The first happened in Goldsboro, North Carolina. On January 24, 1961, just three years prior to the Buzz One Four crash, a B-52 bomber had broken up in midair, due to a major fuel leak in the right wing that caused a weight imbalance; this stressed the wing to the point that it ripped from the plane. When the fuselage broke open, its two hydrogen bombs dropped out and landed in a farmer’s field.7 In this instance, six crew members bailed from the plane and two died in the crash. Obviously neither bomb detonated, because the state of North Carolina still exists, but it had a chilling effect nonetheless. Rather than just being ferried to a new location, these bombs were on a mission when the plane went down, and were ready for action. At the time, it was assumed that redundant safety mechanisms in the bombs had all worked as planned by the bomb designers and catastrophe had been avoided according to script. But further investigation revealed that the arming mechanisms on both bombs had been inexplicably activated, and gone through all but one of the seven steps needed for detonation, because the safety rods of the bombs had somehow been removed. Technically they were still not armed because of the one remaining safety step, but some saw that distinction as semantics. As nuclear historian Chuck Hansen said: “It was like a fully loaded pistol with the safety off, and the hammer cocked—it is not armed until the final safety mechanism, the trigger, is pulled.”8
There was speculation that centrifugal forces, acting on the safety rods when the plane started into its downward spin, had caused the rods to extract from the bombs. The safety rods were intended to prevent the bombs from detonating while still in the plane. The rods were designed to extract only when the bombs were released from the bomb bay. Their removal enables all steps of the bomb‘s automated activation sequence to take place. Unfortunately, the effect of centrifugal forces on the safety rods had not been considered by bomb designers. (It was an unknown unknown.) Thus, the one thing stopping detonation of the bombs was a manual safety switch that remained in the “SAFE,” rather than “ARM” position. Only this single safeguard, a switch controlled solely by the flight crew, prevented detonation.
Bomb experts were able to disassemble and remove one of the bombs from the crash site, but the other had penetrated too deeply into the mud of the waterlogged field to allow its extraction (42 feet; 13 meters). It was deactivated in place, covered back up with dirt, and left in the field. It remains buried at its site of impact to this day.9
A few months later, on March 14, 1961, another B-52 with two hydrogen bombs ran out of fuel during flight and crashed. A sudden loss of cabin pressure had forced the pilot to fly at a lower altitude. The low altitude flying caused increased fuel consumption, and an inflight refueling attempt was only partially successful. The crew then inexplicably refused an offer of an additional inflight refueling, bypassed a potential emergency landing site, and continued flying until the fuel tanks were empty. They ejected safely, and the plane crashed 15 miles west of Yuba City, California. A subsequent investigation found that the crew had been using Dexedrine, a stimulant, to mitigate their fatigue, and this was surmised to be a contributory factor to the accident.10 The people had failed the system, but the technology had not. The bombs did not detonate, thanks to those redundant safety mechanisms.
One might be curious as to why warplanes loaded with multiple hydrogen bombs were flying all around the United States during peacetime. They were all part of operation Chrome Dome.11 This was the strategy used by the Strategic Air Command (SAC), the component of the Air Force that was responsible for maintaining the peace through displays of massive nuclear force.12 Rather than have the bombers and their bombs in fixed, vulnerable ground locations, the thought was to keep one dozen nuclear bombers in the air at all times, fully armed with hydrogen bombs, and ready to attack the Soviet Union on a moment’s notice. This was considered the best way to deal with the hypothetical threat of a preemptive Soviet attack on ground-based bombers. The twelve bombers were continually flying mock bombing missions, with real bombs directed at the Soviet Union. They would approach Soviet airspace, but then veer off before crossing the border, thus maintaining a perpetual game of chicken that was intended to keep the Soviet government constantly intimidated by an ever-present threat of total annihilation.
Even after the Buzz One Four crash in western Maryland, Chrome Dome missions proceeded as usual; this entailed twelve planes with multiple hydrogen bombs flying 24/7. And B-52 crashes continued to occur.
In 1966, a B-52 collided with its refueling plane and crashed in Spain, taking its four Mark 28 hydrogen bombs down with it.13 Three of the bombs landed in the vicinity of the fishing village of Palomares on the Mediterranean coast. The parachutes for two of the three bombs failed and their conventional explosives detonated, but did not cause a fission detonation,14 making a subsequent fusion explosion impossible. (Recall that fission explosions are required to trigger fusion explosions.) Instead, plutonium from the core of the warhead was spread over a 0.75 square mile (2 square kilometers) area, requiring extensive cleanup. The fourth bomb went missing for nearly three months, until it was finally found at the bottom of the sea and recovered.
Then, less than two years later, on January 21, 1968, a cabin fire caused by a seat cushion that was innocently stowed in front of a heating duct caused a B-52 to crash near Thule AFB in Greenland, a territory of Denmark.15 The plane crashed with four MK-39 hydrogen bombs inside. Nothing detonated, but the impact of the crash compromised the integrity of the warheads, and the raging fuel fire helped disseminate the radioactivity over 3 square miles (8 square kilometers). Again, a massive radioactivity cleanup effort, this time hampered by extremely cold temperatures, was required. It took nine months to clean up the site. All of the radioactive material recovered was removed from Greenland and shipped back to the United States at the insistence of the Danish government.16
Since these last two accidents took place on foreign soil, they became international incidents, and the United States had to confess to its allies that it had been routinely flying hydrogen bombs over their countries for years, something it had promised it would not do. Good luck was being pushed to its limits and political pressure was bearing down on SAC.
These five crashes, three in the United States and two on the territory of allies, caused the United States to reevaluate the wisdom of a nuclear strategy based on airborne hydrogen bombs that put its own homeland and that of its allies at risk of accidental destruction. Consequently, the United States abandoned use of airborne nuclear bombers and began to rely on a nuclear defense strategy based on a combination of both ground-based intercontinental ballistic missiles (ICBMs) hardened against preemptive attack within subterranean silos, and submarine-based ICBMs that eluded preemptive attack by constantly moving around under the oceans of the world, like an aquatic version of Chrome Dome. This complex and tightly coupled defense system is designed to keep us safe by the perpetual threat of immediately launching a devastating nuclear counterattack on our enemies as soon as we detect that a launch has been targeted on us.
Now that Chrome Dome is history and ICBMs are the basis for our strategic defense against other nuclear powers, are we safer from accidentally annihilating ourselves? Hard to say. Again, we have to contend with the moving target argument. The systems are completely different now than they were back in the days of Chrome Dome, so our experience with that is irrelevant to our current risk assessment. Nevertheless, declassification of government records regarding our early experience with ICBMs shows that the program had growing pains similar to Chrome Dome’s, the most notable being an explosion in an ICBM silo in Damascus, Arkansas, in 1980, that blew the hydrogen bomb warhead off a Titan II missile (a Mark 53 warhead similar to those in Buzz One Four’s two bombs) and clear out of the silo into the countryside.17 Again, no detonation occurred, but recurring mishaps are certainly a precarious way to keep reassuring ourselves that our redundant fail-safe mechanisms are effective.
The Damascus incident happened in 1980. We might ask, are ICBMs safer now t
han they were back in 1980? After all, we now have 35 more years of experience with ICBMs since the Damascus incident. No one can say for sure. Specific information about our current missile defense system is classified. But we do know that as recently as January 2014, eleven ICBM Air Force officers were charged with illegal drug use and 31 were charged with cheating on ICBM proficiency exams. Later that same year, two Air Force commanders were fired and a third disciplined for leadership lapses and misbehavior. These commanders were all in senior positions at three different nuclear missile bases that collectively control over 300 of the US Air Force’s estimated 450 Minuteman 3 ICBMs. The most senior officer was dismissed specifically because of “a loss of trust and confidence in his leadership abilities.”18 It does not speak well for the safety of the ICBM system when the military officers with their fingers on the nuclear buttons dabble in recreational drugs, cheat on exams, and cannot be trusted to lead.
ON A ROLL
Despite all the unknowns, surely decades of experience handling tens of thousands of hydrogen bombs without an accidental detonation indicates that the risk is negligible or even nonexistent, doesn’t it? What this question is really asking is whether the rule of succession applies.
The rule of succession was the brainchild of mathematician Pierre-Simon Laplace (1749–1827). This statistical principle is based on the notion that the longer that something has not occurred, the less likely it is that it will ever occur.19 Laplace even came up with an equation to calculate the probability of the event’s occurrence with increasing time:
P = 1 − [(n + 1)/(n + 2)]
where P is the probability of an accident, and n is the number of days that have passed without an accident happening. By inspection of the equation, you can see that as time passes the probability keeps getting smaller and smaller toward a limit of zero. Stated simply, if something hasn’t happened in many days, it is extremely unlikely that it will happen tomorrow. Given that more than six decades have passed without an accidental detonation of a nuclear bomb, the rule of succession suggests that the probability of such an occurrence happening tomorrow is virtually zero. It seems logical, and reassuring. But is it true?
Just as the occurrence of the Three Mile Island accident underscored that Rasmussen’s faith in fault trees to capture all nuclear power plant risk was misplaced, the surprise occurrence of other events thought to be extremely improbable has led statisticians to question the validity of the rule of succession. It turns out that there was an unappreciated fallacy of logic underlying the rule of succession that is sometimes referred to as the turkey illusion, after a story that statistician and risk analyst Nassim Taleb tells to illustrate the fallacy. Taleb says that we are deceived into accepting the validity of the rule of succession because we see the world as through the eyes of a turkey.
Here’s how a turkey sees the world. A turkey hatches on a farm, and the farmer takes it from its mother and feeds it. The young turkey wonders if the farmer will return the next day to feed it again. The farmer does. The turkey wonders yet again about its food prospects for tomorrow, and tomorrow again brings the farmer bearing more food. This goes on and on for many days and the turkey eventually comes to the conclusion that its prospects for getting fed the next day are virtually certain, so it stops worrying about starving. But one day the farmer shows up without any food and kills the turkey, because that day happens to be Thanksgiving. The circumstances had changed, and the turkey was unaware that Thanksgiving Day was a day unlike any it had experienced before. This is the turkey illusion. It is the fallacy of logic that assumes circumstances remain static.
The world of today is very different from the world during the time of Chrome Dome (1960–1968). The technology has changed, the weapons have changed, their delivery methods have changed, the military has changed, and even our enemies have changed dramatically. We can gain no comfort from the fact that an accidental detonation of a nuclear weapon has not, so far, occurred. The rule of succession has been debunked. The probability of an accidental nuclear detonation occurring tomorrow might be nearly zero or it might be 100%; we just don’t know what world events will bring us tomorrow. Times have changed, our useful data for predicting accidental detonation risk are practically nonexistent, and our uncertainty level is extremely high. We need to be ever mindful that Thanksgiving Day may be coming.
QUICK COUNT
We have already seen, in the case of nuclear power plants, that when uncertainty is high, we are sometimes better served to move away from risk calculations and instead try to assess the direness of the worst-case outcome. As early as 1966, an attempt was made to do this so we could prepare ourselves for the aftermath of the worst possible event in human history—a total nuclear attack on the United States by the Soviet Union, followed by an American nuclear retaliation on the Soviets. The US Defense Department commissioned a study with the RAND Corporation, a private think tank that the government heavily used for consulting services during that period, to assess the consequences of a massive nuclear weapon bombing of the United States and predict the health outcomes. The report was entitled The Postattack Population of the United States.20
The study took a novel approach. It relied on a computer program called QUICK COUNT, which was designed by RAND as an attack simulator and damage-assessment model. The program took every potential hydrogen bomb target site in the United States (all cities with >50,000 people), and drew a circular perimeter around their population centers that demarcated the limit of percussion waves exceeding 5 psi. It then drew an outer circle, beyond the 5 psi line, where lethal levels of fallout might be expected to occur.21 The program then made the reasonable assumption that everyone within these two circles would die from either the shockwave or radiation sickness. By overlaying these circles on top of 1960 census maps, it was able to provide an estimate of the death toll. The program allowed the user to input various attack scenarios (e.g., number and size of bombs, and their hypothetical targets). It would then do a “quick count” of deaths and spit out the fatality statistics that you could reasonably anticipate from such an attack. Quite a handy tool if you need to deal with the aftermath of the ultimate war.
The study envisaged several different lines of attack that the Soviets might employ, using different combinations of various types of nuclear weapons, with each delivering from 0.3 to 100.0 MT of explosive power. It assumed that as many as 1,200 weapons would be launched, and all of the hypothetical launch scenarios were based on the premise that there would be “a brief nuclear exchange terminated by a peace settlement.” The analysis focused on the condition of the United States population after the attack. The study’s report is mostly concerned with postattack demographics, particularly with regard to age, race, and sex, and serves as a fascinating window into the mindset of Americans in the 1960s.
Noting that the young (<15 years) and the elderly (>65 years) would be less likely to survive the attack than those in the 16 to 64 age range, a bright spot was noted. Since the young and the old would have been a drag on recovery anyway, since they “contribute little to the social product and draw heavily on social resources,” their loss was actually anticipated to speed recovery because the surviving population thus would have been culled down to the most highly productive individuals (16–64 year olds).
Regarding race, there was concern that postattack America might have an altered racial balance. Since blacks tended to be concentrated in cities, an attack strategy that focused on cities would disproportionately kill blacks. But if the Soviet attack strategy was focused primarily against military installations, which tended to be populated mostly by whites, it might tip the racial balance in postattack America in favor of blacks.
Lastly, there was a concern that women would be killed off in higher proportion to men. The thought was, since men worked outside the home and tended to occupy office buildings and other commercial structures that provide a measure of shielding, they would be better able to survive than women who tended to spend their time
in flimsy residential buildings or retail establishments with lesser shielding properties. Thus, women were more likely to be exposed to blast and radiation effects. Also, allegedly being less robust than men, women were thought to be more likely to die from their injuries. The combination of greater exposure and relative weakness of women suggested that there would be a shortage of females in postattack America. This could decrease the birthrate, hinder population recovery, and “even place a strain on the social norms governing sexual liaisons.” It’s odd what people choose to worry about when faced with total annihilation.
The report’s conclusion, based on the 1960 population data maps, was that up to 62% of the American population would likely die within the first three months following an attack. That was the bad news. The good news was that as bomb sizes increase, local geographic variations in racial and gender distribution (the main factors that drive the death disparities) tended to become less pronounced in determining the composition of the postattack population. To put it simply, when the bombs approach sizes where most everything is destroyed, it doesn’t much matter exactly where you happened to be located at the time of the bombing, because your goose is equally cooked. So everyone has about the same risk of dying, regardless of race or gender. Viewed in this way, large hydrogen bombs are egalitarian, blindly killing without regard to race, gender, or any other social demographic.
GLOBAL WARMING
Since the end of the Cold War and the dissolution of the Soviet Union, the prospect of a total nuclear exchange is thought to have greatly lessened. At least progress in nuclear disarmament through weapons reduction treaties has greatly diminished the number of nuclear weapons that we have trained on each other. Presumably, that alone has lowered the risk, although we still rely heavily on a policy of mutually assured destruction (MAD) to lessen the prospect that there is anything to be achieved by starting a nuclear war. MAD is, therefore, considered to be a major deterrent to preemptive attacks on us from our sane enemies.22 But we are not addressing the risks and benefits of total nuclear war here. How close we are to total nuclear apocalypse with our adversaries is a topic for another day. Rather, we just want to know here what the consequences of a single nuclear weapon detonation would be on us personally, whether detonated by an accident of our own military or through an intentional terrorist attack.
Strange Glow Page 41