Hill, I.A., Taira, A., Firth, J.V., et al., 1993Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 131

18. ACOUSTIC ANISOTROPY AND MICROFABRIC DEVELOPMENT IN ACCRETED SEDIMENTFROM THE NANKAI TROUGH1

W. Brückmann,2 K. Moran,3 and E. Taylor4

ABSTRACT

Complete penetration of frontal thrust and décollement of the Nankai Trough accretionary complex in Site 808 during ODPLeg 131 provided a wealth of structural observations and physical property data. In this paper possible mechanisms are discussedthat could be responsible for the development of irregular downhole trends in acoustic anisotropy observed in Site 808. Aftervarious steps of data reduction and screening, a paleomagnetic reorientation procedure is applied to a selected group of physicalproperty data sets. This facilitates the integration of the observed changes in physical properties with the geotectonic frameworkat the deformation front of the Nankai Trough accretionary complex. The paleomagnetic database was employed in the recon-struction of directional properties of acoustic velocities of the Lower Shikoku Basin sedimentary sequence, which is divided bya sharply defined décollement into an accreting and a subducting portion.

P-wave velocity anisotropies derived from paleomagnetically oriented samples in the upper part of this 420-m-thickhemipelagic sequence show maximum values in the direction parallel and normal to the inferred vector of plate convergence(31O°-315°). No preferred orientation of P-wave velocity anisotropy is found in the subducting part of the sequence.

The preferred direction of maximum anisotropy parallel to the convergence vector is also in accordance with the true directionof the observed macro- to mesoscale structural features in Site 808. Microfractures and microcracks forming as stress relief andtensile fractures in cores of semilithified sediment normal and perpendicular to the maximum horizontal stress are discussed ascontrol mechanisms for the development of the observed anisotropy pattern.

INTRODUCTION

General

The downhole development of anisotropies of various physicalproperties in suites of sedimentary sequences has received a great dealof attention; namely, acoustic anisotropies have been investigatedsince the days of DSDP by a number of workers (e.g., Carlson andChristensen, 1977; Carlson et al., 1983; Fuji, 1981; Wetzel, 1986).The reason for this continuing interest is obvious: if the developmentof acoustic anisotropy in different types of sediment and depositionalenvironments could be easily correlated with specific genetic proc-esses or typical microfabrics, it would represent a very fast andeffective analytical tool for these properties. The term anisotropyrefers to the observed quantitative directional dependence of physicalproperties of sediments. It is well known that most physical propertyanisotropies result from modifications of the pore-to-particle orienta-tion during deposition and subsequent burial. Although this generalrule applies also for acoustic anisotropy, there are a number ofwell-known or suspected second-order mechanisms that modify thissimple case. In the following sections all of these will briefly bereviewed and discussed with respect to trends observed at Site 808.

Acoustic Anisotropy

In marine sediments with predominantly horizontal bedding com-pressional (P-wave) velocities generally show a downhole increasethat is stronger in the direction parallel to the bedding plane, thanperpendicular to it. This so-called acoustic anisotropy, Ap, is com-

1 Hill, I.A., Taira, A., Firth, J.V., et al., 1993. Proc. ODP, Sci. Results, 131: CollegeStation, TX (Ocean Drilling Program).

2GEOMAR—Research Center for Marine Geosciences,Wischhofstr. 1-3, Bldg. 4,D-2300 Kiel 14, Federal Republic of Germany.

3 Geological Survey of Canada, Bedford Inst, of Oceanography Box 1006, Dartmouth,Nova Scotia B2Y 4A2, Canada.

4 Woodward-Clyde Consultants, 3440 Bank of California Bldg., 900 4th Ave., Seattle,WA 98164, U.S.A.

monly defined as the difference between velocities in horizontal andvertical direction expressed as a percentage of the mean velocity:

= 200 × (Vpt-Vp])/(Vpt+Vpl)

where Vpl is the P-wave velocity determined in the direction parallel("longitudinal") to the core axis core, Vpt the P-wave velocity in thehorizontal ("transverse") direction. The positive transverse acousticanisotropy normally ranges—depending on lithology and burial his-tory—from values close to 0% near the surface to more than 12% atdepths of several hundred meters.

A primary acoustic anisotropy will be generated in the course ofdeposition in current-dominated environments (Nacci et al., 1974;0'Brien, 1980). Comparing compositionally similar argillaceoussediments of hemipelagic and turbiditic origin from the MississippiFan, Wetzel (1987) found clear differences in patterns of strain andacoustic anisotropy. Turbiditic sediments display a stronger anisot-ropy parallel to the bedding plane as well as normal to bedding thando hemipelagic sediments. This so-called depositional anisotropy isattributed to the preferred orientation of elongated or oblate grainsduring turbiditic deposition. Acoustic anisotropy was also describedas an intrinsic sedimentary property induced or amplified by alternat-ing thin layers of isotropic or anisotropic materials (Postma, 1955;Bachman, 1979; Carlson et al., 1983), although this phenomenon hasnot been quantitatively observed under natural conditions.

An important mechanism for the generation of acoustic anisotropyis the development of preferred orientation of calcite c-axis normalto bedding in carbonate-rich marine sediments (0'Brien, 1990). Thisresults in acoustic anisotropy, as the compressional P-wave velocity,Vp, is lowest parallel to the c-axis of calcite crystals, and highestparallel to their α-axis. Diagenetic reprecipitation of dissolved calcitenormal to bedding, due, for example, to deformation, is frequentlycited as a possible mechanism to create a uniform orientation of calcitecrystals (Carlson and Christensen, 1979; Milholland et al., 1980).Primary deposition of oriented fragments of biogenic calcite crystalshas been discussed as another likely mechanism (Carlson and Chris-tensen, 1979; Milholland et al., 1980; Schaftenaar and Carlson, 1984;O'Brien, 1990).

221

W. BRUCKMANN, K. MORAN, E. TAYLOR

A dependency of acoustic anisotropy in marine sediments on theoverall content of CaCO3 has also been suspected, although no specificgenetic origin is implied. O'Brien (1990) showed for some calcareousclaystones of ODP Holes 603B and 672A a positive correlation of calcitecontent and acoustic anisotropy, while an inverse trend was found byCarlson et al. (1983) for pelagic chalks and limestones.

Compactional anisotropy—parallel alignment of pores and parti-cles normal to bedding due to gravitational compaction under increas-ing overburden—is commonly assumed to be the most importantsingle source for the downhole increase in acoustic anisotropy infine-grained, clay-rich sediments (Hamilton, 1970; Kim et al., 1983,1985; O'Brien, 1990).

The importance of microfabric control has long been acknow-ledged, but rarely quantitatively evaluated. Ramos and Rathmell(1989) demonstrated how the orientation of microfractures and mi-crocracks controls the spatial configuration of acoustic anisotropy insandstone core samples. By continuously mapping the direction ofmaximum acoustic anisotropy, they were able to verify the directionof maximum in-situ compressive stress, which was also independ-ently determined through other methods.

In summary it can be said that acoustic anisotropy in marine sedi-ments has been identified as a common phenomenon, the sources ofwhich are—although theoretically reasonably well understood—rarelyquantitatively determined or constrained through experimental work.

METHODS

Compressional-wave (P-wave) velocity measurements were ob-tained using two different systems on board JOIDES Resolutionduring Leg 131, depending on the degree of lithification of thesediment. P-wave velocities were measured in softer sediment usinga Dalhousie University/Bedford Institute of Oceanography digitalsound velocimeter (DSV). Velocity calculation is based on the accu-rate measurement of the delay time of an impulsive acoustic signaltraveling between a pair of piezoelectric transducers inserted in thesplit sediment cores. The transducers that were used emit a 2-µssquare wave at about 250 and 750 kHz. The transmitted and receivedsignals were digitized by a Nicolet 320 digital oscilloscope andselection of first-arrival and velocity calculations were performed bya dedicated microcomputer. Two transducers were used, separated byapproximately 7 cm, measuring the vertical (along the core axis) andtransverse (normal to the core axis) P-wave velocity. Periodically, theseparation was precisely evaluated by running a calibration procedurein distilled water. At each sampling interval (normally two per sec-tion), the transducers were carefully inserted into the split section andvelocity measured in both directions.

The Hamilton frame velocimeter (Boyce, 1976) was used tomeasure compressional-wave velocities at 500 kHz in discrete sedi-ment samples when induration made it difficult to insert the DSVtransducers without making any perturbations around them and inlithified sediments and basement rocks when insertion became im-possible. Samples were carefully cut using a double-bladed diamondsaw. Sample thickness was measured directly from the velocimeter-frame lead screw through a linear resistor output to a digital multime-ter. Zero travel times for the velocity transducers were estimated bylinear regression of traveltime vs. distance for a series of aluminumand lucite standards. Filtered seawater was used to improve theacoustic contact between the sample and the transducers. The DSVoscilloscope and processing software were used to digitize wave-forms, to calculate velocities, and to store the waveforms for laterattenuation calculations.

An orientation procedure using paleomagnetic data is describedin great detail for structural data in Table 6 in the "Explanatory Notes"section of Taira et al. (1991). Acoustic anisotropies were derived fromP-wave measurements on physical property samples that were cutwith a double-bladed saw from the working-half of the cores. As thetransverse P-wave velocity was always measured in the direction

normal to the surface of the working-half, it is possible to apply anequivalent orientation procedure.

SHIPBOARD RESULTS

General

Leg 131 was targeted at the Nankai Trough accretionary prism,which represents a type example for convergent margins charac-terized by a high input of clastic sediment. This class of accretionarymargins is the most frequently occurring type among the circum-Pa-cific plate boundaries. The working area (Fig. 1 A) is characterized bya regular tectonic style of imbricate thrusting and the shallow depthof the décollement zone in combination with a high regional heat flow,making it an ideal choice for the evaluation of questions pertaining toclastic prisms in general. Studies on the interdependency of deforma-tional processes, physical properties, structural development, andhydrology in accreting and accreted sediment were among the primegoals of this cruise. Many of these objectives could be accomplishedby drilling Site 808 in the frontal part of the toe of the accretionarywedge (Taira et al., this volume). Site 808 provided—for the first timein the history of ocean drilling—a continuous profile through theaccretionary sequence, the décollement, and the subducting sequenceinto the top of the oceanic crust in the toe region of an accretionarywedge (Fig. IB). One of the most striking features of Site 808 is thestrong dichotomy in structural styles and downhole physical propertytrends above and below the décollement. Many deformational struc-tures covering a wide range of scales were found between frontalthrust and décollement, reflecting the compressional tectonic regimein the frontal deformation zone of the prism. Below the décollementvirtually no signs of deformation could be detected, indicating acomplete mechanical decoupling between the accreting and the sub-ducting portion of the sedimentary sequence.

Index Physical Properties

All mass physical properties display an irregular downhole profilewith major discontinuities across the frontal thrust and the décolle-ment. While offsets occurring across the frontal thrust can be ex-plained by the stratigraphic repetition of ca. 160 m, a fundamentallydifferent genetic origin has to be assumed for the décollement. Be-tween frontal thrust and décollement porosity, bulk density and watercontent show steady—albeit lithologically modulated—downholetrends, compatible with a normally consolidated-to- slightly overcon-solidated state of the accreting sediments. The narrow (20 m wide)décollement zone itself is characterized by brittlely deformed, hard-ened sediments of low porosity and high density. Below the décolle-ment another strong offset to higher porosity and lower densitysediment marks the top of the subducting sequence. It has been sug-gested that an impermeable or overpressured décollement restrictsdewatering of this sequence, causing high excess pore pressures andunderconsolidation. This model is in good agreement with porosityand density profiles within the subducting sediments. All physicalproperties are discussed in detail in (Taira et al., 1991).

Compressional-Wave (P-Wave) Velocity Data

Being largely a function of lithology (grain density) and packingstructure of particles (porosity), the downhole development of com-pressional (P-wave) velocity at Site 808 closely matches the observedtrends in sediment mass physical properties. Due to sampling prob-lems in the coarse-grained turbiditic sediments above the frontalthrust, few reliable velocity data (Vpl and Vpt) could be determined inthe accreted sequence above the frontal thrust (Fig. 2A, 2B). Withinthe interval between thrust and décollement (365-945 mbsf), longi-tudinal and transverse acoustic velocities are increasing in a complexcyclic pattern that is likely to be affected by ash content in hemi-pelagic sediments. In the sediments of the outer marginal trench wedge

222

ANISOTROPY AND MICROFABRICS

(409-557 mbsf) and those belonging to the transitional zone (557-618 mbsf) and the upper Shikoku Basin sediments (618-829 mbsf),velocities are scattered over a broad band, that narrows considerablyin the hemipelagic deposits of the Lower Shikoku Basin (824-1243 mbsf). While velocities in the décollement zone itself are widelyscattered with no apparent downhole trend, an abrupt offset to lowervelocities defines the onset of the subducting sedimentary sequence,followed by a regular increase in both Vpl and VpV This agrees wellwith the observed changes in other physical properties.

The large number of physical property data sets with both Vp, andVp, available offers a unique opportunity to study the impact of theaforementioned parameters on acoustic anisotropy. The first reviewof all available data reveals no clear downhole pattern in acousticanisotropy, the expected "normal" trend of increasing anisotropy withdepth can not be found (Fig. 3). On the contrary, a gradual decreasebetween frontal thrust and décollement can be found that revertsbelow the décollement. Unlike other physical parameters, acousticanisotropy does not define the décollement with a clear discontinuity,but rather covers the interval between frontal thrust and base ofsubducting sediments with a band of scatter approximately 20% wide.

ACOUSTIC ANISOTROPY

Evaluation of Possible Mechanisms

In the course of the following sections, the discussed possiblecauses of acoustic anisotropy are reviewed with respect to the Site808 data. To constrain the following analysis to the appropriatelithologies, the first step necessary is a screening process of alllithological descriptions to exclude unusual or unrepresentative lith-ologies from further discussion. It is evident that the high degree ofvariability in acoustic anisotropy can be attributed to compositionalchanges: Units II (556.8-618.5 mbsf) and IVa (618.5-823.7 mbsf)are characterized by frequent ash and tuff layers, resulting in thehighest degree of scatter observed in Hole 808C. Most authors citedin the previous sections have discussed acoustic anisotropy withrespect to a specific lithology, mostly fine-grained, either clayey orcalcareous pelagic to hemipelagic sediment, which is the "target"lithology for the following tests.

The screening of the available data excluded mostly coarsegrained (from sandstone dikes) or monomineralic (pure ash) samplesfrom the accreting section, which is characterized over large intervalsby coarse-grained turbiditic trench sediment with low clay content.Physical property data sets from samples with CaCO3 content higherthan the average background level of 20% were also removed fromfurther analysis. Based on these same criteria, only a few points hadto be removed from the original data set below the décollement, wheresediments are predominantly of hemipelagic origin (Fig. 4).

Graphical and statistical comparison of downhole profiles re-vealed no linear correlation of CaCO3 content and acoustic anisot-ropy, that was demonstrated by O'Brien (1990) to exist in somecalcareous claystones. The correlation coefficient found for this dataset does not suggest a comparable interrelation (r 0.09) (Fig. 5).

Measuring />-wave velocities on standard ODP physical propertysamples to calculate acoustic anisotropies has an inherent source oferror. As physical property sample cubes are cut with faces perpen-dicular and parallel to the surface of the working core half and thecore walls, all />-wave velocities are determined either parallel ornormal to the core axis, which is therefore the reference frame for thecalculated acoustic anisotropy. O'Brien et al. (1989) demonstrated agood correlation between bedding angle and acoustic anisotropy inhemipelagic mudstones from the Japan Trench with acoustic anisot-ropies decreasing from positive values at low degrees of bedding dipto negative values at high angles of bedding dip. As no measurementsof bedding dip angles were made directly on physical propertysamples, the extensive structural data base was employed to comparethe two parameters.

Taking all available data for samples of Unit IVb (lower ShikokuBasin sediment), which are of a similar hemipelagic lithology (Figs.6A, 6B), no correlation of this type could be identified.

Several studies have suggested that primary acoustic anisotropycan be generated through depositional processes, e.g., sedimentationin a depositional regime dominated by currents (see Introduction). Asthe different lithological units encountered at Site 808 are definedunder consideration of their depositional environment, we wouldexpect different levels of correlation in different layers of the se-quence. Judging from the very low correlation coefficients found formost units (Fig. 7), this is obviously not the case. Correlation ofacoustic anisotropy with depth for different lithological units did notindicate a clear downhole trend in any of the Site 808 units except forUnit IVb, which displays decreasing acoustic anisotropy above andincreasing acoustic anisotropy below the décollement (Fig. 7).

Because lithologic Unit IVb in Site 808 is actually dissected bythe décollement, it is possible to directly evaluate the response toaccretion vs. subduction for a defined sediment type. The Unit IVb,termed lower Shikoku Basin sediment, is composed of a fairly coher-ent hemipelagic lithology over a range of about 420 m. Therefore onlythis part of the section at Site 808 is considered for further analysis(Fig. 8). There is no clear correlation of anisotropy with downholedepth—on the contrary a slight decrease in the accreting part of thelower Shikoku Basin sediments vs. a clear increase in the subductingpart of the sequence indicates that this property is related to theprocess of accretion and the resulting compressive forces.

REORIENTATION

To evaluate the possible correlation of acoustic anisotropy direc-tion and the direction of maximum compressive strain, paleomagneticdata were employed; using an existing paleomagnetic data set fromHole 808C, all individual physical property samples on which acous-tic data were obtained, were oriented to yield acoustic anisotropy asa function of geographical direction. This information was then usedto compare accreted vs. subducted sediments in terms of their direc-tional properties (as described in the "Methods" section).

To identify directional information of acoustic anisotropy a polarplot graph was prepared: it depicts a comparison of averaged acousticanisotropies from the lower Shikoku Basin sediments above vs. belowthe décollement (Fig. 9). Acoustic anisotropy is averaged over 10°intervals taking a range of 0° to 180° as a basis. To simplify the visualinterpretation, mirrored averages were also plotted in the respective180° to 360° range. Although the total number of data sets used hereis relatively small (86 data sets cover a range of 120 m above, 102data sets cover a range of 278 m below the décollement), this plotshows a considerable difference between both data sets. While acous-tic anisotropy in the accreting sediments shows peak values in thedirection parallel to the subduction vector (315°) and perpendicularto that, no such preferential orientation can be found in the anisot-ropies below the décollement.

ERROR ANALYSIS

P-wave velocity measurements are inherently difficult, as firstarrivals are difficult to determine (Birch, 1986). Great care musttherefore be taken to evaluate and characterize sources and propor-tions of possible errors. Most earlier attempts to analyze acousticanisotropies from DSDP and ODP />-wave velocity data did notinclude studies of error propagation, although the instrumental accu-racy is in many cases on the order of the observed effects. This isespecially true for data generated during DSDP cruises and to a lesserextent for ODP equipment-derived data.

Using a slightly more cautious approach we have calculated thelargest absolute errors according to Taylor's law as well as the meanabsolute errors according to Gauss's law which both apply to com-

223

W. BRUCKMANN, K. MORAN, E. TAYLOR

30%

Figure 1. A. Plate boundaries and geotectonic reference frame around Japan, and location map of the Leg 13 I/Site 808 working area in the Nankai Troughaccretionary complex (modified from Taira et al., 1991). B. Chronostratigraphy and lithofacies profile of Site 808 (original drawing provided by M. Underwood).

pound formulas that include several independent variables. Thiscalculation was very well implemented for Unit IVb sediment, abovethe décollement only, as this data set is crucial to the validity of ourconclusions. The largest errors are calculated assuming that all errorsof individual variables in a compound formula add up in one direction,while the mean absolute errors are calculated considering the prob-abilistic chance that the effects of errors in individual variables arecounterbalancing each other. Applying this concept to acoustic ani-sotropy, which is calculated according to:

and assuming a 1% precision (Boyce, 1976) for all P-wave velocityin a Hamilton frame velocimeter, Gauss and largest error are calcu-lated as follows:

if Al = I, then 1% = 0.01 I and T = t, then 1% = 0.01 t and thefollowing holds:

Gauss error:

Δ A = V (400 t/(l + t)2)2 (0.01 I)2 +(400 //(/ + t)2 (0.01 t)2

4 / r /where

A = acoustic anisotropy [%]t = P-wave velocity normal to core axis (transverse)/ = P-wave velocity parallel to core axis (longitudinal),

we get as a mean absolute error (Gauss):

Δ A = V ((a A/a 1) Δ i ) 2 + ( ( a A/a t) A O 2

and as the largest absolute error (Taylor):

Δ A = i (a A/a i ) Δ i ) 2 +1 (a A/a o A H .

Defining the partial derivatives {til):

a A 200 ( 1 +1) - 200 ( 1 - 1 )

t)2) l + t)2)

= V 2 ( 4 / 1 / { I + t ) 2 ) 2

largest error:

Δ A = (400 t/{ I + t)2) 0.0 1 / + (400 l/(l + t)2) 0.01 t

= 4 /*/(/ + t)2 + 4 / * / ( / + t)2

2 0 0 ( 1 = 0 - 2 0 0 ( / - 0

l

a Aa / (/ + t ) 2

= 4 0 0 t

(/+ O 2

Errors calculated according to this procedure are given in Table 1.It turns out that maximum and mean absolute errors are on the orderof the observed effects in a number of instances. However, all P-wavemeasurements made on board were normally repeated twice or moretimes if necessary to assure a high degree of reproducibility. Theactual error is therefore likely to be much smaller in all data sets,which seem critical according to our study. Nevertheless, it is obviousthat much more attention should be given to possible errors incomparative analysis of acoustic anisotropies. A critical review oflaboratory procedures as well as of data treatment and the applicationof error propagation theory to studies of this type seems to be critical,especially when only a small number of data sets is used.

ANISOTROPY AND MICROFABRICS

B

Epoch

CDCCDü

BCO

CD

CDüO

CDCCDüO

Age(Ma)

-LOWER SLOPE APRON(hemipelagic mud,

I thin turbidites, slides)

0.28

0.46

1.66

4.9

11.1

13.6

15.6?

FACIESASSOCIATION

UPPER AXIAL TRENCHWEDGE

I (thick-bedded sandturbidites)

LOWER AXIAL TRENCHWEDGE

(thin-bedded sandand silt turbidites)

FRONTAL THRUST

365 mbsf inHole 808C

OUTER MARGINALTRENCH WEDGE(silt turbidites andhemipelagic mud)

TRENCH-to-BASINTRANSITION

UPPER SHIKOKU BASIN

(ash/tuff andhemipelagic mud)

DECOLLEMENT ZONE

945-964 mbsf

LOWER SHIKOKU BASIN

(hemipelagic mud)

ACIDIC VOLCANICLASTICDEPOSITS

BASALTIC BASEMENT

Figure 1 (continued).

UNIT

lla

lib

lie

lie

IVa

IVb

VI

mbsf

10CH120.6

200

500-

700

800823.7

900-

1000

1100

Lithology

sssssssssssss

must, therefore, consider structurally controlled acoustic anisotropyas one option.

It is well established in the literature about the correlation of P-and S-wave velocities and microfabrics, that not only grain fabric butalso microcracks and microfractures can have an important impact onthe development of acoustic anisotropy.

Preferential orientation of acoustic anisotropies has been used todetermine fracture direction in areas of pervasive microfracturing toidentify the orientation of the maximum in-situ stress (Ramos andRathmell, 1989; Yale and Sprunt, 1989). Microcrack orientation canbe directly inferred from the direction of maximum and minimumacoustic anisotropies, as microcracks or microfractures impede anddeflect the propagation of sonic energy, yielding highest P-wavevelocities along strike and lowest perpendicular to strike.

The generation of these structures is discussed in detail by Ramosand Rathmell (1989). If these discontinuities are relaxation featuresthat formed as stress-relief microcracks, their strike will be perpen-dicular to the direction of maximum horizontal in-situ stress (5hmax),while tensile microcracks that may form as precursors to coring-in-duced tensile macrofractures will be parallel to S^^. With the reori-ented anisotropies at hand it is possible to speculate about the char-acter of microfractures leading to the observed correlation of conver-gence direction and acoustic anisotropy.

If microfractures are formed at Site 808 cores as stress relieffeatures perpendicular to the direction of the maximum horizontalin-situ stress (given through the direction of convergence), then Vpt

must be slowest in the direction of 315°. As a corollary, highestacoustic anisotropies should be found in that direction. Microfrac-tures forming in Site 808 cores as tensile fractures parallel to themaximum horizontal in-situ stress should result in highest P-wavevelocities normal to the convergence vector, yielding smallest acous-tic anisotropies parallel to 315°. Because maximum values in acousticanisotropy are found in both of these directions, it is possible that bothmechanisms contribute to the observed pattern. High-resolution elec-tron microscopy might be able to reveal the true configuration ofmicrofractures to support this preliminary interpretation.

It has been argued that grain fabric in the sequence between frontalthrust and décollement was strongly affected by the process of off-scraping and accretion, resulting in unusual acoustic properties. Judg-ing from the data and results presented here, it is not necessary toinvoke such a mechanism to describe the acoustic anisotropies found.

This conclusion is also in good agreement with data from X-raygoniometry studies, which suggest that grain fabric was not signifi-cantly altered by lateral compression in the wedge toe (Behrmann andKopf, this volume).

The normal expected trend of acoustic anisotropy observed belowthe décollement reflects, on the other hand, an inherited property frombasinal processes, where normal gravitational compaction and ac-companying grain rearrangement are the predominant controllingfactors for acoustic properties.

DISCUSSION

As is obvious from our review of possible causes of acoustic ani-sotropy, the commonly assumed depositional, diagenetic, or compac-tional processes cannot be invoked to explain the observed trends atSite 808. Oriented clay minerals, which are highly anisotropic inthemselves, do not contribute to the observed pattern in acousticanisotropy because they maintain a rather uniform background con-centration of 12%-22% throughout the complete section in Site 808.

On the other hand, it is clear from the results of the reorientationprocedure we have utilized here, that acoustic anisotropy is somehowlinked to the stress regime in the toe region of the Nankai Troughaccretionary complex. The accreting sedimentary sequence betweenfrontal thrust and décollement in Site 808 abounds with convergence-related deformational features covering a wide range of scales. We

SUMMARY

Using the extensive shipboard physical property data set from Site808 we have evaluated possible causes of the irregular downholedevelopment of acoustic anisotropies in accreted sediment of theNankai Trough accretionary prism. After various steps of data reduc-tion and screening and testing of various hypotheses proposed byprevious workers to describe the generation of acoustic anisotropy,we were not able to identify among them a single source or a likelymechanism to explain the observed trends. A paleomagnetic reorien-tation procedure was applied to all physical property data sets to yieldtrue geographic bearing of acoustic anisotropies. The results facili-tated the integration of the observed changes in physical propertieswith the geotectonic framework at the deformation front of the NankaiTrough accretionary complex.

225

W. BRUCKMANN, K. MORAN, E. TAYLOR

B o

200 ~

thrust

dec.

1200

1400

1200 "

14001500 30002000 2500 3000 1500 2000 2500

Acoustic velocity longitudinal (m/s) Acoustic velocity transverse (m/sec)

Figure 2. A. Acoustic velocity longitudinal (parallel to core axis) vs. depth at Site 808. B. Acoustic velocity transverse (normal to core axis) vs. depth at Site 808.

While accreted sediment between frontal thrust and décollementshows strongest acoustic anisotropies in the direction parallel andnormal to the direction of subduction, no obvious directional prefer-ence of this property can be found in lithologically equivalent sedi-ment below the décollement. This result supports the conclusion thatthe compressive tectonic regime in the frontal part of the NankaiTrough accretionary wedge alters directional acoustic properties in apredictable manner. Formation of tensile as well as stress reliefmicrofractures parallel and perpendicular to the convergence-relatedcompressional stress defines the observed orientation of maximumacoustic anisotropy. In the absence of this compressional stress no

such control is active below the décollement, where an inheritedbasinal anisotropy pattern prevails. To verify the model proposedhere, further SEM studies and oriented in-situ measurements ofacoustic properties are necessary.

ACKNOWLEDGMENTS

We thank the Shipboard Scientific Party of Leg 131 for the coop-erative scientific effort that made this study possible. We greatlyappreciate the inspiring comments of Tim Byrne, who also providedinvaluable technical help with the reorientation procedures. This study

226

ANISOTROPY AND MICROFABRICS

was financially supported by the Deutsche Forschungsgemeinschaft(DFG). R. Benett and N. I. Christensen reviewed the manuscript andprovided helpful suggestions.

REFERENCES

Birch, F., 1986. The velocity of compressional waves in rocks to 10 kilobars,part I. J. Geophys. Res., 1083-1102.

Boyce, R.E., 1976. Definitions and laboratory techniques of compressionalsound velocity parameters and wet-water content, wet-bulk density, andporosity parameters by gravimetric and gamma ray attenuation techniques.In Jackson, E.D., Schlanger, S.O., et al., Init. Repts. DSDP, 33: Washington(U.S. Govt. Printing Office), 931-958.

Carlson, R.L., and Christensen, N.I., 1977. Velocity anisotropy and physicalproperties of deep sea sediments from the Western South Atlantic. InPerch-Nielsen, K., Supko, PR., et al., Init. Repts. DSDP, 39: Washington(U.S. Govt. Printing Office), 555-559.

, 1979. Velocity anisotropy in semi-indurated deep sea sediments. J.Geophys. Res., 84:205-211.

Carlson, R.L., Schaftenar, CH., and Moore, R.P., 1983. Causes of compres-sional wave anisotropy in calcareous sediments from the Rio Grande Rise.In Barker, PE, Carlson, R.L., and Johnson, D.A., et al., Init. Repts. DSDP,72: Washington (U.S. Govt. Printing Office), 565-576.

Bachman, R.T., 1979. Acoustic anisotropy in marine sediments and sedimen-tary rocks. J. Geophys. Res., 84:7661-7663.

Fuji, N., 1981. Anisotropy in compressional wave velocity and wet-bulkdensities of calcareous sedimentary rocks, Deep Sea Drilling Project Leg62. In Thiede, J., Valuer, T.L., et al., Init. Repts. DSDP, 62: Washington(U.S. Govt. Printing Office), 995-998.

Hamilton, E.L., 1970. Sound velocity and related properties of marine sedi-ments, North Pacific. J. Geophys. Res., 75:4423-4446.

Kim, D.-C, Katahara, K.W., Manghnani, M.H., and Schlanger, S.O., 1983.Velocity and attenuation in deep-sea carbonate sediments. J. Geophys.Res., 88:2337-2343.

Kim, D.-C, Manghnani, M.H., and Schlanger, S.O., 1985. The role of diagene-sis in the development of physical properties of deep-sea carbonate sedi-ments. Mar. Geoi, 69:69-91.

Milholland, P., Manghnani, M.H., Schlanger, S.O., and Sutton, G., 1980.Geoacoustic modeling of deep-sea carbonate sediments. J. Acoust. Soc.Am., 68:1351-1360.

Nacci, VA., Wang, M.C., and Gallagher, J., 1974. Influence of anisotropy andsoil structure on elastic properties of sediments. In Hampton, L. (Ed.),Physics of Sound in Marine Sediments: New York (Plenum), 63-88.

O'Brien, D.K., 1990. Physical, acoustic, and electrical properties of deep-seasediments [Ph.D. dissert.]. Univ. of Hawaii.

O'Brien, D.K., Manghnani, M.H. and Schoonmaker-Tribble, J., 1989. Irregu-lar trends of physical properties in homogeneous clay-rich sediments ofDSDP Leg 87 Hole 584, midslope terrace in the Japan Trench. Mar. Geol,87:183-194.

O'Brien, N.R., Nakazawa, K., and Tokuhashi, S., 1980. Use of clay fabric todistinguish turbiditic and hemipelagic siltstones and silt. Sedimentology,27:47-61.

Postma, G.W., 1955. Wave propagation in a stratified medium. Geophysics,20:780-806.

Ramos, G.G., and Rathmell, J.J., 1989. Effects of mechanical anisotropy oncore strain measurements for in-situ stress determination. In Tomich, J.F.O.(Ed.), Formation Evaluation and Reservoir Geology. Exxon Prod. Res.,Houston, and Soc. Pet. Eng. AIME, 64:1989.

Schaftenaar, CH., and Carlson, R.L., 1984. Calcite fabric and acoustic anisot-ropy in deep-sea carbonates. J. Geophys. Res., 89:503-510.

Taira, A., Hill, I., Firth, J.V., et al., 1991. Proc. ODP, Init. Repts., 131: CollegeStation, TX (Ocean Drilling Program).

Wetzel, A., 1986. Anisotropy and modes of deposition of pelitic MississippiFan deposits. In Bouma, A.H., Coleman, J.M., Meyer, A.W., et al., Init.Repts. DSDP, 96: Washington (U.S. Govt. Printing Office), 811-817.

Wetzel, A., 1987. Sedimentological significance of strain and sonic velocityanisotropy in fine-grained turbiditic and hemipelagic deep-sea sedi-ments—an example from the Mississippi Fan. Mar. Geol., 74:191-207.

Yale, D.P., and Sprunt, E.S., 1989. Prediction of fracture direction using shearacoustic anisotropy. Log Analyst, 30:65-70.

Date of initial receipt: 2 December 1991Date of acceptance: 5 June 1992Ms 131SR-121

200 -

400 -

600 "

1000 -

1200 -

1400

thrust

800 -

dec.

-20 -10 0 10 2D 30

Acoustic anisotropy (%)

all data

Figure 3. Acoustic anisotropy vs. depth at Site 808 using all available data.

227

W. BRUCKMANN, K. MORAN, E. TAYLOR

0

200

400

600

800

1000

1200

1400-10

thrust

dec.

0 10 20

Acoustic anisotropy (%)

below frontal thrust

Figure 4. Acoustic anisotropy vs. depth at Site 808 below the frontal thrust.

-10

400

500

600

700

800

900

1000 ~

1100

1200

Acoustic anisotropy (%)

0 10 20

- dec.

0 20 40

Carbonate (%)

Figure 5. Comparison of CaCO3 content and acoustic anisotropy vs. depth atSite 808.

228

ANISOTROPY AND MICROFABRICS

A 80 B 80

-60 -

-80

-10 200 10

Acoustic anisotropy (%)

above décollement

Figure 6. Acoustic anisotropy vs. bedding angle in the lower Shikoku Basin sediment at Site 808. A. Above décollement. B. Below décollement.

Acoustic anisotropy (%)

below décollement

229

W. BRUCKMANN, K. MORAN, E. TAYLOR

0

200 -

thrust400

600

dec.1000

1200 -1200 ~

1400-10 0 10

Acoustic anisotropy (%)

regression in lithological units

20

1400-10

800 -

thrust

dec.

0 10

Acoustic anisotropy (%)

regression above vs. below décollement

Figure 7. Regression of acoustic anisotropy vs. depth for different lithologic Figure 8. Regression of acoustic anisotropy vs. depth above and below theunits at Site 808. décollement at Site 808.

230

ANISOTROPY AND MICROFABRICS

90c

180c

I U above décollβmβnt

below décollement

Figure 9. Composite graph of acoustic anisotropy in Unit IVb (lower ShikokuBasin sediments at Site 808) above (823-945 mbsf) and below (965-1243mbsf) the décollement, orientation 0°-180°. Data are averaged over 10°intervals; arrow indicates direction of plate convergence (315°).

231

W. BRUCKMANN, K. MORAN, E. TAYLOR

Table 1. Error analysis definitions for Unit IVb sediments above decollement in Hole 808C.

Core, section

131-808C-55R-4555555565656565656565656575757575757575758585858585858585859595959595959596060606060606060

455112233455112233451122334451122334412233445

Depth(mbsf)

823.4824.4825825.6829829.7830831832832.6833.4834.5835.3838.4839.4840840.7841.2842.2843.4844.7847.8848.9849.2849.6850.9851.5852.2852.9853.5857.4858.5859.2859.7860.2860.8861.7862.6866.9869869.8870.5871871.4872.1873.3

Trans.vel.(%)

2105212221412208213522392145212721252138208319732108210220681976203520222057207920572159209120642120207721062082217420892128211720722137213621402039208520952129215420842158217420922077

Long.vel.(%)

2165216721672217216423552165217221572125217120642144219021492169213221952193217422062164220321142167212322802237213821172188213421312146215922412116211221242278222322512177226322062103

Acousticaniso.

2.812.11.210.411.355.050.932.091.49

-0.614.144.511.694.13.849.314.668.26.44.476.990.235.222.392.192.197.937.18

-1.671.332.780.82.810.421.074.613.711.291.376.763.157.70.884.015.31.24

Taylorerror

1.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.99

Rel. Taylorerror

71.1595.3

165.68491.66148.2339.58

215.595.52

133.81327.9248.3244.34

118.148.7552.0421.4342.9424.3331.2244.7528.58

864.638.3183.5591.291.2925.1727.83

119.77150.2171.92

250.0671.22

475.88186.7443.3553.94

155.44145.4829.5463.4225.92

228.1549.8337.68

160.77

Aniso. -Taylor err.

0.810.1

-0.79-1.59-0.653.05

-1.070.09

-0.51-2.612.142.51

-0.312.11.847.322.666.214.42.474.99

-1.773.220.390.190.195.945.18

-3.67-0.670.78

-1.20.81

-1.58-0.932.611.71

-0.71-0.634.761.155.71

-1.122.013.31

-0.76

Aniso. +Taylor err.

4.814.13.212.413.357.052.934.093.491.396.146.513.696.15.84

11.316.65

10.28.46.478.992.237.224.394.194.199.939.180.333.334.782.84.812.423.076.615.713.293.378.765.159.72.886.017.33

Gausserr. abs

1.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.41

Gausserr. rel.

0.50.671.173.481.050.281.520.680.952.320.340.310.840.340.370.150.30.170.220.320.26.110.270.590.640.650.180.20.851.060.511.770.53.361.320.310.381.11.030.210.450.181.610.350.271.14

Aniso. -Gauss err.

1.40.68

-0.21-1.01-0.063.64

-0.490.680.08

-2.022.723.090.282.692.437.93.246.794.993.055.58

-1.183.80.980.780.786.525.77

-3.08-0.081.37

-0.611.39

-0.99-0.343.22.29

-0.13-0.045.351.746.29

-0.542.63.89

-0.17

Aniso. +Gauss err.

4.223.512.621.822.766.462.343.512.910.85.555.923.115.515.26

10.726.079.627.815.888.41.656.633.813.613.69.358.59

-0.262.754.192.214.221.832.496.025.122.72.798.174.579.122.295.436.722.66

232

ANISOTROPY AND MICROFABRICS

Table 1 (continued).

Core, section

131-808C-60R-56161616161616161616162626262626363636363646464646464646465656666666666666667

11223445563344511234122334451211223341

Depth(mbsf)

874876.6877.7878.2879879.8881.2882.3882.8883.5884.4889.6890.6890.8891.7892.4895.9897898.3899.3901.1905.5907.3908908.9909.8910.2911.1912.2915.2917925.1926926.4927.4928929930936

Trans.vel.(%)

205122492265211722702152224221822160211221442203220122212250225521422290222521872297229622212185223222992205221323262217224222772172227923202234226221962235

Long.vel.(%)

221522922258226922922243223723182229230222602377229222382348240121682439234422862271233322672260228323372240238423552254233823192261235024272378231322882305

Acousticaniso.

7.691.89

-0.316.930.964.14

-0.226.043.148.615.277.64.050.764.266.271.216.35.214.43

-1.141.62.053.372.261.641.577.441.241.664.191.834.023.074.516.242.234.13.08

Taylorerror

1.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.991.99

Rel. Taylorerror

25.97105.59646.14

28.82207.3648.28

895.8233.0663.5923.1937.9426.2849.35

262.2946.931.86

165.7631.7138.3745.16

175.68125.197.5659.2588.52

122126.9926.85

161.4120.8347.69

109.4249.7965.1844.343289.748.7264.84

Aniso. -Taylor err.

5.69-0.11-2.314.93

-1.042.14

-2.224.051.146.613.275.62.05

-1.242.264.27

-0.794.33.212.43

-3.14-0.40.051.380.26

-0.36-0.435.44

-0.76-0.342.19

-0.172.021.072.514.250.232.11.08

Aniso. +Taylor err.

9.693.891.698.932.966.141.788.045.14

10.617.279.66.052.766.268.273.218.37.216.430.863.64.055.374.263.643.579.443.243.656.193.836.015.076.518.244.236.15.08

Gausserr. abs

1.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.411.41

Gausserr. rel.

0.180.754.570.21.470.346.330.230.450.160.270.190.351.850.330.231.170.220.270.321.240.880.690.420.630.860.90.191.140.850.340.770.350.460.310.230.630.340.46

Aniso. -Gauss err.

6.280.48

-1.725.52

-0.452.73

-1.644.631.737.23.856.192.64

-0.652.854.86

-0.214.893.83.01

-2.550.180.641.960.850.230.166.03

-0.180.242.780.412.61.653.094.830.822.691.67

Aniso. +Gauss err.

9.13.311.18.342.385.551.197.464.56

10.026.689.015.462.185.687.682.627.716.625.840.283.013.464.793.673.052.998.852.653.075.613.245.434.485.927.663.645.524.5

Note: Site, Core, Sect.: standard ODP core description; trans, vel.: P-wave velocity perpendicular to core axis; long, vel.: P-wave velocity parallel to core axis; acousticaniso.: acoustic anisotropy; Taylor error: Taylor error (see text for discussion); rel. Taylor error.: Taylor error/acoustic anisotropy; aniso. - Taylor err.: acousticanisotropy-Taylor error; aniso. + Taylor err.: acoustic anisotropy + Taylor error; Gauss error: Gauss error (see text for discussion); rel. Gauss error.: Gauss error/acousticanisotropy; aniso. - Gauss err.: acoustic anisotropy-Gauss error; aniso. + Gauss err.: acoustic anisotropy + Gauss error.