Legal & Regulatory

Expert: 90% of U.S. Population Could Die if a Pulse Event Hits the Power Grid

When a large electromagnetic pulse (EMP) or geomagnetic disturbance (GMD) event occurs—which, according to Sen. Ron Johnson (R-Wis.), there is “100% certainty” will happen at some time in the future—as many as 9 out of 10 people in the U.S. could die.

Johnson, chairman of the U.S. Senate Committee on Homeland Security & Governmental Affairs, posed questions to witnesses testifying on Capitol Hill before his committee. He asked R. James Woolsey, chairman of the Foundation for Defense of Democracies and former director of the Central Intelligence Agency, what would happen to society if the electrical grid were to be down for an extended period of time, such as a year or two, following an EMP event?

Woolsey responded, “It’s briefly dealt with in the commission report of [2008]. There are essentially two estimates on how many people would die from hunger, from starvation, from lack of water, and from social disruption. One estimate is that within a year or so, two-thirds of the United States population would die. The other estimate is that within a year or so, 90% of the U.S. population would die. We’re talking about total devastation. We’re not talking about just a regular catastrophe.”

What Are GMDs and EMPs?

Woolsey may have been the most pessimistic of the witnesses, but he wasn’t the only one who was worried. Joseph H. McClelland, director of the Federal Energy Regulatory Commission’s Office of Energy Infrastructure Security, reviewed some of the possibilities in his testimony. He explained that GMD and EMP events are generated from either naturally occurring or man-made causes.

In the case of GMDs, naturally occurring solar magnetic disturbances periodically disrupt the earth’s magnetic field, which in turn, can induce currents on the electric grid that may simultaneously damage or destroy key transformers over a large geographic area.

EMPs can be generated by devices that range from small, portable, easily concealed battery-powered units all the way through missiles equipped with nuclear warheads. In the case of the former, equipment is readily available that can generate localized high-energy bursts designed to disrupt, damage, or destroy electronics, such as those found in control systems on the electric grid.

The EMP generated during the detonation of a nuclear device is far more encompassing and generates three distinct effects, each impacting different types of equipment; a short high energy radio frequency–type burst called E1 that destroys electronics; a slightly longer burst that is similar to lightning termed E2; and a final effect termed E3 that is similar in character and effect to GMD targeting the same equipment including key transformers. Any of these effects can cause voltage problems and instability on the electric grid, which can lead to wide-area blackouts.

Richard L. Garwin, PhD, fellow emeritus at the IBM Thomas J. Watson Research Center, noted in his testimony that “very serious consequences are estimated for such an event of a magnitude that can be expected to occur at random once per century, with greater events occurring with lower probability and lesser events more frequently.”

Has It Ever Happened Before?

McClelland and Garwin provided some examples. The largest event ever recorded—called the Carrington Event—occurred in 1859, during the pre-electric-grid era. McClelland estimated that it was roughly a K8 or K9 level event. The K-index quantifies disturbances in the horizontal component of earth’s magnetic field with an integer in the range 0–9 with 1 being calm and 5 or more indicating a geomagnetic storm. To provide a point of reference, Johnson noted that G3 events (Table 1) are regular occurrences.




Table 1. Geomagnetic storms.
The NOAA Space Weather Scales were introduced as a way to communicate to the general public the current and future space weather conditions and their possible effects on people and systems. Source: Space Weather Prediction Center National Oceanic and Atmospheric Administration

“The only long wires in those days were telegraph wires,” Garwin said. “No grid to bring down, no pipelines, but it did play havoc with telegraph wires—burned up some telegraph offices—and it would be much, much, much worse [now]. It would collapse societies.”

There was another relatively large event in 1921. McClelland did not have a rating level immediately available for the 1921 event, but he suggested it was about 5,000 nanoTesla, which was agreed to have been less than the Carrington Event.

More recently, in 1989, the entire Quebec, Canada, grid collapsed in 93 seconds. About 6 million customers were without power for roughly 10 hours. Cost estimates ranged from $1 billion to $2 billion even though there was very minor equipment damage. McClelland confirmed that the 1989 event was about one-tenth the size of the 1921 event. Other sources have pegged the 1989 storm responsible for the Quebec power outage to be about one-third the intensity of the Carrington Event.

In 2003, a very-low-level event occurred that affected South Africa’s grid. McClelland said it was one-fiftieth the size of the 1921 event, but it stayed on for a period of days. The grid did not collapse. Instead, equipment saw prolonged exposure to the event and, over a period of months, 12 transformers were lost as a result of the event.

Johnson said that Lloyd’s of London estimates about $2 billion worth of damage results annually from G3-type events. He also suggested that a massive corona discharge or solar flare in 2012 missed hitting the Earth by a matter of days.

“On July 23, 2012, there was a Carrington-level event. It missed us by three days,” Woolsley confirmed, attributing the information to a friend who had provided the tip.

What Can Be Done?

“People tend to think of it as science fiction,” Woolsey said. “People get into not wanting to think about it, not wanting to worry about it, because it’s too terrible.”

Even though the consequences of an event seem quite dire, there were suggestions offered to lessen the danger.

Garwin provided four recommendations regarding the bulk power system. He said federal policy and practice are missing programs for the following:

  • Training and equipping utility and transmission operators to bring down within seconds (switch off) transmission lines that are at risk of being damaged.
  • Implementing “rapid islanding” of the grid, to maintain a large fraction of the power consumers in operation by the use of whatever island generation capacity exists; this also facilitates restoring the bulk power system to operation, in contrast with a “black start.”
  • Fitting transmission lines on a priority basis with “neutral current blocking devices” (capacitors) in the common neutral-to-ground link of the 3-phase transformers of EHV transmission systems at one end of the line—whether 3-phase transformers or three single-phase transformers. Where transformers at both ends are autotransformers this may not be possible, in which case series-blocking capacitors in the power lines themselves should be installed (even if shorted until an EMP event is recognized).
  • Alerting grid operators and others to a high-altitude nuclear explosion within milliseconds of the event (by detection of the unambiguous very brief E1 pulse).

When he found out that neutral current blocking devices could be installed for roughly $100,000 per transformer and that roughly 200 critical transformers needed to be protected, Johnson suggested that the price was pocket change for the U.S. Government. Garwin agreed that the price is cheap compared to replacement costs of $7 million per transformer and very cheap compared to the damages that could potentially be avoided.

Noting that very little has been done to correct deficiencies identified in previous studies, Johnson said, “We need to start implementing some protections.”

Aaron Larson, associate editor (@AaronL_Power, @POWERmagazine)

[Ed. correction (Feb. 6, 2022): There was a misspelling of R. James Woolsey’s name in the original article posted on July 22, 2015. His last name was incorrectly spelled Woolsley.]

SHARE this article