The American Government’s Secret Plan for Surviving the End of the World

Newly declassified CIA files offer a glimpse of the playbook the Trump administration will reach for if it stumbles into a nuclear war.

Among the greatest foreign-policy dilemmas faced by former President Jimmy Carter is one that has never been publicly aired but is gaining new relevance. It concerns nuclear war, and how the U.S. government would survive it. Carter’s decisions remain classified, but documents newly declassified by the CIA, along with the archives at several presidential libraries, provide a new window into the White House’s preparations for an imminent apocalypse.

Today, such an apocalypse could be triggered by any number of nuclear-armed states, including North Korea and Pakistan. During Carter’s presidency, such anxieties were focused squarely on the Soviet Union. It was during that period that military planners in both the Soviet Union and United States began to grapple with what until then had been an unthinkable heresy: abandoning the Mutually Assured Destruction catechism that had governed global order since the 1950s and preparing for surviving an all-out nuclear war.

Carter and his White House were interested in more specific questions. If the presidency could survive after a nuclear war, what exactly would it do afterward? How could the surviving commander in chief be identified? Who would identify him? How would he fulfill the three main functions of the presidency: to be the chief executive of the government, the head of state, and the commander in chief of its armed forces?

Carter’s answers came in the form of Presidential Directive 58, which was issued in the final months of his presidency; Ronald Reagan amended those plans with his own presidential directive in 1983. Their contents inform the continuity of government plans that remain in effect for the Trump administration. They have been the object of a multibillion-dollar pastiche of programs and a magnet for conspiracy theorists around the world.

What follows is a glimpse at how the government developed some of its most closely held national-security secrets — and how the Trump administration, or any of its successors, might rely on them to survive the end of the world as we know it.

Pages from "Survival Under Atomic Attack," a U.S. government booklet released across the country in 1950. (Photo credit: City of Boston archives)

When Carter took office, the Soviet Union had a head start on preparing for nuclear war. It had an expensive civil preparedness program; hundreds, perhaps thousands, of underground bunkers; and extensive continuity of government programs.

The United States, for its part, had Ray Derby. Born in Iowa in 1935, Derby became one of the Defense Department’s premier experts on emergency preparedness and disaster response. In Europe, he led noncombatant evacuation drills across NATO, training the trainers who evaluated each unit’s ability to absorb and withstand an attack. In the United States, he led numerous governmentwide task forces on civil defense for nuclear, biological, and chemical accidents. He also developed the standard plan that America’s nuclear bases would use in a disaster. By the time of Carter’s inauguration in 1977, he was in charge of training and operations at the West Virginia Operations Office in the General Services Administration’s Federal Preparedness Agency (FPA).

At the time, the principal federal plan for catastrophic disasters like nuclear war — Federal Emergency Plan D — called for each federal agency to design, develop, and build its own hardened, underground facility. In an emergency, the government would be run from the bunkers. Most agencies did not take the responsibility too seriously.

The FPA tasked Derby with evaluating the implementation of Plan D. The first thing he noticed was that agencies rarely, if ever, rehearsed their respective plans. Few had made any provisions for maintaining vital records — even the laws, regulations, and directives that agencies used in their daily work. Many agency employees didn’t even know whether they were part of the teams that were supposed to evacuate during a disaster.

These failures were the product of systemic neglect since the dawn of the nuclear age. Aside from a brief boom after the Cuban Missile Crisis, during which six federal government relocation centers were built around the nation, the United States did not treat civil defense as a part of its strategic deterrence. The federal government proved eager to spend money to upgrade weapons, not to protect the population or help them survive a nuclear attack. In the 1970s, many government agencies gave up planning post-disaster operations, assuming that the various federal organizations that had the word “preparedness” or “mobilization” in their titles were taking care of it.

Derby thought the problem was larger than funding. The American public said in polls that they wanted a civil defense program. But they lived in peace and didn’t pay attention to its development. They might have assumed that a big program existed somewhere out there, ready to be used in a Soviet attack. By the 1970s, the “duck and cover” years had given way to the comforting dulcet tones of détente. There was no urgency. And presidents were not insisting on any.

There was one other complication: to effectively save the country during a nuclear war, the military had to cross one of its red lines by getting involved in domestic security. After nuclear war, martial law would almost certainly have to be declared, and the military given extraordinary powers to manage resource distribution. But the government also assumed that some sort of martial law would be required before the start of the actual war. As soon as it believed a war might be imminent, the government planned to move significant parts of the population, specifically those who lived near significant strategic military targets, and policymakers knew this might require a degree of coercion, even force. The military did not like to talk about this scenario, and neither did politicians. Plans were therefore developed in secret and classified, ensuring less visibility and public accountability.

And what about members of the government themselves? In an emergency, or a changing of the Defense Condition status, the Joint Chiefs of Staff would order 60 officials to primary relocation sites. The government operated a so-called special facility atop Mount Weather in Berryville, Virginia, where a cadre of top executive branch officials would ride out a nuclear war. Other standby relocation sites were near Hagerstown, Maryland, and Martinsburg, West Virginia, at the Marine base near Quantico, Virginia (for the FBI), and Front Royal, Virginia, near a facility where the State Department was supposed to reconstitute. Still others were hidden at colleges inside or near the Beltway.

But the Army and Air Force had enough helicopters to transport only about a third as many officials as would be required — assuming that aboveground transportation was possible. (Instructions called for them to make their way there in some other unspecified fashion if the helicopters were not available.) And for any top American leaders who managed to make it to Mount Weather, their ability to communicate with federal agencies, other governments, and the American people was questionable, at best, not least because the sites were generally minimally staffed.

Many political leaders disdained suggestions that they practice for war or the very idea they would voluntarily hide themselves in secret bunkers away from the public and their families. (Dick Cheney, a congressman in the 1970s, was one such skeptic.) The assumption in the U.S. government was that the Soviets knew everything anyway. They had even bought land at the base of Mount Weather solely to monitor the comings and goings of emergency personnel. If the apocalypse was coming, warding it off was considered a fool’s errand.

A 1991 bird's-eye view of the above-ground Mount Weather complex. (Photo credit: WILLIAM F. CAMPBELL/The LIFE Images Collection/Getty Images)

By the time Jimmy Carter became president, the country was spending less than $100 million a year on civil defense, compared with more than $30 billion a year to keep its nuclear weapons from becoming obsolete. Congress had identified the value of a unified civil defense program but hadn’t done much to fund it. Carter became the first president since John F. Kennedy to pay significant attention. In September 1978, he declared that civil defense was part of the country’s strategic deterrence, because a population, or government, vulnerable to nuclear attack was more vulnerable to being coerced by threats of an attack. A short presidential decision directive, which was classified at the time, said as much — and little else.

America’s civil defense budget rose modestly at first. But a series of studies acknowledged how weak the country’s civil defenses had become. Fallout shelters built in the 1950s were obsolete and needed to be refurbished or replaced. Executive orders assigning functions to different agencies were widely ignored. There were no federal provisions for evacuating large populations, the lynchpin of any civil defense program. Military exercises all but ignored the hypothetical scenario. These studies also set the path for a whole new policy.

Over the objections of the Pentagon, Carter eventually endorsed a consolidation of the government’s civil defense and continuity programs into one agency and set an ambitious goal. During an all-out nuclear war, the government would aim to have 80 percent of the country survive — and it should prepare to do so on a budget of less than $250 million per year.

On June 19, 1979, the Federal Emergency Management Agency (FEMA) was written into existence. Carter elevated the importance of the agency’s director, assigning to the National Security Council and the Pentagon the task of overseeing civil preparedness. It was now inextricably linked to national security — and strategic nuclear policy. A secret CIA memo said that “trans-attack planning” — how the presidency could function during a nuclear war — was now a part of a national security strategy.

This was sweet music for Ray Derby. FEMA would take charge of the FPA’s Continuity of Government program Derby had been working on, but he would be tasked by FEMA with taking control of the special facility at Mount Weather, becoming, for all intents and purposes, its mayor. Its secrecy would increase, as would its budget and footprint. It would be responsible for the official survival items list, the stockpile of resources that would be needed to rebuild the government after nuclear war.

Meanwhile, the White House was focusing on the hardest challenge of all — providing a mechanism for presidential successors to execute nuclear war orders during and after a nuclear exchange. Early in Carter’s presidency, the director of the White House Military Office, Hugh Carter, convened a small working group to review the White House Emergency Plan, the top secret document that set out how the Secret Service would evacuate the president — and how the White House Military Office would instantiate successors if the president were killed.

The basic blueprint was contained in a series of proposed PEADS — presidential emergency action documents — which National Security Advisor Zbigniew Brzezinski’s military aide Col. Bill Odom then reviewed and revised. Odom, as he wrote in a memoir, found the proposed procedures conspicuously lacking in both imagination and a connection to reality.

One White House memo noted that day-to-day communications between the Pentagon and presidential emergency facilities at Mount Weather, Camp David, and the White House were “satisfactory” under normal conditions. But during a civil disorder, or “uncoordinated sabotage,” it was obvious that satisfactory wouldn’t suffice. And in a conventional or nuclear war, it was determined that the facilities would provide “little protection” — in other words, whichever survivors were stashed there probably wouldn’t survive for long before being targeted themselves.

A seating chart for the special facility atop Mount Weather in Berryville, Va., where a cadre of executive branch officials would ride out a nuclear war.

This meant, in practice, that fixed command posts would not work. A mobile command post was a theoretical option. But the White House couldn’t depend on getting the president or a successor to an emergency escape aircraft in a surprise nuclear attack. Even if they managed to do so, it was impossible to know what kind of staff they would have around them.

Meanwhile, because only Carter and Vice President Walter Mondale had “presidential emergency satchels” — the famed nuclear footballs that verified their identities as commanders in chief — the country’s nuclear command-and-control system would risk coming to a halt if both men were incapacitated or had died, unless there was some other way of identifying presidential successors to the military.

The original solution offered by the White House Military Office were code names uttered by their designees. So if Speaker of the House Thomas P. “Tip” O’Neill Jr. found himself the only surviving successor, all he would need to acquire complete control of the government and its nuclear arsenal would be to offer the surviving Pentagon command center’s emergency action officer a vocal verification of his identity by using the term FLAG DAY. The president pro tempore of the Senate, next in line, was FOUR FINGER. Secretary of State Cyrus Vance would authenticate his identity by calling himself FADE AWAY.

In reality, the Pentagon’s emergency action officer would try to ensure that a survivor was who he said he was. But if Soviet missiles were on the way, the designers of this fragile system had little doubt that the shortcuts would be used, with untold risks for the efficiency of the country’s military response and its basic security.

In light of these challenges, the White House task force tried to conceive of a more flexible, and de-centralized, idea of what it meant for the government to survive. In defining the presidency’s continuity, the task force found itself coming back to three central, if decidedly bureaucratic, concepts: survivability (the president and his basic support team had to ride out a war), connectivity (they had to communicate with one another, the country, and other heads of state), and supportability (people needed other people elsewhere in the country to help them).

This led to a key recommendation: five 50-person “interagency cadres” that would be pre-positioned or pre-deployed during emergencies to support would-be presidential successors. These “presidential successor support teams,” codenamed TREETOP cadres by the Pentagon, would deploy randomly to any one of “several hundred sites, perhaps 2-3 thousand, that would be pre-selected,” allowing for a relocation of institutional knowledge that was “highly flexible and adaptive.”

A hand-drawn schematic of how the governments-in-transit might operate found unredacted in Odom’s notes.

Odom’s team drew up a list of requirements for these teams. The first thing that a deployed team would do was to identify and authenticate the actual president — the ranking successor. The details of the system they developed remain highly classified, but as it was described to me, it involved what might be the first example of “tracking chips” embedded in presidential successor support cards, which would be amplified by radio frequency repeaters. The signals would be collected by FEMA and the National Military Command Center. Other critical technology that would be deployed to assist the would-be presidents would also be tagged and tracked, which might offer a layer of protection against spoofing. (The plan, however, flew ahead of the technology available to make this work. It wasn’t until the George W. Bush administration that the government could passively and electronically track some presidential successors by satellites and the cellular phone system.)

Second: Each team would have to, on its own, help the successor carry out the three main presidential functions: commander in chief, chief of state, chief executive. The team would have to talk to other deployed teams that had survived and securely identify themselves. It would have to talk to the Pentagon, or its surviving elements, to execute the nuclear war plan. It would have to receive intelligence and damage assessments. It would have to talk to state and local governments, too. More prosaically, the 50 people on each team would have to be prepared to function as a stand-alone executive branch without outside help for at least six months.

Odom’s review proceeded slowly. Agencies were asked to weigh in on whether they could carry out a series of secret executive orders. These orders remain classified to this day, but certain public documents offer scraps of information. The Carter White House eventually issued at least 29 PEADs. PEAD 2 dealt with the reconstitution of Congress, a touchy issue for the executive branch, much less mentioned in open correspondence. PEAD 5 was titled “Providing for the Mobilization of the Nation’s Resources.” PEAD 6 dealt with calling an emergency civilian reserve force.

How would Congress be reconstituted? What resources would be mobilized? Who would be drafted? We still don’t know.

People watch a television showing file footage of a North Korean missile launch at a railway station in Seoul on April 5. (Photo credit: JUNG YEON-JE/AFP/Getty Images)

We do know this. When Carter’s presidential directive codifying these changes went into effect in late 1980, the CIA set up its own secret agency, the National Intelligence Emergency Support Office, which would be headquartered in Virginia, receive input from all CIA directorates, and deploy three-person successor support teams to randomly chosen TREETOP locations at a moment’s notice.

We also know, from budget documents, that agencies began to request more money to fund successor support teams. We know that the Pentagon began to test hardened mobile command centers. We know that the Air Force developed plans to add electromagnetic bandwidth devoted to continuity of government to its latest satellites. We know that a designated presidential successor would start being brought to Mount Weather during events that gathered all the branches of government together, like the State of the Union address.

We also know that Reagan found the system inadequate. He was briefed on it before his presidency, but his participation in the 1982 Ivy League war games convinced him that a survivable presidency was untenable and a major gap in the defenses of the country. One of his top aides, Thomas Reed — along with a Marine attached to the National Security Council by the name of Oliver North — persuaded Reagan to permit some modifications to Carter’s system, rather than abandon it.

Flash-forward 35 years. Russia has taken Crimea by force, and an under-resourced NATO worries that an invasion of the Baltic States could bring the alliance to the brink of war. The United States worries that North Korea is on the verge of mating nuclear warheads to intercontinental ballistic missiles that could reach portions of the Unites States. President Donald Trump can still be baited with tweets, has resorted to military force against a sovereign nation backed by Russia after seeing television images of children dying from exposure to sarin gas, has spoken of building a bigger (and not just better) nuclear arsenal, and has not (yet) demonstrated the temperament to respond to a crisis with tact.

Emergency survival plans have evolved since the era of Jimmy Carter. We can safely assume that presidential successors will be authenticated by something more than a whispered code name. But the threats that might prompt their use are still nearer than we all would hope.

Top photo credit: Getty Images/Foreign Policy illustration

Marc Ambinder is editor-at-large for The Week and a contributing editor at GQ. He is author of "Deep State: Inside the Government Secrecy Industry." (@marcambinder)