SEISMOLOGISTS studying how small earthquakes can trigger much bigger ones have detailed the exact sequence of events that could unleash a major quake near Los Angeles, home to 13 million people. The study will help seismologists worldwide make more accurate predictions of regional risks.
Greg Anderson and two colleagues at the US Geological Survey in Pasadena, California, were worried that the Los Angeles basin might be prone to the same kind of cascade effect that unleashed a massive quake at Denali in southern Alaska last year. There, a small earthquake on 3 November created stresses that led to a magnitude-8 monster 11 days later. 鈥淚t got people thinking it could take place in southern California,鈥� says Anderson.
His team attacked the problem by creating three-dimensional computer models of the Los Angeles region, which is criss-crossed by a web of vertical strike-slip faults and sloping thrust faults. The models included the geometry of the faults and the detailed geology beneath the surface. One model allowed them to specify the land masses that would be displaced by a simulated earthquake, and to track how stresses would spread from it. Another calculated how a quake would unfold, second by second. They fine-tuned the models by comparing their results to past earthquakes and to each other.
Advertisement
The models allowed the team to rule out potential quake sequences that had worried them. For example, a rupture in the Sierra Madre-Cucamonga fault system that tracks the edge of mountains north-east of Los Angeles would not affect the huge San Andreas fault or the nearby San Jacinto fault. 鈥淪lip on the Sierra Madre-Cucamonga system is unlikely to jump directly onto either fault,鈥� says Anderson.
But their simulations did turn up one unlikely yet potentially devastating sequence. They found that a quake could jump from the San Jacinto fault to the neighbouring Sierra Madre-Cucamonga system, creating 鈥渁 single large event鈥� right next to Los Angeles. Anderson is quick to caution that this scenario can happen only if both faults are close to failing. So are they? 鈥淭hat鈥檚 the $64,000 question,鈥� he says.
