As Donald J. Trump continues his bromance with equally great statesmen Kim J. Un while giving the cold shoulder to potential romantic partner Khamenei, it is worth remembering that Kim is the one who possesses nuclear weapons. We know this not because the North Koreans told us, though they have, but because seismologists around the world, in different states and working for different organizations, all independently recorded characteristic seismic vibrations for each of North Korea’s six nuclear tests between 2006 and 2017. North Korea’s post-test statements notwithstanding, the most direct source of information we have about North Korea’s nuclear program are those seismic readings.
So what do those readings tell us? Specifically, can these readings tell us whether or not North Korea is capable of producing an advanced “hydrogen bomb,” that is, a bomb that gets most of its energy from fusion (the joining of light atoms like hydrogen) rather than fission (the splitting of heavy atoms like Uranium)? This is not an entirely academic question, a fission bomb’s yield is typically measured in kilotons, that is, how many thousands of tons of TNT its explosive power is equivalent to. A fusion bomb’s yield can be measured in megatons, that is, how many millions of megatons of TNT its explosive power is equivalent to.
Unfortunately, beyond the fact that the tests happened, there is some disagreement among sources as to (a) the seismic magnitude of each test, and (b) how to derive a yield from that measurement. These differences are most pronounced for North Korea’s most recent (and hopefully final) test, which occured September 2, 2017. At that time, a source from “U.S. intelligence” told The Diplomat the U.S. government believes the yield was around 140 kt (about 10x the Hiroshima bomb). NORSAR, a Norwegian seismological organization that monitors nuclear treaty compliance as a side-hustle, put the yield somewhere between 200 and 300 kt. More recently, at group of researchers at U.C. Santa Cruz put the yield at 250 kt, plus or minus 100 kt.
So we have broad agreement that the test was big, but disagreement as to its degree of big-ness (stop me if I’m getting too technical). But if we explore the history of North Korean test yield estimates as well as the history of yield-estimate techniques, we as non-seismologist laypeople can understand both the sources of the uncertainty and why that uncertainty doesn’t matter much from a U.S. policy perspective.
First, let’s look at a chart comparing seismic measurements from the USGS, the Comprehensive Test Ban Treaty Organization (CTBTO), and NORSAR, for each test going back to the first one in 2006. Seismic readings from nuclear explosions are generally measured according to their “short period body wave magnitude” or “mb” — this is one of many types of “magntitude” measurements available for seismic events, the USGS has a helpful glossary of the different types and their purposes. When comparing readings, it’s important to know that “mb” is a logarithmic measure, so the difference between 6.0 and 6.1 is greater than the difference between 4.0 and 4.1.
Right away we can see that, while the the three organizations’ measurements don’t correspond perfectly, they don’t deviate significantly from one another’s either. The CTBTO’s and NORSAR’s are quite consistent, while the USGS measurements are always a bit higher. The quirks of each organization’s instruments and mathematical methods lead to this slight variation in output. However, less important than the specific measurement is the fact that all three organizations’ readings show a general upward trend in event magnitude, with a marked jump in size for the most recent test.
Of those three organizations, only NORSAR publically estimated the nuclear yield for each test. They did so using a fairly simple equation that was developed by the U.S. government decades ago at its nuclear test site in Nevada. That equations runs
where Y is the yield in kt, mb is the seismic magnitude, and k is the “bias correction,” which is specific to the test site where the event occured. NORSAR assigned North Korea’s tests a k-value of 4.3 which, according to NORSAR, “has been advocated” to be the k-value for Russia’s test site in Novaya Zemlya [1]. NORSAR did this because they, as seasoned nordic geologists, believe that the geologic conditions at Novaya Zemlya and those of North Korea’s nuclear test at Punggye-ri are similar enough to justify doing so. Until North Korea allows an independent geological examination of the area, this estimate is about the best we have. Using this bias correction, we can plot yield-estimates derived from each seismological organization’s measurement.
As we can see the USGS’s slightly higher mb measurement for the 2017 test leads to a much bigger yield estimate if we plug it into NORSAR’s equation. We should note, however, that no matter which yield-estimate you are inclined to beleive, all of them indicate a yield increase by an order of magntitude. So are we, laypeople working with open-source data, to make of this?
First, we should again note the trend of increasing yields capped off with a much greater yield. Second, we should remember the “U.S. intelligence” assessment of 140 kt leaked to The Diplomat; that number is almost certainly the work of the Air Force Technical Applications Center, the arm of the U.S. government charged with monitoring the globe for foreign nuclear explosions. As a layperson working from open sources, I have no way of knowing why AFTAC’s yield-estimate is so much lower than the others. But, because this is the Internet, nobody can stop me from speculating, and I suspect that lower estimate has the most to do with (a) a different assessment of the geologic conditions at Punngye-ri, (b) a different seismic magnitude measurement (AFTAC has its own global seismological monitoring network, separate from those of the CTBTO and USGS) and (c), taking the presumed burial depth of the explosion into account. Using open-source digital elevation data and a map released by North Korea showing precise locations of all its nuclear tests since 2006 (see below photo), I was able to make reasonable estimates as to the depth of burial of each test (I was also able to make a fancy 3D model and tour of the test-site) [2]. AFTAC, which has access to the best digital elevation data the U.S. government has to offer, was surely able to do the same thing.
Personally, I am inclined to believe the AFTAC yield-estimate of 140 kt over that of NORSAR, essentially because AFTAC has the greatest well of information and nuclear-testing-specific expertise to draw from. But ultimately, the specific yield estimate isn’t as important as the overall trend, which all of our sources agree on. That trend is steadily increasing yields over time, with the most recent test being an order of magnitude bigger than those previous. That means that the 2017 test explosion was of the order of magnitude we generally associate with fusion or “hydrogen” bombs, so does this prove that it was?
Technically no, but (in this analyst’s opinion) for all practical purposes yes. In order to reach this conclusion we need to look beyond the nuclear yield data to what else we know about the North Korean nuclear weapons program. It is possible to create a bomb with a yield in the 100s of kilotons that only uses fission, the U.S. detonated one in 1952. But such a bomb is heavy and inefficient, that is, it uses a huge amount of nuclear material. To a reasonable degree of certainty, we know that North Korea is under intense nuclear material constraints. We also know that the purpose of North Korea’s nuclear weapons program is not to give seismologists something to do, but to deter the U.S. by threatening its cities with nuclear armed intercontinental ballistic missiles. So it is unlikely that North Korea would use up a huge portion of its nuclear material reserves on an explosive device too heavy to mount on an ICBM.
As a final macabre note, it is worth pointing out that the difference between a 200 kt and a 300 kt nuclear blast isn’t as big as one might think. According to Alex Wellerstein’s Nukemap, a 200 kt airburst in the middle of Manhattan would kill around 1,000,000 people, and a 300 kt airburst would kill around 1,200,000 people. This difference would obviously matter to those 200,000 people, but it doesn’t provide North Korea with any additional deterrence value. The devil is at the strategic level, not in the details.
- I was unable to find any open-source confirmation of anyone, U.S. government or otherwise, advocating a 4.3 k-value for Novaya Zemlya. However this Office of Technology Assessment report (which I also linked to above) has a detailed description of how the U.S. government arrived at its k-values for various Soviet test sites, though it does not list the values themselves.
- The digital elevation data was gathered during a joint US-Japanese mission in the early 90s and is available on Earth Explorer.
