HR: 14:55h
AN: S32D-06 [PDF]
TI: The Prediction of Spatial Aftershock Probabilities (PRESAP)
AU: * McCloskey, J
EM: j.mccloskey@ulster.ac.uk
AF: UNIVERSITY OF ULSTER, SCHOOL OF BIOLOGICAL AND ENVIRONMENTAL SCIENCES
CROMORE ROAD
COLERAINE
NORTHERN IRELAND, COLERAINE, BT52 1SA
AB:
It is now widely accepted that the goal of deterministic earthquake prediction is unattainable in the short term and may even
be forbidden by nonlinearity in the generating dynamics. This nonlinearity does not, however, preclude the estimation of
earthquake probability and, in particular, how this probability might change in space and time; earthquake hazard estimation
might be possible in the absence of earthquake prediction. Recently, there has been a major development in the understanding
of stress triggering of earthquakes which allows accurate calculation of the spatial variation of aftershock probability
following any large earthquake. Over the past few years this Coulomb stress technique (CST) has been the subject of
intensive study in the geophysics literature and has been extremely successful in explaining the spatial distribution of
aftershocks following several major earthquakes. The power of current micro-computers, the great number of local, telemeter
seismic networks, the rapid acquisition of data from satellites coupled with the speed of modern telecommunications and data
transfer all mean that it may be possible that these new techniques could be applied in a forward sense. In other words, it
is theoretically possible today to make predictions of the likely spatial distribution of aftershocks in near-real-time
following a large earthquake. Approximate versions of such predictions could be available within, say, 0.1 days after the
mainshock and might be continually refined and updated over the next 100 days. The European Commission has recently provided
funding for a project to assess the extent to which it is currently possible to move CST predictions into a practically
useful time frame so that low-confidence estimates of aftershock probability might be made within a few hours of an event and
improved in near-real-time, as data of better quality become available over the following day to tens of days.
Specifically, the project aim is to assess the extent to which this is scientifically feasible in terms of our understanding
of the physical phenomena which control the variation of seismicity following a large event due to stress redistribution and
practically possible given present limitations on data availability, data quality and computational or data transfer speeds.
The project is divided into a number of elements designed to reflect the temporal sequence of tasks that must be undertaken
for the prediction of aftershock hazard. These tasks include determining a time-indexed sequence of slip distributions for
both real and synthetic events, calculating a suite of time-indexed stress perturbations and quantitatively comprising
predicted and observed aftershock distributions, and developing techniques for predicting likely strong ground motion from
the predicted spatial distribution of aftershocks.
DE: 0999 General or miscellaneous
SC: Seismology [S]
MN: 2003 Fall Meeting