SHmax−Sv plane. If such fractures were to form in the current stress field and have an appreciable effect on fluid flow in otherwise low permeability reservoirs, it would result in a simple relationship between fracture orientation, stress orientation and permeability anisotropy. Moreover, the simplistic cartoon shown in Figure 5.1a has straightforward implications for using geophysical techniques such as seismic velocity anisotropy, shear-wave splitting, and amplitude versus offset (AVO) to identify in situ directions of permeability anisotropy (e.g. Crampin 1985; Winterstein and Meadows 1995). The subject of the relationships among freacture orientation, stress orientation and shear velocity will be revisited at the end of Chapter 8.
to it (acting to close it). This said, it is not likely that Mode I fractures affect fluid flow because they cannot have significant aperture at depth. As illustrated in Figure 4.21, when Pfslightly exceeds S3, fractures of any appreciable length would be expected to propagate, thereby dropping Pfand causing the fracture to close. In fact, because the static case in the earth is that (Pf−S3)<0, transient high fluid pressures are required to initiate Mode I fractures (as natural hydrofracs), but following initiation, the pressure is expected to drop and the fractures to close. Hence, only extremely small fracture apertures would be expected, having little effect on flow. Let us consider 0.3 MPa as a reasonable upper bound for Pf−S3in a one meter long Mode I fracture because of the relative ease with the fracture would propagate (Figure 4.21). For reasonable values of ν and E, equation (5.2) demonstrates that the maximum aperture of a Mode I fracture would be on the order of 0.01 mm. Obviously, considering a fracture to be only 1 m long is arbitrary (especially because bmaxincreases as L increases), but as L increases, the maximum value of Pf−S3 decreases thereby limiting bmax(equation 5.2). Of course, real Mode I fractures in rock will not have perfectly smooth surfaces so that even when they are closed, a finite aperture will remain (Brown and Scholz 1986) such that in rocks with almost no matrix permeability, closed Mode 1 fractures can enhance flow to a some extent.
Faults (Mode 2 or 3 fractures that have appreciable shear deformation) are likely to be much better conduits for flow than Mode I fractures. Figure 5.2a (modified from Dholakia, Aydin et al. 1998) schematically illustrates how faults evolve from initially planar Mode I fractures, sometimes called joints, or in some cases, bedding planes.
After the passage of time and rotation of principal stresses, shear stress acting on a planar discontinuity can cause slip to occur. In cemented rocks, shearing will cause brecciation (fragmentation and grain breakage) along the fault surface (as well as dila- tancy associated with shear) as well as damage to the rocks adjacent to the fault plane.
Both processes enable the fault to maintain permeability even if considerable effective normal stress acts across the fault at depth. For this reason, faults that are active in the current stress field can have significant effects on fluid flow in many reservoirs (Barton, Zoback et al. 1995). This will be discussed at greater length in Chapter 11. It should be pointed out that the terms fractures and faults are used somewhat informally in this and the chapters that follow. It should be emphasized that it is likely that with the exception of bedding planes, the majority of planar features observed in image logs (next section) that will have the greatest effect on the flow properties of forma- tions at depth are, in fact, faults – planar discontinuities with a finite amount of shear deformation.
The photographs in Figure 5.2a,b (also from Dholakia, Aydin et al. 1998) illustrates the principle of fault-controlled permeability in the Monterey formation of western California at two different scales. The Monterey is a Miocene age siliceous shale with extremely low matrix permeability. It is both the source rock and reservoir for many oil fields in the region. The porosity created in fault-related breccia zones encountered in
S H E A R E D J O I N T
O N S E T O F DA M AG E
B R E C C I AT I O N I N S H E A R Z O N E
Increase in shear deformation increase in permeability
O I L S TA I N I N G b.
Figure 5.2. Schematic illustration of the evolution of a fault from a joint (after Dholakia, Aydin et al.1998). As shear deformation occurs, brecciation results in interconnected porosity thus enhancing formation permeability. In the Monterey formation of California, oil migration is strongly influenced by the porosity generated by brecciation accompanying shear deformation on faults. This can be observed at various scales in core (a) and outcrop (b). AAPGC 1998 reprinted by permission of the AAPG whose permission is required for futher use.
core samples of the Antelope shale from the Buena Vista Hills field of the San Joaquin basin (left side of Figure 5.2a), as well as outcrops along the coastline (Figure 5.2b), are clearly associated with the presence of hydrocarbons. Thus, the enhancement of permeability resulting from the presence of faults in the Monterey is critically important for hydrocarbon production.
Figures 5.1b–d illustrates the idealized relationships between conjugate sets of nor- mal, strike-slip and reverse faults and the horizontal principal stress (as well as the corresponding Mohr circles). Recalling subjects first mentioned in Chapters 1 and 4 related to Andersonian faulting theory and Mohr–Coulomb failure, respectively, Fig- ures 5.1b–d illustrate the orientation of shear faults with respect to the horizontal and vertical principal stresses, associated Mohr circles and earthquake focal plane mech- anisms associated with normal, strike-slip and reverse faulting. Lower hemisphere stereonets (second column) and earthquake focal plane mechanisms (fifth column) are described below. The first and fourth columns (map views and cross-sections) illustrate the geometrical relations discussed in Chapter 4 in the context of equation (4.43). For a coefficient of friction of 0.6, active normal faults (Figure 5.1b) are expected to strike nearly parallel to the direction of SHmax and conjugate fault sets are expected to be active that dip∼60◦from horizontal in the direction of Shmin. Strike-slip faults (Figure 5.1c) are expected to be nearly vertical and form in conjugate directions approximately 30◦ from the direction of SHmax. Reverse faults (Figure 5.1d) are expected to strike in a direction nearly parallel to the direction of Shmin and dip approximately 30◦ in the SHmaxdirection. The Mohr circles associated with each of these stress states (middle column) simply illustrate the relative magnitudes of the three principal stresses (shown as effective stresses) associated with each stress state.
There are three points that need to be remembered about the idealized relationships illustrated in Figure 5.1. First, these figures illustrate the relationship between poten- tially active faults and the stress state that caused them. In reality, many fractures and faults (of quite variable orientation) may be present in situ that have been introduced by various deformational episodes throughout the history of a given formation. It is likely that many of these faults may be inactive (dead) in the current stress field. Because currently active faults seem most capable of affecting permeability and reservoir per- formance (see also Chapter 11), it will be the subset of all faults in situ that are currently active today that will be of primary interest. The second point to note is that in many parts of the world a transitional stress state is observed. That is, a stress state associated with concurrent strike-slip and normal faulting (SHmax ∼Sv > Shmin) in which both strike-slip and normal faults are potentially active or strike-slip and reverse faulting regime (SHmax > Sv ∼ Shmin) in which strike-slip and reverse faults are potentially active. Examples of these stress states will be seen in Chapters 9–11. Finally, while the concept of conjugate fault sets is theoretically valid, in nature one set of faults is usually dominant such that the simple symmetry seen in Figure 5.1 is a reasonable idealization, but is rarely seen.