In this section, a number of aspects of breccias occurring in the Vredefort Dome-Witwatersrand Basin (VWB) and in tectonic fault/shear zone (TFS) settings are compared. Pseudotachylitic breccias in the Vredefort Dome -Witwatersrand Basin can be divided into two groups:
1. the breccias in the goldfields >80 km from the center of the structure and well outside the zone of shock metamorphism;
2. the breccias in the Vredefort dome, where shock-induced features like shatter cones, PDFs and coesite and stishovite indicate the elevated shock conditions that were reached in the rocks of the central uplift.
In the goldfields, the breccias have been dated as syn-Vredefort (Trieloff et al., 1994; see also Friese et al. 2003), and they occupy cataclastic fault zones that display a consistent dip inwards towards the center of the Vredefort Structure and that have consistent normal dip-slip sense of movement (Killick, 1993). The veins are large by fault standards (up to 1 m wide, etc.) but they are still subordinate within fault zones that are considerably wider, like tectonic pseudotachylites. Evidence has been presented that indicates that the Witwatersrand fault breccias include cataclasites and ultramylonites, but at least locally melting has taken place, as deduced from petrographic observations reviewed by, e.g., Reimold and Colliston (1994; see also Reimold et al. 1999b). We conclude that these breccias are likely the result of the collapse of the Vredefort impact crater during the crater modification phase; they comprise, in part, true pseudotachylites, besides much cataclastic and ultramylonitic material. A number of specific aspects, A - F, can be compared:
A. Many breccias from the central part, the Archean basement complex, of the Vredefort Dome - the central uplift of the large Vredefort impact structure, that have been described as "pseudotachylite" in the past are indeed bona fide melt breccia, though of still unconfirmed shock, friction, or shock-friction origin. All pseudotachylitic breccias of the Archean basement core and of the supracrustal collar of the Dome are invariably recrystallised due to post-impact regional metamorphism (Gibson and Reimold 2001a). They were formed in quartzite host rock in the collar of the structure. The matrix represents a very fine-grained, cherty groundmass of silica ± Fe,Mn oxides. And in the innermost part of the core, breccias described by Stepto (1979) and, recently, by Gibson and Reimold (2001b) and Gibson et al. (2002) are also mostly annealed. In these cases it is not possible to state with conviction what kind of breccia these occurrences may have represented. But in the case of the breccias from the center of the Dome, evidence is mounting that they could well be impact (shock) melt breccia instead of pseudotachylite = friction melt (Gibson et al. 2002; Gibson and Reimold 2004).
The Vredefort Dome exposes a wide range of rock types, but can be broadly divided into two zones on the basis of lithology - the collar of the dome is dominated by supracrustal strata and a well-layered structure caused by bedding on a mm to km scale; and the core of the dome is characterized by massive, crystalline gneissic basement. The average shock pressure increases from <10 GPa in the collar to as much as 45-60 GPa in the core (Gibson and Reimold 2004). The lithological differences have a clear effect on the morphology of the breccias - in the collar, veins tend to be more planar than in the crystalline rocks of the core, although they are still highly variable in orientation. A similar effect is seen in the alkali granites in the collar, where the breccias define anastomosing vein arrays. However, another factor is shock pressure.
Gibson et al. (2002) described breccias from the central parts of the dome that display poorly-defined margins against wallrocks and clasts, and a granofelsic matrix texture indicative of high-temperature crystallization and/or annealing (Fig. 7c). They compared these breccias to the so-called lunar granulites, which have been previously interpreted as contact-metamorphosed fragmental breccias, and proposed that both the high temperatures and indistinct boundaries signify breccia formation at shock pressures that were sufficiently high to induce localized shock melting of feldspars and hydrous ferromagnesian minerals. They speculated that frictional heating may have played a role in creating the breccias; however, given the background shock heating, it does not seem necessary to invoke large amounts of frictional heat. More recent work has indicated that thin breccias are commonly compositionally highly heterogeneous along their length, and may even display virtually monomineralic segments corresponding to particular coarsegrained minerals in the wallrocks (also Reimold 1991). This suggests extremely limited melt mobility, something that is not likely to be favoured by a mechanism that requires sliding to generate the melt in the first place.
Only in rare cases have displacements of 10 to 50 cm along veins of pseudotachylitic breccia in the Vredefort Dome been recorded. Other indicators are the lack of evidence of cataclasis in the wallrocks flanking the breccias. This is in marked contrast to tectonic friction melts that typically occur in fault or mylonite zones up to several orders of magnitude wider than the melts themselves. Finally, Gibson and Reimold (2001b, 2004) documented increasing shock effects on a mm-scale adjacent to several thin breccias, which suggests that the breccias are sites of localized shock pressure "spikes". Recent modeling of shock wave propagation in heterogeneous media has shown similar P spikes up to 2 to 3 times the background shock P. Given the heterogeneous nature of almost any geological target, enhancement of the shock P by reflection and/or refraction of the shock wave is not only plausible but more likely the norm. It remains to be seen whether increasing background shock P increases or decreases the amount of enhancement - our feeling is that at low background shock P, reflection and refraction will enhance the shock wave and lead to progressively more melting as the P increases, until a threshold value above which the amount of work being done at the grain scale is sufficient to prevent extreme P fluctuations (as reflectors are destroyed or weakened), and the amount of localized melting decreases but is replaced by more wholesale shock metamorphic damage. Finally, almost all veins show some level of differential slip along their margins - typically at mm, but sometimes up to a few cm, scales. This could be part of the original mechanism of breccia formation: differential acceleration of the shock wave can trigger syn-shock slip on favourably oriented surfaces; however, it is equally possible that the movement may be triggered by stresses during the subsequent crater modification phase acting on the lubricated surfaces.
VWB breccias known from fault zones in the wider Witwatersrand Basin do also include several breccia types, including cataclasites, mylonites, and friction melt. Sometimes a melt phase only constitutes a small portion within mostly cataclastic breccia (Fletcher and Reimold 1989; Killick and Reimold 1990; Reimold and Colliston 1994; Reimold et al. 1999b; Dressler and Reimold 2004).
In tectonic fault- and shear zone related settings (TFS), a wide variety of fault breccia can be observed, including cataclasite, mylonite and ultramylonite, and pseudotachylite (e.g., Passchier and Trouw 1998).
B. In both the Vredefort Dome and Witwatersrand Basin, several generations of breccias have been observed - admittedly quite rarely. Reimold and Colliston (1994) provided evidence for pre-impact brecciation, in the form of narrow breccia veinlets (resembling altered pseudotachylitic breccia) that are intimately deformed together with their host gneisses of the Archean basement complex that experienced tectonic deformation prior to 2.02 Ga (Lana et al. 2003a,b). Multiple generations of breccias are also known from the large-scale bedding-plane parallel fault zones of the Witwatersrand basin (Killick and Reimold 1990 - a sample from Elandsrand gold mine); other samples with crosscutting relationships between various generations of breccia were recently obtained from mine geologists on Elandsrand Gold Mine, and some examples are shown in Figure 8. In addition, Berlenbach and Roering (1992) discussed the possibility that pseudotachylitic breccias as old as the Ventersdorp Supergroup (ca. 2.7 Ga) occurred in the basin. Recently, Friese et al. (2003) reviewed the chronological data on Witwatersrand fault rocks.
Repeated brecciation along TFS should certainly be possible, as such deformation zones represent zones of weakness in the crust that may be prone to reactivation. Still in modern times, mining-induced rockbursts frequently have produced fault rock in the deep-level gold mines of the Witwatersrand (e.g., Stewart et al. 2000, and references therein). However, to our knowledge no friction melt has been described from such locations yet.
C. In the Vredefort Dome, orientations of breccia veinlets can be controlled by pre-existing zones of weakness, such as lithological boundaries between gneiss bands of different rheologies, mylonite zones, and other structural features. However, comprehensive mapping of vein and dike orientations with regard to Archean fabrics and pre-impact deformation features (Dressler and Reimold 2004) has shown that such relationships are by no means the rule. Also, where an apparent relationship between a (sub)planar feature and a breccia vein occurs, it does not necessarily persist in the third dimension. However, it still cannot be ruled out that, at least in some instances, large breccia developments could be controlled by faulting (for example, where conjugate pairs of straight veins with obvious displacements of host rock banding are observed - e.g., some images shown by Reimold and Colliston 1994). In the Witwatersrand basin, large breccia developments are preferentially related to fault zones.
D. In tectonic settings, pseudotachylite is generally faulting/shearing-related. As stated above, pre-existing zones of weakness, such as lithological boundaries or mylonitic shear zones, could be the preferred loci for pseudotachylite formation. Sibson (1975) and Grocott (1981) provided a detailed account of pseudotachylite geometries observed in fault zones. Similar accounts are abundant in the structural geological literature, for example in papers in Magloughlin and Spray (1992). Reimold and Colliston (1994) and the subsequent work of our group have shown that the entire range of these geometries can be observed in abundance in the breccia developments of the core of the Vredefort Dome. However, comparing tectonic and Vredefort impact-induced network breccias, there is a different scale of magnitude involved. Tectonic network breccias rarely exceed several decimeters in width, whereas some network breccias at Vredefort have been mapped for hundreds and, in select cases, even over thousands of meters extent and hundred meter width. Within such network breccia occurrences at Vredefort, there may well occur apparent faults, linear features along which the host gneissosity can be displaced by centimeters or decimeters - and in exceptional cases by as much as 50 centimeters. And yet, even the detailed mapping of recent years of such sites by B.O. Dressler in conjunction with our group (Dressler and Reimold 2004) has not shown any evidence that would lead to an interpretation of such sites as major generation planes. Rather, during compressive "shattering" of a rock mass, faulting and slip along individual planes of weakness or along conjugate orientations takes place, according to strain distribution through the rock mass, possibly locally facilitated by pre-existing planes of weakness (fractures, pre-impact mylonites or cataclasites, lithological banding). In contrast to many tectonic occurrences of pseudotachylite, the long established relationship between slip rate and thickness of a friction melt vein (Sibson 1975) is not valid for impact-generated pseudotachylitic breccias.
It is generally accepted that the process of friction melt formation along a fault surface will initially result in strengthening of a fault due to increase of the friction coefficient. However, once a fault is completely separated, resistance decreases. This hypothesis is, however, based on significant displacement occurring along a fault, with a magnitude of at least meters, which is generally not observed along pseudotachylitic breccia veins in impact structures. Melosh (this volume) investigates how thick masses of pseudotachylitic breccia (although he refers to them as pseudotachylite), as observed for example at Vredefort and Sudbury, could form as a result of friction melting, in the face of the melt-strengthening of rock observed. Melosh proposes that melt is produced on narrow faults but is extruded into adjacent country rock where it could pool. This interesting suggestion must be examined further in the field by continuing efforts of mapping 3D exposures including likely generation planes (narrow shear/fault planes) and apparent sites of melt pooling. Could it be that the massive, straight melt dikes observed locally in the Vredefort Structure could be dilational sites formed during decompression and then infilled with melt generated elsewhere?
E. Distinct chemical relationships between host rock and pseudotachylite have been established for tectonic breccias, as shown by many publications such as those reviewed by Reimold (1991). These relationships are the result of differential fracture toughness of certain minerals and their specific susceptibility to melting (e.g., Spray 1995). The same relationships between chemical compositions of host rock and breccia pairs have been observed for the Vredefort breccias and tectonic pseudotachylite-host rock pairs (Reimold 1991). Dressler (1984, and literature reviewed in there) has already shown that the same applies to the Sudbury Breccias and their host rocks. However, it has also been shown for the Vredefort case that it is possible to determine compositions for spots in Vredefort breccia veinlets (admittedly very thin ones!) that are identical to the mineral compositions in the immediately adjacent host rock. This suggests that at least some breccias were melted in situ and were not formed through injection from a larger generation plane. Furthermore, it indicates that they were not subject to melt fractionation or incomplete mineral melting, or to mixing between minerals or even precursor rock-types.
F. Petrographic characteristics of Vredefort breccias and of many other so-called "pseudotachylites" in impact structures are also similar or even identical to those observed in tectonic pseudotachylite = friction melt. Fine-grained igneous crystallization textures and aphanitic melt matrices are commonly observed, but partial or main presence of cataclastic material also attests to the likelihood that, at least, in some cases, Spray's (1995) model for the generative process of pseudotachylite applies to impact-induced melt breccia as well.
Was this article helpful?