On the internal structure and mechanics of large strike-slip fault zones: field observations of the Carboneras fault in southeastern Spain

Abstract

Deciphering the internal structure of large fault zones is fundamental if a proper understanding is to be gained of their mechanical, hydrological and seismological properties. To this end, new detailed mapping and microstructural observations of the excellently exposed Carboneras fault zone in southeastern Spain have been used to elucidate both the internal arrangement of fault products and their likely mechanical properties. The fault is a 40 km offset strike-slip fault, which constitutes part of the Africa–Iberia plate boundary. The zone of faulting is ∼1 km in width not including the associated damage zone surrounding the fault. It is composed of continuous strands of phyllosilicate-rich fault gouge that bound lenses of variably broken-up protolith. This arrangement provides a number of fluid flow and fluid sealing possibilities within the fault zone. The gouge strands exhibit distributed deformation and are inferred to have strain hardening and/or velocity hardening characteristics. Also included in the fault zone are blocks of dolomite that contain thin (<1 cm thick) fault planes inferred to have been produced by strain weakening/velocity weakening behaviour. These fault planes have a predominantly R₁ Riedel shear orientation and are arranged in an en echelon pattern. A conceptual model of this type of wide fault zone is proposed which contrasts with previous narrow fault zone models. The observed structural and inferred mechanical characteristics of the Carboneras fault zone are compared to seismological observations of the San Andreas fault around Parkfield, CA. Similarities suggest that the Carboneras fault structure may be a useful analogue for this portion of the San Andreas fault at depth.

Introduction

Characterization of the internal structure of fault zones is an essential precursor to understanding and predicting their mechanical, hydraulic and seismic properties. Large-displacement faults generally produce much wider zones of deformation than smaller offset features. This provides much more scope for developing a complex internal geometry, which in turn may lead to significant modifications of the properties of these discontinuities. Indeed, several authors have suggested that transcurrent plate-boundaries such as the San Andreas fault are mechanically weak Lachenbruch and Sass, 1980, Zoback et al., 1987, a characteristic which does not seem to apply to smaller displacement faults (Townend and Zoback, 2000) and which by implication probably do not cut the entire thickness of the continental crust.

Previously, field studies of fault zones and geophysical studies, including modelling of seismic fault zone waves and inferences of the seismic velocity structure, have helped elucidate the internal structure of major strike-slip systems.

For field studies, efforts are hampered by the generally poor exposure that accompanies active deformation on a large fault. Additionally, the types of fault rocks that are produced by brittle deformation, such as friable fault gouges, are easily and quickly degraded by surface weathering effects. For this reason, field studies have concentrated on recently exhumed traces of young but extinct fault zones that are exposed in arid climates Caine et al., 1996, Miller, 1996, Foxford et al., 1998. For large transcurrent fault zones which cut through the entire lithosphere (i.e. plate-bounding strike-slip systems), there are very few places that are suitable for studying the detailed internal structure in the field. Owing to these difficulties, there are remarkably few descriptions of large fault internal geometry from direct observations (but see Chester and Logan, 1986, Chester et al., 1993, Evans et al., 1997, Chester and Chester, 1998, Wibberley and Shimamoto, 2003).

Geophysical techniques have provided an opportunity to study actively deforming large fault zones at depth. The most intensively studied fault zone in the world, the San Andreas fault, has yielded a wealth of geophysical information. Characteristics such as high V_p/V_s ratios have allowed us to speculate on fluid pressures within the fault Johnson and McEvilly, 1995, Thurber et al., 1997. Seismic fault zone wave studies have permitted geophysicists to model fault zone widths Li et al., 1990, Li et al., 1997, although this forward modelling relies somewhat on prior knowledge of, or assumptions about, the fault zone structure (Ben-Zion, 1998). Most recently, improved techniques for the processing of seismic signals have led to high-resolution location of earthquake hypocentres, which has made it possible to map out aspects of fault zone internal structure from microseismicity Nadeau et al., 1994, Nadeau et al., 1995, Rubin et al., 1999, Waldhauser et al., 1999.

The purpose of this paper is to present new observations of the internal structure of a large transcurrent fault exposed in southeastern Spain. The Carboneras fault is inferred to have ∼40 km offset and, as such, must cut through the entire crust and lithosphere. The Carboneras fault differs from other large-displacement strike-slip faults described in the literature because it contains a high percentage of phyllosilicate-rich material. This feature undoubtedly affects the mechanical properties, which in turn determine the internal structure of the fault. The fault has been exhumed from depth during the past 10 Ma and the climate is now semi-arid. For these reasons, the fault has unparalleled exposure and the fault rocks are excellently preserved. The paper first outlines the tectonic and geological setting of the Carboneras fault, then describes detailed field observations of the large-scale internal structure of this fault. Next, the predominant types of fault product, with their macroscopic and microscopic internal structures, are assessed in terms of the likely mechanical behaviour they will exhibit (utilizing the results of previous laboratory studies) and hence the type of seismic slip characteristics it is likely to produce. In the discussion, the fault described is in contrast with previous observations of the internal structure of large fault zones that generally have been derived from phyllosilicate-poor protoliths. Finally, the field observations from the Carboneras area are compared with geophysical observations of the San Andreas fault around Parkfield, CA.