14 DAY TRIAL // Geologists have been trying to develop the ultimate earthquake-forecasting tool for decades, and thus far their efforts have led to the creation of several models that could be used for this purpose. Recently, a team of scientists compared 7 of these models, in an attempt to establish which one is the best.
These simulations are used to help people living in high-risk areas, such as the Pacific Ring of Fire, which encompasses the west coast of the United States, Chile and Japan, among other countries.
Predicting a tremor even half an hour in advance could significantly reduce the human and material damages associated with such an event. But figuring out when the next big one will strike is a tremendously complex science, since there is no way for experts to descend inside fault lines.
These are regions at the boundaries of tectonic plates, where the latter interact with each other. If the two don't flow smoothly against each other, then they might become jammed, while energy accumulates in both of them.
When that stored energy exceeds a certain, variable threshold, the plates snap free from each other, producing earthquakes. The amount of energy released during such an event is directly proportional with the magnitude of the tremor, Our Amazing Planet reports.
All earthquake models revolve around predicting the intricate interactions that underlie these energy releases. The simulations seek to determine when the plates snap free from each other, before the actual event happens.
Though monitoring efforts are in place in Japan, none of the existing tremor models was able to predict the magnitude 9.0 Tohoku Earthquake, which hit the Asian nation on March 11. The 2004 Parkfield magnitude 6 quake that struck California also came without a warning.
However, geologists note that these events do not take place randomly. All models in the Regional Earthquake Likelihood Models (RELM) comparative test were designed to identify the patterns.
The test was conducted by a consortium of 600 scientists called the Southern California Earthquake Center, which operates with funds from the US Geological Survey (USGS) and the National Science Foundation (NSF).
A computer model utilizing a method called pattern informatics was determined to be the most efficient of the group. It operates by scanning for anomalous increases and decreases in seismic activity. When activity levels exceed thresholds set by past tremors, a flag is attached to the troublesome area.
The model flagged 17 of the 22 grid areas that were hit by earthquakes over the past few years. This method was developed by University of California in Davis (UCD) geophysicist Donal Turcotte.
“We are not predicting the occurrence of a specific earthquake. We are giving the relative risk of occurrence of earthquakes,” the researcher explains.