Paleomagnetic and rock magnetic measurements at Site 1205 on Nintoku Seamount focused on assessing the nature of the natural remanent magnetization (NRM) of basement rock and obtaining a preliminary estimate of the paleolatitude of the site during basement formation (a minimum of 55 Ma, as estimated by nannofossil biostratigraphy) (see "Biostratigraphy"). The basement section at Site 1205 is dominated by basaltic lava flows with some units separated by soil horizons or weathered flow tops (see "Physical Volcanology and Igneous Petrology"). Fossiliferous sandstone was recovered in Cores 197-1205A-1R through 4R. The sandstone/basalt contact is represented by a basalt cobble conglomerate. Because of limited recovery, paleomagnetic analysis of the sandstone and conglomerate units was reserved for shore-based study. All shipboard samples for paleomagnetic investigations were collected as drilled minicores.
Low-field volume-normalized magnetic susceptibility was measured on all oriented samples after alternating-field (AF) demagnetization with the Kappabridge KLY-2 magnetic susceptibility meter. In addition to oriented samples, 24 unoriented core chips (~1-5 g/cm3 in size) were measured. Three measurements were taken for each sample and then averaged. Measurements were repeatable to within 1% of the average value. The range of magnetic susceptibilities is 0.6 x 10-3 to 84.4 x 10-3 SI, with an arithmetic mean of 27.5 x 10-3 SI.
The Koenigsberger ratio (Q) was calculated for each sample using a present geomagnetic field intensity of 34.2 A/m (International Geomagnetic Reference Field; Barton et al., 1995). Koenigsberger ratios range from 0.7 to 80.0 (mean = 8.8). Median destructive field (MDF), calculated based on detailed stepwise AF demagnetization data used for constraining characteristic remanent directions (see below), range from 3.1 to 63.3 mT (mean = 21.6 mT).
Figure F27 shows a plot of volume magnetic susceptibility (K), the Koenigsberger ratio (Q), and MDF vs. depth. A peak in Q-ratio values corresponds to basement Unit 10, a sparsely plagioclase-olivine phyric to aphyric basalt lava flow with little alteration. A peak in MDF values corresponds to a number of unoriented oxidized pebbles from Unit 9 that were not used for paleomagnetic studies (see below). These samples also have a relatively low Q-ratio and magnetic susceptibility, possibly corresponding to slight maghemitization of iron oxides in this unit (see "Physical Volcanology and Igneous Petrology").
An estimation of magnetic domain state can be made using the Lowrie-Fuller test (Lowrie and Fuller, 1971). Generally, multidomain grains require larger destructive fields to remove strong-field thermoremanent magnetization (TRM) than a weak-field TRM. The strong-field TRM is approximated with a saturation isothermal remanent magnetization (SIRM); the weak field is approximated with an anhysteretic remanent magnetization (ARM). A caveat for use of this test, however, is that the magnetic mineralogy is uniform. Changes in magnetic coercivity related to different magnetic minerals can mimic changes in magnetic domain size. Isothermal remanent magnetization (IRM) acquisition and backfield IRM acquisition were also measured to estimate the coercivity of remanence of the samples (Table T9).
Samples displaying single-domain (SD) magnetic behavior generally have ARM and SIRM demagnetization curves as shown in Figure F28C, F28D, and F28E. They are characterized by high coercive force (Hcr) and MDF values (Fig. F29C, F29D, F29E). Rapid decay of NRM, ARM, and SIRM and low values of Hcr usually indicate multidomain-like behavior (Fig. F29A, F29B, F29F).
In some samples (e.g., Samples 197-1205A-20R-6, 12-14 cm, and 35R-2, 36-38 cm) (Fig. F30), the NRM has an SD-like decay, whereas the ARM and SIRM demagnetization curves suggest that the behavior should be multidomain. The presence of superparamagnetic (SP) grains with sizes close to the SD/SP threshold (~30-50 nm) (Dunlop and Özdemir, 1997) might explain this discrepancy. In the presence of a magnetic field, the magnetization vector in an SD grain will rotate to align with the magnetic field. When the field is removed, the remanent magnetization will decay with time. The time it takes for an assemblage of SD grains to decay is called the relaxation time. Superparamagnetic grains have very short relaxation times compared to stable SD grains and typically carry an unstable magnetization in the presence of a magnetizing field. The rapid decay of the ARM and SIRM curves for the samples mentioned above might represent the unstable acquisition of magnetization in the SP grains. An alternative explanation for the data is that some samples have a mixture of large multidomain and single-domain grains. Detailed hysteresis measurements planned for shore-based study should help distinguish between these scenarios.
Two types of AF demagnetization behavior observed in analyses of samples at this site merit additional rock magnetic consideration. Univectorial decay of remanence (e.g., Fig. F31F) is often observed, but samples with relatively large magnetically soft components (which are readily demagnetized up to 20 mT) were also seen (e.g., Fig. F31A, F31D). As the stable remanence of this site is characterized by a negative inclination, the NRM intensity may be reduced if the soft component is oriented in a direction roughly antiparallel to the primary remanence, as would be expected for a present-day viscous or Brunhes field overprint (e.g., Cottrell and Tarduno, in press). In this case, the decay curve of remanence with AF demagnetization will not be monotonic, and, therefore, the MDF may not be a useful measure of the hardness of magnetic remanence.
We calculate the vector difference between peak field AF demagnetization steps (here called RM) to show the portion of the magnetically soft component relative to the intermediate coercive component as follows:
We selected the interval between 20 and 50 mT as a basis for comparison because the stable remanent component was often isolated in this interval of AF demagnetization in samples from Site 1205. Figure F32 shows a positive correlation between the logarithm of RM and the compressional wave velocity (VP) measured on paleomagnetic samples with a PWS3 contact probe system (see "Physical Properties"). As VP is a function of density and rigidity of the sample, this may indicate the degree of void space and chemical alteration, both of which may correlate with the amount of clay minerals. Samples having a relatively large magnetically soft component as characterized by the RM factor appear to be less altered. Thermal demagnetization studies are needed to determine the magnetic unblocking temperatures and, hence, the magnetic mineralogy of the samples. Nevertheless, the degree of alteration appears to control the demagnetization behavior of some of the samples.
In general, the Q-ratios, median destructive field values, and results of the Lowrie-Fuller tests suggest that most of the recovered basalt samples preserve a primary magnetization suitable for paleolatitude analyses. The magnetic mineralogy of the Site 1205 basalt samples does not appear to be as variable or as altered as that of Sites 1203 and 1204 (see "Physical Volcanology and Igneous Petrology", and "Alteration and Weathering").
Twenty-five lava flows were recovered and sampled for paleomagnetic analysis. For most lava flows, four or more samples were measured. Where limited recovery occurred or a thin unit was cored (Units 1, 14b, 15b, 21, 22, 28a, 29b, and 30b), fewer samples were taken for shipboard analysis so that sufficient material would be available for high-resolution shore-based studies. Sample inclinations for each unit were averaged using the method of McFadden and Reid (1982).
Minicore samples of basalt from Site 1205 were measured using the 2-G Enterprises superconducting quantum interference device magnetometer. After the measurement of the untreated NRM, samples were progressively demagnetized by AF demagnetization. A 5-mT field increment was used between 5 and 70 mT. In addition, 80 mT was measured. A total of 152 basalt minicores were drilled.
Orthogonal vector diagrams were used to characterize the demagnetization behavior and assess the number of magnetization components represented in the Site 1205 basalt samples. With few exceptions, samples showed stable demagnetization curves. Approximately one-third of the samples had a large normal polarity overprint that was removed after demagnetization to 25 mT (Fig. F33D, F33E). For 45% of the samples, AF demagnetization to 80 mT resulted in the reduction of intensity to <10% of the NRM (Fig. F31). Approximately 50% of samples were not demagnetized to <10% of the NRM intensity after treatment to 80 mT. Nevertheless, the remanence directions after each demagnetization step define a clear trend toward the origin of orthogonal vector diagrams (Fig. F34). We attribute the failure of AF treatment at relatively high peak fields to adequately demagnetize the NRM to the presence of high-coercivity minerals. In some samples (Fig. F33C, F33F), the demagnetization vectors are curved at low-field steps, suggesting that two magnetic components are being demagnetized simultaneously. Five percent of samples did not have a stable remanence that could be fit with principal component analysis (Fig. F33A, F33B).
All samples analyzed were of reversed polarity; one apparent normal polarity sample (197-1205A-7R-2, 43-45 cm) was taken from a core catcher sample. The core piece may have been accidently inverted. The sedimentary column above the basalt flows is in nannofossil Zone NP10, which has an age range of ~53.5 to ~55 Ma (Berggren et al., 1995), corresponding to Chron 24r (Cande and Kent, 1995). Characteristic remanent magnetization directions were fit using principal component analysis (Kirschvink, 1980). In general, the characteristic remanent directions were defined between 20 and 70-80 mT for basalt samples that did not show a large normal polarity overprint. Samples with a large overprint were generally fit between 35 and 70-80 mT. Maximum angular deviations of line fits to the demagnetization data were generally <5°-10°.
Previous drilling on Nintoku Seamount at DSDP Site 432 penetrated 38 m of sandstone and basalt (Jackson, Koizumi, et al., 1980). Because Site 1205 is located ~100 m from Site 432, we can compare our results with those of Kono (1980), who analyzed the paleomagnetic record of DSDP Site 432. The first two basalt units identified at DSDP Site 432, which are the equivalent of a single basement unit at Site 1205 (see "Physical Volcanology and Igneous Petrology"), have inclination values of -65.2° and -66.4°. These yield a mean based on 13 samples that is shallower than that derived from the equivalent basement unit at Site 1205 (-72.0°), although the latter is based on only three samples and has a high uncertainty (±17.5°). We also note that the inclination values for DSDP Site 432 basalt flows were derived from the demagnetization step with minimum dispersion in inclination (Kono, 1980) rather than by principal component analysis. Unit 3 from DSDP Site 432 yielded an inclination (-28.6°) similar to that observed from the equivalent basement unit at Site 1205 (-29.5°).
Lava flows at Site 1205 that are separated by flow tops rather than soils (Units 15b, 16, and 17) (see "Physical Volcanology and Igneous Petrology") may have been emplaced during relatively short periods with respect to time scales typical of secular variation (see "Paleomagnetism and Rock Magnetism" in the "Explanatory Notes" chapter). We combine paleomagnetic data from such lava flows and treat other lava flows separated by soils as independent time units. This results in 22 paleomagnetic inclination groups (Table T10), which together (following the averaging method of McFadden and Reid, 1982) yield a mean inclination of -45.7° (95% confidence interval = +10.5°/-6.3°). The mean inclination indicates a paleolatitude of 27.1° (+10.8°/-5.2°).
The estimated polar angular dispersion (S = 20.4°) (Cox, 1970; Tarduno and Sager, 1995) suggests that secular variation has been sampled. The value derived from the Site 1205 data is greater than expected using global reference curves (McFadden et al., 1991), which predict a value of S = 13.4° (+1.6°/-1.1°). The high dispersion is also apparent in the rather broad distribution of the inclination unit means (Fig. F35).
Several factors could account for the relatively high observed inclination dispersion. The most likely explanation, however, is that AF demagnetization is not adequately defining the characteristic remanent magnetization for some samples. This is supported by the rock magnetic and paleomagnetic evidence for high-coercivity magnetic minerals in some samples and the relatively high directional dispersion in some inclination units (Table T10). We also note that three inclination groups in our preliminary analysis are represented by only one or two samples because material was limited (Table T10). Shore-based analyses using thermal demagnetization should address these limitations.
The preliminary paleolatitude derived from the Site 1205 basalt samples differs from the present-day latitude of the Hawaiian hotspot (~19°). The data are similar to those obtained from paleomagnetic analysis of 65-Ma lava flows sampled at Suiko Seamount (DSDP Site 433) (Kono, 1980), a data set believed to average secular variation. Our data are preliminary, and therefore a determination of whether the paleolatitude of Nintoku Seamount during its formation is offset from the paleolatitude of Suiko Seamount is not possible; we expect the answer to this question will come from shore-based work employing thermal demagnetization procedures. Nevertheless, the paleomagnetic results from Site 1205 together with those from Suiko Seamount (Kono, 1980) and Detroit Seamount Sites 884 (Tarduno and Cottrell, 1997; Cottrell and Tarduno, in press), 1203, and 1204 form a consistent data set compatible with southward motion of the Hawaiian hotspot since the Late Cretaceous.