HR: 0830h
AN: S41C-0083    [PDF]
TI: Forward modelling of Coulomb stress: Does mu matter?
AU: * Steacy, S
EM: s.steacy@ulster.ac.uk
AF: Geophysics Research Group, University of Ulster, Coleraine, BT52 1SA Ireland
AU: Marsan, D
EM: David.Marsan@univ-savoie.fr
AF: Laboratoire de G‚ophysique Interne et de Tectonophysique, Universit‚ de Savoie, Le Bourget-du-Lac, 123 France
AB: The clear correlation between Coulomb stress perturbations and the spatial distribution of aftershocks suggests that it may be possible to make real time Coulomb maps delineating areas likely, or unlikely, to experience aftershocks. As part of the PRESAP project investigating this possibility, we have assessed the importance of the coefficient of friction in computing these maps. Although the static coefficient of friction measured in the laboratory is consistently between about 0.6 and 0.8; Coulomb modellers have suggested widely varying values of "effective friction" (i.e. including some unknown fluid effects) ranging from 0.0 to 0.8. Here we investigate the effect of mu on the spatial correlation between the Coulomb map and the aftershock distribution for a number of mainshocks and for varying assumptions of the planes upon which the stress should be resolved. For each model, we compute a Boolean correlation coefficient between the stress field and the aftershock distribution, a method that avoids the problem of nodal plane ambiguity. Preliminary results for the Landers earthquake show that the Coulomb maps are insensitive to the coefficient of friction and visual inspection of the maps support this assessment. If this result is supported by the remainder of the analysis, it will be very encouraging from a forward modelling perspective as it means that the predictive capacity of Coulomb maps is independent of the choice of a poorly constrained coefficient of friction.
DE: 7223 Seismic hazard assessment and prediction
SC: Seismology [S]
MN: 2003 Fall Meeting