Command and Control

Sergeant Ramoneda wrapped his baker’s apron around his head for protection from the flames and returned to the burning plane, searching for more survivors. Moments later, the high explosives in the Mark 4 detonated. The blast could be heard thirty miles away. It killed Ramoneda and five firefighters, wounded almost two hundred people, destroyed all of the base’s fire trucks, set nearby buildings on fire, and scattered burning fuel and pieces of molten fuselage across an area of about two square miles. Captain Steffes and seven others on the plane escaped with minor injuries. Twelve crew members and passengers died, including General Travis, in whose honor the base was soon renamed. The Air Force told the press that the B-29 had been on “a long training mission,” without mentioning that an atomic bomb had caused the explosion.

An accident could involve more than one weapon. On July 27, 1956, an American B-47 bomber took off from Lakenheath Air Base in Suffolk, England. It was, in fact, on a routine training flight. The plane did not carry a nuclear weapon. Captain Russell Bowling and his crew were scheduled to perform an aerial refueling, a series of touch-and-go landings, and a test of the B-47’s radar system. The first three touch-and-go landings at Lakenheath went smoothly. The plane veered off the runway during the fourth and slammed into a storage igloo containing Mark 6 atomic bombs. A SAC officer described the accident to LeMay in a classified telegram:

The B-47 tore apart the igloo and knocked about 3 Mark Sixes. A/C [aircraft] then exploded showering burning fuel overall. Crew perished. Most of A/C wreckage pivoted on igloo and came to rest with A/C nose just beyond igloo bank which kept main fuel fire outside smashed igloo. Preliminary exam by bomb disposal officer says a miracle that one Mark Six with exposed detonators sheared didn’t go. Fire fighters extinguished fire around Mark Sixes fast.

The cores were stored in a different igloo. If the B-47 had struck that igloo instead, tearing it open and igniting it, a cloud of plutonium could have floated across the English countryside.

* * *

The engineers at Sandia knew that nuclear weapons could never be made perfectly safe. Oskar Morgenstern — an eminent Princeton economist, military strategist, and Pentagon adviser — noted the futility of seeking that goal. “Some day there will be an accidental explosion of a nuclear weapon,” Morgenstern wrote. “The human mind cannot construct something that is infallible… the laws of probability virtually guarantee such an accident.” Every nation that possessed nuclear weapons had to confront the inherent risk. “Maintaining a nuclear capability in some state of readiness is fundamentally a matter of playing percentages,” a Sandia report acknowledged. In order to reduce the danger, weapon designers and military officials wrestled with two difficult but interconnected questions: What was the “acceptable” probability of an accidental nuclear explosion? And what were the technical means to keep the odds as low as possible?

The Army’s Office of Special Weapons Developments had addressed the first question in a 1955 report, “Acceptable Military Risks from Accidental Detonation of Atomic Weapons.” It looked at the frequency of natural disasters in the United States during the previous fifty years, quantified their harmful effects according to property damage and loss of life — and then argued that accidental nuclear explosions should be permitted on American soil at the same rate as similarly devastating earthquakes, floods, and tornadoes. According to that formula, the Army suggested that the acceptable probability of a hydrogen bomb detonating within the United States should be 1 in 100,000 during the course of a year. The acceptable risk of an atomic bomb going off was set at 1 in 125.

After Secretary of the Air Force Quarles expressed concern about the safety of sealed-pit weapons, the Armed Forces Special Weapons Project began its own research on acceptable probabilities. The Army had assumed that the American people would regard a nuclear accident no differently from an act of God. An AFSWP study questioned the assumption, warning that the “psychological impact of a nuclear detonation might well be disastrous” and that “there will likely be a tendency to blame the ‘irresponsible’ military and scientists.” Moreover, the study pointed out that the safety of nuclear weapons already in the American stockpile had been measured solely by the risk of a technical malfunction. Human error had been excluded as a possible cause of accidents; it was thought too complex to quantify. The AFSWP study criticized that omission: “The unpredictable behavior of human beings is a grave problem when dealing with nuclear weapons.”