Previous   |    Next

doi:10.2204/iodp.proc.314315316.216.2012

Results

Bulk analysis

Although variations in the anisotropy of a physical property may often, at least qualitatively, be attributed to strain, subtle lithologic changes may greatly affect observed anisotropies through contrasting contributions from individual components. Upon changes in their relative amounts, a binary mixture of two components with contrasting intrinsic properties and anisotropies may cause fluctuations of the total effective anisotropy without this anisotropy being at all related to strain. In the case of magnetic susceptibility, because K_m is usually considered a proxy for rock composition, comparison between K_m and P_j can serve as a test of the mineralogical control over magnetic anisotropy (Borradaile and Henry, 1997). Figure F3A is a semilog plot of P_j vs. K_m for all of our samples. Overall, the values for K_m vary considerably from ~80 × 10^–6 to ~5000 × 10^–6, and an apparent positive correlation was observed past a transition range between 200 × 10^–6 and 300 × 10^–6, below which no systematic effect of K_m over P_j is apparent. The high susceptibility values observed in some units at Sites C0006, C0007, C0008, and to a lesser extent at Site C0004, were attributed to the presence of numerous volcanic ash and sand layers (see the “Site C0004,” “Site C0006,” “Site C0007,” and “Site C0008” chapters [Expedition 316 Scientists, 2009a, 2009b, 2009c, 2009d]). These variations in magnetic susceptibility, which were found to closely parallel natural magnetic remanence, are more specifically caused by variable contents of magnetite (Kitamura et al., 2010). The P_j vs. K_m semilog plot suggests a value of ~250 × 10^–6 for the divide between paramagnetic-dominated and ferromagnetic-influenced signals, which is consistent with reference data obtained for paramagnetic rocks and minerals (Rochette et al., 1992; Borradaile and Henry, 1997; Martín-Hernández and Hirt, 2003).

The same plot as that drawn for magnetic susceptibility is shown in Figure F3B for P-wave velocity data. In this case, one could expect P-wave velocity anisotropy to show contrasting sensitivity to various components, a good example of which would be the presence of a network of microcracks. For our samples, however, no particular pattern is evident. Except for Site C0002, all data remain in the same anisotropy range regardless of the average P-wave velocity measured on the sample.

Physical property fabrics

A common approach to interpreting AMS ellipsoids in terms of strain consists of plotting Jelinek parameters P_j and T_j against one another (Tarling and Hrouda, 1993). Whereas the eccentricity value provided by P_j is an indication of the extent to which a given fabric is developed, the shape parameter T_j, which is basically the relative difference between the foliation parameter F = K₂/K₃ and the lineation parameter L = K₁/K₂, defines that fabric as either linear (prolate) or planar (oblate). For sedimentary rocks, and mud rocks in particular, progressive overprinting of a tectonic fabric on an initial sedimentary fabric is often qualitatively described by a path initiating at the top of the oblate field (T_j ~ 1), decreasing in both P_j and T_j to reflect the destruction of the initial sedimentary fabric by subhorizontal shortening and then, after reaching a minimum in P_j and T_j, increasing for both parameters with the formation of a new foliation oriented perpendicular to the initial one (Borradaile and Henry, 1997; Parés, 2004). Such a T_j vs. P_j plot is shown in Figure F4A for our AMS data. Overall, the two parameters seem to roughly correlate, although the artifact on P_j associated with variable content in magnetite considerably affects the plot and precludes comparison between drilled sites in term of strain intensity.

We show in Figure F4B a T_j vs. P_j plot for P-wave velocity data. As for AMS, most data are located in the oblate domain. Values from all sites seem, however, to participate in the same initial stage of the strain path defined earlier (joint decrease of the two parameters initiating from a highly oblate fabric). Three zones can be distinguished within this plot: (1) a domain that comprises mostly samples from Site C0002 with P-wave anisotropy > 1.2 and clearly oblate fabrics with T_j > 0.4, (2) a domain where most samples from other sites have anisotropy values of P_j ≤ 1.2 and still oblate fabrics with 0 ≤ T_j ≤ 0.6, and (3) a low-anisotropy and prolate fabric domain with P_j ≤ 1.2 and T_j < 0. Following similar reasoning as for AMS, the first domain seems to correspond to samples that have been vertically compacted with low net disturbance by horizontal shortening. On the other end of the path, where the values of T_j become negative, samples may be considered to be most affected by horizontal shortening. Although such representation and corresponding interpretation have, to our knowledge, never been applied to P-wave velocity data, one can expect from APV a response similar to AMS in poorly consolidated material with high porosity. In the samples studied here, two major sources of P-wave anisotropy should be taken into account: the degree of alignment of the constituting phyllosilicates (see Johnston and Christensen, 1995) and the preferential orientation of pore space (an example of such an effect can be found in Louis et al., 2003). Therefore, the conceptual mechanisms of rotation of platy particles under uniaxial stress essentially apply to both AMS and APV, with the difference that preferential orientation of the pore space may cause the P-wave velocity to exhibit higher anisotropies than AMS, as seems to occur in Figures F3 and F4. The fact that (1) both assumed sources of anisotropy for P-wave velocity are expected to evolve in conjunction upon vertical compaction or horizontal shortening and (2) from a poroelastic viewpoint the microstructure is devoid of features capable of overriding the background signal, as is the case for the magnetic mineralogy, suggests that the conceptual strain path may in fact be better fitted with APV data than with AMS data when considering the entire structure. In this framework, Figure F4B suggests gradual progression from globally compacted sediments (of which the sample batch at Site C0002 is a good example) to material increasingly affected by lateral shortening. Using the terminology of Borradaile and Henry (1997), samples from NanTroSEIZE Expeditions 315 and 316 exhibit sedimentary to transitional fabrics.

Tectonic implications

Figure F5 depicts equal-area lower hemisphere projections of the AMS and APV eigenvectors corresponding to the maximum (solid black squares) and minimum (solid red circles) values after paleomagnetic reorientation for all sites. Sample groups are generally defined according to the lithologic units, except for Sites C0004 and C0006, where subgroups were used. For every (sub)group, the depth interval covered by the samples is reported on the top left corner of the AMS plot. On each plot, the average value of T_j is also provided, together with the direction of maximum horizontal stress derived for the same intervals by Chang et al. (2010) using borehole breakouts, when available. Open symbols for eigenvectors correspond to samples with either very low anisotropy or evident misorientation for both AMS and APV. Some intervals such as the hanging wall of the splay fault at Site C0004 could not be included because samples were obtained from randomly oriented core fragments.

Using AMS plots as a reference, three types of fabrics can be observed: sedimentary, transitional with sedimentary relicts (K₁ oriented parallel to the structural axis and K₃ oriented perpendicular to the bedding), and transitional with K₃ axis rotating toward the shortening direction. Not all projections exhibit straightforward patterns. However, some general observations can be made concerning their apparent relationship to deformation. At Sites C0002 and C0008, slope sediments display essentially sedimentary fabrics with subvertical K₃ directions and no clear grouping of K₁ directions. On the opposite end of the spectrum, most evolved fabrics are found at Site C0001 in the hanging wall of the megasplay fault and within the Shikoku Basin facies sediments at Sites C0006 and C0007 (Unit III). This last observation is consistent with the findings of Kitamura et al. (2010). Finally, intermediate fabrics with K₁ grouping approximately parallel to the structures and subvertical K₃ directions are observed in the underthrust sediments at Site C0004 and in Unit II of Sites C0006 and C0007. These fabrics are virtually identical to the ones observed by Owens (1993) at Ocean Drilling Program (ODP) Site 808. In all cases where clustering of K₁ is observed, there is a good geometric agreement with the maximum horizontal stress direction of Chang et al. (2010).

The one-to-one comparison between AMS and APV plots is inconclusive at this stage. Generally, minimum P-wave velocities are found along the vertical direction, and in some cases, clustering of the maximum velocity (V₁) directions is observed (Site C0004 and mid-interval at Site C0006), albeit with certain angular offsets with respect to AMS, that could be related to the presence of some localized deformation feature. Although this result suggests that APV may not be a strong directional marker, the T_j values obtained from APV and AMS for the intervals shown in Figure F5 are in good agreement, as shown in Figure F6. Here, T_j values are averaged per interval with standard deviation for the error bar. The values for P_j were not used because of their observed sensitivity to the magnetic mineralogy (Fig. F4A). In Figure F6, T_j values for APV and AMS data are located close to the 1:1 correlation line. The zones most likely affected by lateral compaction at Site C0001 (megasplay hanging wall) and in the Shikoku Basin facies sediments of Sites C0006 and C0007 (frontal thrust), are confirmed by low average T_j values for both AMS and APV. In contrast, the two Site C0002 units (solid red circles) on top of the plotted data are apparently the ones with the most preserved sedimentary fabrics. Comparison of values for Site C0004 underthrust sediments with Site C0008 slope sediments suggests that splay fault activity did not result in further vertical compaction. On the contrary, the shallower of the two Site C0004 subunits exhibits a relatively low T_j value suggestive of horizontal shortening, which might be due to presently acting horizontal stress underneath the megasplay.

Top of page   |    Previous   |    Next