Chapter 6 Conclusions

The observed microscopic fabrics are typical of specific fault core components. The embryonic fabric is typical of the breccia zone, the intermediate fabric is typical of the portion of the breccia zone lying near the gouge zone, and the mature fabric is typical of the gouge zone. Occurrences of lenses and pockets of breccia within the gouge zones (e.g., Billi et al., 2008), show that gouge zones usually develop from pre-existing (i.e., early) breccia zones by grain size reduction (e.g., Engelder, 1974; Billi and Storti, 2004; Sammis and King, 2007). It follows that the observed embryonic, intermediate, and mature cataclastic fabrics probably represent three subsequent evolutionary stages of developing carbonate fault cores. Using this assumption, the microscopic observations presented in this paper are synthesized in the following temporal model for the evolution of carbonate fault rocks:
- In the early deformational stages, large angular grains are in contact with one another and therefore cannot, or can barely, rotate. Lithostatic and tectonic stresses are transmitted through a limited number of contact points among large grains.
- This grain-to-grain contact configuration leads to development of indentation stresses, which, combined with the presence of possible pre-existing weak surfaces or flaws (e.g., stylolites and crystal boundaries), cause the fragmentation of the grain mainly by intragranular extension fracturing and, occasionally, by shear fracturing. In the case of intragranular extension fracturing by subparallel fractures, fractures orthogonal to the long axis of the elongated fragments are probably induced by the marked shape anisotropy.
- The elongated fragments, in fact, may easily be flexed during shear deformation and, subsequently, be fractured perpendicularly to the long axis and to the flexure-related tensile fibre stress (e.g., Turcotte and Schubert, 1982; Engelder, 1987; Billi et al., 2003a). This process significantly increases the fracture connectivity and, therefore, the fluid transmissibility through the rock.
- With the disappearance of several large grains by intragranular extension fracturing, and with the development of an embryonic fine matrix, the degree of spatial freedom of the large grains increases. In particular, the absence of large neighbouring grains allows the large survivor grains to rotate and roll under the effect of the fault slip.
- Grain rotation and the rolling, together with grain sliding, are probably the main causes of chipping and subsequent rounding (i.e., by abrasive wear) of large grains (e.g., Heilbronner and Keulen, 2006; Storti et al., 2007).
- Shear fractures seem more frequent when a fine matrix develops (i.e., mature cataclastic fabric). This may be connected with a different state of stress undergone by the grains and with the fact that the development of a fine matrix allows the tectonic and lithostatic stresses to be transmitted through a greater number of contact points among grains. The above-discussed model of fault core development is somewhat supported by recent numerical models of fault gouge evolution (Guo and Morgan, 2007, 2008). In these models, the evolution of gouge zones with slip is split into two subsequent stages. In the early and shorter stage, shear strain is accommodated primarily by extensional fractures and related grain comminution (intragranular extension fracturing). During this stage, the shear zone attains a peak friction value and porosity varies greatly (i.e., either increasing or decreasing) as a function of normal stress. In the subsequent and longer stage, grain rolling and sliding become more significant and fault gouge accumulates (chipping). During this stage, the sliding friction, porosity, and average grain size tend to decrease with slip that is mostly accommodated in the gouge zone. The general abundance of dissolution structures in the samples studied suggests that the pressure solution played its part in the brecciation process of the fault rocks studied. the absence of these structures in some samples is probably due to the local presence of further, albeit small, shear areas which have decreased the porosity and permeability of the rock at that point and which have therefore inhibited the passage of mass and the dissolution process . We have carried out particle size distribution and particle geometry analysis on three different cataclastic fabrics, namely the embryonic cataclasite, the intermediate cataclasite, andthe mature cataclasite to constrain the cataclastic processes associated with the various stages of strain localisation along the fault. Following the genetic model of brecciation processes proposed by J´ebrak (1997), in the Dr-Ds diagram the obtained results fall in the field of corrosive wear, which is thus indicated as the main process leading to cataclasis in the studied samples. In turn, this suggests comminution of carbonate fragments assisted by chemical alteration in the presence of a reactive percolating fluid, which operated through the dissolution of grains at the grain-fluid interface. Microstructural data show that at least two sets of calcite veins formed within the ori ginal primary fabric, and that they are truncated and partially dissolved along pressure-solution seams. Therefore, these fine-grained calcite veins are interpreted as the oldest preserved structures. A younger set of veins, filled by coarse-grained calcite, cuts across the pressure-solution seams and the other structural features, thus documenting multiple events of veining during the faulting history. Accordingly, multiple episodes of fluid ingress accompanied by calcite deposition took place within FRB. Cataclasites consist of a fine-to very-fine-grained matrix hosting poorly-sorted fragments of veined limestone, as well as early cataclased fragments. Pressure-solution seams represent the preferential sites where cataclasis initiated. Hence, we conclude that cataclasis and mechanical comminution postdate and overprint the early sets of calcite veins and pressure-solution seams. With progressing deformation, the foliated fabric forms at the expense of the early veined and cataclased structural facies producing gently anastomosed foliation planes and localised simple-shear dominated structures. In the massive samples, this foliated fabric seems to exploit inherited structural anisotropies, as in the case of the described C shears that rework earlier pressure-solution seams and S surfaces exploiting calcite vein walls. In later stages, clay-rich fault gouge lenses and discrete slip surfaces rework both the non-foliated and the foliated structural facies described above, representing the overall and progressive embrittlement of the system. This final proposed mechanical evolution of the FRB is consistent with a model whereby nucleation and progressive evolution were steered by
I. multiple cycles of deformation evolving from volumetric to localised
II. progressive narrowing of the fault core
III. fluid ingress and competing calcite-precipitating and calcite-dissolving processes
IV. oscillations between strain hardening and strain softening conditions.
These events could have occurred during pulses of coseismic rupturing accommodated by fluid- assisted cataclasis. Most likely, cyclicity continued as long as the general stress field and the local fluid supply conditions remained stable. Our results suggest that strain localisation in carbonate rocks is a dynamic process, which changes in space and time during faulting, and is tightly connected with the evolution of key fault fluid properties steering precipitation vs. dissolution of carbonates. Different brittle structural facies within a fault core may provide key information to reconstruct the progressive sequence of deformation events and, in turn, the general temporal evolution of a fault zone. During progressive localisation, each structural facies sets the ground for the following increments of accommodation.