Warning the president consisted of a series of steps, each of which consumed precious minutes. It took one or two minutes for the missile to break through the clouds and be seen by the satellite. It took another minute for the satellite to identify it and alert a ground station in Karlsruhe, Germany. It took another two minutes for the crew in Karlsruhe to confirm the alert and send it to the North American Aerospace Defense Command in Colorado Springs. NORAD then took another two minutes to confirm the attack and call the president’s crisis coordinator—his chief of staff.
Francis’s military assistant was informed of the launch at 12:09. He called Francis at 12:10. The chief of staff now had three minutes to decide whether to tell the president. At 12:13, feeling sure there was an attack under way, he told the president.
Everything had gone perfectly, but eleven minutes had slipped away, leaving slightly more than half an hour to move the president to safety.
Francis knew that much of the remaining half-hour was already spoken for. It was one thing to tell the president that a missile was on its way, but it was quite another thing to do something about it. In this case, the president would need enough time to get from Mar-a-Lago to the airport, board Air Force One, and get airborne. The trip from Mar-a-Lago to the Palm Beach International Airport was a ten-minute drive in a motorcade in the best of times with a police escort and no traffic. But the Southern Avenue Bridge was now clogged with people trying to get as far away from Mar-a-Lago as they possibly could. The president could use the helicopter again, Francis thought, but it would still take a few minutes to walk to the helicopter, several more minutes to get to the airport, and then several minutes more to board Air Force One and take off. Assuming the helicopter was a viable option, Francis concluded, they needed to be in it within the next ten minutes—by 12:20 PM. “I thought we had enough time,” Francis recalled later, “but we couldn’t dilly-dally.”
President Trump was having lunch on the terrace at Mar-a-Lago and refused to move. The Secret Service was adamant that Mar-a-Lago was no longer a safe location for him. “We need to get you to Air Force One and get you airborne,” a Secret Service officer remembered telling Trump. But the president was reluctant to leave. No one could tell him where they were going—clearly Washington was not safe—and his staff found that it was difficult, if not impossible, to persuade the president that North Korea’s missiles could reach Mar-a-Lago. According to one participant, Trump kept repeating a line they had all heard before: “all his rockets are crashing.” Kellogg struggled to explain that not every rocket had crashed and that, regardless, they would be far safer from all kinds of threats, rockets or not, once they were in the air. But Trump remained adamant that the Situation Room beneath Mar-a-Lago was the safest place for them to be. “We did tests, and the foundation is anchored into the coral reef with steel and concrete,” a staff member recalled Trump boasting. “That sucker’s going nowhere!”
At 12:18, after the discussion drew the interest of other diners, Trump agreed to go down to the Situation Room, but he was still adamant about remaining at Mar-a-Lago. Another argument ensued. Keith Kellogg, who, according to participants, was shouting by now, said that if the secretary of defense had been present, he too would insist that Trump leave for the airport immediately: “Jim would say the same thing,” Kellogg pleaded. “He would say get on the goddamned plane.”
The remark about Mattis seemed to upset the president. “That’s enough,” he said curtly and stood up. “I have an idea. A really brilliant idea. You just keep having your meeting by yourself. I’m done.” With that, he walked out of the room, accompanied only by his Secret Service detail.
Francis recalled being as confused as anyone else in the room. The president’s staff had seen him get angry and walk out of interviews before, but never meetings. Francis looked at his watch: 12:23. They could leave as late as 12:25 and still make it to the airport, although it would be close. He decided to give Trump two minutes to come back—but no more. He was prepared to ask the Secret Service to carry the president to the helicopter if necessary.
When Trump had not returned one minute later, Francis stepped out of the Situation Room and walked down the corridor to find him.
In the corridor outside the SCIF, Francis ran into a Mar-a-Lago employee. “One of the kitchen staff looked at me and just said, ‘Mr. Trump said you can find your own way home,’” Francis recalled. “I ran back up the stairs and saw his helicopter taking off.”
The president departed Mar-a-Lago at 12:25 PM, on schedule, although not quite as Francis had planned.
Weekend Warriors
As the North Korean missiles climbed into space, they came into full view of the radars linked to the only US line of defense: the missile defense system sitting in Alaska. Consisting of forty-four ground-based interceptors (GBIs), mostly located in Alaska, the system was crewed by members of the Alaska Army National Guard. It would be the responsibility of these Guardsmen to shoot down the missiles. “There’s no big red button to put that [interceptor] in play,” Colonel Kevin Kick, the unit’s commander, explained. “It all happens with a click of a mouse.”
The Alaska National Guard had a narrow window for those mouse clicks. A long-range missile is nearly impossible to hit during the period when its engines are firing; the limitations of time and space are simply too great. And once the warhead begins to fall back to earth, it reaches twenty times the speed of sound, making it virtually impossible to intercept.
But between those two moments is a window in which the hot nuclear warhead is coasting silently through the freezing cold of outer space. It is in these few moments, during the middle part of its journey—called “midcourse”—that the warhead is theoretically vulnerable. But to stop it the nation’s defenders must act quickly. Within minutes of detecting an enemy missile launch, American missile defenses must fire their interceptors to catch the warhead in this critical stage of its flight. Each interceptor releases a small “kill vehicle” whose only purpose is to slam into the nuclear warhead and destroy it, by impact alone, in what is called a “hit to kill.” Of the forty-four interceptors the United States maintained, forty had been placed in Alaska to ensure the best possible shot at any missiles fired from North Korea.
The Alaskan missile defense system—which has been temporarily mothballed pending an independent review of its technical performance—was called Ground-based Midcourse Defense (GMD) to distinguish it from other missile defenses, such as those based on ships or those designed to intercept missiles in other phases of their trajectory. But the most distinguishing feature of GMD was that, in 2020, it was the only missile defense system that could defend the United States against missiles from North Korea.
Even among missile defense advocates, however, there was little love for the system as it was deployed in Alaska. Like so many other defense decisions, the GMD system was a product of compromises and technical limitations. A panel of eminent scientists and engineers had been asked by the Defense Department to review the state of American missile defenses. The panel proposed replacing the system in Alaska entirely, calling the defense it offered “fragile” and mocking the “hobby shop” approach of the people who built it. They proposed designing all-new interceptor missiles and all-new radars and implementing an entirely new concept for how the system would operate. And these were the system’s friends.
It should not be surprising, given these reviews, that GMD had a terrible test record. In actual flight tests, its success rate was right at 50 percent—a coin flip. In nearly twenty years of testing, the system had worked only nine out of eighteen times. Only one of those eighteen tests involved shooting at an ICBM like the ones North Korea was firing that day. Even worse, these estimates had been padded by excluding tests that had to be scrubbed for bad weather or technical malfunctions. The system’s defenders—that is, those who wanted to fix it—knew that it needed many, many more tests. “There’s no way to prove out the design—let alone its reliability—without more fl
ight tests,” one of the panel members complained. “It’s stupid.”
But tests were expensive, costing nearly $300 million each. And there was another, bigger problem: failed tests sent the wrong message. They showed that the system was not working. Defense officials were worried that Kim Jong Un would be emboldened as he realized how ineffective the system was. Administration after administration had scaled back testing, resorting to classifying ever more information about failures. One document plainly admitted that the “deterrence value” of the system was “decreased by unsuccessful flights.”
In short, the only missile defense system charged with protecting the United States against Kim Jong Un’s ICBMs simply did not work—but the Defense Department refused to fix it, worrying that the tests required to do so would simply show the North Koreans that the system was a failure. This explains why the Obama and Trump administrations both spent billions of dollars to increase the number of GMD interceptors, but far less to actually improve them.
Instead, the Department of Defense came up with another plan: to fire a lot of interceptors at every missile and hope one of them got lucky. In March 2020, this concept of operations meant that each mouse click sent four interceptors flying into space for every North Korean missile coming at the United States. The reasoning was simple math: If each interceptor had a fifty-fifty chance of stopping an ICBM, four tries should be enough. What are the odds of flipping a coin and getting heads four times in a row?
Critics noted that this approach was based on an impoverished understanding of statistics. “I think this reasoning is flawed,” explained one scholar who would later be interviewed by the commission. “It assumes that the failure modes of the interceptors are independent of one another. But, in practice, if one interceptor fails because of a design flaw, say, it’s much more likely that others will do so too for the same reason.”
And indeed, GMD had plenty of design flaws, many of which were well known prior to March 2020. One test failed because of a manufacturing defect—a faulty circuit board that short-circuited when bits of solder or wire shook loose during the violent process of launching the interceptors into space. More than thirty of the forty-four interceptors had these defective circuit boards, but although this defect was publicly reported years earlier by the Los Angeles Times, the Missile Defense Agency decided that “no corrective actions” were needed for the old circuit boards, while deciding to change the manufacturing process for the new ones. That left more than thirty interceptors with the old boards. “If there’s a foreign object in one unit, it’s sort of whistling past the graveyard to assume that that’s a once-in-a-lifetime event,” one of the panelists who had reviewed the system told the Times reporter.
Another problem that confronted the nation’s missile defenders on the morning of March 22 was also an issue of simple math. A total of forty-four interceptors was enough to defend against eleven North Korean missiles. But North Korea had fired thirteen missiles at the United States. And to make matters worse, each missile released a pair of decoys to fool the “kill vehicles” trying to find the warheads in space. A system designed to deal with only eleven North Korean nuclear warheads now faced the task of finding thirteen real ones hidden among more than two dozen fakes.
There is no agreement, even among experts, as to whether the GMD system successfully intercepted any North Korean missiles on March 21. This is a problem that has plagued previous assessments of missile defense effectiveness. The real world is not a video game, even if the computer screen in the fire control unit displays gamelike messages like “probable kill” if it believes the intercept worked. At best, analysts can sift through reams of data from sensors and reach a probabilistic judgment as to whether a missile was intercepted or simply broke up on its own. There is always the danger, of course, of seeing what one wants to see in a pile of confusing and sometimes contradictory data.
We do know that, of the thirteen missiles North Korea fired at the United States, six failed to deliver nuclear weapons against our homeland. These missiles either broke up in flight or had warheads that disintegrated as they reentered the earth’s atmosphere. Some experts claim that these six missiles may not have been armed at all.
According to the Department of Defense, the fact that six of the thirteen nuclear warheads failed to reach the United States was powerful evidence that the nation’s missile defense system functioned well under demanding conditions. Indeed, they claimed that the system worked better than could reasonably have been expected, managing to pick out and destroy six of the eleven nuclear warheads it had the capacity to intercept, for a success rate of more than 50 percent. What the system did, explained one official, was “give us a limited capability to deal with a relatively small number of incoming ballistic missiles, which is better than nothing.”
Critics disputed, however, that any North Korean missiles were intercepted at all. It is far more likely, they argued, that the missiles broke up in flight or that the warheads disintegrated upon reentry. They noted that some North Korean missile tests had failed, including those whose mocked-up warheads burned up as they reentered the atmosphere.
Our investigators commissioned an independent review of the data provided by the Pentagon, but this review was not able to confirm any intercepts. The author of the review noted that “the data used by the Army raise many questions that create uncertainty over how much confidence can be placed in what the Army used to assess warhead kills.
“The method used by the Army to assess warhead kills appears reasonable on first inspection but on closer scrutiny serious questions can be raised,” the report continued. It concluded that “the Army does not appear to have sufficient data to assign high confidence to its claims.”
“A Protocol No-No”
President Trump’s helicopter landed on the tarmac at Palm Beach International Airport at 12:28 PM, three minutes after departing Mar-a-Lago and nineteen minutes before the anticipated impact of the North Korean ICBM. The engines of Air Force One were already running—a sign of urgency that one official traveling with the president characterized as “normally a protocol no-no.”
Air Force One would wait for the president, but no one else. “My boss,” the pilot explained to us later, “called and told me to depart as soon as the president got on board.” A small number of presidential staffers who had been staying nearby at the Hilton Palm Beach Airport Hotel had managed to board the plane before Trump arrived; Francis had arranged for his military assistant, who had been staying at the hotel, to get as many people on board as possible. One of the staff members remembered the unusually tight security protocol, with the Secret Service checking identification in addition to badges and bomb-sniffing dogs. “They were drooling all [over] the luggage,” she recalled. “I had dog spittle all over my bags.”
After the president’s helicopter landed, he walked across the tarmac and then carefully ascended the ramp stairs leading up to the airplane. Before he settled into his seat, he told the pilot, “Let’s get the hell out of here.”
Air Force One was wheels up at 12:37 PM. The pilot and crew recalled the particular violence of the takeoff. “It was a full-thrust departure,” one recounted, “up like a rocket.” The passengers were shaken by the sudden ascent. “We were climbing so high and so fast,” said another, “I started to wonder if we’d need oxygen masks.”
The pilot was racing to get the plane up high enough, and out far enough over the ocean, to ensure a safe distance from any nuclear explosion. He also directed everyone on board to close the aircraft’s window shades so as not to be blinded by the flash.
The president’s blind was still up when, at 12:48 PM, a 200-kiloton thermonuclear weapon detonated over Jupiter, Florida. By chance, he escaped being blinded, but it was due to more than luck that Air Force One itself survived. The plane was designed to be far more survivable in a nuclear environment than any commercial airliner, and the pilot had, in any event, gotten the aircraft far enough away from the epicenter of the
blast.
After the flash caught his eye, President Trump looked out the window, watching the fireball form slowly and rise up into the atmosphere. He observed the scene for a few seconds before the pilot banked the plane hard and headed farther out to sea.
“The president just said, ‘Absolutely beautiful,’” recalled a staffer. “I started to cry.”
10
Black Rain
Saturday, March 21, 2020, marked the worst catastrophe in American history.
North Korea’s nuclear strikes were carried out over the course of six hours. Of the thirteen nuclear weapons fired against the United States, seven delivered their powerful thermonuclear weapons onto the country and its citizens.
One nuclear weapon, aimed at Mar-a-Lago, struck Jupiter, Florida. Another nuclear weapon destroyed Pearl Harbor. A third struck Manhattan. Two nuclear weapons, probably aimed at the White House, missed wildly but fell on northern Virginia, their explosions separated by about an hour. Two more exploded off the California coast near San Diego, missing the port by enough that no one was killed.
In attempting to create a definitive and official account of this unparalleled national tragedy, we discovered that written accounts failed to convey accurately for millions of Americans what that day was like. The same event can mean many different things to many people. The commission found as many perspectives on the horror as there were survivors for us to interview.
The 2020 Commission Report on the North Korean Nuclear Attacks Against the United States Page 18
