The seismic events in central Alaska on 23 October last year, and then on 3 November when a magnitude 7.9 earthquake ruptured more than 300 km of the Earth's surface, won't rate much mention in human history. Such was the remoteness of the area that, happily, no one was killed. And despite the forces involved, as seen in this image of ground movement along a highway, and in the occurrence of many huge landslides, damage to infrastructure was comparatively light.

Now, however, analyses of the earthquakes are being recorded for scientific history. One of the first reports appears in Geophysical Research Letters (doi:10.1029/2002GL016724; 2003), where Greg Anderson and Chen Ji describe their modelling of stress transfer during this earthquake sequence. They conclude that the magnitude 6.7 earthquake of 23 October transferred 30–50 kilopascals of stress to the hypocentre of the second (the Denali fault mainshock) — thereby, as the authors put it, “encouraging the occurrence” of the mainshock. From other studies, it seems that levels of 10–20 kPa of stress transfer can trigger further seismicity.

Both of the earthquakes occurred along the Denali fault system, in a mountain range some 90 miles south of Fairbanks. This is a 'strike-slip' fault, where two tectonic plates are in transform movement past one another, akin to the San Andreas fault that so exercises Californians. For context, the Denali fault mainshock was about the same magnitude as the San Francisco earthquake of 1906, with surface offsets across the fault of up to 9 metres.

Like the San Andreas, the Denali fault system is complex, with many regional faults. Anderson and Ji also calculate the combined effect on those faults of the events of 23 October and 3 November. The resulting picture is varied, with minor stress release or increase on most fault segments, but in a couple of cases a massive increase of more than 400 kPa. This is well over the commonly cited threshold for triggering further seismic activity — showing that the threshold is perhaps not a 'hard floor', and that these fault segments might prove especially lively subjects for monitoring.