Recent years have seen an extraordinary amount of epically powerful earthquakes, and in Sumatra, Chile and Japan Propulsion approaching or 9.0. This has led some to assume that such disasters are the clusters that we have entered a time in that period, and to be feared.
A new analysis of the data showed the U.S. Geological Survey, yes, a long gap between the current and previous major earthquakes statistically unlikely, but it does not prove that it is not a coincidence.
Peter Shearer of the University of California at San Diego and Philip Stark of California, Berkeley (both — the U.S.) pay attention to each earthquake of magnitude above 7.0, what happened from 1900 to 2011. In order to cut off the aftershocks and local deformation caused by the previous earthquake, they established the rule: do not consider any earthquake with a magnitude smaller, which took place over three years, and within a thousand kilometers from the original. This is a very loose definition of aftershocks. It is, for example, prompted the scientists to discard two earthquakes of magnitude 8.5 near Sumatra that took place not far from the previous push. On the other hand, it is consistent with modern scientific ideas of how to increase the pressure of a major earthquake in areas distant from the epicenter of a fairly long distance.
The authors justify their exclusion by the fact that in reality they are interested in global seismic activity, and local events would not affect the analysis. In addition, it is noted that the statistics in a way that is insensitive to the selection criteria: you can shorten the distance by a third and get similar results.
From the analysis on such terms makes clear is that the overall level of strong earthquakes (above 7.0) is now close to the average. Just happens more aftershocks above magnitude 8.0. Such earthquakes currently registered record a lot, but this record is just over the previous figures.
Whether there is something unusual? The authors randomly assigned 100 thousand times earthquake catalog in order to find out whether it is possible to talk about clustering disasters. 85% of these clusters contain randomization earthquakes magnitude 8.0 and above, which are reminiscent of the one that we live in 2004, and 97% — clusters shock strength of 8.5 and above. A number of additional tests showed that the observed distribution corresponds to a Poisson process, that is it by chance.
The researchers then conducted a stepwise analysis of randomization step magnitude 0.1 from 8.0. As a result, three-quarters of a set of randomization containing clusters, the chances that the occurrence did not exceed 5%. Almost a third of them had clusters with a probability of less than 1%.
This is important, since earthquakes of magnitude above 8.5 was almost forty years — since the earthquake in Alaska in 1965. The probability that this could happen, was based on analysis of only 1.3%. But the authors note that a different rule-off rate could be different. One way or another, but "unlikely" does not mean "impossible."
In general, researchers have come to the conclusion that global earthquakes are not clustered. Clusters can be local, but scientists already know how to explain their appearance, while geological mechanism that could lead to a global cluster, science is not known.
On the other hand, it is only a statistical analysis of what seismologists registered for a miserable century.
The results are presented in the journal PNAS.
Prepared according to Ars Technica.