Abstract - At a previous Hedberg Conference the question was raised as to whether the problem of predicting fault seal potential is better approached by interpreting data on the basis of a sound theoretical understanding or, less ambitiously, by relying on empirical methods. While the theoretical approach of traditional science has obvious appeal to academics it has to be acknowledged that many natural systems have chaotic properties and are therefore not susceptible to prediction of specific instances or at specific places. A case for empirical methods can be made by considering the processes by which fault zones are generated. As these processes are essentially non-linear, their products are also highly non-linear and therefore unpredictable other than in a very general way.
Lithological juxtapositions across a fault can be predicted by construction of Allen diagrams. However, faults are not notional surfaces across which displacements have occurred but consist of fault zones of varying thicknesses and properties, which strongly modify the hydraulic effects of juxtaposition geometries. The two basic processes which contribute to the thickness and content of a fault zone at a point on a fault both generate multiple slip surfaces. These bifurcation processes operate either at the propagating tip-line of a fault, tip-line bifurcation, or at points on an existing fault surface, asperity birfurcation. Both processes operate on a wide range of scales. Even if fault rocks remained lithologically similar to their source rocks, the existence of multiple slip surfaces within a fault zone strongly modifies the Allen diagram. Two 50m throw faults, for example, give rise to different juxtapositions than does a single 100m fault. As the number of slip surfaces varies rapidly and unpredictably over a fault surface, the juxtapostions must also vary unpredictably.
But the content of a fault zone also varies in a highly unsystematic manner. This variation is mainly due to the continued operation of asperity bifurcation processes throughout the active life of a fault. Asperity bifurcation occurs when an irregularity, on either the hangingwall or footwall bounding slip surface, is mechanically eroded by creation of a new slip surface which isolates it from the wallrock. While irregularities of the bounding slip surfaces may originate early in the fault development, by refraction of the original fracture across boundaries between mechanically contrasting layers, they are also continuously generated throughout the active life of a fault by bed-parallel slip, for example. The bed-parallel slip may take the form of flexural slip in the wallrock, which accommodates footwall uplift and hangingwall subsidence, or may be accommodating the shear deformation which is characteristic of wallrock between closely spaced faults. The bed-parallel slip generates saw-tooth asperities on the fault zone margins which are repeatedly eroded by further fault displacement. This process is a major contributor of wallrock clasts into the fault zone.
The chaotic variations of both content and thickness of a fault zone which would be expected to result from the bifurcation processes, are evident in outcropping fault zones. Shale smears, which represent an additional element of complexity, are often associated with individual slip surfaces within fault zones. The number of shale layers or smears in fault zones varies as rapidly and unpredictably as the number of slip surfaces. Perhaps the most crucial observation is that where the wallrocks contain significant shales, a fault zone will contain at least one shale layer. This one relatively constant element, as quantified by the various shale smear algorithms in use, accounts for the relative success of predictions based on shale gouge ratios, with values of SGR >15-20 fairly consistently associated with sealing faults.
While empirical methods may provide a less than ideal solution to fault seal assessment, they should be viewed in the light of what is possible given the nature of the natural system, rather than judged by the standards of traditional scientific rigour. Specific prediction has been shown to be impossible in many other natural systems even though the systems are well understood and their parameters quantified.
Abstract of talk given to:
Reservoir Scale Deformation - Characterization and Prediction, AAPG Hedberg Research Conference, Bryce, Utah, June 22-28 1997