PAVIA, Italy — Sumatra 2004. Wenchuan 2008. Haiti 2010. Japan 2011. In each case, the story was the same: An earthquake struck without warning, taking thousands of lives and smashing buildings like sand castles.
Geoscientists still can’t predict when a major quake will strike, and many have given up trying. But many do try to issue more general forecasts of hazards and potential damage. This week, researchers added a potentially powerful new tool to their kit: the largest seismic database of its kind ever constructed, based on tens of thousands of earthquake records stretching back more than 1,000 years. Together with a new global map of strain accumulation at plate boundaries, the data sets will form the core of an international public-private partnership intended to reshape the science of earthquake forecasting.
The Global Earthquake Model (GEM) Foundation, which has been developing the seismic risk platform since 2009, unveiled key components of it at a meeting here, called Reveal 2013.
“To advance the science of earthquake forecasting we must enhance the data,” GEM co-founder Ross Stein told the opening session. Stein, a seismologist with the U.S. Geological Survey, and a handful of other disaster experts proposed the project in 2006, with the aim of bringing a systematic and open approach to a discipline that had previously been hampered by disparate methods. “Everyone knew we needed to do this,” Stein says. “No one was willing to put the money up. GEM did.”
To develop the new maps, researchers re-figured the magnitudes and locations of nearly a thousand historic earthquakes dating back to 1000 C.E. according to strict, uniform new criteria. Using modern algorithms, they also recalculated seismogram records of 20,000 earthquakes over the past 100 years. To get a handle on the driving force behind most earthquakes, other geoscientists reappraised the movements of Earth’s tectonic plates to estimate the rate of deformation at plate boundaries. In all, they deduced 20,000 velocities from measurements at 70,000 stations around the world. GEM says that the software used to analyze the data, known as OpenQuake, will be publicly released next year to set a uniform and open standard in hazard calculations.
GEM hopes to develop better ways to calculate both seismic hazard-the probability that earthquakes will occur over the next 50 years-and seismic risk, the casualties and economic losses likely to result. Risk is of particular interest to the insurance companies that fund much of GEM’s research. At the meeting, researchers unveiled an ambitious new project to assess the quality of building stocks around the world, using national data sets, satellite imagery, and on-the-street censuses recorded by smart phone apps.
Reveal 2013 participants say that they are impressed by the partnership’s rapid progress. “We do have our own world map of natural hazards,” says Anselm Smolka of the insurance firm Munich Re. “But with a team of just 20, we cannot do more. We need cooperation. GEM gives industry access to a global network of scientists.”
Seth Stein, a geophysicist at Northwestern University in Evanston, Illinois, and a longtime critic of quake forecasting, says that he welcomes the attempt to bring clarity to the field. But he warns that no amount of data collection can overcome the deep uncertainties inherent in Earth faulting processes. For example, he says, even a complete 2000-year history of seismic shaking in China would have given planners no clues to the three most devastating events that have struck since 1950. “GEM is doing a good job of improving the things that can be improved,” says Seth Stein (who is not related to Ross Stein). “So a better record will help us some. But whether it will ever be adequate to allow us to fully characterize the probability of events, my guess is probably no.”
Ross Stein acknowledges the need for humility in the face of Earth’s complexities: “We must tell the public what we know; and we must tell them what we don’t know-in language they understand.”
This is adapted from ScienceNOW, the online daily news service of the journal Science. http://news.sciencemag.org
© 2013, ScienceNOW
