Future Tense

The United States Is Safest When All Nuclear Tests Are Virtual

Trump wants to restart live nuclear testing, but he’s totally unprepared for the fallout.

A mushroom cloud in the desert
The first-ever nuclear weapon detonation, which happened on July 16, 1945, took place at the Trinity Site in New Mexico. Department of Energy

We do not know the names of the first victims of nuclear fallout, but we know their nationality. They were Americans, mostly living in rural New Mexico, and when the U.S. Army detonated the world’s first nuclear weapon at the Trinity Site on July 16, 1945, these people became the first downwinders.

The nuclear age made its official entry when the United States dropped an atomic bomb first on Hiroshima, and then on Nagasaki, killing between 130,000 and 256,000 people. But while the effects of fallout on the surviving populations of Hiroshima and Nagasaki were extensively studied by the United States, the range of health impacts on Trinity downwinders remains unknown to this day.

What began at Trinity would continue until Sept. 23, 1992, when a detonation in an underground chamber at the Nevada Test Site marked the United States’ 1054th and final live nuclear test. The cumulative result of nearly half a century of regular nuclear testing was an extensive data set on atomic detonations, an arsenal that remains staggeringly powerful to this day, and thousands of civilians unwittingly exposed.

As the Trump administration floats a restart of live nuclear testing, it threatens to make the whole planet into downwinders, while jeopardizing an existing American advantage in nuclear science and weapons—all in the name of restarting a nuclear arms race the United States might not be able to win.

The end of live nuclear testing did not mean the end of nuclear modernization. The objective of the last live nuclear test, named « “Divider,” » was, as documented by the National Nuclear Security Administration, “to ensure the safety of U.S. deterrent forces.” It was a quality check by technicians, one that only made its way into history because of the abruptness with which the program stopped. Divider was not a demonstration of anything new, but part of the iterative maintenance of the engines built to end the world.

The infrastructure of the Manhattan Project was designed solely to get to Trinity and usable nuclear weapons. The atomic infrastructure built up after it, the complex of labs and greater government coordination, spent the Cold War both innovating and iterating on nuclear weapon design. Even after detonations went underground in 1963, with the advent of the Partial Test Ban Treaty, these tests continued to design new warheads.

After the Trinity test, New Mexico would host just two more live nuclear tests. Colorado would host two, and Alaska and Mississippi each held one. The remainder of the nuclear tests conducted by the United States took place either on Pacific islands or at the Nevada National Security Site. These tests created downwinders at every use, as fallout drifted to where people were living. In the Pacific, the people downwind lived on the Marshall Islands, and despite being exposed to fallout from 67 nuclear tests, Islanders received just $150 million in compensation, or less than the purchase price of two F-35As.

That some fallout from the Nevada tests was carried through the atmosphere to New Mexico is one complication in sorting out the extent of harm done to Trinity downwinders. Among other effects, it seems likely that fallout from the Trinity test contributed to the anomalously high infant mortality rate New Mexico experienced in the second half of 1945.

The government knew there would be collateral damage in the areas surrounding Trinity, but they didn’t seem to grasp its magnitude. Before Trinity, the Army stationed 20 government agents in towns up to 100 miles away from the blast site, specifically so that the Army would have evidence to counter any lawsuits about property damage from people in the area.
While calculations done in the weeks before the test indicated a real risk of fallout, “The specter of endless lawsuits haunted the military, and most of the authorities simply wanted to put the whole test and its after-effects out of sight and mind,” as a 2010 CDC study of Los Alamos documents noted.

But it was geopolitics, not health concerns, that ultimately brought about the end of live nuclear testing in the United States. In 1991, mere months before the final dissolution of the Soviet Union, Soviet Secretary General Mikhail Gorbachev declared a unilateral moratorium on tests involving nuclear detonations, in part as a response to overtures of arms reduction from the United States. In 1992, President George H. W. Bush followed suit, and every president since has extended this moratorium—though the Trump administration might be the first to change that.

Since the moratorium, Los Alamos, Sandia, and Lawrence Livermore National Laboratories have turned to complex and sometimes even classified mathematics to look inside the existing nukes, without having to crack them open. Live testing isn’t necessary to craft new weapons or maintain old ones, especially when the labs can draw on simulations informed by rigorous study of the data collected from 1054 prior tests. The result is a kind of warship of Theseus: While no new warheads are built, the labs still design and build new bombs and missiles to house the existing stockpiled warheads.

Computer modeling isn’t a concession in the absence of live testing—from the start, simulations of nuclear problems were as much a part of the program as field tests. In 1952, Los Alamos built an ENIAC upgrade named MANIAC (for Mathematical Analyzer, Numerical Integrator, And Computer) and used it to run calculations needed for the development of the hydrogen bomb. Supercomputing grew up in the labs in sync with live testing.

The modern switch to an entirely simulated research paradigm, aided by decades of development in powerful and sophisticated supercomputers, has resolved some seemingly intractable problems.“For decades, the nuclear weaponeers puzzled over why the [live] test results appeared to break from [the conservation of energy] principle,” reported the Washington Post in 2011. Moving to simulations and mathematical models made the problem solvable. “Then, the ‘energy balance’ problem, as it was known, was solved by a Livermore physicist, Omar Hurricane, who won the 2009 E.O. Lawrence Award from the Department of Energy for his work, which remains classified.”

Testing bombs involves more than just testing the boom inside them. There are still live tests of some nuclear weapon components, like airdropping inert bomb casings from fighter jets, while leaving the actual atom-sundering part of the tests up to the computers. Relying on simulations and modeling for weapon maintenance has also had the effect—and, since the late 1990s, the stated goal—of reducing the complexity of the nuclear arsenal. Fewer kinds of warheads means more time spent maintaining the types in service, and building a deeper knowledge base on what is most useful.

But now, driven by a concern that China’s modest nuclear arsenal is not bound by the same treaties as the massive stockpiles still held by the United States and Russia, the Trump White House has floated the idea of restarting nuclear tests as a way to bring China into a new trilateral treaty.

Restarting live nuclear testing carries with it a whole host of risks, from arms races to newly harmed downwinders to potential catastrophic accidents. It also ignores the reality of how the labs have adapted to nuclear stewardship for a whole generation. The expertise that came with nearly 50 years of live testing has all but retired, and while there are records available for people to attempt to replicate how nuclear testing was once done, it is no longer an innate skill of the nuclear lab workforce.

Even discounting computer simulations, the United States had already conducted hundreds more nuclear tests than any other nation. This is true especially in comparison to newer nuclear states, like India, Pakistan, and North Korea, whose finite programs and small arsenals stand as products divorced from the Cold War efforts of the United States and Russia.

Siegfried Hecker, as director of Los Alamos National Labs from 1986 to 1997, begrudgingly accepted the turn to virtual nuclear testing. But now, noting that the United States had conducted over 20 times as many nuclear tests as China, Hecker said that with a restart of live testing, “we lose more than we gain.”

If the restart of nuclear testing is driven by a desire to demonstrate national strength, the renewed nuclear arms race it sets in motion will have the exact opposite impact. For as much as the administration seems intent on setting this in motion, there is no clear plan to handle the fallout.

Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.