Scientists anxious about other big quakes

9.4 earthquake (more than 3,000 times stronger than the one in Haiti) occurred off the coast of Sumatra. Data from computer models, published in the journal Science, warned that the gigantic tremors had increased the amount of stress on two other two areas of the fault and on another fault to the east. In March of 2005, this eastern fault experienced a magnitude 8.7 quake.

The smaller Haitian quake on the Enriquillo fault is unlikely to have such widespread consequences but could affect the nearby Septentrional fault in the north, which runs through Haiti and the Dominican Republic and has been quiet for more than 800 years, according to USGS scientist Prentice.

To determine how the strains in both of these faults have changed this week, seismologists like Ross Stein of the USGS are analyzing measurements that show where and how violently the earth shook on Tuesday and feeding this data into computer programs that model how stress in other regions of the fault has changed. But the scientists admit that the models cannot perfectly predict what will happen next. “We have to be really humble as earth scientists,” Stein said. “We have a colossal record of failing to predict [earthquakes in advance].”

Haiti may be particularly challenging because seismologists like Stein would normally check local seismometers — instruments that record data about earthquakes — for smaller “aftershock” earthquakes in the stressed regions their models predicted. For Haiti, this may be impossible because the island has very few seismometers and a crude catalog of the aftershocks at best. Even the data that described the large original earthquake came largely from distant instruments in the United States. “We have a poorer idea of how the fault slipped than we would if this were an earthquake in the United States or Japan, and we have a poorer record of the aftershocks that follow it,” said Stein. “We have one or more hands tied behind our back.”


No time

It will be weeks before scientists know whether these models give us a better idea of new risks in the Caribbean. Stein said, though, that even when models do identify areas of risk — whether in the Caribbean, Turkey or California — they cannot predict when the next earthquake will come.

To predict when earthquakes will occur, some scientists try to examine the historical record of when and how often quakes have occurred in the past. This too poses challenges, said Stein, because only the very tiniest temblors — of magnitude 1 or 2 — tend to occur in regular, repeating intervals.

There has been some progress in long-term earthquake forecasting where we estimate the likelihood in a period of years to decades of an earthquake occurring,” said Michael Hamburger of Indiana University in Bloomington. “But that is not the same as short-term earthquake prediction — to say that an earthquake is happening in the next few weeks, and we can evacuate a town.”