Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
10 SUPPLEMENTARY NOTE ON ROCK MECHANI CS PROBLEMS AT ACTIVE GINS Important deformation mechanisms include fracture, frictional sliding, cataclastic flow, pressure solution, diffusion creep, and dislocation creep. For any given rock type, one deformation process will be dominant under a particular set of conditions. As either the conditions, the rock type, or the grain size change, the operative process may change with an accompanying change in the Theological behavior. Several examples are instructive for understanding the range of Reformational behavior that might be expected in continental margins. It is commonly believed that the transition from the upper to the lower crust is characterized by an abrupt increase in strength because of a change in rock type from a more silicic to more mafic composition. This occurs because at the temperature and pressure of the transition, the upper crustal rocks -can flow by dislocation creep at moderate deviatoric stress, whereas the stress required for flow to operate in the lower crustal rocks is much higher. Thus, the lower crustal rocks may deform by brittle processes of fracture and frictional sliding because the stress required for this process-is lower than that for dislocation ~ creep e Qualitatively, this idea is very useful, but we lack sufficient knowledge of the constitutive parameters for likely rock types to permit useful quantitative calculations. For example, the constitutive relationship for wet quartzite is often used to represent the upper crust even though it is unlikely-that upper crustal rocks are actually wet quartzites. If they were granitic, for example, the expected stresses would be different because the rheology of granite under these conditions is not the same as that of quartzite. - Local fracture and pervasive cataclastic flow result reduction in grain size. This lower grain size may allow processes, such as pressure solution, diffusion creep, or dislocation creep, to become the dominant Reformational mechanism. In this way, the relationship between stress and strain rate change with time from mean stress-dependent plasticity to linear or nonlinear viscosity. 109
Sliding on faults may occur either by earthquakes or by stable sliding, depending on the constitutive behavior of the rocks in the fault zone. If the constitutive behavior is such that an increase in salting velocity results in a reduction in shear resistance (velocity weakening), then unstable sliding is possible; if it results in an increase in resistance (velocity strengthening), sliding should be stable. This appears to explain the transition from earthquakes at shallow depths on strike-slip faults to stable sliding at greater depths, since an increase in temperature can cause a transition from velocity weakening to velocity strengthening. It might also help to explain the fact that large subduction earthquakes occur at some convergent margins and not at others, because there could be systematic differences in the rock types from one boundary to another e One possibility is that: the constitutive behavior for subducted sediments might be very dif ferent than for serpentinite or for mafic rocks. To know the rheology of various parts of continental margins, it is necessary to have two different data sets. One includes the rheological behavior of important rock types under a wide range of conditions, and the other includes knowledge of the rock types and ambient conditions that exist in the continental margins. The first of these comes largely from laboratory experiments, and the second, from a variety of sources. Laboratory Experiments Experiments must~be done under a wide range of physical conditions, from the low temperatures and pressures found in accreti-onary wedges, to the high temperatures and pressures associated with partial melting deep in subduction zones. In addition, it is important that experiments be done with controlled fluid pressures, because fluids man affect the rheology, both by the purely mechanical fluid pressure effects embodied in the effective pressure concept and through chemical effects. It is also important that experimental-data be obtained at large strains because the strains are often large in continental margin deformation, and it is important that the experimental results can be shown to apply to this situation. The wide ranges in-experimental requirements present challenges to experimental design. A single design of an experimental deformation apparatus cannot be used over the entire range of conditions mentioned above. For any deformation apparatus, digital data collection and computer control offer significant advantages and should be included in the design. Because the study of low permeability clay-rich sediments is important, such apparatus must be designed for conducting experiments that; may last weeks or months. It could be useful to 110
employ rotary shear so that deformation in shear zones can be studied to high strain. Defo Cation of consolidated rocks requires apparatus capable of attaining higher pressures and temperatures. This means that for a given sample size the entire apparatus must be larger than for experiments at lower temperatures and pressures. However, it is not easy to design an apparatus that can cover the entire range of interesting conditions. A number of apparatus exist that have some of the desired capabilities, but design and construction of new, more sophisticated equipment is necessary. Only a small number of laboratories exist that are sufficiently well equipped to conduct the necessary research. Conditions in Continental Margins To apply laboratory experimental results to continental margins, it is necessary to know both the rock types and the physical conditions found in this setting. Relevant physical . conditions include temperature, mean stress, deviatoric stress, and fluid pressure. Deformation processes are highly sensitive to these factors; so unless the operative process is known, it is not possible to predict the appropriate constitutive behavior. Paradox of Stress Magnitudes on Faults In divergent, convergent, and strike-slip continental margins, faults commonly appear to slip at shear stress magnitudes that are lower than one would expect to be possible. Our expectations are based on the simple concept that frictional sliding can only occur if the shear stress is equal to the normal stress times a coefficient of friction. If the normal stress and the coefficient of friction are known, the shear stress can be calculated. The existence of low angle normal faults in extensional margins, the apparently low shear stress on the-San Andreas fault, and the existence of nearly horizontal detachment faults in a number of places, such as the Los Angeles basin, are all phenomena that do not fit this simple mechanical view. What is wrong? It might be tempting to say that, for some reason, laboratory experimental friction coefficents do not apply to natural faults. However, this is not a satisfactory explanation, because there are situations in which natural faults behave exactly as expected based on laboratory measurements. One good example is the field experiment done at Rangely, Colorado, in which fault slip was repeatedly started and stopped by changing the pore fluid pressure through pumping water into and out of boreholes drilled for Chat purpose. The example of Rangley is 111
particularly valuable because it is a location in which the stresses were known by hydraulic fracturing measurements, the pore pressures were directly measured, and laboratory friction measurements were made on rock samples that were known to exist where the faulting was occurring. In this well-characterized case, the frictional behavior was exactly what was expected from Mohr-Coulomb failure analysis using the effective pressure concept. Thus, the paradox that some faults seem to move with low ratios of shear stress to normal stress is not solved by dismissing the results of laboratory experiments and known mechanical relations e There may be some reason that these concepts are not applicable to low stress faults. For instance, some other deformation mechanism might operate instead of frictional sliding If so, we need to discover what it is. In some cases, the rock types might be unusual, with low coefficients of friction, and if so we need to determine what these rock types are and conduct appropriate laboratory experiments on them. The magnitudes and distributions of fluid pressure may be very different from what we now believe is possible. The stress magnitudes may be higher than they appear to be. If we know the operative processes, the mechanical properties of the relevant rocks associated with these processes, and the actual conditions existing on the faults, we should be able to solve the paradox. Our current knowledge of one or more of these factors is inadequate, and we do not even know now in which area the inadequacy lies. Whatever the explanation, we have the opportunity to resolve the paradox. In so doing, we will undoubtedly discover that our current view of the mechanics of these faults is somehow incorrect. Discovering the correct explanation will be a fundamental advance. The implications of this new understanding could be far reaching in many areas of earth science. To solve this paradox, we must combine inputs from many different geological approaches, all focused on the fundamental problem of deformation at continental margins. 112
