International Network of Engineers and Scientists Against Proliferation

Analysis and Interpretation of the North Korean Nuclear Test

Martin B. Kalinowski, Ole Roß

After the Democratic People’s Republic of Korea (DPRK, North Korea) had informed China of its intent to conduct a nuclear test six days in advance, the state-run Korean Central News Agency announced that the DPRK performed a successful nuclear explosion on October 9, 2006.[1] Several seismic observatories all over the world recorded an event that took place in the north-east of the country at 1:35 UTC on that Monday with a seismic body wave magnitude of 4.1 + 0.1.

Explosion Characteristics

Seismic analysts can in principle conclude whether a seismic event was caused by an explosion or by an earthquake. Waveform parameters can be used to distinguish between these different sources of seismic energy. Tectonic shifting releases energy along extended plate contact areas, while a sudden burst produces spherical-symmetric waves. Thus signals of earthquakes start more gently and extend over a certain time. An explosion causes a more abrupt and shorter deflection in a seismogram.

But these typical characteristics of an explosion were hard to detect in the weak signals related to the event in North Korea over long distances. Only the closest station recorded the characteristic signal shape (see Figure 1).

A strong evidence for an anthropogenic cause of the signal arises from the fact that the north-east of North Korea is a seismically stable region. Figure 2 shows the distribution of seismic activity in the region since 1990 only in the western Sea of Japan that extends to the North,. whereas seismic activity in the region of the event of October 9 has been negligible. Therefore, it is highly unlikely that the seismic signals were caused by an earthquake.

Seismic event can be very accurately located (see Table 1). In this case, the maximum deviation from the average of three independent analysis results is less than 0.1 degree, which corresponds to a distance of about 8 km at the latitude in question. As shown in the satellite images in Figure 3, the suspected test site is in a mountainous area. According to Google Earth, the elevation of the suspected tunnel entrance is 1,400 m above sea level; close summits exceed 2,000 m.

Institution Origin Time Latitude Longitude Stations Magnitude

Low Yield

First preliminary analyses of the explosion gave magnitudes over a wide range. Within two days, quality-controlled and more sophisticated evaluations showed a seismic body wave magnitude range of 4.0 to 4.2.

The body wave magnitude mb can be used to assess the explosive yield Y. Figure 4 shows some nuclear test events with confirmed yields and measured magnitudes. The linear fit to these events is mb = 4.16+0.88 log Y. The variance is 0.2. This indicates the uncertainty of yields that are determined from seismic body wave magnitudes.

The relation between the yield of an explosion and the resulting earthquake magnitude depends strongly on the surrounding geology and the regional attenuation properties.[5] If the test is conducted in a big cavity, e.g. a tunnel, less energy is transferred to the rock than in a compact containment.[6]

Due to these uncertainties and different assumptions about the seismic magnitude, the first yield estimates that were published after the North Korean test spread broadly. In early statements the reported numbers ranged from 0.2 up to 15 kt TNT. Within a few days, more reliable analyses calculated yields from 550 to 800 tons TNT. Figure 4 helps to put these assessments into perspective with other tests. As a conclusion, the upper range appears to be closer to reality – our fit leads to a yield between 650 und 1100 tons.

Regardless of these uncertainties the yield of the explosion was quite small compared to other nuclear tests. All first tests of other countries were much stronger, from 4-5 (India) up to 60 kt TNT (France). The first test ever, called Trinity and conducted on 16 July 1945, reached a yield of about 21 kt TNT. There are only very few tests of India and Pakistan and some US tactical weapons tests reported with comparably low yields in the sub-kiloton range.

A Pre-detonated Plutonium Device?

The low yield raises the question whether the explosion was caused by chemical explosives or by a nuclear one that most likely did not achieve the expected result.

The likely scenario is that North Korea tested a nuclear explosion that reached its criticality before the fission material was optimally compressed. As a result, the increase in nuclear energy causes the assembly to drive apart before a significant portion of the nuclear material has undergone fission.

This pre-detonation scenario is called a “fizzle” and is technically considered a failed test although the released energy is still enough to render the device a weapon of mass destruction that fulfills the political and military goals of an emerging nuclear weapon state. A pre-detonation can be caused by the high rate of neutrons from spontaneous fission of plutonium-240. Another possibility is insufficient plutonium compression by the conventional ignition mechanism.

Uranium isotopes do not produce a neutron rate that results in an early start of the nuclear chain reaction. Accordingly, the low yield tells that North Korea has most likely detonated a plutonium device. This is consistent with open source information according to which North Korea has not yet acquired a significant quantity of highly enriched uranium. The North Korean stocks of separated plutonium are estimated to range from 20- 53 kg. The total amount of plutonium produced is 43-61 kg. The separated stock is enough for 4-13 nuclear weapons, assuming a consumption of 5 kg per weapon and for 3-7 based on the significant quantity defined by the International Atomic Energy Agency of 8 kg. The production rate of the 5 MW nuclear power plant in Yongbyon is 5-7 kg plutonium per year.[7]

Chemical Explosion?

On the other hand, it is technically possible to reach a yield of more than 1 kiloton TNT with chemical explosives, even though a big effort is necessary. In the Non Proliferation Experiment at Nevada Test Site in 1993, a 1,400 t mixture of ammonium nitrate and fuel oil was detonated with the explicit aim to simulate the impact of a nuclear weapon test. Its purpose was to provide calibration data for seismic analysis. Several other chemical test explosions with up to 4 kt TNT equivalent were performed in the United States.[8] North Korea would have to have brought dozens of truck loads with chemical explosives to the test site, but no indication of such a huge logistic undertaking has been detected.

There is no possibility to distinguish between nuclear and conventional explosions through analysis of the seismic signals. To proof undoubtedly the nuclear origin of an explosion it is necessary to detect radioisotopes produced in the nuclear fission processes.

Required for Proof: Radioactivity Release

In order to become detectable over larger distances, sufficient quantities of radioactivity must escape from deep underground and be released into the atmosphere. Even if sophisticated engineering methods are applied to ensure the complete containment of underground explosions, there is a significant risk of substantial quantities of radioactivity to be emitted.

An investigation of reported releases from about 700 underground test explosions at the Nevada Test Site shows that after roughly every second test radioactivity had been detected above ground.[9] Particle-born radioactivity had a lower emission probability because it is more likely to be filtered out on their path to the surface. Fission noble gases were most frequently observed.

Most of those tests at Nevada were conducted in several hundred meters deep and thoroughly plugged bore holes. Some were done in tunnels driven into mountains. In general, they can be considered to be better contained than the suspected tunnel used in North Korea.

Even if radioactivity reaches the surface, it is not sure that the plume concentration is high enough to be detected in samples taken at a distance. This depends not only on the initial concentration but also on the drastic impact of atmospheric dilution and the fairly rapid radioactive decay. ¹³³Xe has a half-life of 5.2 days, ¹³⁵Xe of only 9.1 hours.

Radioactive Plume Detected

The doubts whether North Korea has in fact tested a nuclear device or otherwise has tried to make the world believe it did by setting up a huge conventional explosion were silenced on October 16, 2006, when the Director of the U.S. National Intelligence Public Affairs Office gave the following statement:

“Analysis of air samples collected on October 11, 2006 detected radioactive debris which confirms that North Korea conducted an underground nuclear explosion in the vicinity of P’unggye on October 9, 2006. The explosion yield was less than a kiloton.”[10]

Unfortunately, no details were provided about what exactly the US airplane had collected. Reportedly, an official told the Agence France-Presse that xenon was among the detected substances, but no details were given. As a result, the international community is not able to interpret the data independently. Later, the government of South Korea also announced to have measured Xenon concentrations from 0.9 to 6 mBq/m³ between October 12 and 16, 2006.

It is crucial to determine whether the measured xenon isotopes provide a clear proof or are simply consistent with the assumption of a nuclear explosions while at the same time they could possibly be caused by a legitimate civilian source.

Measuring ¹³³Xe alone would still be ambiguous, because it could originate from one of the regular emission pulses of a regional nuclear power plant. Only the supplementary measurement of additional xenon isotopes or other specific radionuclides allows for a clear distinction between an explosion and a reactor source. In particular, a certain combination of ¹³⁵Xe, ¹³³mXe, ¹³³Xe and ¹³¹mXe measurements allows for a robust identification of the source characteristics.[11]

Blown in the Wind

Air trajectories and plume dispersion starting at the location and time of the North Korean test or shortly afterwards have been calculated with the HYSPLIT Model which is made available for online use by the U.S. National Ocean and Atmosphere Administration (NOAA).[12] One of the main difficulties is to find the appropriate starting time and altitude. Small variations may have big effects on the location of the plume after just a few days (see Figure 5).

Investigations of the releases from underground explosions in Nevada have shown that the most likely time of an uncontrolled release of radioactivity is within the first hour after the explosion.[13] Although underground tests can cause venting up to 3 km height, the most likely release is at ground level.

While on the day of the explosion the main wind was north-eastbound, the ground level part of the plume turned around one day later and blew south on the following days. Accordingly, particles in the plume took different directions depending on their height in the plume. For a particle on a higher altitude, the trajectory leads towards the north-east. On the other hand, the part of the plume near the ground reached the South Korean coast three days after the test, which is indicated by the published results of the South Korean government.

As there are many unknown parameters, it is hard to draw unambiguous conclusions. But one can calculate the minimum amount of ¹³³Xe that must have been released into the atmosphere in order to cause a concentration above the detection limit of 1 mBq/m³ on 11 October, 2006, when the US airplane took samples.

HYSPLIT simulations were used to determine the dilution of the plume at 500-3,000 m height early on the day of the US sample taking reported as successful. It shows the concentration to be diluted in the order of 10^-16 within 48 hours after the explosion. The minimum source strength for ¹³³Xe that would result in a concentration higher than the minimum detectable concentration above the Sea of Japan on October 11 is found to be in the order of 10 TBq. This is in the percent range of the total ¹³³Xe activity directly produced by a 1 kt nuclear explosion and therefore realistic. At a height of 0-500 m, the concentration was still one order of magnitude higher (see Figure 6), so that at ground level detection was possible with a source-strength of about 1 TBq.

Detecting ¹³⁵Xe after two days is more difficult because it has a similar detection limit but decays more rapidly and there remains less than 3% of the initial amount after 48 hours. On the other hand, the initial yield is 50 times larger for ¹³⁵Xe than for ¹³³Xe. Therefore, it should also have been possible to detect it in air samples of October 11, if several percent of the total amount were released.

Two days after the explosion, the concentrations of the isomers ¹³³mXe and ¹³¹mXe are two to three orders below that of ¹³³Xe. As the detection limit is even higher, it seems hardly possible that these isomers were detected on October 11.

Finding ¹³³Xe and ¹³⁵Xe can only be rated as consistent with a nuclear explosion but not as unambiguous proof. The annual total emission of 133Xe from a light water nuclear power plant is typically in the order of a few TBq. Single pulse emissions may rarely reach the 0.1 TBq level.

In this assessment a one hour emission is assumed shortly after the test. As there are still more xenon isotopes produced from precursors through the decay chain, it might have been even easier to measure some of the isotopes two days after test in a wider area, in case the release continued.

Conclusions

Shortly before the explosion on October 9, North Korea informed China of a forthcoming test of a nuclear device equivalent to 4 kt TNT. But not even this low yield expectation was not achieved, which raised the question whether the explosion was nuclear or chemical.

The definite proof of a nuclear explosion could only be achieved through on-site inspections, which are foreseen as an option under the Comprehensive Nuclear- Test-Ban Treaty (CTBT) once it enters into force but were not feasible in the case of North Korea.

Nevertheless, according to U.S. National Intelligence, the air measurements strongly indicate that the explosion was nuclear. The remarkably low yield hints to a technical failure, probably a pre-detonation of a plutonium device.

The seismic control network of the CTBT worked as designed, although only half of it is in operation to date. But the corresponding network of 80 stations for radionuclide measurements is not yet far enough developed to detect the relevant traces of atmospheric radioactivity. For this task, national means are still required, but these results are often kept secret to other states as well as to the public and are thus not accessible for independent analyses.

The obvious conclusion, namely that the North Korean test failed at least partially, raises the suspicion that North Korea may want to undertake another nuclear test. The international monitoring community will observe the country closely. It is a challenge for diplomacy to convince the North Korean leadership not to test again. North Korea should sign the CTBT as 176 other countries did before. Acknowledgement: This study was made possible by a grant from the German Foundation for Peace Research (DSF).

---

1.   Korean Central News Agency, DPRK Successfully Conducts Underground Nuclear Test, October 9, 2006; www.kcna.co.jp/item/2006/200610/news10/10.htm#1.
2.   Informational Processing Center, Geophysical Survey, Russian Academy of Sciences.
3.   US Geological Service, National Earthquakes Information Center.
4.   International Data Center, Provisional Technical Secretariat of the Preparatory Commission for the Comprehensive Nuclear-Test Ban Treaty Organisation (CTBTO), as published at www.ceme.gsras.ru.
5.   J. Schlittenhardt, Seismic yield estimation using teleseismic P- and PKP-waves recorded at the GRF-(Gräfenberg) array, Geophysical Journal 95, 1988, p. 163-179.
6.   Richard L. Garwin and Frank N. v. Hippel, Arms Control Today, Nov. 2006.
7.   Institute for Science and International Security, The North Korean Plutonium Stock Mid-2005, September 7, 2005, Issue Brief; www.isis-online.org/publications/dprk/dprkplutoniumstockmid05.pdf.
8.   Nuclear Information Project, Divine Strake: Global Strike Low-Yield Nuclear Simulation, Nuclear Brief April 3, 2006; http://nukestrat.com/us/stratcom/gs-divinestrake.htm.
9.   Martin B. Kalinowski, Characterisation of prompt and delayed atmospheric radioactivity releases from underground nuclear tests at Nevada as a function of release time,submitted to Health Physics.
10.   Office of the Director of National Intelligence, Public Affairs Office, Statement by the Office of the Director of National Intelligence on the North Korea Nuclear Test, Press Release, October 16, 2006; http://www.dni.gov/announcements/20061016_release.pdf.
11.   Martin B. Kalinowski and Christoph Pistner, Isotopic signature of atmospheric xenon released from light water reactors, Journal of Environmental Radioactivity, Volume 88, issue 3, 2006, p. 215-235.
12.   National Ocean and Atmosphre Administration (NOAA), Air Resource Laboratory (ARL), NOAA ARL HYSPLIT Model; www.arl.noaa.gov/ready/hysplit4.html.
13.   Kalinowski, op.cit.
14.   David Albright and Paul Brannan, ISIS Imagery Brief: North Korean Site After Nuclear Test, Institute for Science and International Security, October 17, 2006; www.isis-online.org/publications/dprk/dprktestbrief17october2006.pdf.