Note: This news release is available in Japanese. ( Adobe Acrobat Reader is required.)
"Major faults near Istanbul are likely late in their earthquake cycles," and the rapidly growing Turkish city is likely to suffer a large earthquake within the next 30 years, according to an April 28 study in the journal, Science.
Turkish, Japanese and U.S. authors of the Science report were quick to caution, however, that they cannot predict earthquakes. Rather, they calculated the odds and compared their estimates with the long record of earthquakes near Istanbul. The researchers urged a calm response to their findings, adding that the work should help Turkish government officials minimize hazards.
"An earthquake could happen near Istanbul tomorrow, or in many years," says geophysicist Tom Parsons of the U.S. Geological Survey (USGS), lead author of the Science study. "But, in all likelihood, we have time to reengineer many buildings to better protect the people who live and work there. We hope that our research will prove useful to Turkish leaders and scientists as they evaluate various options."
His comments were echoed by coauthors Aykut Barka of Istanbul Technical University; Shinji Toda of the Earthquake Research Institute at the University of Tokyo; and USGS scientists Ross S. Stein and James H. Dieterich.
"Our statistical analysis of earthquake risks should serve as a planning tool," says Barka, a professor of geology. "We are actively sharing our research with planners, scientists and engineers. We hope that their response will help prevent injuries and property damage in the event of an earthquake."
On August 17, 1999, an earthquake measuring 7.4 magnitude on the Richter scale shook Izmit, Turkey. Three months later on November 12, the city of Düzce suffered a similar quake of 7.1 magnitude. Some 18,000 people were killed, and at least 15,400 buildings collapsed, resulting in $10-25 billion worth of damage.
Over the next 30 years, the Science authors report, Istanbul faces a 62 percent probability of more "strong shaking," comparable to the recent disasters in Izmit and Düzce. They calculate their margin of error for this estimate to be 15 percent, so that the risk could easily range from 47 to77 percent.
They further calculate that the risk of a major earthquake is roughly 50/50 for the next 22-year period, or a 50 percent probability, give or take 13 percent. Over the next 10 years, the probability was estimated to be 32 percent, plus or minus 12 percent.
The Izmit and Düzce events were only the latest in a string of large earthquakes to rock the North Anatolian fault since 1939. Located northwest of this quake-prone region, Istanbul has been "heavily damaged" at least a dozen times over the past 1,500 years.
To better explain the series of quakes rupturing toward Istanbul, Parsons and his coauthors focused on "stress triggering." Whenever an earthquake occurs at one point along a fault, Stein explains, it may create higher stress levels at another spot. By taking this mechanism into account, the research team sought to create more accurate estimates of future hazards. Traditional models, based mainly on the average time between events, underestimate earthquake hazards on the North Anatolian fault because they don't reflect the buildup of stress as earthquakes have progressed down the fault.
First, the researchers assembled a "catalog" of major earthquakes occurring along the Yalova, Izmit, Prince's Islands and Marmara Sea faults (all part of the North Anatolian fault system), using damage descriptions compiled by Nicholas Ambraseys and Caroline Finkel of Imperial College, London. Estimating the approximate epicenter, magnitude and rupture lengths of each disaster meant studying damage descriptions of Turkish earthquakes since 1500 A.D. From such documents and instrument readings from more recent events, the team pieced together the past behavior of various faults in the Marmara Sea.
Because some damage descriptions were exaggerated, while other events went unreported, the researchers had to mathematically account for such uncertainties. Still, Turkey's long history proved a boon to the project. "The basic idea was to convert historical descriptions into an idea of where earthquakes happened and how big they were," Stein explains. "The resulting earthquake record that we developed, however inadequate it may be, is far more complete than we will ever have for the U.S. cities of Los Angeles, San Francisco, or Seattle because we have better records for Istanbul. In California, we have few records from before the gold rush of 1849. By contrast, we have a magnificent record for the past 1,500 years in Istanbul."
Ultimately, the Science authors derived slip rates, which were then used to estimate earthquake probabilities from historical records and their calibration modeling. (This technique differs from currently used methods, which depend more on data from global positioning satellites to estimate slip rates.)
Two earthquakes occurred on the Izmit fault in 1719 and 1999, the researchers report. Three quakes ruptured the Yalova fault in 1509, 1719 and 1894. But, the central Marmara fault has been silent since 1509, and the Prince's Islands fault last broke in May 1766. Based on this history, the Science paper concludes, "at least two of the four faults are likely late in their earthquake cycles."
The research team recommends a conservative response to their work. "We want people to be aware that we are using novel methods of risk calculation that have not been fully tested," Stein points out. "It's also important to remember that a probability model is simply playing the odds. The bottom line is that we believe Istanbul's risk of another major earthquake in the relatively near future is high."
ORDER PARSONS ARTICLE: "Heightened odds of large earthquakes near Istanbul: An interaction-based probability calculation" by Tom Parsons, Ross S. Stein and James H. Dieterich of the USGS in Menlo Park, CA; Shinji Toda of the Earthquake Research Institute at the University of Tokyo, Japan; and Aykut Barka with Istanbul Technical University, Turkey.
CONTACTS: Tom Parsons at 650-329-5074 (phone), 650-329-5190 (fax), tparsons@usgs.gov (e-mail); Aykut Barka at 90-212-285-6299 (phone), 90-212-285-6210 (fax), barka@itu.edu.tr (e-mail); Shinji Toda at 81-3-5841-5763 (phone), 81-3-5802-3391 (fax); toda@eri.u-tokyo.ac.jp (e-mail). For copies of this article, please email scipak@aaas.org, call 202-326-6440, or fax this form to 202-789-0455.
NAME __________________________________________________
TITLE & ORGANIZATION __________________________________________________
ADDRESS __________________________________________________
PHONE & FAX __________________________________________________
E-MAIL __________________________________________________
Journal
Science