Richard Carrington was drawing sunspots.
On the morning of September 1, 1859, the British amateur astronomer was at his private observatory in Redhill, Surrey, making a routine pencil sketch of the solar disk projected onto his screen, when he observed something he had never seen before and had no framework to understand. A brilliant white light erupted from within a complex sunspot group, intensifying rapidly for approximately five minutes before fading. He noted the time: 11:18 AM. He sketched what he saw and sent an immediate report to the Royal Astronomical Society.
That evening, the global telegraph network began to behave strangely. Operators across North America and Europe found that their equipment was generating spurious currents of its own, without any battery connected. Some lines caught fire. Some operators received electrical shocks when they touched their keys. Others discovered that they could transmit messages using the atmospheric current alone, with all power disconnected from their systems.
The northern lights that night were visible from Cuba, the Bahamas, and northern Mexico. Miners in the Rocky Mountains woke in the middle of the night convinced that dawn had arrived. The auroras were bright enough to read newspapers by in Boston.
What Carrington had observed at 11:18 AM on September 1 was a solar flare, the first ever witnessed and documented by a human observer. The electromagnetic storm that followed was the largest geomagnetic event in the instrumental record. It lasted approximately 17 hours. The telegraph network, the most sophisticated electromagnetic technology of 1859, was the only human infrastructure vulnerable to its effects.
In 2024, the equivalent infrastructure includes power grids covering billions of people, hundreds of operational satellites, GPS systems integrated into transportation, banking, emergency services, and agriculture, submarine communications cables connecting the global internet, and every electronic device whose operation depends on the grid that supplies it.
The 2008 National Academy of Sciences report Severe Space Weather Events: Understanding Societal and Economic Impacts estimated the cost of a Carrington-scale event hitting current infrastructure at more than two trillion dollars in the first year alone, with recovery times measured in years rather than months due to a and irreplaceable bottleneck in the supply chain.
The Physics of a Coronal Mass Ejection
A solar storm of the Carrington type begins with a solar flare, a rapid release of electromagnetic energy from the Sun’s surface driven by the sudden reorganization of magnetic field lines in an active sunspot region. The flare’s electromagnetic radiation, traveling at the speed of light, reaches Earth in approximately eight minutes. This component of the event, intense X-ray and ultraviolet radiation, is the warning.
What follows the radiation is the coronal mass ejection, a billion-ton cloud of magnetized plasma ejected from the Sun’s corona at speeds between several hundred and several thousand kilometers per second. The Carrington CME is estimated to have traveled at approximately 2,300 kilometers per second, reaching Earth in approximately 17 hours, significantly faster than the typical two to three days for an average CME.
When a CME’s embedded magnetic field interacts with Earth’s magnetosphere, the consequences depend on the orientation of the CME’s magnetic field relative to Earth’s. If the CME’s field is oriented southward, opposite to Earth’s northward-pointing magnetosphere, it can connect directly to the magnetosphere through a process called magnetic reconnection, driving massive currents through Earth’s ionosphere and into the ground below.
These ground-induced currents are the mechanism by which a CME destroys power grid infrastructure. High-voltage transformers are designed to handle alternating current at frequencies. The slow oscillations of geomagnetic storms, at periods of minutes rather than the 60 cycles per second of grid power, saturate transformer iron cores, causing intense heating. A transformer heated beyond its design limits can fail catastrophically, and large high-voltage transformers, the 500-kilovolt units that form the backbone of continental power grids, cannot be repaired in the field. They must be replaced.
The replacement supply chain is the bottleneck that makes a Carrington-scale event potentially more disastrous than any other non-nuclear catastrophe. Large high-voltage transformers are manufactured by a small number of specialized facilities worldwide. Lead times for custom units run to 12 to 18 months under normal production conditions. The transformer manufacturing capacity is sufficient for routine replacement of aging equipment but not for mass replacement of storm-damaged units across an entire continental grid simultaneously.
Thomas Berger, former director of NOAA’s Space Weather Prediction Center, has been about this vulnerability in documented public statements: because most electrical technology is currently ground-based, the copper coils of the transformers at the heart of power distribution systems would melt, possibly even reaching cause a planetary blackout. Some consequences could be felt for years.
The 2008 NAS report’s scenario, which the subsequent 2013 update by John Kappenman for the Oak Ridge National Laboratory refined further, describes a North American transformer damage scenario in which between 140 and 340 large high-voltage transformers fail simultaneously, producing blackouts affecting up to 130 million people whose duration could extend to four to ten years in the most affected regions.
The 2012 Near-Miss
The most important documented data point for understanding the Carrington Event’s contemporary relevance is an event that most people have never heard of, because it did not hit Earth.
On July 23, 2012, the Sun ejected a coronal mass ejection of exceptional intensity. Solar physicists Daniel Baker of the University of Colorado and Janet Luhmann of UC Berkeley, whose analysis was published in the journal Space Weather in 2013, measured the CME’s properties using data from the STEREO-A spacecraft, which was positioned in Earth’s orbit approximately 120 degrees ahead of Earth at the time. Their measurements showed that the July 2012 CME was comparable in intensity to the Carrington Event, possibly exceeding it in some parameters.
Earth missed the CME by approximately nine days. If Earth had been at the position it occupied nine days earlier in its orbit, the CME would have been a direct hit.
Baker’s published assessment: if the eruption had occurred one week earlier when the active region was facing Earth, it would have hit us. He estimated the probability that Earth will be hit by a Carrington-scale event in the next decade at approximately 12 percent, based on the observed frequency of extreme solar events over the instrumental record.
12 percent per decade is not a small probability. A 12 percent chance per decade means approximately a 50 percent cumulative probability over a 40-year period. Whatever the number, the July 2012 near-miss establishes that Carrington-scale events are not uniquely historical curiosities but are recurring features of the solar activity cycle whose Earth-hitting probability is calculable and non-negligible.
The 2003 Halloween Storms and Their Documented Effects
The October-November 2003 solar storm sequence, named the Halloween Storms for their timing, provides the most extensively recorded modern example of solar storm infrastructure effects and the closest available proxy for what a Carrington-scale event’s consequences would look like in contemporary infrastructure.
The Halloween Storms produced a series of X-class flares, culminating in an X28+ event on November 4 whose intensity exceeded the X-class scale’s calibrated range. The documented infrastructure effects were and measurable:
Power grid failures in Sweden. On October 30, 2003, a geomagnetically-induced current event caused a power failure in the Malmö region of southern Sweden affecting approximately 50,000 customers for approximately 20 minutes. The Swedish grid operator documented the event in detail, providing one of the most precise post-event analyses of geomagnetically-induced current effects on a modern high-voltage grid.
Satellite effects. More than 47 satellites reported anomalies during the Halloween Storms, with two satellites suffering permanent damage. The ADEOS-2 Japanese Earth observation satellite lost approximately 15 percent of its power generation capability. Satellite operators documented degraded performance across multiple orbital regimes.
GPS degradation. Wide Area Augmentation System GPS accuracy was degraded for approximately 30 hours during the peak storm periods, affecting precision navigation applications including aviation and surveying.
Polar route aviation disruption. Airlines rerouted aircraft away from polar routes during the storm period to avoid the heightened radiation environment and communications degradation. The extra fuel costs and schedule disruption were documented by major carriers.
These documented effects occurred during a storm whose intensity was significant but substantially below the Carrington Event’s estimated intensity. The documented Halloween Storm effects, scaled to Carrington intensity through the NAS report’s analysis, produce the multi-year grid failure and cascading infrastructure collapse scenario that makes the Carrington Event the risk that national security planners and infrastructure engineers treat as the most consequential natural hazard in the risk portfolio.
the Infrastructure Vulnerabilities
The NAS report’s infrastructure analysis is worth developing in terms rather than in general impact language because the specificity is what makes the risk assessment compelling rather than abstract.
Power grid vulnerability is the primary concern because the modern economy’s other critical systems, water treatment, communications, transportation, financial systems, medical systems, all depend on continuous electrical power. The cascade mechanism: transformer failures produce blackouts, extended blackouts disable water treatment and pumping systems, water system failure creates public health emergencies, fuel supply chains dependent on electronic payment and GPS logistics fail, hospitals exhaust backup generator fuel, communications dependent on powered infrastructure fail progressively.
The timeline of cascading failures under an extended grid blackout has been modeled by researchers including the authors of the NAS report and by the Foundation for Resilient Societies, whose published analysis of extended blackout scenarios documents the cascade points. The first 24 hours are the period in which backup generator fuel provides continuity for critical systems. The first week is the period in which fuel resupply logistics, themselves dependent on electronic payment and GPS, begin to fail. The first month is the period in which the urban food supply, dependent on continuous cold chain maintenance and electronic payment, becomes critical.
Satellite vulnerability is the secondary concern. GPS satellites, communications satellites, and Earth observation satellites are exposed to the direct particle radiation of the CME before the geomagnetic effects reach the ground. Satellite damage or loss of function removes GPS navigation from the transportation, emergency response, and precision agriculture systems that have been built around GPS availability as a permanent infrastructure assumption.
Communications vulnerability follows from satellite damage and grid failure. High-frequency radio communications, which do not depend on satellites or grid power, are degraded by the intense ionospheric disturbances of the geomagnetic storm itself, removing even the most resilient backup communications channel during the storm’s most intense phase.
The Solar Cycle Context
The Carrington Event occurred on September 1, 1859, near the peak of Solar Cycle 10. The most powerful solar events historically occur during the rising and peak phases of the solar cycle, when the Sun’s magnetic activity is at its maximum and the most complex sunspot regions, capable of producing the most energetic flares and CMEs, are most prevalent.
Solar Cycle 25, which began in December 2019 at solar minimum, has been more active than the predictions of many solar forecasters. The Solar Cycle 25 Prediction Panel, convened by NOAA and NASA, initially predicted a below-average cycle. By 2023, actual sunspot counts were tracking significantly above the panel’s predictions, and several X-class flares occurred in 2023 and 2024 that exceeded the predicted activity level.
The solar maximum of Solar Cycle 25 is expected to occur between 2024 and 2026, placing the period of highest Carrington-scale event probability at the current time.
Whether the current solar maximum will produce a Carrington-class event is not predictable with current solar physics capabilities. The magnetic configurations required to produce the most extreme events can develop on timescales of days, and current solar monitoring provides approximately 15 to 45 minutes of warning between CME detection and Earth impact, depending on the CME’s speed.
Fifteen minutes of warning is the operational constraint that makes infrastructure protection against Carrington-scale events require advance preparation rather than reactive response. The transformers that would fail in a Carrington event cannot be protected in real-time. The satellites that would be damaged are not hardened against the relevant particle flux in most commercial designs. The GPS systems whose failure would cascade through navigation-dependent infrastructure cannot be backed up in 15 minutes.
What can be done in advance, and what most national grid operators have not fully done, is the combination of transformer protection devices, emergency stockpiling of replacement transformers, and the operational procedures that allow controlled de-energization of vulnerable grid sections during the warning period.
The United States Federal Energy Regulatory Commission issued a directive in 2013 requiring grid operators to develop geomagnetic disturbance mitigation plans. The implementation of these plans has been recorded as incomplete by subsequent regulatory reviews.
The vulnerability documented in 1859 by the failure of the telegraph network appears in 2024 by the incompleteness of the regulatory compliance that a Carrington recurrence would expose.
The storm that missed Earth in July 2012 is still in the data record.
The probability that the next one misses is calculable.
It is not 100 percent.
