Deep Signals May Hold Clues for Understanding How Quakes Start
Scientists studying earthquakes on a large fault in Alaska have identified types of seismic activity that can precede a
The observation raises the possibility that real-time monitoring of similar precursory signals from large faults could
one day be used as a tool for forecasting damaging earthquakes.
The research team, which included seismologist Yoshi Kaneko of GNS Science, studied a number of earthquakes on the
180km-long Minto Flats Fault Zone in central Alaska.
They found that two quakes some years apart, of magnitude 3.7 and 3.8, were preceded by a ‘nucleation process’
characterised by a radiation of high- and low-frequency waves lasting about 20 seconds.
The research was published in Nature Geoscience this week under the title ‘Earthquake nucleation and fault slip complexity in the lower crust of central Alaska’.
The process, which is predicted only by computer models and laboratory analogue experiments at this stage, involved
numerous smaller earthquakes and a very-low-frequency earthquake occurring at about 19km deep that transitions into a
The generally accepted view is that earthquakes start abruptly, with no evidence of a precursory process.
However, the Alaskan research has shown there are complex processes at play deep within the seismogenic zone of some
Dr Kaneko said the study benefitted from a temporary deployment of earthquake recording instruments in the region that
was not well monitored previously. The instruments enabled very precise measurement of timing, depth, and location of
the precursory signals that led up to the earthquakes.
To explain the nucleation process, the study proposed a computer model in which slow slip on a fault transitions to fast
slip that results in a normal earthquake.
Dr Kaneko added that more work would be needed to confirm the validity of the computer model of the Alaskan nucleation
“Research from around the world suggests these type of precursory signals are very rare. We haven’t rigorously searched
for this type of signal in New Zealand yet, but we plan to do this in the future, " he said.
It was not known if faults in New Zealand would show the same kind of precursory activity as the Alaskan fault.
“As of now, we know that most earthquakes start abruptly, without any nucleation signals. This fact makes it very
difficult to predict earthquakes.”
This research has been done as a part of a long-standing collaboration between seismologists Dr Carl Tape from The
University of Alaska Fairbanks and Yoshihiro Kaneko of GNS Science. Dr Kaneko's research on this project was supported
by a Rutherford Discovery Fellowship funded by the Royal Society of New Zealand.
It pays to be well prepared for natural hazards such as earthquakes. Information on being prepared is available at this