Kevin Vranes has a phud in Physical Ocean- ography and Cli- matology. He now studies sci- ence policy and politics at the If any of you guys can see the North Korea test (01:35 GMT / Oct 9) on available seismograms, please email me or put it in the comments. None of the LCSN instruments show anything and while the USGS has a page up on the shaking generated, they don't have the waveforms available.
One delicious bit of science-politics I can already see brewing is the battle over the test size. Pulled out of the NY Times article covering this event are the following different assessments of the explosion size:
Russia's defense minister, Mr. Ivanov, said that the Russian military had confirmed the test and estimated its force at somewhere between 5 and 15 kilotons -- much larger than estimates from South Korea. France estimated it as merely the equivalent of about 500 tons of TNT, and did not confirm that it was the result of a nuclear device, The Associated Press reported.
and
"We have assessed that the explosion in North Korea was a sub-kiloton explosion," said the intelligence official, speaking on condition of anonymity because of the sensitivity of the issue. He added, "We don't know, in fact, whether it was a nuclear explosion." He spoke as intelligence analysts in Washington were in the early stages of assessing the explosion.
Can it be possible? Would different countries come to different conclusions on the size of the explosion for political reasons? Nah.... they wouldn't do that, would they? Of course they would - and will. I am definitely not a seismo expert, but my understanding is that the link between energy created in a test and seismic energy picked up instrumentally is not cut-and-dried.
Perhaps the U.S. has a choice to make here. Play up the size to bring stronger world condemnation and hopefully stronger cohesion for action against the Democratic Peoples Republic of Korea? [Funny joke, huh? But the Dear Leader really does have N. Korea calling itself that.] Or play down the size to humiliate the DPRK in a sort of "hahaha, your dick is sooo little!!" kind of way?
If the science stays a bit uncertain on the size of the test, the Bush Administration gets to play it whichever way they want. Will it matter to those of you deeply concerned about the Bush Admin's war on science? I mean, let me give you a scenario:
Seismologists come to the general conclusion that there is a 80% chance it was a big test and 20% chance it was a small test. Yet the Bush Admin decides to publicly maintain it was only a small test, going against the "consensus" of seismologists. Does the Bush Admin then get your derision for misusing science? Or do they get the benefit of the doubt for saying whatever they feel they need to say in order to pursue their international diplomacy goals here (regardless of how you feel about those goals)? Do they get derision at the same level as for ignoring climate change threats, misrepresenting stem cell numbers, or mucking with endangered species numbers? Or is this a different situation?
Comments
# 1 | Lab Lemming | October 9, 2006 4:24 PM
Japanese station waveforms are available here, but they seem to be running 3 days behind:
http://www.fnet.bosai.go.jp/freesia/waveform/prev.html
And at least the US has the report on the USGS home page. The CTBTO, despite being formed to monitor these things, has nothing but UN red tape on their homepage (update- overnight they put up a letter condemning the test- but still no data).
My take here:
http://lablemminglounge.blogspot.com/2006/10/seismology-of-nuclear-test.html
# 2 | John Fleck | October 9, 2006 4:38 PM
From another one of those blogging academics:
http://www.armscontrolwonk.com/1230/nork-data-it-was-a-dud
(Drill down in the comments for explanations of why it's not as cut-and-dried as Jeffrey initially argues)
IRIS data here:
http://www.iris.edu/news/special.htm
# 3 | John Fleck | October 9, 2006 4:41 PM
And more:
http://www.washingtonmonthly.com/archives/individual/2006_10/009701.php
# 4 | J-Dog | October 9, 2006 4:43 PM
Based on this Administration's past performance, whatever size they get their tame seismologists to sign off on, just take the polar view, and you'll be spot on!
BTW - Are they blaming this on Clinton yet?
# 5 | bigTom | October 9, 2006 5:51 PM
I'm more than a bit rusty on these matters (my last seismology was grad school >30 years ago), but I seem to recall the greatest uncertainty was in the test area geology. But I think the uncertainty is only about +/- a factor of two, not enough to explain the wildly different Russian estimate from the smaller Korean, USGS, French estimates. I'm leaning towards the dud theory myself. Japan should be in a good position to supply their own estimate (which hopefully isn't influenced by US politics).
# 6 | John Wilkins | October 9, 2006 5:55 PM
Try contacting the Seismology and Tsunami watch in Hawaii. I can't recall the exact title, but they're on Oahu. I visited them last year and they showed me extensive seismic survey traces of events in the Pacific and I'd be surprised if they didn't have this on file.
# 7 | Matt | October 9, 2006 5:58 PM
If they publicly maintain it was a small test and do not acknowledge the 80%-20% split then I would say they are misusing the science.
Politically interesting, also, is why N.Korea tested this device now (assuming it was indeed a nuclear device - could it simply have been a large conventional explosion?). I would bet it is directly linked to visits of the new Japanese prime minister to China and S.Korea over the weekend and today. A little reminder from N.Korea - don't forget us!
Haven't seen any waveforms yet.
# 8 | John Fleck | October 9, 2006 7:27 PM
Matt -
A convention device sufficient to create a half kiloton blast (the bottom end of the range being discussed) would require many, many truckloads of ANFO. In other words, it would likely be detectable in satellite imagery. Divine Strake, the Defense Threat Reduction Agency test that was to have been done this year in Nevada, would have required 60 truckloads of ANFO to create a similar blast.
# 9 | DanR | October 9, 2006 8:22 PM
IRIS: http://www.iris.edu/news/special.htm
via ArmsControlWonk
# 10 | Western Geologist | October 9, 2006 9:06 PM
IRIS has a page up with some seismograms from China, South Korea, and Japan:
http://www.iris.edu/news/special.htm
Apparently you can download recordings from any of their stations here:
http://www.iris.edu/cgi-bin/wilberII_page3.pl?evname=20061009_013527.0.spyder&qtrxyrad=current///all
I've never tried this before, but I just submitted an automated request for stations MDJ in China and MAJO in Japan. To my untrained eye those look like the best records.
# 11 | Matt | October 9, 2006 9:16 PM
John - point taken - I ignored the obvious
# 12 | bigTom | October 9, 2006 9:47 PM
Matt, I believe it was the day after the Great Leaders ascension to power anniversary. Thats the most likely explaination of the timing.
We should be able to get topological differences (before/after test) to determine the volume of the underground cavity, as an independent measure of yield.
One of India's test devices had a seismic mag of 5.1, this one was
3.6 (S. Korea) or 4.2 USGS, so the subkiloton estimate seems to be reasonable. I don't know how the Russians came up with their much higher estimate.
# 13 | Western Geologist | October 9, 2006 10:42 PM
I've uploaded a couple of seismograms from the North Korean detonation and two regional earthquakes here.
# 14 | julie | October 9, 2006 11:07 PM
Spinning the science of how big a test n Korea did is a hugely different situation than spinning the science of, say, climate change. The test is a closed loop of politics: the test is making a political point, we use scientific consensus re: how big the test was to asses the threat, and the way that the US government uses the evidence of test size will be politically calculated. The scientific judgment is just throwing another piece of information into that closed political loop. The whole point is politics-of course the US govnt is going to spin the test- and we would expect nothing less of them, that is their job.
With the earth sciences, especially re: climate change, the scientific consensus being spun politically has broad and grave consequences because the question is about non-political issues. So the spin is politicians pursuing their own interests (the interests of their underwriters) at the expense of the interests of the entire f-ing earth. That is not their job.
# 15 | kevin v | October 9, 2006 11:45 PM
Julie, you could say that the N. Korea business feeds into something of far graver threat than climate change -- some feedback of processes that leads to nuclear exchanges that leads to outright nuclear war. (Perhaps unlikely, but not impossible.) Your point still holds, but what if the spinning of the science in the nuke case is what ultimately leads to the nuclear exchange feedback?\
WG - thanks. one major difference is the focus depth -- the test essentially happened on the surface. I'd also expect the manmade blast to be more "impulsive" because essentially it is -- for the same energy released, the blast energy is all released from one point source whereas in an earthquake the energy is released along the entire fault plane that ruptures (which can be tens to hundreds of kilometers long), which spreads the energy release in both time and space.
# 16 | Andrew Dodds | October 10, 2006 9:09 AM
The simple difference between an explosion and a Natural event is that an explosion has a 'Push' first motion along the direction of the wave in every direction, wheras an earthquake will give a 'Pull' first motion over half the planet. Obviously you need data from several stations to tell the difference.
Getting the magnitude is extremely hard - it strongly depends on how well coupled the device was to bedrock, especially for lower yields.
The point about the test quite possibly not even being nuclear is quite interesting - imagine for the moment that you are the scientist-in-chief of NK's nuclear project, and you have been ordered to carry out a test, but you know that you don't have a working weapon. You can either go to your leader (who has just promised the world a n-test) and tell him you can't do it, leading to your drawn out and unpleasant death, or you can stuff a thousand tonnes of TNT and assorted ammo into the test, set it off and hope no one notices..
# 17 | Western Geologist | October 10, 2006 9:17 AM
You're very welcome Kevin. I initially expected the nuclear blast to be more impulsive for exactly the reason you mentioned, but it doesn't really look more impulsive than the 71 km-deep Kyushu event to me (on the filtered data at least). The North Korean blast certainly looks more impulsive on the unfiltered data (I really wish I knew why it's appropriate to apply a filter).
I think the rupture length accounts for some of the differences in the later arrivals, but I'm not sure how much that influences the first arrivals. The rupture will propagate along the failure plane, so the actual portion of the fault that's moving at any time during the earthquake will be much smaller than the total rupture area. So I can think of reasons why the later arrivals on a seismogram from an earthquake and from a nuclear blast will be different, but I'm not sure why the first arrivals would be different.
# 18 | Western Geologist | October 10, 2006 9:22 AM
Andrew,
So the focal mechanism from a blast would not have any extensional areas? It'd look like a filled circle?
# 19 | John Fleck | October 10, 2006 11:04 AM
So I noticed a very interesting thing this morning as I listened to the various news accounts. Almost no one was discussing the technical questions: how big was the blast and what might that tell us about North Korea's military capabilities. It's as if a simple on-off switch has been flipped to the "on" position: North Korea has X, where the details of X need no discussion. In fact, it seems to me that very little has changed. We knew North Korea had some sort of nuclear weapons program, and were unsure about the details. We still know that North Korea has a nuclear program, and are unsure about the details. We were concerned about the possibility of the North Koreans sharing nuclear weapons technology with others. We are still concerned.
The rituals of nuclear weapon states largely unhitched from the physical reality of the devices and how they might be employed.
# 20 | julie | October 10, 2006 12:15 PM
'what if the spinning of the science in the nuke case is what ultimately leads to the nuclear exchange feedback?'
war and war posturing is the realm and responsibility of politicians. if politicians spin the facts such that they aggravate a civilization-destroying nuclear exchange we wouldn't hypothetically look back and say 'they should have paid better attention to the scientific consensus' we would say they used the wrong rhetoric or political maneuvers. The truth about the size of the threat will NOT save us from nuclear war-the way the governments respond to and spin the threat might. Vs the case of climate change, for example, where paying attention to scientific consensus is what might save us.
# 21 | Brian S. | October 10, 2006 12:34 PM
I think being accurate is what's mostly likely to save us from any potential disaster. Under Kevin's hypothetical, the Bushies would get my derision for saying it was a big test, if they didn't add there's a significant level of uncertainty. No difference between this and a pure science issue.
# 22 | kevin v | October 10, 2006 12:37 PM
alright, Julie wins that one. Any of the rest of you care to comment on the politics/science angle? That was the far more interesting angle to me, anyway.
JF - that's an excellent observation of the situation -- best I've seen so far. I would guess that the thinkers in the national security apparatus are right there with you, while the cableTV talking heads (both the hosts and the guests) blather on about the less important, less subtle issues. (Such as "how can we blame Clinton [Bush] for f'ing this up?") As you point out, the problem and the possible solutions haven't changed much at all. Unfortunately the solutions remain exceedinly problematic. Military action was already a non-starter from Seoul's and Tokyo's standpoints and NK has had an economic noose around its neck for decades already. Tightening it ain't going to make much difference.
# 23 | kevin v | October 10, 2006 12:40 PM
Andrew -- good observation, thanks for reminding me about that. For reasons that John Fleck pointed out, it must be very unlikely that this was a chemical explosion, leaving the possibility that this was a successful test of a very small plutonium mass or else a failed test that went "sub"nuclear?
# 24 | kevin v | October 10, 2006 12:48 PM
WG - as for filtering, I think most seismic filtering is bandpass. Bandpass filtering is done in the frequency domain so that you can isolate a range of frequencies you are interested in, thus separating out your signal from the surrounding noise. (I've done this quite a bit for oceanographic data and never for seismic data, but the concept is the same.) Seismic instruments are obviously extremely sensitive, picking up P- and S- waves from halfway around the globe. Being so sensitive, they also pick up trucks driving by, elk bugling in the meadow across from the instrument, etc. The frequency of vibrations caused by the truck driving by is distinctly different from the P and S wave eq-caused frequencies, so you filter to deemphasize the noise (truck) and emphasize the signal (body waves). Bandpass filtering is essentially running a high-pass filter (keeping the high-frequency energy and killing the low-frequency stuff) and low-pass filter (killing the high-frequency signal) at the same time.
# 25 | kevin v | October 10, 2006 12:54 PM
Brian - your comment came in before mine got posted, so I when I said 'Julie wins,' I meant she beat me. 8-) Anyway, I'm still inclined to view it as Julie is -- that this is not a purely scientific issue. The science is a tool to give some pieces of information that feed into political calculations. The other tools at play here are intelligence, psychology, etc. If psychologists were advising the B Admin to downplay the size of the test for psychological reasons that feed into the diplomacy goals, I'd be fine with that.
# 26 | gengar | October 10, 2006 12:56 PM
The reason the first arrivals are different is that explosions generate lots of P (compressional) waves and not many S (shear) waves, whereas earthquakes tend to do the opposite; and P waves travel faster. For an explosion, you'd therefore expect most of the energy to arrive in the first few seconds, which is what you see here. The focal solution would indeed be a filled circle, for the same reason.
# 27 | kevin v | October 10, 2006 1:08 PM
hey y'all - I just found five comments in the junk pile, which means that when you sent it you got a screen that said something like, "This is being held for moderation." I was never notified that I got these comments so they just stagnated in the junk pile, which I rarely check. I'll try to start checking it more frequently, but if you get that "held for moderation" screen, please email me right away and tell me that you commented so I can go in and approve it. danke!
# 28 | Western Geologist | October 12, 2006 11:13 AM
Kevin, thanks for the information about filtering (it's bringing back memories of talking about Fourier transforms during grad school). Just to make sure I understand, when you bandpass filter data you wind up with two outputs, one of the high frequency signal and one of the low? This makes the filters at the IRIS site make more sense too. Their regional filter was 1-10 Hz while their "teleseismic long" filter was 0.02-0.5 Hz. Near an earthquake it'd be possible to earthquake-related high-frequency waves, while further from the event they'd be lost in the noise and only the lower frequency signal would be useful. Is my reasoning OK?
Gengar, thanks for the explanation.
I've uploaded a couple other seismograms here if anyone's interested.
I found a very informative page at Lawrence Livermore National Lab, Monitoring Clandestine Nuclear Tests. It's good a good explanation of the waves generated by a nuclear explosion and some of the difficulties in interpreting seismograms from those events.
# 29 | yami | October 12, 2006 11:43 AM
I've got some seismograms. Except for the nearest GSN station in northern China, they're pretty crappy-looking. As soon as I get some time I'll post a bit more about how we can estimate an event's magnitude, but it'll take me a bit to get that time...
Gengar: The filled circle focal mechanism isn't just because explosions generate a different proportion of P waves. The starting motion of an explosion is outward/compressive in all directions. An earthquake, though, is a push-pull (or actually, two coupled push-pulls), and the pulling portion generates P waves that arrive with the extensional part of the cycle first.
# 30 | kevin v | October 12, 2006 1:28 PM
WG - when you're band-pass filtering, you're allowing a range of frequencies (those in the "pass band") to come through while you're rejecting (attenuating, really) those outside of the pass band. It's really a way to filter out both the high freq and the low freq signals in your time series because you might consider them to be noise. I think you're right about different freqs being important depending on your proximty to the focus. The teleseismic energy has to pass through the mantle and core where as the near-field body waves only through the crust, so that will affect the energy, as will the time spent in contact with the solid earth.