Published online 2 June 2010 | Nature | doi:10.1038/news.2010.263


Earthquake risk calculator goes global

Model should enable researchers to reduce vulnerability to seismic shocks.

Clifford Berrette, 11, sits on the rubble of his neighbor's home in Port-au-Prince, Haiti, on Wednesday, January 27, 2010A model could help people to calculate the risk of a local tremor.C. Cole/Los Angeles Times/MCT/NEWSCOM

A global project to predict the risk from earthquakes to different communities is edging closer to reality.

The Global Earthquake Model (GEM) will calculate seismic risk based on the geological characteristics of an area, as well as the types of building and the building regulations in place.

Results of the pilot phase are being announced at an outreach meeting in Washington DC on 3-4 June. So far, the project has produced technical tools needed to calculate risk, and software to help scientists work with the data. These tools, and the results they produce, will now be carefully checked before the first verified estimates of seismic risk — and global maps showing these risks — are released. All the models, the code used to write them and the resulting data will eventually be freely available.

But the real challenge, say some observers, will be making this data relevant to the people who need it most — the inhabitants of some of the world's poorest and earthquake-threatened countries, such as Haiti, which suffered a catastrophic earthquake in January this year.

Seismic success

The GEM project, which began in 2006, involves partners from national governments, international bodies such as the World Bank and private sponsors. It will be fully working by the end of 2013, if not sooner, says GEM's secretary general Rui Pinho, a structural engineer at the University of Pavia in Italy.

During the pilot phase, seismic-risk and hazard specialists gathered information from existing models, and extracted the best parts to develop their own risk and hazard 'engines'. These engines are computer programs designed to estimate the risk that seismic activity poses to a particular region.

“The results are clearly demonstrating that we have something capable of calculating hazard for the entire globe.”

Helen Crowley, a civil engineer at the European Centre for Training and Research in Earthquake Engineering in Pavia, led the seismic-risk-engine team. She says that they've come much further than expected during the pilot phase. Her team has developed the engine and tested it with real data, first for a small area a few kilometres square. At first, the program took half an hour to compute the estimated probability of the loss of life over 50 or 100 years. Now the engine can do the same calculations in just a few seconds. "We didn't think we'd get that far," says Crowley.

Marco Pagani, a seismic engineer at the Swiss Federal Institute of Technology (ETH) Zurich, ran the part of the project that attempted to predict the 'hazard' — the probability that the ground will move more than a given amount in a given period of time. So far, Pagani is confident that his team's models work well. "The results are clearly demonstrating that we have something capable of calculating hazard for the entire globe," he says.

Less slick, more smart

But at the other end of the complex computer models is the nitty-gritty of life in an earthquake zone. Communicating risk can be as big a problem as calculating it. Recently invited to help with GEM is Brian Tucker, the founder of GeoHazards International, based in Palo Alto, California, who works to reduce the earthquake vulnerability of communities in developing countries. He is concerned that the needs of these end users were not considered in the first round of GEM.

The final model is going to be "slick", says Tucker, and that will greatly benefit technical researchers. "But my hunch is that what is needed is not such a sophisticated approach," he says. Tucker hopes that at the meeting in Washington DC he will be able to impress on the GEM leadership the importance of identifying who they want to use their model, and build it accordingly. "What a homeowner needs is not just different to what a researcher or insurer wants — it could be in some ways antithetical," says Tucker.


The long-term aim of GEM is to allow local officials or members of the public to add data about their local area using a communication portal, and run a calculation to get an idea of the seismic risk the area is subject to.

The tests of the first round of models against real-life situations "demonstrate what unfortunately was shown to us; that Haiti was an area of high seismic risk", says Pinho. The next stage is to check the seismic risk and hazard data produced so far, and by 2011 the portals where the tools can be accessed should be ready for release, says Pinho. Pinho hopes that in developing countries GEM could help to bring in simple rules for sound construction, in areas where many of the local population remain unaware of the seismic risk to which they are vulnerable. "In developing countries the focus needs to be on saving lives," he says. 

Commenting is now closed.