Earthquakes within tectonic interiors may be the product of long aftershock sequences, rather than present day deformation, researchers say. Ian Randall reports on a controversial hypothesis
Geoscientist Online 9 December 2009
The claim, by Seth Stein (Department of Earth and Planetary Sciences, Northwestern University) and Mian Liu (Department of Geological Sciences, University of Missouri) casts doubt of the whole basis of intra-plate earthquake predictions.
At plate boundaries, known tectonic motions enable reliable estimates to be made of the likely locations of and intervals between seismic events. However, far from such boundaries, problems arise. As Stein & Liu succinctly put it: “the interiors of ideal plates should not deform”.
However, at present, earthquakes within continental interiors are regarded as indicating continuing deformation, as at plate boundaries. If this is true, then past seismic activity can be used to predict future motion, and to forecast where and when future events might occur.
In the region of New Madrid and Charleston, in the United States magnitude seven earthquakes in the 19th Century create, says Tom Parsons (US Geological Survey): “…
two large ‘bull’s-eyes’ on the US National Seismic Hazard Maps”. Far from indicating continuing strain accumulation, however, Stein & Liu propose that these earthquakes may actually have been aftershocks from previous, large-scale events. For this reason they believe that an inverse relationship exists between aftershock duration and the rate of intraplate deformation.
Aftershocks are a phenomenon that commonly follows large-scale seismic events - the product of stress changes induced by the main movement. Gradually the magnitude of aftershocks decays back to the normal background level of seismicity typical of the region. This decrease usually takes around a decade following the main shock, for most large-scale earthquakes. At plate boundaries, the regular motion quickly reloads faults with new stresses, drowning out the effect of the main earthquake. Within plate interiors, however, with deformation occurring at rates below one millimetre per year, reloading occurs far more slowly, allowing aftershocks to go on occurring hundreds of years after the event.
Aftershock sequences decay exponentially with time; therefore it becomes possible to calculate aftershock durations by studying the changing magnitudes of progressive events towards the background rate. For longer sequences, this becomes less easy once the aftershock duration exceeds the available historic earthquake record, which (especially in the USA) only up to about two centuries at most.
In New Madrid and Charleston, if the seismic events there form part of a declining series of aftershocks, this must mean that the overall potential hazard from earthquakes is falling – and not increasing as would be the case at a plate boundary. This would imply that time and resources devoted to earthquake protection in these areas might be excessive.
Stein and Liu’s argument has come in for criticism from those who say that the window of observation on these supposed aftershock sequences is be too small to enable scientists to make valid forecasts. Tom Parsons says: “there is still a tendency to place undue emphasis on known past events in hazard assessments when trying to see the future.”
Dr Michael Ellis, (British Geological Survey) a specialist in landscape evolution and active tectonics and formerly Professor at the Center for Earthquake Research and Information in the University of Memphis, agrees. He told
Geoscientist: “In any steady-state deformation, there will be a natural variability about some mean. Stein and Liu are chasing the mean value and asking if it’s steady or declining. Our historical window, though, is likely too short to characterize the variability, and so we cannot yet identify the mean value.”
Furthermore, it has been suggested that we need to seek more information for regions beyond Charleston and New Madrid before we can begin to use this theory to draw conclusions about the nature of the earthquake risk in a given region.
Ellis also points out that one should expect within continental interiors to see deformation that is more variable, in both space and time: “Unlike plate boundaries, where almost all of the deformation has essentially collapsed onto a two-dimensional plane… deformation within continents (and more generally, within plates) tends to be three-dimensional, and therefore more variously-oriented faults are involved.”
In addition, it may be unrealistic to assume that earthquakes in regions such as New Madrid and Charleston are caused solely by the slow build-up of regional deformation. Such events could as easily be generated by localised seismic pressures, or larger scale processes, such as crustal rebound due to deglaciation.
Regardless, Stein & Liu propose that, in order to overcome the uncertainty they believe exists within earthquake prediction, greater focus should be given to studying fault networks that lie within continental interiors.
Reference
Nature 462, 87-89 (5 November 2009) Long aftershock sequences within continents and implications for earthquake hazard assessment, Seth Stein & Mian Liu
Further reading
Geoscientist 15.7, pp5-7 (July 2005) The largest earthquakes in the lower 48. Ellis, Michael Alexander.