ISSN:
1420-9136
Keywords:
Rock friction
;
constitutive behavior
;
earthquake prediction
;
fault mechanics
;
fault creep
;
strainmeters
;
creepmeters
Source:
Springer Online Journal Archives 1860-2000
Topics:
Geosciences
,
Physics
Notes:
Abstract Laboratory experiments show that the frictional resistance of rocks depends on the velocity of sliding and the state of the sliding surface as well as on the normal stress. Although the dependence on velocity is small in magnitude, and consequently difficult to measure with high accuracy, this dependence plays a major role in whether frictional sliding is stable or unstable. Constitutive descriptions of laboratory results involve a characteristic distance of sliding, over which the frictional resistance evolves following a step change in sliding velocity. Interactions occur between the elastic response of laboratory testing machines and the change in resistance with slip resulting from this evolution and these interactions are responsible for the stability of sliding. In most situations, materials that show an increase in steady state resistance with increases in sliding velocity (velocity strengthening) will slide stably, while the opposite velocity dependence (velocity weakening) can result in either stable or unstable sliding. The laboratory results suggest that changes in strain and velocity should occur prior to earthquakes. Extrapolation of laboratory results to the earth requires knowledge of how to scale the laboratory constitutive parameters. Use of laboratory constitutive laws to aid in understanding natural fault behavior also requires numerical models to deal with spatial variations of constitutive parameters on fault surfaces and of elastic strain in the adjacent rock. The model of strike slip faulting presented byTse andRice (1986) employs laboratory based constitutive laws and is used in this paper to explore the implications of laboratory results for designing a field monitoring program for earthquake prediction. The results of one of their model simulations are used to calculate the temporal and spatial variation of displacement and strain during an entire model earthquake cycle, with emphasis on the changes that occur in the time period imminent to an earthquake. The premonitory changes in strain that occur are quite small in magnitude near the earth's surface and although detectable with existing shallow borehole instruments, they may only be distinguishable from environmental noise within one month of the earthquake. Strain changes at focal depths of several kilometers would be at detectable levels for a longer time if measurements in suitably deep drill holes could be made. Premonitory changes in velocity of points at the earth's surface are predicted to be of sufficient magnitude that they should be measurable with two-color geodimeters and useful for earthquake prediction, especially if points at distances from a fault equal to focal depths are included. Predicted premonitory displacements at depth are significant; thus it would be valuable to develop techniques for precise surveying of deep drill holes.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF00879010
Permalink