Department of Geoscience, Stony Brook University

Scientists pinpoint North Koreaís 2009 nuclear test site to a 140-meter precision



Since May 25, 2009, when the Democratic People's Republic of Korea (North Korea) announced that it conducted a second nuclear test, the exact location of the test has remained elusive to international organizations, federal agencies, military and the scientific community. Until now. A new study from Stony Brook University has determined North Koreaís 2009 test site to be at (41į.17'38.14"N, 129į4'54.21"E) with a geographic precision of 140 meters (153 yards). The study reaches the precision with a strategy of utilizing the scientific forensic evidence registered by North Koreaís 2006 nuclear test. This page serves as a general introduction to the process of ground-based determination of a nuclear test site, its challenges and the new findings in the study.

North Korea's nuclear test sites (red star), seismic stations (triangles), and seismic waveforms plotted color-coded with the associated station and labeled with the year of the test.

To understand the ground-based determination of a nuclear test site, let's imagine a scenario that a person speaks and a group of people are constantly listening. Now, you ask the listening group to figure out where the speaker is. In this analogy, the speaker is North Koreaís 2009 nuclear test and the listening group are the seismometers around the globe. And, of course, the nuclear test generates seismic waves propagating inside the Earth as opposed to that the speaker transmits sound waves travelling in the air. The location pinpointing is usually done in a standard way that the listening group measure the arrival times of the voice signals they receive. For each listening person, the possible locations of the speaker would lie on a circle centered at the listening person, with a radius equal to the arrival time of the sound signal he/she receives times the sound speed. If you have three people (or more) listening, you will pin down the location of the speaker, which is the overlaying point of the circles centered at each of the listening person.

Sounds an easy task, right? Well, our only problem is that the speaker speaks in a jungle. Sounds come out of the jungle no longer along straight lines. They are instead bent, reflected and scattered through the jungle before you hear them. Without perfect knowledge of the jungle, their actual paths canít be well calibrated and the determination of the speakerís location will have a large error. This is exactly the situation confronting scientists when trying to locate North Koreaís 2009 test. The Earthís interior, where the nuclear seismic waves travel, is no less complex than a jungle.

So how to improve the accuracy of locating the speaker in a jungle? The obvious approach is to better understand the complexities of the jungle. For the "jungle" inside the Earth, this is being done by using lots of dense seismic observations for better imaging the Earthís interior by seismic tomography, a technique similar to the CAT-scan used by the doctors to image a patientís body. With the continued effort of improving our knowledge about the complexities of the Earthís interior, the horizontal uncertainty of the 2009 event location reported by the United States Geological Survey (USGS) narrows down to about 7.6 kilometers (4.8 miles). The uncertainty reflects our imperfect knowledge of the Earth's interior.

Determined location of North Korea's 2009 test site (big circle, with an uncertainty of 140 meters) plotted on a Google Earth map centered at North Korea's 2006 test site. The sizes of the circles are proportional to the relative yield.

The study from Stony Brook University employs a very different strategy and reaches a new level of precision. Letís go back to the speaker-jungle analogy. If the speaker has already spoken once in the jungle, the situation is entirely different. It is not that sounds would not travel a complex path anymore in the jungle, they still do. The difference now is that the speaker left forensic imprints of the jungle complexities to the listening group at the very first time of the speaking. The next time the speaker speaks in a close proximity, because the sound waves would experience same complex paths in the jungle as in the first speaking, the forensic evidence of the jungle complexities the speaker left at the first time can be used to remove the effect of the jungle complexities in the recording of the subsequent speaking. When you take the differences of the arrival time signals in the listening group between the speakings, the signals associated with the complexities of the jungle are cancelled out and the residual signals are only sensitive to the locations. As you can imagine now, once the jungle complexities are taken care of, an extremely high precision in reached.So, the strategy is not to try to fully understand the complexities of the jungle, but to take advantage of the forensic evidence of the jungle complexities that are imprinted in the recordings of the first speaking. This is why the new study, by using the forensic evidence registered by North Korea's 2006 nuclear test, is able to locate North Korea's 2009 nuclear test in an extremely high precision.

So, a speaker in a jungle already spoke once and was exposed. Any tricks of avoiding being caught in the future? There is one, the only one. Donít speak again.

For press coverage of the study: Yonhap (South Korea)
For press coverage of the study: Xinhua (China)

For the scientific publication of the study, click here
For the event information at USGS .
For more information about North Korea's 2009 nuclear test reported by the Comprehensive Nuclear-Test-Ban Treaty Organization: click here
For more information about Comprehensive Nuclear-Test-Ban Treaty Organization:
For more information about Professor Lianxing Wen:

Copyright Stony Brook University, 2010.