The Carrington Event (1859): The Largest Geomagnetic Storm in Recorded History.

What the Carrington Event was, in a paragraph.

The Carrington Event — named for the English astronomer Richard Christopher Carrington (1826–1875), who at the moment the event began was independently observing the same phenomenon as the Kew Observatory's amateur observer Richard Hodgson — was the most intense geomagnetic super-storm in the recorded history of human instrumentation. It comprised at least two and probably three closely-spaced coronal mass ejections from a complex active region on the solar disc visible from Earth in late August and early September 1859. The first and largest CME, ejected at approximately 11:18 GMT on September 1 in association with the white-light flare that Carrington and Hodgson independently observed, traversed the Sun-Earth distance in approximately 17.6 hours — faster than any subsequently-measured CME — arriving at Earth's magnetosphere at approximately 04:00 GMT on September 2, 1859. The geomagnetic storm that followed produced ground-induced currents strong enough to overload the telegraph systems then in service across the United States, Britain, continental Europe, and parts of Asia and Australia. Telegraph offices reported sparks jumping between operators and equipment; in some cases operators disconnected their batteries entirely and continued to transmit and receive messages on the induced current alone, an effect documented in contemporary newspaper reporting and the operator's logs preserved by the American Telegraph Company and Western Union. Aurorae — the visible signature of the storm at low altitudes — were observed at unprecedented low latitudes: Honolulu, Cuba, Jamaica, Panama, central Colombia, and the Hawaiian Islands, well within the tropics. Contemporary observers in Boston reported being able to read the newspaper by aurora light at midnight. The event was systematically documented at the time and given its first comprehensive synthesis by the Scottish physicist Balfour Stewart in 1861. Modern reconstruction of the storm intensity from the surviving Colaba (Bombay) magnetic-observatory trace, which uniquely captured a substantial portion of the storm without going off-scale, places the minimum disturbance-storm-time (Dst) index at approximately -1750 nanoTesla, roughly three times the intensity of the largest 20th-century event (the March 1989 storm that collapsed the Hydro-Québec grid). The "mystery" of the Carrington Event, in the conventional sense, is not the event itself — the physics is well understood — but the infrastructure-vulnerability question it raises: what a comparable storm, hitting a fully electrified industrial civilization with continental high-voltage transmission grids and orbital satellite infrastructure, would do. The July 23, 2012 CME, of broadly comparable magnitude to the Carrington event, missed Earth's orbital position by approximately nine days; the magnitude and trajectory of that near-miss are the strongest modern evidence that Carrington-class events are not statistical aberrations.

The documented record.

The Carrington-Hodgson observation

On the morning of September 1, 1859, Richard Carrington was conducting his customary sunspot survey at his private observatory at Redhill, Surrey, using a four-and-a-half-inch refracting telescope with a solar projection apparatus. Verified At 11:18 GMT (Carrington's recorded time) he observed two patches of "intensely bright and white" light appear at the edges of an active sunspot group he had been sketching. He recorded their movement across the projected solar disc over the next five minutes; they then faded. Carrington's account, published in the Monthly Notices of the Royal Astronomical Society in November 1859, is the first documented observation of a solar flare in the visible-light continuum (a "white-light flare") and the founding observation of the field of solar flare physics [1]. Independently and almost simultaneously, the amateur English astronomer Richard Hodgson observed the same flare from his own equipment in Highgate, near London, and published his confirming account in the same journal issue [2].

The geomagnetic instrument record

The September 1859 storm was registered at multiple magnetic observatories that had been in continuous operation since the 1830s and 1840s as part of the international "Magnetic Crusade" coordinated by Edward Sabine, Carl Friedrich Gauss, and others. Verified The Kew Observatory near London, the Royal Greenwich Observatory, the Helsingfors (Helsinki) Observatory, the Colaba Observatory in Bombay (now the Indian Institute of Geomagnetism's Alibag observatory complex), and observatories at St. Petersburg, Rome, Trevandrum, and Hobart all produced magnetograph traces during the September 1–3 window. Many of those traces went off-scale at the height of the disturbance; Colaba's, uniquely among the well-instrumented stations, captured the main phase with a horizontal-component decrease of approximately 1,600 nT. This Colaba trace is the most important single instrumental record for modern reconstruction of the storm's intensity and is the basis for the conventional ~-1750 nT Dst estimate [3]. The records are preserved in the National Archives of Great Britain (Kew), the Royal Greenwich archives at Cambridge University Library, and the Indian Institute of Geomagnetism.

The telegraph effects

Across the global telegraph network of 1859, the storm produced unprecedented operational disturbances. Verified The most-cited contemporary accounts: at the Boston station of the American Telegraph Company, on September 2, operators reported sparks jumping from the equipment, paper burning, and stations unable to transmit through the induced currents. At about 8:30 AM on September 2, after the batteries had been disconnected, the Boston operator and the operator at Portland, Maine, discovered that they could continue to send and receive messages on the induced current alone for approximately two hours. The "Boston-Portland" exchange — widely reprinted in 1859 newspapers and in the American Journal of Science — remains the most-quoted single piece of evidence of the storm's electromagnetic intensity at ground level [4]. Comparable battery-free operation was documented on lines in France between Lyons and Paris, and contemporary reporting from British telegraph offices describes operator shocks, equipment fires at the Norwegian Telegraph Office, and the suspension of service across substantial portions of the network for periods of hours to a day [5].

The auroral observations

Aurorae visible from the Caribbean, central America, the equatorial Pacific, and tropical Asia are the most-cited unusual signature of the storm. Verified The auroral observations recorded in the contemporary newspaper and scientific literature include: bright auroras visible from Havana, Cuba, and from Kingston, Jamaica, on the nights of September 1–2 and 2–3; auroras visible from Panama and from Cartagena, Colombia (~10°N geomagnetic latitude); auroras observed from the Hawaiian Islands and recorded in the missionary journals at Honolulu (~21°N); auroras reported by ships in the equatorial Pacific, with one widely-reprinted account from a vessel near the Samoan group. In the northern hemisphere, the auroras of September 1–3 were variously described in newspaper reports from Boston, New York, Washington, San Francisco, and many smaller cities as "of intense brilliancy," with the September 2 New York Times reporting that "the streets were lighted as if by full moonlight" [5][6]. The auroral oval, in normal conditions confined to high latitudes, was displaced equatorward by approximately 30–40 degrees during the storm's main phase.

The Balfour Stewart synthesis (1861)

Balfour Stewart of the Kew Observatory published the first comprehensive scientific analysis of the September 1859 events in the Philosophical Transactions of the Royal Society in 1861. Verified Stewart's paper compiled the Kew, Greenwich, Helsingfors, and other magnetic-observatory records, placed them in their broader 1859 solar context, and argued for a causal connection between the Carrington-Hodgson white-light flare and the geomagnetic storm that began approximately seventeen hours later. Stewart's argument was the first significant scientific case for a Sun-Earth electromagnetic linkage of this kind [7]. The full physical mechanism — coronal mass ejection of magnetized plasma, propagation through the heliosphere, interaction with Earth's magnetopause — was not understood for another century; CMEs as a class of phenomenon were not identified instrumentally until the Skylab coronagraph observations of 1973–1974.

The 1859 ice-core signature

Polar ice cores from Greenland and Antarctica preserve a measurable nitrate (NO3) signature associated with very large solar particle events. Claimed The 1859 layer in multiple ice cores has been examined for an associated SPE signature. The initial reports (McCracken et al., 2001) of an unusually large 1859 nitrate spike have been substantially qualified by subsequent work (Wolff et al., 2012; Sukhodolov et al., 2017), which has argued that the nitrate-spike method is unreliable as a proxy for individual large SPEs and that the 1859 signature is not unambiguous. The cosmogenic-isotope record (Be-10, C-14), by contrast, has been used to identify several very large solar-particle events in the pre-instrumental record (notably 774–775 CE, 993–994 CE, 660 BCE) that may have been substantially larger than Carrington in particle output, though their geomagnetic-storm signatures cannot be directly measured. Disputed [8][9]

The 1989 Quebec storm and the calibration of "Carrington-class"

The March 13–14, 1989 geomagnetic storm — the most intense well-instrumented storm of the modern era — produced a minimum Dst of approximately -589 nT. Verified The induced currents in the Hydro-Québec long-distance transmission grid produced a cascade failure that left approximately six million people without power for nine hours and damaged a transformer at the Salem Generating Station in New Jersey. The Quebec event has since become the calibration benchmark for the question "what would a Carrington-scale event do?" The relevant scaling is approximately a factor of three in disturbance intensity, but the relationship between storm intensity and infrastructure damage is non-linear and not directly observable [10].

The July 2012 near-miss

On July 23, 2012, a coronal mass ejection of broadly Carrington-class magnitude was emitted from the Sun. Verified The CME was directly measured by the STEREO-A spacecraft, which happened to be in the CME's trajectory. The measured plasma and field parameters were used by D. N. Baker, Bruce Tsurutani, and colleagues in a 2013 Space Weather paper to argue that, had the CME struck Earth, the resulting storm would have been at least comparable to the 1859 event and potentially larger. Earth's orbital position at the time placed it approximately nine days ahead of the CME's trajectory; the CME missed [11]. The 2012 event is the strongest modern evidence that Carrington-class CMEs are an ongoing class of solar phenomenon, not a 19th-century anomaly.

The modern infrastructure-impact projections.

The Lloyd's / AER 2013 report

The most-cited modern estimate of the economic consequences of a Carrington-class event is the 2013 joint study by Lloyd's of London and Atmospheric and Environmental Research (AER), Solar Storm Risk to the North American Electric Grid. Claimed The report estimated that a Carrington-class event hitting the contemporary U.S. grid could leave 20–40 million people without power, for periods of weeks to months in the most heavily affected regions, with first-order economic losses of $0.6–$2.6 trillion. The model assumed extra-high-voltage transformer damage as the principal infrastructure-loss mechanism and assumed transformer-replacement times of 12–18 months given existing global manufacturing capacity [12]. The report's methodology has been the subject of subsequent technical debate; subsequent peer-reviewed work has supported the broad order-of-magnitude conclusion while disputing specific scenarios.

The NAS 2008 / 2009 study

The National Research Council's 2008 workshop and 2009 report Severe Space Weather Events: Understanding Societal and Economic Impacts presented a parallel analysis. Claimed The NAS report's principal contribution was to detail the cascade-failure pathway for the modern electrical grid: ground-induced currents in transmission lines drive saturation of EHV transformers, which produces local overheating and harmonic distortion that can damage the transformer windings; large enough numbers of transformer failures across a grid produce cascade failure of the grid as a whole; the most heavily affected transformers are difficult-to-replace custom EHV units with manufacturing lead times measured in years [13]. Subsequent industry response, coordinated through the North American Electric Reliability Corporation (NERC) and the Federal Energy Regulatory Commission (FERC), has produced standards (TPL-007 series) requiring utilities to assess and where necessary mitigate geomagnetic-disturbance risk.

The 2020 PROSWIFT Act

The Promoting Research and Observations of Space Weather to Improve the Forecasting of Tomorrow Act (PROSWIFT, Public Law 116-181), signed into law in October 2020, formalized federal responsibilities for space-weather forecasting and infrastructure protection. Verified The act assigns operational forecasting to the National Oceanic and Atmospheric Administration's Space Weather Prediction Center, research to NASA and NSF, and coordination of national-security implications to the Department of Defense. The PROSWIFT Act is the principal federal legislative response to the Carrington-class risk and an explicit acknowledgment that the risk is treated as a serious infrastructure question [14].

The satellite and communications question

A Carrington-class CME's effect on modern satellite infrastructure is, on the available analyses, more uncertain than its effect on the terrestrial grid. Claimed The principal mechanisms are: (a) single-event upsets and latch-up failures in spacecraft electronics from solar-energetic-particle exposure; (b) atmospheric heating and consequent expansion of the upper atmosphere increasing drag on low-earth-orbit satellites; (c) charging differentials on satellite surfaces in geostationary orbit. The February 2022 loss of approximately forty Starlink satellites to a modest geomagnetic storm (Dst ~-66 nT) was the largest LEO infrastructure loss from space weather in the satellite era and provided the first dramatic operational data on the atmospheric-drag mechanism. Verified The implications for the now ~7,000 satellites in low Earth orbit during a Carrington-class event are not well-constrained.

The minority "smaller threat" position

A minority of researchers, principally in the electrical engineering and grid-operations communities, have argued that the catastrophic-grid-collapse scenario is overstated. Disputed The principal arguments: (a) industry standards have substantially improved since 1989, with TPL-007 mitigation now in place; (b) the EHV transformer-failure modes assumed in the Lloyd's and NAS reports may be more conservative than the empirical record warrants; (c) the protective response time of modern grid operators is faster than the 1989 cascade illustrates. The mainstream space-weather community has not been persuaded by these arguments; the question is, by its nature, not directly testable until a Carrington-class event actually arrives.

The unresolved questions.

The actual Carrington Dst

The conventional -1750 nT figure rests primarily on the Colaba (Bombay) magnetic-observatory trace from September 2, 1859. Disputed Subsequent reanalyses have produced estimates ranging from approximately -850 nT (Cliver & Dietrich, 2013) to approximately -1760 nT (Tsurutani, Gonzalez et al., 2003), depending on the treatment of the local ionospheric current contribution to the recorded H-component disturbance and the calibration of the Colaba magnetograph relative to modern standards. The 2013 Cliver-Dietrich reanalysis specifically argued that the Carrington storm, while severe, was probably not three times larger than the 1989 Quebec event but rather closer to twice as large. The reanalysis does not change the qualitative conclusion that Carrington was the largest in the instrumental record; it does change the scaling factor for modern infrastructure projections [15].

The recurrence interval

How frequently Carrington-class events occur is poorly constrained. Disputed The conventional figure of "approximately once per 100–500 years" is derived from a combination of the 167-year instrumental record (one event, the 1859 storm itself; the 1921 storm and several other large 20th-century events were smaller), the auroral-record statistics from older historical observations, and the cosmogenic-isotope record. Riley (2012) used the cosmogenic-isotope record to estimate a probability of approximately 12% per decade for a Carrington-class event; Love (2012) produced a substantially lower estimate. The wide range reflects the very small sample size of well-characterized events.

The interaction with modern infrastructure

The principal uncertainty in the Lloyd's and NAS scenarios is the actual response of modern EHV transformer populations to ground-induced currents at Carrington magnitudes. Unverified No transformer population has been subjected to a Carrington-class event since the technology came into widespread use. The 1989 Quebec event tested a particular transformer-population configuration; the response of contemporary grids depends on parameters (geomagnetic latitude, ground conductivity, transformer age and design, mitigation hardware in place) that vary substantially across grids. The PROSWIFT Act, the TPL-007 series, and the corresponding work in Europe, Japan, and South Korea are responses to a class of vulnerability whose actual magnitude under a Carrington-class event remains projection rather than measurement.

The pre-1859 record

Several historical events — the auroras of 1770 (recorded in detail in East Asian sources including the personal journals of Korean astronomers), the storms associated with the 1859 active cycle but in earlier or later months, and the 1872 "Civil War aurora" — are candidates for Carrington-class status in the pre-instrumental or partially-instrumented record. Disputed The Hayakawa et al. (2017–2020) sequence of papers reconstructing the 1770 event from East Asian observations argued that it may have produced auroras at lower latitudes than 1859, though the magnetic-storm intensity cannot be reconstructed for an event predating systematic instrumentation. The 1872 event has been argued (Hayakawa et al., 2018) to have been geomagnetically comparable to 1859.

Primary material.

• The Royal Astronomical Society archives hold Carrington's original 1859 observation notes and the manuscript drafts of his November 1859 Monthly Notices paper.
• The Kew Observatory magnetograph records, held with the Met Office historical collection and at the British National Archives, contain the September 1–3, 1859 traces and the surrounding-week baseline records.
• The Indian Institute of Geomagnetism at Alibag holds the original Colaba magnetograph traces — the most important single instrumental record of the storm's intensity.
• The Royal Greenwich Observatory archives, at Cambridge University Library, contain the parallel Greenwich magnetic traces.
• The National Oceanic and Atmospheric Administration / National Centers for Environmental Information consolidates digitized versions of the 1859 magnetograph traces and provides modern reconstructions of the storm parameters.
• NASA's Goddard Space Flight Center and the Solar Physics community archives at the High Altitude Observatory (Boulder) hold the principal modern-reanalysis literature on the event.
• The American Antiquarian Society in Worcester, Massachusetts holds the 1859 newspaper microfilm collections used by modern researchers to reconstruct the geographical extent of the auroral observations and the telegraph-system disturbances in the United States.

The sequence.

1. Late August 1859 A complex active region develops on the visible solar disc. Multiple smaller flares are observed by European solar observers in the preceding week.
2. August 28–29, 1859 An earlier moderate geomagnetic storm produces auroras across Europe and North America. Telegraph operators report minor disturbance. This is likely the result of a precursor CME from the same active region.
3. September 1, 1859, 11:18 GMT Richard Carrington at Redhill, Surrey, and Richard Hodgson at Highgate independently observe the white-light flare. Carrington records its movement across the projected solar disc for ~5 minutes. The associated CME is launched.
4. September 2, 1859, ~04:00 GMT The CME arrives at Earth's magnetosphere. Sudden commencement registered at Kew, Greenwich, Helsingfors, Colaba, and other observatories.
5. September 2, 1859, overnight UTC Main phase of the storm. Telegraph systems across the Atlantic-basin and European networks fail. Aurora observed from Cuba, Jamaica, Hawaii, the equatorial Pacific. The Boston-Portland battery-free transmission is documented at ~8:30 AM Eastern.
6. September 3, 1859 Storm decays. Aurora persists into the night of September 3–4.
7. November 1859 Carrington and Hodgson publish their independent observation accounts in Monthly Notices of the Royal Astronomical Society. American Journal of Science, Nature, and continental scientific journals carry the early synthesis of the storm.
8. 1861 Balfour Stewart publishes the first comprehensive scientific analysis. The Carrington-Hodgson observation is identified as causally connected to the geomagnetic storm.
9. 1872 The "Civil War aurora" storm of February. Subsequently argued by Hayakawa et al. (2018) to have been comparable in geomagnetic intensity to 1859.
10. May 1921 The "New York Railroad Storm." The most intense geomagnetic storm of the early 20th century, with significant telegraph and railway-signal disturbance.
11. March 13–14, 1989 The Quebec storm. Hydro-Québec cascade failure; modern grid-collapse calibration.
12. 2003 (Tsurutani et al.) The Carrington Dst is reconstructed at ~-1760 nT from the Colaba trace.
13. 2008–2009 National Research Council reports on severe space weather.
14. July 23, 2012 A Carrington-class CME is launched. Earth's orbital position misses by ~9 days. STEREO-A directly measures the event in transit.
15. 2013 (Lloyd's / AER) Solar Storm Risk to the North American Electric Grid published.
16. 2013 (Cliver-Dietrich) Reanalysis suggests the Carrington Dst may have been ~-850 nT rather than ~-1750 nT.
17. October 21, 2020 The PROSWIFT Act (P.L. 116-181) becomes law.
18. February 4, 2022 SpaceX loses approximately 40 Starlink satellites to a modest geomagnetic storm. First dramatic operational data on the atmospheric-drag mechanism.
19. May 2024 The "Gannon Storm" of May 10–11 produces auroras at unusually low latitudes (Mexico, Italy, parts of India). The largest geomagnetic storm since 2003; substantially smaller than Carrington but the largest direct experience of modern grids with a strong storm.

Cases on this archive that connect.

The Tunguska Event (File 016) — a different atmospheric extreme; the largest airburst in recorded history paired here with the largest geomagnetic storm. The shared theme is the inadequacy of the small-sample historical record for predicting the recurrence of extreme natural events.

The Wow! Signal (File 036) — another case where a single observed event sits at the edge of the instrumental capacity that recorded it and the central scientific question is "how unusual is it, given that we only have one example?"

More related files coming. Planned: the 774–775 CE Miyake event (a probable extreme solar-particle event preserved in the cosmogenic-isotope record); the 1989 Quebec storm as its own file; the 2024 Gannon storm as a contemporary analog.

Full bibliography.

1. Carrington, R. C., "Description of a Singular Appearance seen in the Sun on September 1, 1859," Monthly Notices of the Royal Astronomical Society, Vol. 20, November 1859.
2. Hodgson, R., "On a curious Appearance seen in the Sun," Monthly Notices of the Royal Astronomical Society, Vol. 20, November 1859.
3. Tsurutani, B. T., Gonzalez, W. D., Lakhina, G. S., and Alex, S., "The extreme magnetic storm of 1-2 September 1859," Journal of Geophysical Research: Space Physics, Vol. 108(A7), 2003. Includes the reconstruction of Dst from the Colaba trace.
4. "The Aurora Borealis — Influence on the Telegraph Lines," The New York Times, September 5, 1859. The principal contemporary account of the Boston-Portland battery-free exchange.
5. Loomis, E., "The Great Auroral Exhibition of August 28th to September 4th, 1859," American Journal of Science, Vol. 28, 1859. The principal contemporary American synthesis.
6. Green, J. L., and Boardsen, S., "Duration and extent of the great auroral storm of 1859," Advances in Space Research, Vol. 38(2), 2006.
7. Stewart, B., "On the great magnetic disturbance which extended from August 28 to September 7, 1859, as recorded by photography at the Kew Observatory," Philosophical Transactions of the Royal Society, Vol. 151, 1861.
8. McCracken, K. G., Dreschhoff, G. A. M., Zeller, E. J., Smart, D. F., and Shea, M. A., "Solar cosmic ray events for the period 1561–1994: 1. Identification in polar ice, 1561–1950," Journal of Geophysical Research, Vol. 106(A10), 2001. The initial ice-core SPE-signature paper.
9. Wolff, E. W., Bigler, M., Curran, M. A. J., et al., "The Carrington event not observed in most ice core nitrate records," Geophysical Research Letters, Vol. 39, 2012.
10. Bolduc, L., "GIC observations and studies in the Hydro-Québec power system," Journal of Atmospheric and Solar-Terrestrial Physics, Vol. 64, 2002. The technical post-mortem on the 1989 cascade.
11. Baker, D. N., Li, X., Pulkkinen, A., et al., "A major solar eruptive event in July 2012: Defining extreme space weather scenarios," Space Weather, Vol. 11, 2013.
12. Lloyd's of London and Atmospheric and Environmental Research (AER), Solar Storm Risk to the North American Electric Grid, 2013.
13. National Research Council, Severe Space Weather Events—Understanding Societal and Economic Impacts: A Workshop Report, The National Academies Press, 2008. (Final report, 2009.)
14. Promoting Research and Observations of Space Weather to Improve the Forecasting of Tomorrow Act (PROSWIFT), Public Law 116-181, October 21, 2020.
15. Cliver, E. W., and Dietrich, W. F., "The 1859 space weather event revisited: limits of extreme activity," Journal of Space Weather and Space Climate, Vol. 3, 2013.