In 2000, the International Code Council released a new building code, stemming from the USGS map, that suggests that buildings within the New Madrid seismic zone (NMSZ) should be built, or retrofitted, with earthquake mitigation standards similar to those in southern California. State and local officials within the NMSZ are currently deciding whether or not to adopt the new code, IBC2000, which was developed under the supervision of the Federal Emergency Management Agency.
According to an article in the May 13 issue of Eos, adopting the code would be a mistake. The earthquake hazard in NMSZ is much smaller than the USGS maps suggest, and mitigation costs would far outweigh the benefits, the authors say.
The USGS map quantifies earthquake risk in the NMSZ using two factors, says Art Frankel, a geophysicist at the USGS in Denver and lead developer of the National Seismic Hazards Map. The first is the frequency of large earthquakes. Deposits of liquefied sand from past earthquakes indicate that magnitude-7.5 to 8.0 earthquakes hit the region once every 500 years or so, Frankel says. The second factor is the degree of ground shaking that a quake induces. Earthquakes of a similar size cause a lot more shaking in the NMSZ than they do in California due to differences in underlying geology, Frankel says. Therefore, while the rate of quakes is lower in NMSZ, the total risk -- the product of earthquake frequency and its impact on groundshaking, averaged over time -- is just as high in the NMSZ as it is in California.
Andrew Newman, co-author of the Eos article and a seismologist at the Los Alamos National Laboratory, argues that a magnitude-7.5 to 8.0 quake simply could not have occurred in the region once every 500 years. He and colleagues used GPS to precisely measure how quickly different parts of the ground moved, relative to on another, across the NMSZ from 1991 to 1997. They found almost zero motion, suggesting that there could not be enough of a driving force to cause large earthquakes at any regular frequency. For earthquakes to occur, generally two plates, or two segments of plate, have to move in relation to one another to generate the forces causing faults to slip, Newman says. In California, for example, the North America Plate and the Pacific Plate are moving past one another at rate of about 46 millimeters per year, Newman says -- driving the frequent quakes along the San Andreas fault. In the NMSZ, it is much more likely that magnitude 8.0 quakes occur once every 4,000 or 5,000 years, Newman concluded in a 1999 Science paper.
But intraplate earthquakes are unique, says Stanford geophysicist Paul Segall, and it may not be appropriate to apply theories developed from earthquakes at plate boundaries to intraplate quakes. He and his colleagues developed a physical model, published in Science in 2000, indicating that frequent, strong earthquakes would not necessarily require large movements across the NMSZ.
Another piece of evidence suggesting that the USGS map overestimates the quake hazard lies in the Gutenberg-Richter Relation, says Seth Stein, a seismologist at Northwestern University in Evanston, Ill., and lead author of the Eos article. That relation says that for every single increase in the magnitude of an earthquake, say from 5.0 to 6.0, the frequency of the earthquakes goes down by a factor of 10. Based on that relationship, and the fact that lower magnitude earthquakes are quite rare in the NMSZ, Stein concludes that magnitude-8.0 earthquakes could not occur once every 500 years.
But the paleoseimological record argues against applying the Gutenberg-Richter Relation to intraplate quakes, Frankel says. "The dominant opinion is that the paleoseismological data are the best evidence of recurrence time. You can't assume that Gutenberg-Richter fits in the New Madrid."
Even if the risks are as high as the USGS map make them out to be, the financial costs of requiring buildings to meet California standards outweigh the benefits, Stein says. In 1999, FEMA used the USGS map and data on the buildings in Memphis to estimate that, over the long term, earthquake damage will cost the city approximately $17.2 million a year. That figure spreads the costs of expected, future quakes over the years between the quakes, yielding an annualized earthquake loss. Joseph Tomasello, a structural engineer at the Reaves Firm in Memphis, Tenn., and co-author of the Eos paper, estimates that new construction in Memphis would be about 10 percent more expensive if all the new buildings had to comply with the new IBC 2000 standards. Memphis spends approximately $2 billion annually in new construction costs, the authors argue, so the total cost for the new measures would be more than $200 million -- dwarfing the annualized earthquake loss of $17 million. "Roughly speaking, you would be spending about 10 to 20 dollars to save 1 dollar in damage," Stein says.
The authors argue that the money that might be spent on mitigation could be put to other important uses, such as hiring new teachers (rather than physically strengthening the school) or providing insurance to fewer uninsured patients (rather than upgrading a hospital).
Frankel argues that the 10 percent price tag for earthquake mitigation is much
too high. He estimates those measures would cost only an additional two or three
percent, which significantly tips the scale in the cost-benefit analysis. Also,
he suggests, the codes are intended to save lives, not to maximize financial
benefits. Segall echoes those concerns: "Let's make sure the buildings
are safe, and if we overspend a bit, it's not a bad investment."