1 One thing to keep in mind - a nuclear bomb requires an initial explosion to initiate the nuclear reaction. That may explain the 'ramp up' of seizmic activity. The PRK may have simply not known how to make an efficient nuclear device.

Another thing - there are two types of nuclear weapons: fusion (the Atom Bomb) and fission (Hydrogen Bomb). Only six countries— United States, Russia, United Kingdom, People's Republic of China, France, and possibly India—are known to possess hydrogen bombs.(Re wiki entry on Nuclear Weapons)

We would have to pull up seismic data on both types of bombs to see if the PRK profile fits. It may be that the PRK detonated an atom bomb type device and that it failed.

Your analysis does have some merit though - in order for a nuclear device to be efficient there can be no 'ramp up'. The initial explosion has to a) be shaped to direct all explosive energy to the nuclear core (think soccer ball) and b) immense to facilitate the fission/fusion of the nuclear core.

At least thats how my layman's mind understands things

2 Actually fusion is Hydrogen, and fission is atomic (uranium/plutonium).

However, the seismic activity seems to indicate that a small explosion happened first, followed by a larger one. As you stated, this would not be indicative of a nuclear blast. Possibly conventional explosives in which there was a delay in detonation. 550 tons (as the NKs are quoting) would certainly not be outside the range of normal explosives.

3 Got the two mixed up - thanks for catching.

4 Dan,

Do you have the time-scale for the x-axis of the 1998 graph? If the total duration of the 1998 plot is on the scale of 20-30 hours, then "zooming in" on section around the spike could produce a result consistent with the CNN graph.

5 Before passing judgement, I'd want to see a labelled horizontal time scale for the two examples to compare with the CNN sample (which shows the North Korean event lasting at least 5 minutes).

I'm suspicious specifically because 9/11 deniers are quite fond of a seismic plot with overcompressed horizontal resolution which makes the WTC 1 and 2 collapse look like it was initiated by an explosion; in reality, the "spike" they show lasted on the order of 10-20 seconds and comes from the debris hitting the ground.

6 Yes, without a time scale, these data are practically meaningless.

7 I think the CNN graph might not show what we're all thinking. I think it's raw data that needs to be analyzed before they make a simpler graph. I say this because if you look at the USGS Live Internet Seismic Server, the graphs there make it look like there are nukes going off all over the planet. Tiksi, Russia, looks like it should be a smouldering ruin. I think that graph needs to be professionally interpreted before we can even guess what it actually means.

8 I believe the seismograph trace you have posted here is not related to the N Korean test but rather a 6.0 mag quake between the Philippines and Taiwan. The time of 10am UTC checks out. As you have mentioned, this trace looks very much like a natural earthquake, and apparently it is.

Link to the USGS Page on this quake:
http://earthquake.usgs.gov/eqcenter/recenteqsww/Quakes/ustqag.php#details

9 Just looking at the plot of the NK event, the "ramp up" looks to have started a full 3-4 minutes before the most severe shocks. I've heard mixed reports that it was an underground detonation. Could it be that the rampup was a large conventional or fizzling nuke that caused a more severe subterrainian collapse?

10 Actually, that seismograph is related to the test. The USGS has a list of recent earthquakes. Number 2 is the NK test, and there's a link to the very same graph, which is being updated.

Looking at it more, I think it is showing up on multiple stations' graphs. The labelling on the vertical axis threw me, I thought it was an hourly graph, and apparently it is not.

11 Not only is it possible to fake an underground nuclear detonation, the United States actually did it on 22 September, 1993 as part of a proliferation detection proof of concept. The source was 1.3-kiloton of chemical explosive 400 meters underground. It produced a 4.1 Richter Earthquake (USGS catalog query may take a moment to load). I'm not a physicist or anything but I wonder what the seismograph comparison between the two looks like.

12 The USGS page I linked to lists an earthquake in NK and 3 others in the Phillipines. So the smaller spike could be from that earthquake (those earthquakes?).

13 That'd be "them earthquakes" brother.

14 The seismograph linked above simulates a paper record loaded on a drum. It lists time in hours along the vertical scale and time in minutes along the horizontal scale. The line represents one station (Incheon) over a 24 hour period. As the drum rotates the line advances. Things start looking shakey at ~10:06am if I'm reading it right and the time has been set right.

15 Some thoughts:

I'm as fascinated by the labels on the x-axis of the displayed seismograph output as by the event data. The pre-printed vertical bars seem to indicate 10 min. intervals. I'm going to assume this is not correct as it would make the "foreshock" trace about 4 min. in duration and the main shock event better than 5 min. - both numbers being absurd in terms of explosion mechanics. 10 millisecond intervals I would believe.

Dan is correct that any kind of fission bomb requires an initiating chemical explosive. Plutonium bombs require what are called "high-brisance" explosives. These are very fast-burning explosives whose detonation wave moves at 5 or more miles per sec. Given that the relevant parts of a plutonium-based weapon core are probably less than a foot across, these explosives do their jobs in microseconds. Then the really big bang happens and that only takes microseconds to go to completion too.

Hence the energy vs. time curve exemplified by the Indian nuke test data shown. On the time scale used by seismographs, the relatively puny energy release of the initiating chemical explosive trigger is swamped by the enormously greater energy release of the main explosion. Both taken together happen too fast for a typical seismograph to show more than one really tall trace covering the interval during which the bomb is actually going off. Everything after that is fracture shocks from the expanding sphere of engery ripping through the native rock and then rebound shocks as the distrurbed strata are hit with reflected portions of the bomb's energy bouncing back toward the origin point or moving at some other refraction/reflection-determined angle in the general vicinity. Nukes are rude and noisy damned things.

The data for the NORK "test" shows a definite precursor event of relatively brief duration that stands out fairly well from the main shock and shows significant decay, in its own right, before the main shock really gets going. The main shock is also rather wide for a nuclear blast (assuming that those vertical bars on the graph really are on 10 msec. centers). Such a distribution of energy seems to me to better fit a large conventional explosion of a comparatively slow, low-brisance compound such as ANFO (ammonium nitrate-fuel oil). ANFO needs to be set off by a small high-brisance charge and then does a relatively leisurely job of exploding itself the rest of the way. ANFO is cheap, safe to handle and is routinely used in multi-ton lots by the mining and construction industries of all nations so the ability of the NORKS to stack up 500+ tons of the stuff in one place is no stretch.

This analysis is far from definitive. The main energy release just might be a fizzled or partially-fizzled plutonium implosion weapon, though the magnitude-vs-time thing for the main shock looks too pokey.

If it was an attempt at a plutonium implosion detonation that failed to entirely work, the interesting question is why the NORKs would risk possible failure when they could - or so we have assumed - just do a comparatively slam-dunk test of a gun-type uranium device with much less likelihood of a glitch. Political point gets made either way, right?

The only answer that seems plausible is that the NORKs are not, in fact, anywhere near as far along with their uranium enrichment program as we have been lead to believe. Uranium bombs are comparatively easy to make, but highly-enriched uranium is not easy to make. Plutonium is easy to make, but the bombs that use it are much tougher to fabricate than those based on uranium.

Assuming an at least partial fizzle result here might well indicate the NORKs haven't really mastered the hard parts of building either kind of fission bomb yet. I gotta figure they would have done a U-235 bomb if they had the stuff to make one - ergo, they don't. Plute, they have, but perhaps they haven't mastered the construction of the necessay "match" to light it.

If true, these circumstances would give us the luxury of more time in hand when dealing with the NORK regime as they would not, in fact, yet constitute an established nuclear threat on even the scale of the immediate post-WW2 U.S. If the now-inevitable sanctions can be put in place fast and hard enough, we and our allies may be able to successfuly implode the NORK regime before the regime can successfully implode its plutonium.

Putting on my tin-foil hat now, it is at least conceivable that the NORKs deliberately fizzled a plutonium bomb in order to fake us into making exactly the analysis I just sketched out. In that case, they must know sanctions are coming. If there is some plan afoot to use already secretly extant uranium bombs in some kind of near-term strike on the U.S. homeland or on U.S. interests elsewhere in the world - or to covertly assist their de facto Iranian allies in doing either or both - the advantage of the misdirection might just outweigh the sanctions consequences in the mind of a not-conventionally-rational Kim and/or Ahmedinejad if the nuclear sucker punch can be delivered soon enough. I think such a plan unlikely unless Kim has already come to the conclusion that his rule is doomed fairly soon even if the U.S. does little or nothing to push it over. In such a case, he might just figure it makes sense to (mixed metaphor alert) "go all in" on a Hail Mary play.

Given the miserable state of the non-technical parts of the U.S. intelligence gathering apparatus these days, I suspect we cannot avail ourselves of the necessary data from within North Korea to accurately discriminate between these possible near-future scenarios. The most important function of an intelligence service is to effectively gauge the actual nature of a threat so that one does not have to entertain genuinely paranoid-seeming hypotheses simply because they do not violate the laws of physics and cannot be definitively ruled out based on objectively verifiable evidence. I have no confidence the CIA, et al, are able to do this where the NORKs are concerned.

16 Dick,

Please consider:

Seismic waves don't all travel at the same speed, but the seismograph records both types (Compressional and Shear) on one needle. The rate of speed is difference between the faster and slower waves is around 8 km/second. In this case this is about a minute difference between arrival times given the approximate distance between the event and the station of 470 km. I would not assume that the horizontal scale records milliseconds.

17 Dan Irving,

The time from explosive detonation to nuclear explosion is at most two or three milliseconds. Seismic recorders can at best discriminate on to 50 to 100 millisecond basis.

On top of that the explosive start is caused by at most 100 lbs of explosive. i.e. lost in the noise level.

18 The event shown on the chart is the M 6.0 that occurred at 10:01:45 UTC, at a depth of 10 km, under the ocean in the Philippine Island Region. The location of the sensor that recorded the event is Inchon, South Korea.

The event in North Korea occurred at 01:35:27 (about 8 hours 25 minutes earlier).

The seismogram does NOT protray the North Korean test.

You can see more information at the following site

http://earthquake.usgs.gov/eqcenter/recenteqsww/Quakes/quakes_all.php

19 Okay, so now I am confused. The Phillipine Earthquake hit the Incheon station at 10:05:53.33 UTC which fits great. The nKorea event was at 01:35:27 UTC why does that time not show anything? If its not depicted, why not, or am I misreading the chart? The Phillipine earthquake was much, much larger but I find it hard to believe a 4.1 quake wouldn't show up at incheon.

20 D.B.R.: The vertical timescale shows one line per hour, and the horizontal timescale on the image runs from minute 00 to maybe minute 32 or 33 or so, so it looks like the right half of the plot was cropped -- an event starting at at 01:35 would have been just to the right of the right edge of the image.

21 Bill: I've been looking at the full plot from the USGS site not the cropped image above. It does not show any visible activity on the full live plot: same for the seismograms from Hawaii and Japan. In another couple of hours the plot will scroll past 01:35:27 UTC 9 Oct so if anyone is interested they should capture it now. I thought that the time on the machine might be off but as Bob Hancock pointed out the big event matches the Phillipine Earthquake.

22 The time on all seismometers is in GMT and as several comments have mentioned the time on the seismogram above is wrong. I looked at data right after it was reported. The expected activity was not there. I don't know what went on, but I am very skeptical of the whole situation.

23 Scott B: I think the USGS is showing the event in UTC which can be slightly different than GMT but equal at the scale shown on these seismograms (less than a pixel difference). Like you pointed out at your link the explosion does not appear when it should. Very odd, indeed.

24 As Scott B has pointed out, this particuar seismogram did not detect the NK test explosion, although it does cover the time period of the test. There are likely two reasons. First, this was a small event, as has been reported from various authorities. Secondly, these instruments are "tuned", if you will, to detect different sorts of events. Natural earthquakes, especially the large ones, produce some *very* low-frequency signals, and instruments intended for detecting large quakes at a distance will filter out the high-frequency signal to reduce noise. Other instruments are set to detect higher frequency signals, but there is more noise to handle there. What the USGS is publishing on the web are the traces of instruments designed to detect large quakes worldwide, no doubt on the assumption that this is more interesting to a US and world audience.

The usual rule of thumb is that you measure large quakes at a distance but the small stuff from local instruments. In this case we are trying to detect a small event from a distance, which is difficult, and probably why we are being told it will take some days of analysis to get a solid determination.

About the only thing we can say, based on the publicly available data, is that this was not a big bomb. I remember watching the seismic signals from US tests in the 70's as they arrived (very exciting to have a pre-announced earthquake to measure.) They were very easy to spot.