INTRODUCTION
Systematic logging of fractures in thecore and the drill hole walls of CRP-212A is being undertaken to document the brittle deformation patterns and crustal stress regimes associated with sifting along the Transantarctic Mountains Front at Cape Roberts. Because the Cape Roberts Project drill sites are located along the Transantarctic Mountains - Victoria Land rift basin structural boundary, abundant fracturing of the sedimentary strata was expected. The objectives of the fracture analysis are to acquire an age- controlled structural record of the kinematic and dynamic history along the Transantarctic Mountains Front and to obtain the first information on the contemporary in situ stress field within Antarctica.
Fracture logging of core was completed for CRP-2 and CRP-2A cores. A large population of fractures was present and a total of 2 017 core fractures were logged. In the more indurated portion of the cored sequence, fracture surface features were well developed, yielding kinematic and dynamic data. A range of both natural fractures and induced fractures were identified based on these surface fracto- graphic features and on morphological characteristics of the core fractures. Preliminary descriptions of the core fractures are provided below. Down-hole logging using bore hole televiewer and dipmeter tools was completed for portions of the drill hole (see section on Down-Hole Logging) and will be used to map fractures in the drill hole walls, to search for any drill hole break-outs, and to provide orientation for portions of the CRP-212A core.
FRACTURE STUDY PROCEDURES
Fracture logging was carried out on the whole core at the Drill Site science laboratory. Core fractures were numbered sequentially downward from core top and depths to the top and base of each fracture were recorded. Dip angle and direction were measured with respect to an arbitrarily placed red line scribed along the length of the core. Where sequential core runs could be fitted together directly, the red scribe line was matched between runs. For HQ core, c. 10% of the core runs could be fitted together, whereas 56% of the NQ core runs could be matched. The number of consecutive core runs that could be fitted together increased down core, culminating in a 42 m interval between 576 and 618 mbsf that was fitted and scribed continuously. Core fracture logging included observation and photography of fracture morphology, fracture surface features, fracture fill material, fracture terminations, and cross-cutting and abutting relations between fractures. Procedures for logging fractures in
cores and criteria used tor distinguishing natural and induced fractures generally followed Kulanderetal. (1990).
After fracture logging was completed, the whole core was scanned using CoreScan@ equipment leased from DMT, Germany. The Corescan@ obtains digital images of the entire core circumference by rotating the whole core on rollers. line-scanning, and digitally joining these into unrolled' core images up to a maximum of 100 cm in length. Due to the poor induration of core material, it was only possible to carry out wliole-core scanning of c. 2% of the CRP-212A sequence down to 83 mbsf. There were sufficiently coherent sections of the core between 83 and 200 mbsf to allow c. 60% of the whole core to be scanned.
Between 200 and 624.15 mbsf, 82% of the whole core was scanned. The slabbed face of the entire working half of the core was also scanned after the core was split and placed in the core boxes. For Box 30 only, the archive half of the split core was scanned instead of the working half.
In CRP-2A, down-hole logging with dipmeter was conlpleted from 65 to 170,200 to 255, and 280 to 624 mbsf and with bore-hole televiewer from 64.7 to 163.7 (unosiented) and from 200 to 441 mbsf (oriented). All the dipmeter logs were oriented. Analysis of these down-hole logs will beused to map fractures in the drill hole walls and will also be used to orient the CRP-2A core by matching drill hole wall and core fractures (cf. Nelson et al., 1987).
In particular, we will match the bore hole televiewer images of the drill hole walls with the whole-core scan images to provide core orientation (e.g. Schmitz et al., 1989; Weber, 1994). A single attempt to orient the core directly using a core orienting tool was unsuccessful. W e will use core intervals oriented by matching with bore hole televiewer imagery to test whether orientation of core based on palaeomagnetic vectors is reliable for CRP-212A core and, if so, the palaeomagnetic data will be used to orient additional core intervals, as was done for CRP-l (Paulsen & Wilson, 1998). Here we provide an example of orienting core based on comparison of bore hole televiewer and Corescan@ imagery. Until further analyses can be carried out, however, the fracture data obtained from the core is only oriented with respect to an "arbitrary north"
defined by the red scribe line, which differs between unmatched core runs.
FRACTURE DISTRIBUTION AND DENSITY
Fractures are present in all portions of the core. Fracture density, plotted as fractureslm, is shown in figure 2.1. Note that breaks in the histogram with no fractures reflect intervals where no core was recovered. Fracture densities in the Quaternary/Pliocene section average 2.29 fractureslm and range from 1 to 4 fractureslm. In the Miocene section, fracture densities range from 1 to 9 fractureslm and average
Fig. 2.1 -Fracture density plot showing the number of fractures per 5 m intervals ofcore withrelation lodepthandlilhological and ageboundaries within the CRP-2A core. P-wave velocity boundary modified from figure 2.25.
about 3.55 fractureslm. In the upper Oligocene section, fracture densities range from 1 to 10 fractureslm and average about 2.93 fractureslm. In the lower Oligocene section, fracture densities range from 0 to 10 fractureslm and average about 3.21 fractureslm.
The marked increase in fracture density across the contact between Pliocene and Miocene strata is only partly attributable to increased induration, as much of the upper Miocene section is also poorly indurated. Fractures commonly occurred in well-indurated diamictites, sandstones, and siltstones, and less commonly in conglomerate units and unlithified sand intervals. Fracture density commonly increases across P-wave velocity boundaries that have been correlated with seismic reflectors, suggesting that fracturing was influenced by mechanical differences between rock types andlor related to degree of induration. Further analysis of fracture density vs lithology and grain size will reveal any consistent relations within the core. Pronounced peaks in fracture density occur at 520 and 540 mbsf and mark intervals containing abundant normal faults and veins.
FRACTURE TYPES IN CRP-2/2A CORE
To interpret the mode of origin and significance of CRP-2/2A core fractures, it is essential to differentiate
natural fractures from induced fractures. Natural I ' n n ' ~ u n ~ are those that existed in the crust prior to drilling iiinl were intersected by coring. Natural fractures in t h e ('HP ? A core include microfanits. vcins,clastic dykes, anil possible sub-vertical joints. Other types of natural def.oi~i~iiition, incl~~ding soft-sediment folding, brecciation, aild plin'iiit shearing, arc treated elsewhere in this volume (see section on Deformation). Induced fractures form in respoiisc to drilling- or coring-related perturbations of the S ( rcss firlil.
or due to subsequent handling of the core. In sonic ciisrs such induced fractures are indicative of the con~cmpoi~iii~y maximum and minimum horizontal stress directions in tlic crust. CRP-212A core contains abundant drilling--. coring , and handling-induced fractures.
Natural Fractures
Microfaiilts. Several varieties of n~icrofaults occur in the CRP-2A core. Only discrete fault planes inferred to lie of brittle origin are described here. Open, brittle microf;ui Its with approximately down-dip slickenlines a n d in some cases having polished slickenside surfaces, occur sporadically through the core between 51 and 549 inbsf (Fig. 2.2). Dips of these faults typically range between 55 and 7S0 and, where offset can be observed, they have, normal-sensedisplacement. Withgreater depth, fault planes are typically sealed by mm-scale bands of material, sonic of which is as yet unidentified. Sealed microfaults appear at c. 300 mbsf and are abundant in intervals from c. 3 10 to 325 mbsf and c. 5 l0 mbsf to the base of the cored interval.
In the vast majority of cases where offset of either bedding or veins was observed, the faults have normal-sense displacement (Fig. 2.3). Some rare reverse-sense faults are also present. Although not restricted to any particular lithology, closed microfaults are abundantin strata showing
Fig. 2.2 - Brittle microfault with down-dip slickenlines. Fault dip 68O.
depth = 480.36-480.47 mbsf. core diameter = 45 mm.
Fig. 2.3 -Closed microfaults with normal-sense offset of bedding. Fault dips=55and58o,depth=315.9Oto316.0Ombsf.corediameter=45mm.
evidence of substantial soft-sediment disruption, suggesting a possible genetic relationship. The closed microfaults typically have dips ranging from c. 60 to 80°
although a few lower-angle faults are present. In many portions of the core, the fault planes have a conjugate geometry, with approximately equal and opposite dips (Fig. 2.4). In several cases, clear cross-cutting relations
Fig. 2.4 -Veins on 'unrolled' Corescan@ image of whole core. Negative image shows hairline white calcite veins as thin black lines (outlined by thin white dash lines). Note conjugate geometry (equivalent strike and opposite dips) and normal-sense offset of one vein by a second. This vein-filled normal fault also displaces bedding. Vein dips are 65 and 72O.
The vertical lines on the core image are the scribedreferencelines. Depth
= 542.88 - 543.18 mbsf. unrolled core circumference = c. 142 mm.
bctween members ofthe con.jugate sets indicate they are coeval (Fig. 2.4).
Based on macroscopic observation, (lie material filling the Sault planes appears to be of several different types.
Dark bands along surfaces that truncate and offset bedding i r e common in core helow about 350 nibsf. One such surface parted, revealing well-developed surface polish and slickenlines within the dark fault zone material, indicating a brittle-shear origin. The dark bands thus may consist of cataclastically crushed material formed due t o fault surface shearing. It is also possible that someof these bands lack shear movement and instead represent injected fine-grained sedimentary material. A second type of fault fill material is pale grey in colour with a granular appearance. and appears to be sedimentary material lining the fault surfaces. Carbonate cementation along normal fault planes occurs in sandstone intervals and in places appears similar to the grey sedimentary fill material. The cement either forms continuous sub-planar zones of grey cement or forms patches or spheres of cement aligned along the fault planes (Fig. 2.5). Vein material, consisting of thin calcite veins or layered calcite and pyrite, is also common along normal fault planes (Fig. 2.4). Future thin- section examination of the materials along fault planes will be used to clarify their type and origin and to identify any textures of kinematic significance.
Veins. Mineralized veins were first identified at c. 332 mbsf and become abundant deeper in the sequence.
beginning at c. 440 mbsf. The veins commonly contain calcite and many also contain pyrite. Swarms of very thin, hairline calcite veins occur at several levels below c. 520 mbsf. Although vein dips range from sub-horizontal to sub-vertical, the vast majority of the veins have dips between 60 and 80' (e.g. Fig. 2.4). As discussed above, some calcite veins follow normal fault planes with conjugate geometry and show mutual offsets (Fig. 2.4). In other cases, veins with similar steep dips are compound,
Fig. 2.5 - Cemented normal fault on 'unrolled' CorescanCC image of whole core. Note carbonate cement along normal fault planes that displace bedding and distributed spherules of the same cement. Niormal faultdips 70° depth = 5 1 1.49 - 5 11.62mbsf. unrolledcorecircumference
= c . 142 mm.
W Initial Rcpoi l o n ('RP-2/2A consisting of multiple, thin strands. and have en echelon
segments that overlap or coalesce. This type of geomet ry is characteristic of opening-mode tension veins. At present i t is not possible systeniatically to differentiate veins following shear planes from tensile veins, but mineral textures determined from thin-section examination may help to clarify this key issue.
Clastic Intrusions. Several planar clastic dykes with thicknesses of 1-2 cm O C C L I ~ in the core between c. 337 and 600 mbsf (see also section on Deformation). These dykes typically have pale rims of carbonate cement and some associated pyrite mineralization (Fig. 2.6). Clastic dykes range in dip angle from 75 to 86'. Additional mm-scale sedimentary intrusions also occur and, as noted above.
locally follow normal fault planes. These pli~iiiir 10
anastarnosing, thin sedimentary veins are p:ii.ticuli~ily common between c. 3 15 and 365 mbsf.
Induced Fractures
PetalandPetal-Centreline Fractures. Petal :iiiO pi.-tiil- centreline fractures form below the drill bit in response 10
stress induced when the weight on the bit temporarily increases (Lorenz et al., 1990; Li & Schmitt, 1997). 'l'lw.se induced fractures have curving shapes that f'ollow [lie stress trajectories radiating below the bit and this (list i iic~ive curviplanar geometry allows them to be conl'idently identified in core (Kulander et al.. 1990). Over 100 curviplanar fractures interpreted as petal a n d pclal- centreline fractures are present in the CRP-212A core ancl occur at all depths between c. 19 and 587 mbsf. Altlioiii:li not restricted to a particular lithology, the petal-centre1 ine fractures most commonly occur in fine-grained strata. In cemented strata the petal-centreline fracture siirl'i~ccs displayed distinctive fractographic features including hackle plumes and arrest lines that indicate down-core fracture propagation directions. A sub-population of this fracture set formed shallow, scoop- or spoon-shaped flakes along the core margins and had sub-vertical dips (Fig. 2.7).
Such fractures are common at the top of core runs, suggesting they formed as the drill string 'tagged' the bottom at the start of a new core run. consistent with models for the genesis of petal-centreline fractures (Kulander et al., 1990; Lorenz et al., 1990; Li & S c h n i i ~ ~ , 1997). Significantly, preliminary examination of tlie bore hole televiewer imagery reveals sub-vertical fractures in the drill hole walls over the same depth intervals as some of the petal-centreline fractures logged in the core. This will allow us to obtain orientation for the petal-centreline fractures and map themaximum horizontal stress direction in the crust, which has been shown to parallel the strike of petal-centreline fractures (Plumb & Cox, 1987; Kulander et al., 1990).
Disc Frnct~wes. Disc fractures form normal to the core axis when unloading produces axial tension within the core. Sub-horizontal to low-angle (<30Â dip) fractures are Fig. 2.6 - Clastic dyke on 'unrolled' Corescan@ image of whole core. abundant in CRP-2/2A core. In many cases these fractures Note ,.in, of cen,ent, ~~k~ dips 750. depth = 337.68 . nucleated on fossils or lonestones that served as stress 337.87 mbsf, unrolled core circumference = c. 142 mm. concentrators (Fig. 2.8). Some of these fractures have
Fig. 2.7- Petal-centreline fracture at depth 258.17 - 258.45 mbsf. Note sub-vertical dip and curviplanar. 'spoon' shape of fracture. Second petal fracture curves to merge with the main petal-centreline fracture.
b. 2.8 - D i s c fracture nucleated on macrofossil. Note origin point at fossil m a r ~ i n and hackle plume radiating across fracture face (dash line).
These fealurcs demonstrate an induced. tensile origin for lhis horizontal fracture. Dcptli = 448.26 mbsf. core is 45 m m diameter.
distinctive curviplanar 'saddle' shapes typical of disc fractures formed in areas where the horizontal stresses are anisotropic (Bankwitz & Bankwitz, 1995; Bell, 1996).
Below c. 200 mbsf, these core fractures, where developed in fine-grained lithologies. commonly showed surface fractographic features including origin points, hackle plumes and twist hackle along core margins (Fig. 2.8), proving an induced, tensile fracture origin. Measuren~ents of plume axis trends were taken and, in some core intervals, the plumes showed consistent trends. This suggests that the plume axes on the disc fractures may have formed parallel to the trend of the maximum horizontal compressive stress direction, as demonstrated elsewhere (K~ilander et al., 1990).
Other Induced Fractures. Planar to irregular induced fractures in the core formed due to a variety of other factors. Sub-horizontal tensile fractures formed due to the upward pull on the drill string exerted when the hydraulic system elevated the chuck during a coring run, when the core lifter mechanism latched on to the base of the core, and when the drillers broke the core from the bedrock at the end of a run. These types of fractures commonly had well-developed hackle plume structures characteristic of tensile fracture propagation. Another type of fracture formed at or close to the bit during drilling when the core was broken off along a subhorizontal plane and then spun when the drill bit re-engaged, producing finely etched circular lines and gsooves on interlocking surfaces, resembling bowls shaped on a potter's wheel. Torsion fractures, consisting of irregular or helical breaks where the core was twisted, occurred sporadically in the core.
They were most common near the base of core runs in clay-rich lithologies, where the core resisted sliding and the slow entry into the core barrel subjected the core to torque from the rotating drill bit. In some cases, up to 30 cm of core at the base of the run would be pervasively broken by wavy gaping tension fractures arranged in en echelon patterns consistent with the clockwise torque exerted by the rotating drill bit. Handling-induced fractures of many varieties developed during core processing and transport. Some such fractures mimic the geometry of
fixct lire types previously described and may have nucleated on incipient tlrilling o r c o r i n ~ - i ~ ~ d i ~ c e d fractures that had not propagated completely through the core.
Fractures present i n the drill hole walls must either be prc-existing natural fracturesordrilling-inducedfractures, such as petal-centrelinc types that propagate into the rock proximal to the drill bit. Coring- and handling-induced fractures will not be present. Thus comparison of fractures logged in core with those visible in the televiewer imagery canconstrain interpretations of fractureoi-igiin. The oriented televiewer images allow fracture attitudes tobedetermined, with dip calculated from the amplitude of the sine wave and dip direction given by the orientation of the wave trough. In addition, where fractures can be matched between core and televiewer images. segments of the core can b e oriented and true geographic coordinates can be assigned to measurements made with respect to the arbitrary core reference lines. Much of the fracture mapping and core orientation based on the bore hole televiewer imagery will be carried out when high-resolution digital images are obtained from the original analogue televiewer data. As an example, however, figure 2.9 shows televiewer data from the drill hole between c. 327 to 33 1 mbsf, consisting of a sequence of merged Polaroid photos taken at 1 m intervals for on-site interpretations. The same depth interval of core reconstructed in digitally 'stitched' scan images of the unrolled circun~ference of the core. resized and reoriented to match the drill hole image, is also shown in figure 2.9.
A set of steep SE-dipping fractures is clearly imaged on the televiewer log and can be matched directly to steep core fractures interpreted as being induced petal-centreline fractures. Steep NW-dipping core fractures, also probable induced petal-centreline fractures. are only barely discernible on the televiewer imagery. Natural fractures, consisting of faults and clastic intrusions, dip at moderate to low angles to the NW. These are present, though not as sharp, on the televiewer imagery. Coring-induced and handling-induced fractures, visible on the scanned core, are not seen on the bore hole televiewer imagery. Overall the match between fractures in the drill hole wall and in the core is excellent in this interval.
A set of steep SE-dipping fractures is clearly imaged on the televiewer log and can be matched directly to steep core fractures interpreted as being induced petal-centreline fractures. Steep NW-dipping core fractures, also probable induced petal-centreline fractures. are only barely discernible on the televiewer imagery. Natural fractures, consisting of faults and clastic intrusions, dip at moderate to low angles to the NW. These are present, though not as sharp, on the televiewer imagery. Coring-induced and handling-induced fractures, visible on the scanned core, are not seen on the bore hole televiewer imagery. Overall the match between fractures in the drill hole wall and in the core is excellent in this interval.