fact that the average inclination derived from azimuthally unoriented core samples is less than the true inclination (Cox and Gordon, 1984). The actual geocentric axial dipole inclination (expected value) for Site U1301 is 66.5°. However, applying the correction for inclination-only data bias reduces the expected value only to 64.2°. The average inclination of the samples above 450 mbsf, where the apparent remagnetization has been least, is only 55.7°, a difference of nearly 9°. The cause of this discrepancy is unclear and requires further research.

Top of page   |   Previous   |   Next