SELF-ORGANIZED CRITICALITY AND FLUID-ROCK INTERACTIONS IN THE BRITTLEFIELD

Citation
Ig. Main et al., SELF-ORGANIZED CRITICALITY AND FLUID-ROCK INTERACTIONS IN THE BRITTLEFIELD, Pure and Applied Geophysics, 142(3-4), 1994, pp. 529-543
Citations number
NO
Categorie Soggetti
Geosciences, Interdisciplinary
Journal title
ISSN journal
00334553
Volume
142
Issue
3-4
Year of publication
1994
Pages
529 - 543
Database
ISI
SICI code
0033-4553(1994)142:3-4<529:SCAFII>2.0.ZU;2-8
Abstract
The concept of self-organised criticality (SOC) has recently been sugg ested as a paradigm for the long-term behaviour of earthquakes, even t hough many of the currently-proposed models require some tuning of the state variables or local conservation rules to produce the universall y-observed Gutenberg-Richter frequency-magnitude distribution with a b value near 1. For example, a systematic negative correlation is predi cted between model b values and the degree of conservation of local fo rce after the slip of a single element in an elastic spring/block/fric tional slider model. A similar relation is described here for a cellul ar automaton model with constitutive laws based on fracture mechanics. Such systems, although critical phenomena in the sense of producing o rder on all scales, are clearly not universal, and may not in general even be true examples of SOC. Nevertheless they adequately reproduce b oth the observed power-law (fractal or multifractal) scaling and its r eported short-term fluctuation. We also present experimental and field evidence for similar systematic variations in b value with the degree of force conservation (expressed in terms of a normalised crack exten sion force) during subcritical crack growth involving the physical and chemical influence of pore fluids during a single cycle of failure bo th in tension and compression. We find that the level of conservation is strongly influenced by fluid-rock interaction under stress, allowin g energy partition into processes such as: physico-chemical stress cor rosion reactions; the dissolution and precipitation of mineral species on crack surfaces; and the purely mechanical phenomenon of dilatant h ardening. All of these are known to occur in the Earth on a local scal e, but few have been explicitly included in automaton models of seismi city. The implication is that over long time periods pore fluids may e xert a strong physical and chemical influence on the universal state o f SOC which the system evolves in a complex interplay of local feedbac k mechanisms keeping the system near criticality, perhaps most strikin gly due to the 'valve' action of faults. In the short term, crustal fl uids might nevertheless be responsible for systematic local fluctuatio ns about this average state.