Shear Stress in Fault Zones Magnitude Slower Than Typical Laboratory from Barry Kean Atkinson Strain Rates

Home , Fault (geology), Shear (geology), Shear stress

708 Nature Vol. 296 22 April 1982 orders of magnitude of some tectonic strain rates but some five orders of Shear stress in fault zones magnitude slower than typical laboratory from Barry Kean Atkinson strain rates. Since then this technique has been applied to the study of stress levels on SINCE estimates of the shear stress acting on example, some surficial rocks have higher water-weakened faults in a range of crustal fault zones at major plate boundaries vary values of deviatoric stress than are found in rocks10 • by at least an order of magnitude, such the western US, perhaps implying shear For faults in many wet crustal rocks a fundamental issues as the driving stresses of several tens of MPa at depth. In dramatic weakening occurs at strain rates mechanism of plate tectonics and the addition, the moment of earthquake below I0-7 s- 1 , probably due to some energetics of crustal faulting remain faulting for a given length of fault break is combination of stress corrosion and entirely unresolved1• The shear stresses in about one order of magnitude higher in diffusional mass transfer. The effect is in the Earth's crust cannot be measured some intraplate regions of the US than in addition to any strength reduction due directly, except very close to the surface, so western California. The argument has, merely to the mechanical effect of high likely stress levels have to be inferred from however, recently been demolished by pore fluid pressures. Additional support is various secondary lines of evidence. Raleigh and Evernden8 - for the US at given to the stress corrosion hypothesis by Laboratory studies of rock friction least. They point to some little appreciated the observation that in quartz, the rate of generally predict that shear stresses on facts. crack growth increases by five orders of earthquake faults must be at least as high as First, attenuation rates of horizontally magnitude on raising the temperature from 100 MPa, unless pore fluid pressures propagating seismic waves with 20 to 200°C (ref. II). approach that due to the overburden. Such frequencies relevant to intensity Some of these experimental results can high pore fluid pressures may occur observations (1-4Hz) are grossly different be extrapolated to conditions thought to transiently during earthquake faulting throughout the US. Second, the energy of occur at depths down to 15 km along the through a shear-heating-induced thermal intensity-relevant frequencies for San Andreas Fault Zone. For sandstone expansion of water2 but an impressive earthquakes in the conterminous US is and quartzite, shear stress for sliding under array of data (summarized in ref. 3) shows solely dependent on fault length and has no slow, tectonic strain rates of lQ- 11 to lQ- 14 that pore pressures much greater than relation to attenuation. Finally, for a fixed s- 1 is of the order of 10 MPa, even if pore hydrostatic will generally only develop in length of fault break there is a very water pressure never exceeds hydrostatic. crustal rock sequences containing thick substantial increase in seismic moment for For granite, however, a modest pore water blankets of impervious rocks. a change in attenuation from that typical of overpressure is required to lower shear A different level of shear stress is western California to that typical of eastern stress for sliding into the lO MPa range. indicated by seismological studies. The US. Thus, moment values have to be The explanation for diversity of shear average stress drop in crustal earthquakes explained within models of the Earth that stresses on crustal fault zones may lie in is usually of the order of 10 MPa, or Jess, are highly heterogeneous. water's different mechanical, ther irrespective of the strength of the These facts can be explained if modynamic and, especially, chemical earthquake source, implying that ambient earthquakes in eastern US occur along properties. There is sufficient variation stress levels are also of this order. The key fault zones that constitute soft inclusions in both in the time scale of fault zone evidence of long-term, low shear stress, an otherwise rigid and strong crust/mantle processes and in the material properties of probably around lO MPa, comes from the system. In order to calculate the true crustal rocks which, given the diverse absence of a heat flow anomaly around the moment of an earthquake, allowance must physico-chemical actions of water, could San Andreas Fault. A recent be made for the effective dimensions of a explain the range of shear stress estimates comprehensive review4 of the latest heat volume in which stress relaxation occurs. for fault zones. There is a strong case8 that flow data for the western US further Relatively high moments for short fault the environment in which many earth supports this view. However, unless breaks are only achieved in association quakes occur is of low shear stress, but with relative plate velocities are unreasonably with a large volume of relaxation which can stronger asperities of varying sizes and high (tens of em yr1), the metamorphic be larger than the inclusion. There is no density of distribution along the fault gradients and K-Ar ages of rocks along the physical anomaly implied by this result, surface which account for Jess common, Alpine Fault of New Zealand are consistent merely a more heterogeneous earth model local shear stress of several tens of MPa. with shear-heating due to a much higher than is usually used in seismology. Future work must concentrate on shear stress5 , of at least 100 MPa. Raleigh and Evernden8 also review the working out the details of potential Geological evidence from fault rock evidence for low shear stress on chemical weakening effects of pore water textures suggests that both high shear stress Californian fault zones and conclude that on rock strength under simulated crustal (100 MPa) and low shear stress (10 MPa) fault zones in all regions of the US are sites conditions. Some reliable means will also faulting occur in the upper crust, with the of low ambient shear stress and low stress be required for extrapolating laboratory latter more common6 • High pore fluid drop. experiments to the larger strains typical of pressures are clearly involved in some low Fault motion at low shear stresses has slip on earthquake fault zones. D stress crustal faulting and higher shear been supposed difficult because high pore stresses are usually associated with reverse fluid pressures have been thought to be the faulting on dry immature fault zones (the only way to reduce the frictional resistance I. Hanks, T.C. & Raleigh, C.B. J. geophys. Res. 85, 7 6083 (1980). Alpine Fault may be exceptionally dry ). of a fault. Recent research has shown, 2. Sibson, R.H. Nature143, 66 (1973). Even if earthquakes along plate however, that the chemical effects of pore 3. Brace, W .F. Int. J. Rock. Mech. Min. Sci. & Geomech. boundaries often occur in an environment Abstr. 17,241 (1980). water may also have a significant effect. 4. Lachenbruch, A.H. & Sass, J.H. J. geophys. Res. 85, where ambient shear stress is low, it could It is notoriously difficult to run 6186 (1980). be argued that intraplate earthquakes are deformation experiments in the laboratory 5. Scholz, C.H., Beavan, J. & Hanks, T.C. J. geophys. Res. 84,6770 (1979). different and occur where the ambient at strain rates comparable with those in 6. Sibson, R.H. J. geophys. Res. 85, 6239 (1980). shear stress is higher. In the eastern US, for tectonically loaded fault zones. Yet, such 7. Sibson, R.H., White, S.H. & Atkinson, B.K. Bull. R. Soc. N.Z. 18, 55 (1979). experiments are essential if the range of 8. Raleigh, C. B. & Evernden, J. Am. geophys. Un. Geophys. potential water-weakening reactions are to Monogr. 24, 173 (1981). be assessed. A few years ago, it was shown 9. Rutter, E.H. & Mainpricc, D.H. Pageophysics 116, 634 Barry Kean Atkinson is in the Department of (1978). that stress relaxation experiments9 could be 10. Atkinson, B.K. & Dennis, S.M. Trans. Am. geophys. Un. Geology, Royal School of Mines, Imperial 62, 1036 (1981). College, Prince Consort Road, London SW7 used to assess strain rates on fault zones II. Meredith,P.G.&Atkinson,B.K. Trans.Am.geophys. Un. 10 1 2BP. down to about I0- s- , within one or two 62, )040 (1981).