(June 27, 1998)

An Algorythmic Tell-Tale
by Tiziana Lanza

Studies on earthquake prediction in Italy are finally making some headway, despite the fact that the tracking network is still rough. In Russia, however, where seismic activity monitoring is more efficient and widespread, a system is being developed which bases prediction on algorhythms which identify anomalies in the behavior of the Earth's crust. The initial results are quite promising

Nine months have gone by since the Umbrian earthquake, and there's no end in sight. The tremors early this month in Gualdo Tadino, 4-5 on the Mercalli scale, reveal that the sequence hasn't ended yet. Seismologists are thinking things over, since no one predicted the September quake or the one which followed. Even so, retrospective analyses are useful to study the earthquakes and find a way to predict them. However, it's important to make the distinction between different kinds of prediction: one which predicts date, time, place, and size, currently impossible, and the forecast which gives a rough idea of the time, space, and size defining a future seismic event.

For years now, seismologists have played with algorhythms picking our anomalies in the behavior of the earth's crust in order to find indicators of future earthquakes. The Russian scientist Yuri Tyupkin, head of the Geophysics Center at the Moscow Academy of Sciences, at the end of a seminar held recently in Rome at the National Geophysics Institute, explained his findings from his team's algorhythm. The data they used were, however, retrospective.

An algorhythm is based on the idea that there are two phases at the epicenter of a future earthquake: quiescence and premonitory seismic activation. Seismic quiescence is a time interval marked by low seismic activity. For years now Russian scientists from the National Geophysics Institute have proposed modifying the algorhythm for Italy. It was in fact created for major earthquakes in the Caucasus, which are up around 7.0 on the Mercalli. In Italy, it's rare to have such a big quake. Over the past 2500 years, there have only been three so big. The most recent was the one in Messina in 1908, which hit 7.1.

Retrospective analysis of the Umbria quake revealed marked seismic quiescence between January and April 1997, separated from the main tremor by a period of intense seismic activity. In May, there were tremors in the central Appenines, and on September 4 there was a 4.5 tremor in Colfiorito followed by a hundred small aftershocks. These premonitory shocks preceded the 5.6 one on September 26. Further analysis is needed before a prediction method can be found which can reliably predict time and space intervals for upcoming quakes.

The Italian seismologist Rita di Giovambattista, who works with Yuri Tyupkin, notes that "In Italy we're still at the start as far as predicting earthquakes is concerned. We still don't have statistics which allow us to give the right importance to observed anomalies. We need to gather as much data as possible and constantly monitor the territory to be able to judge if a seismic quiescence is a necessary and sufficient condition for a specific zone to have an earthquake." For his part, Yuri Tyupkin finished the seminar saying that the algorhythm is currently just a good tool for study and research. Even in Russia, where prediction has been studied for years, observed anomalies in a certain region are passed only to technicians and never to the general population.

Translation by Jonathan Chaloff