Thanks, I suspect, to Glenn’s concern, Technorati shows 13 links to The Knowledge, Technology Review‘s article about the dangers of biological weapons — but only one to Assessing the Threat, the companion piece that casts doubts regarding the same subject. None of the links I found via Technorati lead to any lengthy commentary, and Glenn — whom I regard as a national treasure, so this is not meant to convey disrespect — doesn’t seem to get beyond saying “this is scary, and we ought to do something about it.” The blogosphere, it would seem, has its limitations, even in the face of mortal threats.
By way of starting somewhere, then, I have a framework to offer that might at least help us determine how much trouble we’re in …
I am frequently frustrated by blogospheric commentary, even when I know myself to be in near-perfect agreement with the authors, written by people who obviously don’t know the first thing about risk management. We don’t have to guess about this stuff: methods exist to qualitatively and quantitatively assess risk, and there are known strategies (admittedly at a high, conceptual level) among which to select for managing any given risk.
For my risk-assessment framework, I’m going to shamelessly steal from an earlier post of mine here on Chicago Boyz, entitled 61 Years Ago Today, in which I developed a four-quadrant classification system for the projection of military force. My point in that earlier post was that while hitting a target anywhere in the world in less than an hour is technically feasible, we can only do so with WMDs, that is, nuclear-armed ICBMs or SLBMs. I regard it as desirable for the US to develop the capability of applying discriminate force on the same timescale — to see how we could have had that capability 20 years ago or more, you’ll just have to go back and read the post.
Now for another set of quadrants:
|
107
mortality
100 |
I
|
III |
|
II
|
IV |
|
|
sec |
100 |
107 |
The independent variable (x-axis) is still time, ranging from roughly 1 second to several months — but please note that this is time of effect, not time of delivery, much less time of preparation. The dependent variable, mortality, is the death toll from a terrorist attack, ranging from 1 to 10 million.
Existing, conventional methods all fit in Quadrant II — time of effect ranging from seconds to hours, and death toll ranging from 1 to several thousand.
Nuclear or chemical terror attacks push into Quadrant I — time of effect is still short, but death tolls are orders of magnitude higher, ranging from thousands to millions.
At least three types of attacks could fit into Quadrant III — time of effect greatly lengthened, up to months, with death toll as high as that of a nuclear bomb detonating in Manhattan:
- EMP – To quote myself: “Take the microprocessors away, and a modern city becomes a ghost town. No electric power, no motorized transport, no telecommunications; everyone would have to leave. On foot.” The ultimate example, assuming technology known to be in the possession of North Korea and nearly in the possession of Iran, would be a ship-launched Scud putting a Hiroshima-sized nuke into the ionosphere above (say) the Eastern Seaboard of the US. Tens of millions of people would have no more than a few days’ supply of food and purified water. Malnutrition, waterborne diseases, and (depending on the season) exposure could easily run the casualty count into seven figures before the evacuation was completed. But note that there would be relatively few prompt casualties.
- Biological attacks – It’s redundant to quote from the “pro” TR article, and I won’t deny that some of its scenarios, eg a virus that’s as easy to catch as chickenpox and renders its victims completely senile, are terrifying. The important point here is the latency involved — only a handful victims for days or even weeks, but millions after a few months.
- Dirty bombs – Apologies in advance for simply directing readers to my Reassessing Dirty Bombs (II); I note that there is some controversy about the supposed effectiveness of this technique.
(Not to overlook the obvious, Quadrant IV attacks seem like a waste of time, though I suppose an exotic assassination attempt involving a virus tailored to attack the President’s DNA could fit in this category.)
The important thing, I think, about Quadrant III attacks is that they would represent a fundamental shift in emphasis on the part of the bad guys. Up till now they’ve gone for the dramatic: explosions, shootings, crashes, fires; highly visual stuff that kills people immediately. But what if the bad guys’ whole M.O. changed, to (to quote myself) “an attitude that it’s better to kill, say, 10,000 infidels 10 years from now than 1,000 infidels today[?] [H]ad four jetliners been crashed into water works in DC and NYC on 9/11/01, the death toll by now would almost certainly have been many times the number murdered that day.”
What I’m suggesting is that we keep an eye out for any attack indicative of delayed gratification. If the bad guys start going in for that sort of thing, we’re in a whole new ball game, one that may well require the kind of societal mobilization undertaken in past “secular crises” (terminology).
One of the most frustrating aspects of our current situation is that terrorist incidents in the US have not followed a power-law distribution — had they done so with their number N ∝ M-1, for example (M for mortality), then in addition to 9/11, we would have experienced 10 incidents with death tolls between 100 and 1,000, and 100 incidents with death tolls between 10 and 100. Instead, there’s been essentially nothing — a few clumsy attempts which, even had they worked, would have resulted in single-digit mortality figures. On the one hand, this lack of any remotely similar activity (whatever its cause) has saved thousands of lives; on the other, it has left us in a kind of experiential vacuum of “unk-unks” (unknown unknowns) — in which we hear speculation ranging from “al-Qa’eda doesn’t exist” to “we’ve got to deport ten million illegal immigrants to stop terrorism” — greatly increasing the risk of a serious misdirection of resources which may itself result in high body counts.
I therefore further suggest that we look elsewhere for proxy data — first, to non-terrorist incidents, of which my favorite example is Milwaukee in 1993. But mainly to vulnerable environments outside the US but in some way connected to it. Like, say, a big, crowded, proverbially unhygienic Third World city occupied by American troops. Baghdad is the perfect testbed for bioterror, or indeed any Quadrant III attack. Whatever the ethnicity or sect of the victims, any incident would be perceived as having winners and losers, and any sufficiently large-scale unpleasantness can be spun to make the Americans look bad.
None of this directly keeps us from getting killed. It may, however, help us select among risk-management strategies and develop plans for how to direct our resources. I don’t think we’re in that much trouble yet. But we need to know how to know when we’re in trouble, and chip away at those unk-unks.
(Readers who really like my stuff a whole lot may be sufficiently motivated to read yet another tangentially-related post on Arcturus, subtitled the Mathematics of Terror.)
Jay, I take your point, but I’m a little a suspicious of EMP as a major threat. Electromagnetic pulse is a well understood phenomenon (since the late 50’s, IIRC) and is not terribly hard to defeat.
When an electromagnetic wave passes across a conductor (or the conductor moves through the wave) a voltage is induced in the conductor. If there is a complete circuit, the voltage will force a current to flow through the conductor. That’s the basic principle that makes an electric motor work. But there are quite a few high hurdles to overcome to make EMP a weapon:
1. Voltage does not equal current. Voltage is potential energy, not kinetic, and is analogous to water pressure. You can have pressure in a pipe, but if the valves are closed no water flows. Also, if there are overpressure vents or absorbers in the plumbing system, transient pressure surges are contained or quickly relieved by venting. Diodes and transient voltage suppressors (tranzorbs) serve this function in electrical circuits. They shunt overvoltages directly to ground. Surge protectors sold for computers and home entertainment systems serve this function. Also, many of the high-end electronics (military, industrial) systems have substantial surge protection built in.
2. All noncoherent (meaning non-laser type) radiated energy follows the inverse square rule, approximately stated as Intensity = 1 / Distance Squared. Move away from a radiating source to twice the distance and the intensity drops to 1/4 of what it was, move away 10 times as far and it drops to 1/100th as intense. You would need a very intense source like, say, a 10-20 megaton hydrogen bomb, to create enough initial energy to propagate a wave of energy over enough distance to damage very much. And if you have a 10-20 megaton hydrogen bomb, EMP is just an ancillary effect. I’m more worried about that really damaging sun tan I’m gonna get from it.
I think the big threats come from bioweapons and nerve agents. Nerve gas is incredibly deadly stuff and can be manufactured in any plant that makes pesticides.
I’d love to see figures on the effective radius of nuclear EMP. I know that a megaton-yield weapon can blanket nearly the entire coterminous US with upwards of 100,000 volts per meter and infer that a significantly smaller weapon could similarly affect an area a few hundred miles across, and therefore believe that such use could kill far more people than a surface blast. Of course, shielding is possible, but cumbersome, so we should hope for (in telecommunications in particular) greater use of optical rather than electrical transmission. The real way to manage this risk, needless to say, is with missile defense.
Interestingly, the record of nerve gas as a terrorist weapon is poor; see the limited effect of the Aum Shinrikyo attacks in Tokyo. The question is, what went “wrong”?
From what I’ve read, ‘what went wrong’ in Tokyo was their distribution system for the gas didn’t work very well; didn’t put out enough volume fast enough to get a really good effect.
Unfortunately, simple to fix.
Michael Hiteshew,
I think your underestimating the impact of an EM pulse originating in space. When detonated above the ionosphere a nuclear bomb interacts with the earth’s magnetic field to turn ionosphere into a giant radio broadcaster saturating the area underneath with the pulse. A substantial percentage of the energy of the bomb is converted into the pulse.
Surge protectors sold for computers and home entertainment systems serve this function. Also, many of the high-end electronics (military, industrial) systems have substantial surge protection built in.
The speed of the pulse (zero to zenith to zero) is several hundred times faster than that of lightening which is what most surge suppressors are designed to handle. The pulse will be there and gone long before the vast majority of suppressors even begin to trip. Worse, the pulse can evoke destructive currents inside chips themselves. An EM pulse can kill electronics setting unplugged on the self.
Hardened systems work by shielding the entire system, including the power supply, from the pulse. New solid state suppressors are supposed to be able to stop the pulses but the vast majority of civilian systems won’t have them.
Jay Manifold,
Interestingly, the record of nerve gas as a terrorist weapon is poor; see the limited effect of the Aum Shinrikyo attacks in Tokyo. The question is, what went “wrong”?
Aum Shinrikyo didn’t invest any effort into their delivery system. They basically used a system of perforated bottles that slowly oozed the sarin into balled up tissue paper. They needed a system that would let the people deploy the weapon get away.
If they had used a simply compressed air aerosol system or suicide attackers, the death toll could have been a couple of orders of magnitude higher. Fortunately, the attack was not well planned having been rushed into place to create a diversion from an expected police raid.
Shannon wrote:
A substantial percentage of the energy of the bomb is converted into the pulse.
The numbers I’ve seen are under 1/2% of the bomb’s energy. That’s hardly “a substantial percentage” Shannon.
The speed of the pulse (zero to zenith to zero) is several hundred times faster than that of lightening which is what most surge suppressors are designed to handle.
The rise time of the surge is only important if you’re depending on heating to melt fuses or mechanical circuit breakers to trip. Diodes don’t care about rise times. When the exceed voltage is reached, the diode becomes a path to ground.
There’s also a lot mythology surrounding EMP. The exoatmospheric tests in the above the Pacific in the 50’s (Starfish?) supposedly caused lots of power outages in Hawaii. Turns out, not true! There were minor effects, a few street lights went out, a few burgular alarms, etc.
New solid state suppressors are supposed to be able to stop the pulses but the vast majority of civilian systems won’t have them.
Diodes and transzorbs ARE solid state suppressors. There’s nothing new about them. They’re ubiquitous, even in consumer grade electronics, which are typically made to the lowest reliability classification.
Worse, the pulse can evoke destructive currents inside chips themselves.
That would depend on the intensity of the pulse.
Consider this: your average PC is subjected to electromagnetic radiation all the time, from radio waves to cell phones to EMI of various frequencies being emitted by motors, heaters, mechanical switches, etc. Does everyone’s PC die every day? Do all the internet servers die when there’s a thunder storm somewhere? No.
Consider this: once upon a time solar flares wreaked havoc on early telecom satellites. Not any longer. We’ve learned to deal with it effectively. To this day though, whenever there’s a solar flare announced, you’ll read predictions that we should expect “major” disruptions in the world’s telecom systems and power grids. Rarely is there a even problem.
Consider this: power companies routinely deal with surges on their sytems in excess of 1 million volts from lightning. There have been lightning strikes in excess of 5 million volts that were effectively clamped and caused no damage.
As I said, the idea that EMP is a potential WMD is unconvincing. If someone has a multimegaton thermonuclear weapon, they’re likely to set off over New York, where the blast and heat will kill thousands to millions of people, not 50 miles up over Kansas in hopes of frying our computers.
If I had to write this post all over again, I’d have shortened the part about EMP, since it wasn’t supposed to be the point. I need an editor. Anyway …
I concur that anyone with a megaton-range weapon would probably not use it on EMP, but I think that someone with one or a few 10-kiloton range weapons would be strongly tempted to try an EMP shot. What I do not know is the effective radius of such a weapon, or rather, how the affected radii scale with yield. If they scale the way blast effects do, with the cube root of yield, then a 10-kT nuke would have fully 1/10 the EMP radius of a 10-MT one. If it’s the square root, then the radius would be only about 1/30.
This could be surprisingly important. Actual recovery from such an attack would be via a sort of dry-land Dunkirk. If the vehicles used in evacuation could get in and out of the affected area on a single tank of gas, such an operation becomes much easier to carry out.
Enough EMP. If I managed to say anything important in this post, it was probably the notion that Baghdad is the place to watch for a bioterror incident.