The study, published in Frontiers in Astronomy and Space Sciences, dissects every negative excursion of the SYM-H index, a one-minute-resolution measure of the low-latitude magnetic disturbance, between January 1981 and May 2026. That record spans nearly five complete solar cycles and contains 78,096 separate disturbance events. Rather than measuring each excursion from threshold crossing to recovery as one undivided duration, the researcher split every event at its minimum: the growth lifetime runs from the downward crossing of a threshold to the deepest point of the disturbance, while the recovery lifetime runs from that minimum back up to the recovery crossing.

The split matters because the two halves of a storm are governed by different physics. Growth is driven from outside, as enhanced solar-wind convection injects energetic ions into the ring current and depresses the equatorial magnetic field. Recovery is controlled from inside, as charge exchange, Coulomb collisions, wave-particle scattering, and dayside outflow slowly drain the trapped population. The minimum of the excursion is the single instant at which accumulation gives way to loss, making it the natural dividing line between the two regimes.

When the two duration ensembles were fitted with a finite-range power law — a distribution whose slope is set by an exponent and whose reach is set by a temporal cutoff — the two channels turned out to be strikingly different. Growth lifetimes follow an exponent of 1.079 with a cutoff near 669 minutes, while recovery lifetimes follow a steeper exponent of 1.305 with a cutoff near 3,328 minutes, roughly five times longer. The separation between the two exponents, about 0.226, holds at every threshold tested, from quiet-time ripples at the median disturbance level all the way up to superstorm amplitudes near minus 300 nanotesla.

The asymmetry is not subtle. In the archetypal superstorm of September 1989, the field spent about 10 hours descending to its minimum and more than 63 hours climbing back — a factor of six in a single event. Across all 89 one-year windows in the record, the recovery exponent exceeded the growth exponent in every single window, a consistency that survives even when the analysis is restricted to the 45 non-overlapping windows that share no data.

The most surprising result emerges when the exponents are tracked against solar activity. The growth exponent rises significantly with sunspot number, with a correlation coefficient of about +0.56 in independent windows and a shift of roughly +0.10 from solar minimum to solar maximum. The recovery exponent, by contrast, is statistically flat across all five cycles, with any potential solar-cycle modulation bounded to roughly a tenth of the growth response. The solar cycle therefore imprints itself selectively on disturbance timing: it touches the growth slope and the temporal cutoffs of both channels, but leaves the recovery exponent untouched.

This decoupling resolves a long-standing puzzle. Earlier work had shown that the undivided burst lifetime of SYM-H follows a single power law that is robust to the choice of threshold, which seemed to suggest a unified underlying process. The new analysis shows that this apparent invariance is an emergent property. Because the recovery cutoff is roughly five times the growth cutoff, the tail of the combined burst distribution is dominated by configurations in which nearly the entire burst is recovery. The burst exponent is therefore pinned by the recovery channel, whose exponent is threshold-stable and cycle-flat, while the body of the distribution carries the growth signature. A single power-law fit to the composite returns an effective exponent that falls between the two channel values, reproducing the earlier result without requiring a single physical process.

Peeling back another layer, the recovery channel itself decomposes into a dwell — the contiguous run of samples tied at the minimum value — and a climb back to threshold. The dwell is steep, short-scaled, and flat across the solar cycle, while the climb transit carries a cycle-responsive exponent that mirrors the growth side. The composite recovery exponent is a mixture of these two sub-processes, and its flatness is a consequence of the standard storm-timing convention that assigns the cycle-flat dwell to the recovery channel.

The findings carry practical weight for space weather forecasting. Recovery-timing priors can be treated as cycle-independent provided their range is conditioned on solar phase, while growth-timing priors must be conditioned on cycle phase throughout. For modelers, the decoupling sets a falsifiable target: any stochastic description of SYM-H excursions must now produce two distinct exponents from one time series, route the solar-cycle imprint through the growth channel alone, and reproduce the symmetric cycle response of the underlying transit statistics.

The analysis applies specifically to low-to-mid-latitude disturbances, where the SYM-H index is most representative. Higher-latitude geomagnetic activity is known to follow different and more variable scaling behavior, shaped by complicated ionospheric effects that dominate closer to the poles. Extending the two-channel decomposition to those regions remains an open challenge for future work.

Subject of Research: Statistical scaling of geomagnetic storm growth and recovery timing across five solar cycles

Article Title: Growth–recovery decoupling in the timing of geomagnetic disturbances across five solar cycles

Article References: Wanliss, J. A. (2026). Growth–recovery decoupling in the timing of geomagnetic disturbances across five solar cycles. Frontiers in Astronomy and Space Sciences, 13, Article 1948617. https://doi.org/10.3389/fspas.2026.1948617

Image Credits: AI Generated

DOI: 10.3389/fspas.2026.1948617

Keywords: geomagnetic storms, SYM-H index, power-law distributions, solar cycle, ring current, space weather, magnetosphere, burst-lifetime statistics, Dst index, solar wind, Growth, recovery

Surface flux transport, or SFT, models are among the most important tools solar physicists use to monitor and forecast the behavior of the Sun. These models track how magnetic flux emerges on the solar surface and how it is subsequently carried around by large-scale flows, including the differential rotation of the Sun and the meridional circulation that transports magnetic material toward the poles. By simulating the evolution of the surface magnetic field over weeks, months, and years, SFT models allow researchers to understand how solar activity builds up and decays across the roughly eleven-year solar cycle. That understanding, in turn, feeds into predictions of sunspots, flares, and coronal mass ejections, the explosive events that can send bursts of radiation and charged particles toward Earth and disrupt communications, navigation systems, and electrical infrastructure.

The central challenge the SwRI-led team set out to solve is a fundamental observational one: at any given moment, half of the Sun is facing away from Earth. Unlike the Moon, which is tidally locked and shows us the same face forever, the Sun rotates completely on its axis roughly every 27 days, so every patch of the solar surface eventually swings into view. But between those appearances, active regions can form, evolve, and decay entirely out of sight. Earlier SFT models, lacking simultaneous data from the hidden hemisphere, were often caught off guard. In some cases, as the Sun rotated, a sunspot would emerge unexpectedly from the far side, forcing researchers into quick recalculations of their magnetic maps and degrading the reliability of forecasts built on them.

To fill this observational gap, scientists have since the 1970s turned to helioseismology, a technique that studies the Sun’s interior and its far side using acoustic waves. Just as geologists probe what lies deep beneath the Earth’s crust by analyzing how seismic waves travel through the planet, solar physicists analyze sound waves reverberating through the Sun’s interior. Active regions on the far side perturb these acoustic signals in characteristic ways, allowing researchers to infer the presence of magnetically active areas before they rotate into view. Helioseismic far-side imaging has led to vast improvements in space weather forecasting, giving forecasters days of advance warning about emerging sunspot groups that would otherwise appear without notice.

Yet even helioseismic models have been hampered by bad information. The acoustic signatures that betray an active region are not always what they seem. Dr. Bibhuti Jha, a research scientist in SwRI’s Solar System Science and Exploration Division in Boulder, Colorado, and the lead author of the new paper, described the problem in vivid terms. Acoustic data can sometimes indicate an active region on the far side, he explained, but when the area finally comes into view, nothing is there. It is almost like a false positive, and the team calls these phantom features ghost active regions. When such false detections are folded into an SFT model, they inject spurious magnetic flux that then propagates through the simulation, contaminating the map of the solar surface and any forecasts derived from it.

The new framework tackles this problem directly by introducing a probabilistic treatment of far-side detections. Rather than accepting every helioseismic detection at face value, the model systematically evaluates the accuracy of each candidate active region before it is incorporated into the surface flux transport simulation. Dr. Lisa Upton, an SwRI lead scientist also based in the Boulder office and a co-author of the paper, explained the approach: before far-side helioseismology data are added to the model, the team associates a probability with each far-side active region based on its properties. Detections with characteristics that closely match confirmed active regions earn high confidence, while ambiguous or weak signals are weighted accordingly or filtered out. The result is a model that weeds out ghost active regions and delivers a more accurate map of the entire solar surface at any given time.

To build and validate this framework, the team studied near-simultaneous maps of the whole solar surface from 2010 to 2014, a period that spans the rise to the maximum of Solar Cycle 24. During this window, observations from Earth-based instruments could be combined with data from space-based assets positioned to view the opposite hemisphere, providing ground truth against which far-side helioseismic detections could be checked. By comparing what the acoustic methods predicted with what was actually observed when each region rotated into view, the researchers could quantify the reliability of far-side detections and calibrate their probabilistic assignments. The resulting SFT models are more complete and accurate than any previous effort, according to the team.

The significance of this work extends well beyond the technical details of magnetic cartography. Solar activity drives space weather, and space weather is increasingly recognized as a practical hazard with economic and security implications. Strong solar flares can degrade high-frequency radio communications used by aviation and maritime operations, while coronal mass ejections can induce currents in long power lines and damage transformers, as demonstrated by historic geomagnetic storms. Satellites in Earth orbit face elevated radiation risks during active periods, and crewed missions beyond low Earth orbit require accurate forecasts to protect astronauts from energetic particle events. A more faithful representation of the Sun’s global magnetic field, including the half that cannot be seen directly, gives forecasters a firmer foundation for anticipating when and where the next eruption might occur.

The probabilistic framework also provides a practical pathway for incorporating helioseismic data into SFT models more broadly, something that has been difficult to do rigorously in the past because of the noise and false positives inherent in far-side imaging. By treating each detection as a probability rather than a certainty, the method allows models to benefit from far-side information without being misled by its errors, a balance that previous approaches struggled to strike. Jha characterized the advance as a step toward a more realistic modeling of the magnetic field of both the visible and invisible hemispheres of the Sun, a goal that has motivated solar physicists for decades.

Looking ahead, the team hopes that better surface maps will translate into better predictions of the solar cycle itself. As Jha noted, modeling the surface of the Sun to see how the magnetic field behaves helps scientists understand how solar activity varies over the cycle, and if the next solar cycle can be plotted out in advance, researchers will have a much better idea of how the Sun is going to behave in the coming years. The study, titled A Probabilistic Framework for Incorporating Helioseismic Far-Side Active Regions into SFT Models, was published on October 8, 2026, in The Astrophysical Journal Supplement Series, and it arrives at a moment of growing public and governmental attention to space weather readiness. As society’s dependence on technology vulnerable to solar storms continues to grow, tools that make the Sun’s hidden half visible, even probabilistically, represent a meaningful stride toward a more predictable star.

Subject of Research: Probabilistic incorporation of helioseismic far-side active region data into solar surface flux transport models

Article Title: SwRI-led team develops improved model for mapping the Sun

Article References: SwRI-led team develops improved model for mapping the Sun. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: Sun, solar magnetic field, surface flux transport, helioseismology, far-side imaging, space weather, solar cycle, sunspots, SwRI, solar forecasting, active regions, The Astrophysical Journal Supplement Series

Auroras occur when charged particles from the Sun, funneled along Earth’s magnetic field lines, collide with atoms and molecules in the upper atmosphere. Oxygen atoms glow green and red at altitudes between roughly 100 and 500 kilometers, while nitrogen contributes blues and purples. These displays are most common inside the auroral ovals that encircle the magnetic poles, but during powerful geomagnetic storms, when coronal mass ejections slam into Earth’s magnetosphere, the ovals expand equatorward and the lights become visible at middle and even low latitudes. Because the frequency and intensity of such storms follow the roughly 11-year solar cycle, historical aurora sightings serve as a proxy record of solar behavior, complementing the sunspot number, the longest-running direct dataset of solar observations.

The Extremadura newspaper, founded in 1923, is one of the oldest and most recognized media outlets in its region, and it was never intended as a scientific instrument. Yet its complete run, digitized in PDF form with help from the newspaper’s staff, allowed the team to run optical character recognition searches using more than a dozen geophysically relevant keywords, including the Spanish term for the phenomenon, aurora, along with related words such as phenomeno and the less common phrase for northern lights. The same systematic approach had previously yielded records of a super-bolide, a meteorite fall, and an exceptional month of thunderstorms, demonstrating that regional newspapers can be mined as surprisingly rich documentary archives.

The recovered articles fall into two categories: 12 items describing direct observations and 19 offering general reports and scientific explanations. The direct sightings cluster around the great geomagnetic storms of the 20th century, with notable events in 1926, 1938, 1950, 1956, 1957, 1958, and 1991. Crucially, when the researchers plotted these dates against the smoothed sunspot number, the auroras lined up with the maxima of solar cycles 16 through 19 and cycle 22. That correlation matters, because no systematic catalog of auroras observed in the Iberian Peninsula during the 20th century previously existed; the newspaper collection now fills that gap.

The earliest event, the geomagnetic storm of January 1926, generated two news items. One briefly reported a brilliant aurora seen for several hours over the United States, while a longer piece published that July described a formidable magnetic storm that disrupted telegraph, telephone, and submarine cable communications for hours. The article recounted how the Norwegian physicist Carl Störmer measured auroral rays reaching 72 degrees of elevation in the Oslo sky, implying altitudes of about 503 kilometers. It described the spectacle unfolding from a yellow-greenish arc in the north into an immense corona of dazzling red rays, visible from Sweden and Denmark to France, and noted that observatories in Meudon and Ebro recorded the accompanying magnetic disturbance. The sunspot group responsible was estimated to span roughly 100,000 kilometers.

Perhaps the most dramatic event in the collection is the great aurora of 25 to 26 January 1938, remembered as the Fatima Storm because some Roman Catholics linked it to the Fatima prophesies. The newspaper reported that the Astronomical Observatory of San Fernando confirmed an aurora of varying intensity recorded from 23:00 until 02:00, with maxima at 21:00 and around 23:30, and that magnetic instruments in the city had shown alterations for days beforehand. Dispatches from Berlin, Rome, and Warsaw described the display stretching from Scandinavia to the Mediterranean, with multicolored lines on a dark red background. Modern estimates place the storm’s intensity at roughly minus 336 nanotesla on the Dcx index, with a maximum Kp index of 9. The storm cut transatlantic radio communications and even delayed trains in England whose signaling apparatus malfunctioned. In war-torn Spain, citizens and soldiers unfamiliar with the phenomenon interpreted the blood-red sky as a presage of greater tragedy.

The 1950s, near the peak of the strongest solar cycle ever recorded, produced a flurry of sightings. In February 1950, the sky turned red over the Extremaduran town of Hervás, and the newspaper noted that some timorous residents considered it a harbinger of war. In January 1957, red glows were reported from Seville, Huelva, Ávila, Lugo, Monforte de Lemos, and Salamanca, lasting about 20 minutes in Seville, while in Lugo firefighters mobilized for hours believing the aurora was a large fire. That storm reached a minimum Dst index of minus 250 nanotesla. In February 1958, the crew of the German ship Beate Bolten sighted an aurora near the Panama Canal, the closest to the Equator recorded during the International Geophysical Year, and the newspaper noted that German captains had spotted about a dozen auroras around Spain during that period. A September 1958 storm, with a Dst minimum near minus 302 nanotesla, produced spectacular rayed bands and pulsating red surfaces across Germany and France.

The most recent major event in the archive is the great red aurora of November 1991, one of the largest geomagnetic storms on record with a Dst peak of minus 354 nanotesla. The newspaper described luminous phenomena in the Huesca Pyrenees, Lleida, and Zaragoza that residents repeatedly reported to the fire brigade as fires, a telltale signature of red auroral light at low latitudes. An astronomer at a local agrometeorological center explained that even trained observers initially mistook the glow for flames. The analysis shows that such misunderstandings are not anecdotal quirks but recurring social responses: in 1957, 1991, and likely 1950, emergency services or frightened citizens mistook auroras for conflagrations or omens. For social scientists, the archive offers a century-long record of how geomagnetic storms rippled through public opinion.

Not every sighting survived scrutiny. A 1952 item described concentric rainbow-colored arcs over Hervás, but the researchers judged it highly improbable to be an aurora, since concentric arcs are uncharacteristic of the dynamic, changing forms auroras display, rainbow hues are physically implausible at Extremadura’s latitude of 38 to 40 degrees where red is the expected color, and no geomagnetic storm is documented around that date. The general reports in the collection also trace the evolution of auroral science itself, covering explanations by Spanish meteorologists in the 1950s and 1970s, Soviet and French attempts to create artificial auroras in 1974, NASA’s THEMIS mission of 2007, and aurora-hunting expeditions to Greenland, Iceland, and Norway in the 2010s.

The broader significance of the work lies in what it says about reconstructing the Sun’s past. Space weather is not merely a curiosity; severe geomagnetic storms can damage power grids, satellites, and communication systems, and understanding their historical frequency is essential for assessing future risk. Because instrumental magnetic records extend back barely a century and a half, historians of the Sun rely on proxies such as aurora sightings, sunspot drawings, and even radiocarbon spikes in tree rings. The Extremadura study demonstrates that local newspapers from low-latitude regions, where auroras are rare precisely because only exceptional storms produce them, are especially valuable: every sighting they record is, by definition, evidence of an extraordinary solar event. A comparison with previous surveys of newspaper space-weather reporting suggests the coverage in Extremadura is entirely typical, meaning similar archives almost certainly await discovery in other regional papers worldwide, each yellowed page a potential data point in the long history of our stormy star.

Subject of Research: Historical aurora records from a Spanish newspaper as proxies for 20th-century solar activity

Article Title: Aurora records in the Spanish newspaper Extremadura for the period 1923–2017

Article References: Sánchez Romero, C., Díaz-Condiño, L., Tovar Hernández, I., Pérez Aparicio, A. J., Sánchez Carrasco, V. M., Gallego Herrezuelo, M. C., & Vaquero, J. M. (2025). Aurora records in the Spanish newspaper Extremadura for the period 1923–2017. History of Geo- and Space Sciences, 16(2), 13-22. https://doi.org/10.5194/hgss-16-13-2025

Image Credits: AI Generated

DOI: 10.5194/hgss-16-13-2025

Keywords: aurora borealis, geomagnetic storms, solar activity, sunspots, space weather, newspaper archives, Extremadura, Iberian Peninsula, solar cycle, history of science, International Geophysical Year, Aurora

The Ap index, together with its close cousin Kp, was introduced by Julius Bartels in 1949 and 1957 and has since become a standard measure of the strength of magnetospheric convection, the vast circulation of plasma and magnetic flux driven by magnetic reconnection on the day and night sides of the magnetosphere. This convection penetrates deep into the inner magnetosphere, shaping electric fields that influence everything from satellite orbits to the ring current that circles the Earth. Because Ap is derived from three-hourly measurements, it captures a remarkably wide dynamic range, from quiet conditions below 20 nanotesla to disturbed values exceeding 60 or even 70 nanotesla during major storms.

Previous studies had already hinted that the Ap record contains rich periodic structure. Analyses of monthly Ap power spectra going back to 1868 have revealed a period of roughly four years associated with a double-peaked structure in geomagnetic activity, daily spectra dominated by harmonics of a six-month period, and additional signatures tied to the 27-day rotation of the Sun. Other researchers, applying singular spectrum analysis to the Dst index, which tracks the ring current more directly, identified significant components at 6 months and at 47, 22, 11, 10.6, 5.5, and 1 year. What the new study adds is a moving, time-resolved picture: instead of computing a single spectrum over an entire record, the team computed Fourier coefficients continuously, window by window, so that the spectral content of Ap could be watched as it evolves in step with geomagnetic activity.

The method is elegant in its simplicity. The authors treated the Ap index as a function of time and decomposed it into a Fourier sum with a fundamental period of 27 days, corresponding to one solar rotation. Because Ap is reported every three hours, each 27-day window contains 216 data points, and the team computed up to 108 harmonics within each window. Rather than relying on a fast Fourier transform of raw samples, they used linear interpolation between the three-hourly values and evaluated the Fourier integrals analytically over each three-hour segment, effectively summing the areas of trapezoids weighted by sines and cosines. This approach, they note, allows a Fourier series reconstruction that typically outperforms a straightforward Fourier transform, and it yields both amplitudes and phases for every harmonic at every moment in time.

The reconstruction itself is strikingly successful. For February 2001, a relatively quiet month with a maximum Ap of about 40 nanotesla and no abrupt peaks, the original data and the Fourier reconstruction are practically indistinguishable for most of the month. February 2003, by contrast, was a violently disturbed interval, with Ap peaking near 70 nanotesla and rapid transitions from smooth behavior to abrupt jumps; even here the reconstruction tracks the data closely, with only minor deviations. February 2017, a quiet period near solar minimum, rounds out the sample, and again the reconstructed curve hugs the observations. The message is that even a jagged, three-hourly geophysical index can be captured almost perfectly by a modest set of harmonics whose amplitudes and phases drift smoothly in time.

Plotting the harmonic amplitudes as contour maps, with time on the horizontal axis and harmonic number on the vertical axis, revealed a clear hierarchy: the strongest power sits at low frequencies, and amplitudes fall steadily as frequency rises. Among the low-frequency structures, the team identified pronounced lines at harmonic numbers 27 and 54, corresponding to diurnal and semidiurnal periods, the unmistakable fingerprints of atmospheric tides. These tidal signatures were especially sharp in the quiet February 2017 data and in February 2003, but notably weaker in the more active February 2001. The interpretation is intriguing: geomagnetic activity is strong enough to bury the magnetic signatures of atmospheric tides and planetary waves within the Ap time series, so the very disappearance of tidal lines becomes an indicator of a disturbed magnetosphere.

The phases of the harmonics told an equally disciplined story. After making the phases continuous and normalizing them by harmonic number, the researchers found that they decrease nearly monotonically, and almost linearly, with time. During quiet February 2017, most of these normalized phase curves are essentially straight lines, while deviations from linearity appear precisely during intervals of heightened geomagnetic activity. When the team computed the root mean square of the normalized phases in several frequency bands, they found that during the solar-maximum years 2001 and 2003 the lowest band, spanning harmonics 1 through 19, started near 90 degrees, whereas in quiet 2017 it started near 0 degrees, a systematic difference that tracks the phase of the solar cycle.

By constructing a complex counterpart of the Ap index from the amplitudes and phases, the team also plotted hodographs, the looping paths traced by the tip of a rotating complex vector. These loops turned out to encode the level of disturbance in a visually intuitive way. February 2001, at solar maximum, produced large, widely spread loops reflecting intense storm activity; February 2003 yielded more compact trajectories with moderate disturbances; and quiet February 2017 drew fewer, tighter loops spread uniformly over the complex plane. The authors suggest that such concentration patterns near the origin during disturbed periods, versus uniform spreading during quiet times, may be a general property worth exploring further.

The central surprise came when the team tested a natural hypothesis. They expected that during disturbed periods, higher-frequency harmonics would intensify, mirroring the abrupt, jagged changes in the Ap index. To quantify this, they computed the ratio of the root mean square amplitude of harmonics 15 through 20 to that of harmonics 5 through 10 for every time window. No one-to-one correspondence with geomagnetic activity emerged, but one event stood out dramatically: the substorm of 13 to 14 February 2001, in which a Dst decrease was followed by recovery and Ap climbed above 20 nanotesla. During that event, the ratio dropped sharply, meaning the higher harmonics were suppressed relative to the lower ones, the exact opposite of the anticipated behavior.

The authors are careful about what this means. They had hoped that abrupt changes in Ap might be predictable from the behavior of the Fourier coefficients before the changes occur, but the analysis did not support that hope. The suppression of high frequencies during the 2001 substorm might be a general feature of geomagnetic storms, a property of only some class of substorms, or simply a coincidence, and the team argues that it deserves dedicated future analysis. If high-frequency suppression does turn out to be typical of most storms, they note, it would carry a profound implication: the auroral index Ap would be governed mainly by processes intrinsic to the magnetosphere rather than by direct solar forcing. Either way, the study demonstrates that a familiar, half-century-old index still conceals structure that standard analyses have missed, and that watching its harmonics evolve in time offers a new lens on the dynamics of near-Earth space.

Subject of Research: Fourier spectral analysis of the planetary geomagnetic index Ap during disturbed and quiet periods

Article Title: Harmonic content of Ap index

Article References: Vandas, M., Romashets, E., Imam, T., Hasan, T., Majumder, P., & Karki, S. (2026). Harmonic content of Ap index. Annales Geophysicae, 44(2), 689-696. https://doi.org/10.5194/angeo-44-689-2026

Image Credits: AI Generated

DOI: 10.5194/angeo-44-689-2026

Keywords: geomagnetic index, Ap index, Fourier analysis, space weather, magnetosphere, substorm, Dst index, solar cycle, atmospheric tides, geomagnetic storms, magnetospheric convection, spectral analysis

The archive, assembled by Jānis Kauliņš of the Institute of Astronomy at the University of Latvia and described in the journal History of Geo- and Space Sciences, documents systematic visual and photographic observations of mesospheric clouds carried out between 1957 and 1983. The campaign began during the International Geophysical Year of 1957, a landmark global effort to coordinate Earth science observations, and continued uninterrupted for more than a quarter of a century under the auspices of the Latvian branch of the All-Union Astronomical and Geodesic Society, later renamed the Latvian Astronomical Society. Because the observations followed a uniform, internationally recognized methodology throughout, they constitute a methodically consistent data series of unusual duration, something that exists for only a handful of locations on the planet.

The scientific stakes are considerable. Until the 1980s, very few instruments could reach or study the mesosphere, the atmospheric layer between roughly 50 and 85 kilometers, leaving ground-based cloud observations as one of the only practical windows into its state and dynamics. In recent decades, interest in noctilucent clouds has surged dramatically. The clouds have been appearing at latitudes where they were never seen before, and several studies have debated whether their increasing brightness, frequency, and shifting heights are connected to climate change. Rising greenhouse gases cool the mesosphere even as they warm the lower atmosphere, and a colder mesopause favors the formation of the ice crystals that make these clouds shine. Some researchers report a clear downward trend in the average height of the clouds’ lower boundary that has accelerated since the 1960s, while others analyzing cloud statistics and physical parameters find no significant long-term trends, arguing that solar-cycle effects can mask underlying changes. Long, homogeneous historical series like the Latvian one are precisely what is needed to settle such disputes.

The Latvian observations were conducted from four sites: Riga, Sigulda, Baldone, and Lielauce. The most enduring of these was Sigulda, where a dedicated observation pavilion, nicknamed the Cloud Hut, was built in 1956 on land granted to the astronomical society. Its opening faced north, the direction where noctilucent clouds appear against the dark twilight sky, and it housed an impressive array of equipment: aerial photo cameras mounted on a masonry pillar with a deep, stable foundation, a wide-angle theodolite, a geodetic theodolite, binoculars, marine and aviation chronometers, and a radio receiver for capturing precise time signals. Observations there ran from 1957 until 1983, managed throughout by Matīss Dīriķis, a scientific associate of the University of Latvia’s Astronomical Observatory. In Riga, observations began on the roof platform of a six-story residential building before moving to a tower of Riga Palace in 1958 and later to the grounds of the university observatory.

The observation logs themselves are fascinating documents of scientific practice under constraint. Observers recorded the visibility and typology of the clouds at the beginning of every quarter of an hour, with times noted to the minute, alongside brightness ratings, morphological classifications, twilight-segment visibility graded from completely clear to completely obscured, and total and low cloudiness on a zero-to-ten scale. Air temperature and pressure were added at Sigulda from 1964 onward. Timekeeping was a serious business: a marine chronometer made by J. Bruce and Sons around 1900 kept time at Sigulda, later joined by a similar Thomas Mercer instrument, with corrections checked against radio time signals. By 1981, a household quartz clock had entered service, and in 1983 an amateur-built quartz clock constructed by electronic engineer Valdis Gedrovics was used. One curious quirk of the archive is that observers were required to write only in pencil, a rule introduced during the International Geophysical Year that persisted for decades.

The photographic collection is the archive’s crown jewel. It contains 2,106 large-format negatives, most with frames of 180 by 130 millimeters and 84 at 180 by 240 millimeters, all remarkably well preserved. The workhorse camera was the AFA-IM, a semi-military aerial camera developed in 1938 for the Soviet armed forces, fitted with a four-lens anastigmatic Industar-51 lens of 210-millimeter focal length. The large frame size was no accident: it allowed photogrammetric processing, meaning cloud heights could be measured by triangulation from simultaneous exposures taken at different sites. Fifty-three pairs of synchronously captured photographs from Sigulda with Riga or Lielauce survive, and these could yield 53 independent measurements of cloud height for the 1961 to 1964 window, a dataset that could directly test whether the lower boundary of noctilucent clouds has been sinking as some climate models suggest. Exposure times were calibrated against the depth of the Sun below the horizon using an experimentally derived table, eliminating the subjective factor from brightness estimates.

Digitizing the archive was a painstaking exercise in historical rescue. Some logs were lost in the 1990s when the University of Latvia museum changed location and custodians several times, and all photo negatives taken after 1967 have not survived. Kauliņš and volunteer students converted the surviving journals into structured spreadsheets, with one file per year containing four sheets: an explanatory front page, a summary of observation nights and observers, the full night-by-night record of cloud detections and meteorological parameters, and a catalog of photographs. Early entries from 1957 to about 1960 were written in Russian and had to be translated, and the Cyrillic alphabet occasionally tripped up Latvian-speaking observers, producing ambiguities between visually similar letters that correspond to different twilight-segment grades. The digitized logs and appendices are now openly accessible through the University of Latvia’s e-resources repository.

The statistical picture emerging from the logs shows the mid-1960s and 1970s as periods of particularly intense observing activity, with hundreds of nights documented, while 1957 yielded only a single dubious sighting. The records also capture unexpected environmental detail: early Riga observers frequently noted smoke from factory chimneys drifting from the industrial district of the Port of Riga as a factor disturbing their observations, an incidental record of mid-century urban air pollution that, according to the author, has no modern equivalent in the city. Such contextual notes, preserved alongside the quantitative data, give the archive a richness that pure instrumental records often lack.

What happens next could matter for climate science. The Institute of Astronomy at the University of Latvia plans to scan the full negative archive with a professional high-performance scanner, the same class of equipment used to digitize the Schmidt telescope glass-plate library at Baldone Observatory, using a custom positioning device built at the university’s satellite laser ranging station. Beyond height measurements, analyses of cloud movement and morphology could illuminate the dynamics of the upper atmosphere, its relationship with the solar cycle, and its long-term changes. A preliminary statistical evaluation even hints at an apparent inverse relationship between cloud observations and solar activity as measured by the Wolf sunspot number, though the author cautions that recent observations cast doubt on this assumption and that further processing is needed. With so few comparably long noctilucent cloud series in existence, this rescued Latvian archive stands to become a benchmark for understanding how the edge of space has changed over six decades.

Subject of Research: Historical ground-based observations of mesospheric (noctilucent) clouds in Latvia from 1957 to 1983 and their digitization for climate and mesosphere research

Article Title: Observations of mesospheric clouds in Latvia in 1957–1983

Article References: Kauliņš, J. (2026). Observations of mesospheric clouds in Latvia in 1957–1983. History of Geo- and Space Sciences, 17(2), 113-122. https://doi.org/10.5194/hgss-17-113-2026

Image Credits: AI Generated

DOI: 10.5194/hgss-17-113-2026

Keywords: noctilucent clouds, mesosphere, Latvia, International Geophysical Year, climate change, photogrammetry, observation archives, aerial cameras, solar cycle, citizen science, upper atmosphere, data digitization

The technical foundation of the work is transfer learning, a technique that repurposes neural networks already trained on vast general image collections rather than building one from scratch. The team first experimented with GoogLeNet, a comparatively compact network of about seven million parameters, before settling on InceptionV3, a larger architecture with roughly 24 million parameters that performed most evenly across all four classes. The final classification layers of the pretrained network were replaced so the output became a probability vector over the four sky-condition categories. Training images were rescaled to the network’s required 299 by 299 pixel input and augmented with random horizontal and vertical flips and rotations of up to ten degrees, a standard practice that prevents the model from simply memorising its training examples. On a desktop computer equipped with an NVIDIA GeForce RTX 4090 graphics card, training the final network took about twelve minutes, and classifying the entire decade of images took just 14.2 hours, or roughly six milliseconds per image.

What makes the study unusual is not the network itself but the painstaking construction of its ground truth. Supervised learning lives or dies by the quality of its labelled examples, and auroral images are notoriously ambiguous. Faint aurora blends into thin cloud, moonlit clouds mimic red emission bands, and frost on the camera dome can masquerade as a celestial structure. Partamies labelled thousands of images by hand, then ran several iterative rounds in which an unfinished classifier proposed labels, random samples were re-checked by the human expert, and any image whose content could not be judged unambiguously was banished to an ‘Unclear’ category and discarded. Masking out the lowest ten degrees of elevation near the horizon, where mountains, instrument domes and metadata clutter the frame, proved essential for producing clean labels. The resulting training set of carefully vetted images, spanning every winter season from 2016 to 2025, is itself published as a reference dataset for future method comparisons.

Validation was equally rigorous. The team divided the labelled data into training, validation and independent test sets, and additionally drew 2,000 random images per class from the full classified archive to estimate real-world performance. ClearAurora and CloudyNoAurora images were recognised with 97 percent success, CloudyAurora with 93 percent, and CloudyNoAurora with 90 percent in testing. The most common confusions occurred near class boundaries, as expected: images with barely half-sky cloud cover, or aurora so faint it barely registers. A separate cross-check with the observatory’s older Nikon DSLR camera system, which overlapped with the current Sony setup in early 2016, showed that the classifier transfers to different hardware with match rates between 92 and 97 percent, suggesting the discrepancies stem from instrument properties such as white balance rather than changing sky conditions.

Because clouds dominate the dataset so thoroughly, the authors validated their cloud statistics against a completely independent instrument: a co-located cloud sensor operated with University College London, which infers sky clarity from the temperature difference between the sky and the sensor itself. Monthly values from the two methods correlated with a coefficient of 0.86, a strong agreement given how differently the two measurements are produced. This matters because the cloudiness statistics are not a side product but a central scientific result. Svalbard sits under some of the cloudiest winter skies in the Arctic, and previous long-term meteorological studies had reported winter cloudiness around 60 percent with an increasing trend between 1981 and 2010. The new image-based statistics confirm the high cloud fraction and add, for the first time, a decade of consistent nighttime cloud measurements from a location where the Sun does not rise for months.

The auroral findings are equally consequential. Auroral occurrence over Svalbard averages about 25 percent of imaging time and shows no correlation with the solar cycle, despite the dataset spanning the rise from solar minimum toward the recent maximum. In 2019, a year of very low sunspot activity, auroral occurrence peaked at 33 percent; in 2023, a year of higher solar activity, it fell to just 16 percent. The controlling variable in both cases was cloudiness, which was exceptionally low in 2019 and exceptionally high, at 78 percent, in 2023. The result echoes earlier Finnish studies based on visual inspection of all-sky film archives, which likewise found no obvious link between auroral occurrence and solar cycle evolution at auroral oval latitudes, even though the complexity and intensity of auroral structures do vary with geomagnetic activity.

The diurnal pattern is equally distinctive. Expressed in Magnetic Local Time, auroral occurrence over Svalbard peaks in the late morning hours around 07:00 to 11:00 MLT, with a secondary maximum in the pre-midnight sector around 17:00 to 21:00 MLT, and minima in the early afternoon and around magnetic midnight. This double-peaked distribution held steady across all ten years. January and December emerge as the months with the highest auroral occurrence and the clearest skies, while November is decisively the cloudiest month, making it the worst time to hunt for the lights over Svalbard. The observatory’s unique position near the poleward boundary of the auroral oval, combined with the polar night during which optical observations are possible around the clock, means the dataset also captures the dayside aurora, a phenomenon largely inaccessible at lower latitudes.

The practical payoff extends well beyond statistics. KHOnet2026 now runs in real time at the Kjell Henriksen Observatory, classifying each new image as it arrives and continuously expanding the labelled archive. For researchers studying the fine structural evolution of aurora, the ability to prune away the roughly two-thirds of images obscured by cloud saves enormous amounts of computing time and human attention. The classified sky conditions also serve as ready-made metadata for the dozens of other optical instruments hosted at the observatory, and the occurrence climatology will help planners of sounding rocket campaigns and EISCAT radar experiments schedule observations when skies are statistically most likely to cooperate. The team notes that future refinements could employ self-supervised learning to expand the labelled set, or Vision Transformers as an alternative architecture, and that adding labelled images from the older DSLR archive could extend the classified time series by another seven years.

There are honest caveats. All manual labelling was performed by a single expert, introducing an unavoidable subjective element into the class boundaries, and twilight and moonlit conditions, which the authors deliberately retained in all classes, add a layer of ambiguity that a human cannot always resolve either. Tests excluding twilight and bright-moon images shifted the occurrence statistics by several percentage points but preserved the overall diurnal behaviour, and the published data files include solar and lunar elevation angles so that future users can filter the dataset to their own standards. Even so, the study stands as a demonstration of how machine learning, applied with unusual care to its weakest link, the training labels, can transform a decade of raw camera data into a coherent scientific record. For a phenomenon as famously unpredictable as the aurora, the message from Svalbard is unexpectedly mundane: above all, it is the weather that decides what you see.

Subject of Research: Automatic classification of auroral and cloud occurrence in a decade of all-sky camera images from Svalbard using convolutional neural networks

Article Title: High-latitude auroral and cloudiness occurrence from automatic image classification

Article References: Partamies, N., & Syrjäsuo, M. (2026). High-latitude auroral and cloudiness occurrence from automatic image classification. Annales Geophysicae, 44(2), 1003-1022. https://doi.org/10.5194/angeo-44-1003-2026

Image Credits: AI Generated

DOI: 10.5194/angeo-44-1003-2026

Keywords: aurora, all-sky camera, Svalbard, machine learning, convolutional neural network, cloud occurrence, auroral occurrence, transfer learning, polar night, solar cycle, Kjell Henriksen Observatory, space weather

The study, published in the journal Solar Physics, was led by Partha Chowdhury of the University of Calcutta together with Jagdev Singh, V. Muthu Priyal, and Belur Ravindra of the Indian Institute of Astrophysics in Bengaluru. The researchers analyzed digitized images taken in the light of singly ionized calcium, the Ca ii K spectral line, from the Kodaikanal Solar Observatory. These images capture the chromosphere, a thin layer of the solar atmosphere sitting just above the visible photosphere, where magnetic fields heat the gas and make bright regions known as plages blaze in ultraviolet-adjacent wavelengths. Plages are the chromospheric ghosts of sunspots: where magnetic flux crowds the surface, the calcium emission brightens, even when no dark spot is present.

Because plages are visible even when sunspots are not, they offer a more complete census of the Sun’s magnetic activity than sunspot counts alone. The Kodaikanal archive, spanning Solar Cycles 14 through 22, roughly the first nine decades of the twentieth century, is one of the longest continuous records of chromospheric behavior anywhere in the world. After careful digitization and homogenization, a process that corrects for changes in photographic plates, instruments, and observing conditions over the decades, the team measured the total area covered by plages on every available day and compared it with three standard yardsticks of solar activity: the international sunspot number, the total sunspot area, and the 10.7 centimeter radio flux, a microwave emission that tracks coronal magnetic heating.

The correlations the team found are strikingly strong. Across all nine cycles, the chromospheric plage area tracks each of the photospheric and coronal activity indices with correlation coefficients exceeding 0.85, a level of agreement that leaves little doubt that the Sun’s upper and lower atmospheric layers are driven by the same underlying magnetic engine. This matters for more than just bookkeeping. Because plage records extend further back in time than some modern instruments, they can serve as a reliable proxy for reconstructing solar activity across the twentieth century, including the ultraviolet output that subtly influences Earth’s upper atmosphere and climate system.

But the study went beyond simple correlations. The researchers examined whether the plage areas obey two well-known but poorly understood features of the solar cycle. The first is the Gnevyshev-Ohl rule, an empirical oddity noting that in most pairs of consecutive solar cycles, the odd-numbered cycle is stronger than the even-numbered one that precedes it. The team confirmed that plage areas follow this rule, suggesting that the chromosphere inherits the same cycle-to-cycle memory that governs sunspot production. The second feature is the Gnevyshev gap, a temporary dip in activity that often splits the peak of a solar cycle into two humps. The analysis revealed clear Gnevyshev gaps in the plage data during several cycle maxima, confirming that this double-peaked structure is not merely a sunspot quirk but a genuine, whole-Sun phenomenon imprinted on the chromosphere as well.

To dig deeper into the timing of these variations, the team turned to a powerful mathematical technique called Morlet wavelet analysis, which can detect oscillations whose strength and period change over time, something a traditional Fourier analysis cannot do. They supplemented this with wavelet coherence, a method that measures not just whether two signals share a rhythm, but whether those rhythms stay in step with each other, drifting in and out of phase like two musicians who occasionally lose the beat. The technique is widely used in geophysics and climate science, and it is ideally suited to the Sun, whose magnetic activity is famously irregular.

The wavelet analysis uncovered two families of intermediate-term periodicities hiding inside the dominant eleven-year cycle. The first are Rieger-type periods, oscillations of roughly 130 to 190 days. These were first discovered in 1984, when researchers noticed that gamma-ray flares from the Sun seemed to cluster with a period of about 154 days. Since then, similar periodicities have been found in sunspots, flares, and coronal mass ejections, and they are widely interpreted as the signature of magnetic Rossby waves, vast, planet-scale waves of magnetized plasma rolling around the Sun’s interior shear layer known as the tachocline, where the Sun’s differential rotation winds up its magnetic field. The second family comprises quasi-biennial oscillations, or QBOs, with periods of one to four years, which are thought to reflect a secondary magnetic cycle operating in parallel with the main eleven-year one.

Here the story takes an intriguing turn. The Rieger-type periods and QBOs show considerable variability from one solar cycle to the next, and the wavelet coherence analysis reveals that at these intermediate timescales, the chromospheric plage areas and the photospheric activity indices frequently fall out of phase with one another. The rhythms exist, but they do not always march together. This asynchrony suggests that the shorter-period variations may arise from processes that affect the chromosphere and photosphere differently, or from instabilities in the tachocline whose surface manifestations depend on the details of each individual cycle, including the strength of the interior magnetic field at the time.

At the fundamental scale, however, the picture is one of remarkable unity. Within the 9 to 12 year periodicity belt that defines the solar cycle itself, the team found a stable phase synchrony between chromospheric plages and every photospheric and coronal index they examined. In other words, over the long haul, the Sun’s chromosphere and its visible surface rise and fall together like two ends of the same magnetic tide. This tight decadal coupling is exactly what modern solar dynamo models predict: magnetic fields generated by the rotation of plasma at the tachocline buoyantly rise through the convection zone, emerge as sunspots and active regions in the photosphere, and simultaneously light up the overlying chromosphere as plages. The new results provide hard empirical constraints that any credible dynamo model must now reproduce.

The findings also align with theoretical work on tachocline instabilities and magnetic Rossby waves, which have been invoked to explain everything from the double-peaked shape of solar maxima to the so-called seasons of space weather, bursts of intense flaring that come and go over months. By confirming Rieger-type and quasi-biennial signals in a chromospheric dataset spanning nine cycles, the study strengthens the case that these waves are a persistent feature of the solar interior rather than a fluke of a few well-observed cycles. For space weather forecasters, the implications are tantalizing: if the phase relationships between these periodicities and the activity indices can be pinned down well enough, the hidden rhythms of the tachocline might one day help anticipate the timing of the most active and hazardous phases of the solar cycle. For now, the century-old glass plates of Kodaikanal have once again proven that some of the best windows into the Sun’s deepest secrets were captured long before the space age began.

Subject of Research: Long-term chromospheric plage evolution and its relationship to solar activity indices across Solar Cycles 14 to 22

Article Title: Long-term Evolution of Chromospheric Plage Areas from Kodaikanal Observatory Ca ii K Images and Their Relation to Solar Activity Indices

Article References: Chowdhury, P., Singh, J., Priyal, V. M., & Ravindra, B. (2026). Long-term Evolution of Chromospheric Plage Areas from Kodaikanal Observatory Ca ii K Images and Their Relation to Solar Activity Indices. Solar Physics, 301(10), Article 149. https://doi.org/10.1007/s11207-026-02733-y

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02733-y

Keywords: solar physics, chromosphere, Ca ii K, solar plages, solar cycle, sunspots, Rieger periodicity, quasi-biennial oscillations, solar dynamo, tachocline, Rossby waves, Kodaikanal Observatory

The stakes of this archival detective work are higher than they might appear. French solar observations form the primary data source covering the Maunder Minimum, the period from roughly 1645 to 1715 when sunspots became vanishingly rare and Europe shivered through what climatologists call the Little Ice Age. In the 1990s, the astronomer Elisabeth Nesme-Ribes extensively analyzed the Parisian Observatory archives and published a series of influential reconstructions. Tragically, the solar data she compiled have since been lost; only a single figure of average sunspot latitudes survives. That loss made a fresh, independent reduction of the original documents not merely desirable but essential, and the new study delivers exactly that, digitizing measurements that had previously existed only in ink on paper.

The technical challenge of converting seventeenth- and eighteenth-century observations into modern solar coordinates is formidable. La Hire did not simply note that a spot existed; he measured the times at which spots crossed the solar meridian, recorded angular distances from the disk’s limbs, and drew the spots in engravings that captured their shapes and relative positions. To transform these numbers into latitudes and longitudes, the researchers needed the angular size of the solar disk on every observing day, which changes through the year as Earth moves along its elliptical orbit. They interpolated routine meridian measurements published posthumously by Le Monnier in 1741, recovering disk diameters to within a few arcseconds for dates spanning the entire eight-year window. Combining transit timings with these disk sizes, they traced each sunspot’s path across the visible hemisphere and computed daily heliographic coordinates for every individual spot and group.

The reconstruction demanded careful correction of errors scattered through the historical record. The researchers found typos in dates, mislabeled authorship, and arithmetic slips in both the journals and the printed Mémoires. On 31 May 1703, for example, La Hire made an error in a sunspot transit measurement that could only be resolved by consulting his own journal. On 26 June of that year, an engraving appears to have been accidentally reversed and flipped compared with the other drawings, a conclusion supported by the relative positions of the umbrae and by agreement with a drawing by Giovanni Domenico Cassini. On 2 January 1707, a measurement taken seven minutes after noon required a positional correction of about 3.6 degrees toward the solar Equator. Each such fix, the authors note, matters when the goal is a machine-readable catalogue that modern solar physicists can trust.

The Parisian records also proved decisive in resolving puzzles among contemporary observers across Europe. The reconstructed positions resolved an inconsistency in the observations of Johannes Hoffmann and showed that the observer Müller was using a different type of telescope, a finding that changes how his measurements should be interpreted. The team cross-checked La Hire’s data against records from the Kirch family, Hoffmann, Eimmart, Müller, Becker, Hertel, Sturm, Wideburg, Manfredi, and Derham, weaving a continental network of observations into a coherent whole. In one striking episode from January 1704, Giacomo Filippo Maraldi reported two widely separated sunspot groups on the same day, remarking that it had been a long time since such widely separated spots had been seen together, a vivid reminder of how unusual even modest solar activity had become during those years.

Among the technical fruits of the study is a new estimate of the Sun’s sidereal differential rotation during this quiet epoch. For a synodic rotation period of 27.5 days, the sidereal rotation rate works out to 14.08 degrees per day, slightly slower than the 14.18 degrees per day of the modern Carrington frame that serves as the standard reference for solar rotation today. The researchers derived rotation rates from pairs of measurements separated by hours, days, and full disk passages, tabulating them for individual spots and groups. The persistence of Joy’s law, the well-known tendency of sunspot groups to tilt with their leading spots closer to the Equator than their trailing spots, was also verified in these ancient data, suggesting that the underlying dynamics of magnetic flux emergence were operating much as they do now, even at the tail end of a grand minimum.

The time-latitude diagram constructed from the new catalogue, a butterfly-like plot showing where spots appear as the solar cycle progresses, is consistent with the previously available results from the analyses of the Parisian observations, including Nesme-Ribes’s surviving figure of average latitudes. This agreement is reassuring on two fronts: it validates the lost reconstructions of the 1990s and confirms that the new, independently derived dataset can safely replace them. The researchers also documented the Wilson effect in an unexpected place, noting that on 3 June 1703 La Hire observed a sunspot at the very edge of the disk appearing as a small depression in the limb, a phenomenon now understood to arise because spots are shallow structures rather than flat markings on a perfect sphere.

The day-to-day texture of the observations brings the era to life. La Hire frequently worked with telescopes of 16 feet and longer, switching instruments depending on conditions; on 18 January 1704 he could not discern a spot with a 3-foot tube but saw it clearly with a 16-foot one, likely because of poor contrast near the limb. Clouds, fog, and haze repeatedly interrupted the observing campaigns, and the journals candidly record blank Suns, vanished spots, and rare double appearances. In October 1705, two sunspot groups were seen simultaneously, an event the observers flagged as a rarity, and in November 1707 Cassini marveled that a spot had appeared in the northern hemisphere at a latitude of about 13 degrees, an occurrence he called extraordinary. Such details, once mere color, now feed quantitative reconstructions of the solar cycle’s recovery from its deepest recorded lull.

All of the reconstructed sunspot parameters, including daily latitudes, longitudes, areas, and counts, are provided in electronic supplementary materials accompanying the paper, giving modelers of the solar dynamo a fresh, high-resolution window onto the Sun’s behavior as it emerged from the Maunder Minimum. Understanding how the cycle recovered from that grand minimum is not just an exercise in historical astronomy; it informs how scientists assess the Sun’s capacity for prolonged lulls in activity and what such lulls might mean for climate and space weather on timescales of centuries. La Hire, sketching spots with quill and micrometer in the fog above Paris, could not have imagined that his notebooks would one day anchor twenty-first-century dynamo simulations. Three hundred years on, his patience is paying scientific dividends once again.

Subject of Research: Reconstruction of sunspot positions, areas, and counts from Philippe de La Hire's observations at the Paris Observatory during 1703 to 1710, in the late Maunder Minimum.

Article Title: Sunspot Observations by Philippe de La Hire from 1703 to 1710

Article References: Zolotova, N., & Vokhmyanin, M. (2026). Sunspot Observations by Philippe de La Hire from 1703 to 1710. Solar Physics, 301(9), Article 133. https://doi.org/10.1007/s11207-026-02719-w

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02719-w

Keywords: sunspots, Maunder Minimum, solar cycle, solar physics, Philippe de La Hire, Paris Observatory, solar rotation, Joy's law, historical astronomy, sunspot records, solar dynamo, space climate

The study, conducted by Owen Giersch, analyzed X-ray flares of M class and higher—the moderate to extreme events that matter most for space weather operations. The research took two complementary approaches. First, flares were sorted into logarithmically spaced intensity categories and compared across sunspot cycles. Second, flare rates were tabulated according to the McIntosh classification of the sunspot group that produced each event, allowing a direct test of whether a given type of sunspot region flares at the same rate regardless of which cycle it belongs to. Both tests delivered results that depart from the comforting assumption of cycle-to-cycle consistency.

The McIntosh classification scheme, which underpins the second part of the analysis, has a long pedigree. Sunspot classification began with Alfred Cortie’s scheme of 1901, which is no longer used. Observers in Zurich developed an alternative system in the 1930s, using nine categories to describe the evolution of sunspot groups. In the 1960s, Patrick McIntosh modified this Zurich system, removing two classes and adding parameters describing the penumbra of the largest sunspot in a region and the compactness of spots within the region. The resulting Modified Zurich Classification—now universally known as the McIntosh system—combines the Zurich class, the penumbral character, and the compactness into 60 permitted classes. Because sunspot classification serves as a proxy for the magnetic complexity of an active region, it remains the backbone of most flare forecasting practice.

The data underpinning the analysis came from two archival sources. For flares prior to 1996, records were obtained from the National Centers for Environmental Information, and from 1996 onwards from the Space Weather Prediction Center, both drawing ultimately on the GOES satellite series that has monitored solar X-ray output continuously since cycle 21. Sunspot region reports came from United States Air Force Solar Observing Optical Network sites, whose daily observations include quality flags ranging from 1, indicating very poor seeing conditions, to 5, indicating excellent conditions. Because pre-1996 reports were listed from up to five individual sites without being reduced to a single classification, the author developed a method to consolidate multiple daily reports: retaining the highest-quality report, then adopting the most frequently occurring class, with random selection only in the rare case of irreconcilable ties.

The data cleaning itself illustrates the messy reality of half a century of observational records. Numerous invalid classifications appeared in the archive—classes with missing or implausible parameters—and a set of explicit rules was applied to correct them, drawing on the physical logic of the McIntosh scheme. After all corrections, only 193 regions could not be assigned a valid class, and these were omitted from the analysis. For the flare-rate comparison by sunspot class, only flares with an assigned region were included, while the raw cycle-to-cycle flare counts used all reported events. This careful curation matters, because the study’s conclusions hinge on comparisons between data streams that were assembled by different teams, with different instruments and practices, across 50 years.

The statistical machinery was deliberately conventional, designed to be transparent rather than exotic. To compare flare intensity distributions between two cycles, a chi-squared test was employed, with categories combined wherever either distribution contained fewer than five flares, and p-values computed so that distributions could be deemed similar at the 95 percent confidence level. To compare flare rates by McIntosh class, both chi-squared tests and linear regression were used. Under the regression approach, if two cycles behaved identically, plotting one cycle’s flare rates against the other’s should yield a line with a gradient of one, an intercept of zero, and a correlation coefficient close to one. Departures from that ideal indicate genuine differences in how sunspot classes translate into flares.

The results on intensity are nuanced but troubling. At the 95 percent confidence level, the flare intensity distribution of cycle 21 cannot be considered similar to those of cycles 22 through 25—a discrepancy the author attributes in part to the fact that cycle 21 was the first cycle during which the GOES satellites operated, raising the possibility of calibration or measurement issues. The comparisons between cycles 22 and 25 and between cycles 23 and 25 also failed the similarity test. A broad pattern emerged: the greater the separation between two cycles in time, the more dissimilar their flare intensity distributions tend to be. The striking exception is the pairing of cycles 22 and 24, which produced the highest p-value of any comparison, suggesting those two cycles flared in remarkably similar proportions despite differing in overall amplitude.

The second finding is more fundamental. When flare rates were compared by McIntosh class across cycles, chi-squared tests found essentially no pairs of cycles that could be considered similar. The sole nominal exception—a comparison of cycle 21 as the observed distribution against cycle 22 as the expected distribution—flipped to dissimilar when the roles were reversed, a signature of a statistical false positive. The regression analysis reinforced the picture: while the intercepts of seven out of ten cycle comparisons were close to zero, as expected for the least active sunspot classes, only the comparisons involving cycles 25 and 22 produced gradients close to one. In plain terms, the rate at which a given kind of sunspot group produces M-class or larger flares varies significantly from one solar cycle to another.

What explains this variability? The study is candid that the answer remains unclear. One possibility is observer bias: sunspot classification is a human judgment, and many researchers have documented problems with sunspot group analysis, particularly with sunspot areas, which form one input to the McIntosh parameters. If area measurement techniques—and therefore the resulting values—have drifted over decades, other derived parameters such as extent and compactness may have drifted too, contaminating any cycle-to-cycle comparison. The alternative is more provocative: that the difference reflects a genuine physical change in how the Sun organizes magnetic energy from cycle to cycle. Intriguingly, recent work by other researchers found that as the McIntosh Zurich class increases from A to F, the total magnetic energy of active regions rises, but the free energy—the portion available to power eruptions—does not change significantly, hinting that classification alone may not capture what makes a region flare-ready.

Despite the ambiguities, the analysis closes with something immediately useful for forecasters. Assuming flares follow a Poisson process, the probability of at least one M-class or larger flare occurring in a given region over a specified interval can be computed from its flare rate, and with multiple regions on the Sun—common near solar maximum—the combined probability follows from multiplying the individual exponential terms. The full dataset generated for the study has been released in a public GitHub repository, and the author emphasizes that further analysis of the USAF sunspot classifications is needed to determine whether the cycle-to-cycle variation is intrinsic to the Sun or an artifact of evolving observational practice. Either answer carries weight: one rewrites the physics of flare statistics, the other a cautionary tale about the archives on which space weather science is built.

Subject of Research: Statistical analysis of X-ray solar flare intensity distributions and flare rates by McIntosh sunspot class across solar cycles 21 to 25

Article Title: Analysis of X-ray Solar Flare Rates from 1976 to 2025

Article References: Giersch, O. (2026). Analysis of X-ray Solar Flare Rates from 1976 to 2025. Solar Physics, 301(9), Article 145. https://doi.org/10.1007/s11207-026-02743-w

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02743-w

Keywords: solar flares, X-ray flares, sunspots, McIntosh classification, solar cycle, space weather, GOES satellites, flare forecasting, solar physics, statistical analysis, sunspot classification, active regions

The chain of causality that makes such reconstructions possible begins far beyond Earth. The Sun’s magnetic field, dragged outward by the solar wind, fills the heliosphere, the vast protective bubble surrounding the solar system. Galactic cosmic rays, high-energy particles arriving from outside, must fight their way through this magnetic shielding to reach Earth. When the Sun is magnetically active, the shielding strengthens and fewer cosmic rays penetrate; when the Sun quiets, the floodgates loosen. Upon entering the atmosphere, cosmic rays collide with nitrogen, oxygen and argon, spawning showers of secondary particles that forge rare radionuclides such as carbon-14 and beryllium-10. Carbon-14, once formed, is incorporated into carbon dioxide, absorbed by trees during photosynthesis and permanently archived in annual growth rings. Because a magnetically active Sun suppresses cosmic ray influx, the abundance of these isotopes in wood and ice is inversely correlated with solar activity, giving scientists a decipherable record stretching back thousands of years.

Deciphering it, however, is harder than it sounds. Previous reconstruction efforts relied on statistical regressions that mapped isotope-derived quantities onto sunspot numbers, and those regressions carried hidden dangers. As the new study’s authors, led by Chitradeep Saha of the University of Reading, point out, even a regression with an impressively high correlation coefficient can misfire if the data contain non-linearities, zero-level offsets, or uneven variance across amplitudes. The most notorious failure mode appears during grand minima, the extended intervals when solar activity collapses, such as the Maunder minimum of the seventeenth century. When regression equations calibrated on modern data are extrapolated to these unusually quiet conditions, they can yield sunspot numbers below zero. Values are often simply clipped to zero, but that crude fix distorts cycle averages and inflates estimates of the Sun’s total irradiance, which in turn muddies attempts to understand the Sun’s influence on past climate.

The Reading-led team, which also includes Mathew Owens, Mike Lockwood and Luke Barnard, together with colleagues at ETH Zurich, the University of Lancashire, the University of Oslo, Lund University and the University of Groningen, took a different path. Rather than inverting the physics with a regression, they ran it forward, over and over, in a Monte Carlo framework. The method, implemented in publicly released code under the name PRISM, begins by generating an ensemble of thousands of hypothetical sunspot cycles drawn from statistical priors: cycle amplitudes sampled from a log-normal distribution, cycle lengths from a Gaussian distribution centred on 10.5 years, and a random start offset for each window of time. Each trial cycle is then passed through a sequence of two semi-empirical forward models that translate sunspot number into open solar flux, the total magnetic flux threading the outer boundary of the corona, and then into the heliospheric modulation potential, a quantity describing how much energy cosmic rays lose as they traverse the heliosphere.

The forward models rest on decades of established solar physics. The first couples sunspot number to the emergence of new magnetic flux through a empirically optimised source function, balanced against a phase-dependent loss rate derived from solar cycles 13 through 24. The second model computes the modulation potential from the open flux together with the tilt and polarity of the heliospheric current sheet, following a formulation calibrated by Owens and colleagues in 2024. Crucially, the open solar flux evolves with memory: it accumulates from past sunspot activity and decays through magnetic reconnection, introducing a hysteresis that makes the inverse problem fundamentally non-unique. Multiple distinct sunspot histories can produce statistically indistinguishable modulation records, which is precisely why simple deterministic inversions break down.

To handle that non-uniqueness, the team employed Approximate Bayesian Computation, a statistical technique that sidesteps the need for an explicit likelihood function. In each sliding ten-to-fifteen-year window, ten thousand Monte Carlo realisations of sunspot cycles are propagated through the forward models and compared directly against the observed modulation potential using a weighted Euclidean distance. The best two percent of candidates, some two hundred realisations, are retained as samples from the approximate posterior distribution, and their spread provides rigorous, quantified uncertainty bounds reported as 68 percent highest-density intervals. Because the sunspot cycle amplitudes are constrained to be non-negative by construction, the resulting reconstruction can never produce the negative values that haunted earlier regression-based approaches, and it requires no post-hoc correction.

To test the method, the researchers applied it to two annual-resolution records of the modulation potential. The first, spanning 1845 to 2020, was derived from geomagnetic observations of open solar flux by Owens and colleagues. When used as the inversion target, the method recovered sunspot numbers and open flux in close agreement with the direct instrumental record maintained by SILSO, with a mean absolute error of just 19.25 megavolts, about three percent of the mean modulation potential. That success validated the technique and justified applying it to the second, far longer dataset: a radiocarbon-based modulation potential record covering 971 to 1932, reconstructed by Nicolas Brehm of ETH Zurich and colleagues from tree-ring carbon-14 measurements. Before feeding the tree-ring record into the inversion, the team cross-calibrated it against the geomagnetic record over their overlapping decades, applying an additive correction of 65.76 megavolts that statistical tests showed produced a near-perfectly symmetric, homoskedastic residual distribution.

The millennial-scale reconstruction that emerged is rich with detail. It captures the familiar grand minima and maxima of the past thousand years, including the Maunder minimum, the Spörer minimum, the Dalton minimum, and the double-peaked grand maximum the Sun passed through between 1900 and 2020. Crucially, the reconstructed open solar flux never falls to zero. Even during the deepest quiet of the Spörer minimum, the unsigned open flux dropped only to about 1.21 times ten to the fourteen webers in 1443, the lowest value in the entire record and well below anything observed in the telescopic era, yet still decisively nonzero. During the Maunder minimum the flux averaged around 2.63 times ten to the fourteen webers. This confirms that the solar dynamo never fully shuts down during grand minima but instead idles in a reduced, finite activity state, a conclusion consistent with flux transport dynamo models in which meridional plasma circulation sustains weak cycles and eventually drives recovery.

The new record also documents some striking extremes at the other end of the scale. The peak annual open solar flux of the twentieth century, reached in 1958, was about 10.96 times ten to the fourteen webers, the highest since the year 1200. Only three years in the entire millennium exceeded it, the largest peaking at roughly 14.21 times ten to the fourteen webers in 981 CE, shortly after the record begins. In other words, the era of telescopic observation has sampled a large fraction, but not all, of the Sun’s true dynamic range; deeper minima and higher maxima both occurred before instruments existed to see them. The reconstruction also flags three intervals around 993, 1052 and 1279 CE where proposed Miyake events, extreme solar particle storms recorded as abrupt radiocarbon spikes, contaminate the cosmic-ray-based record, and the authors conservatively mask these windows from their results.

Beyond its intrinsic appeal as a thousand-year diary of solar magnetism, the work has immediate practical value. Annually resolved, physically consistent sunspot numbers feed directly into reconstructions of total and spectral solar irradiance, which in turn constrain climate models exploring the Sun’s role in terrestrial temperature variability over past centuries. The uncertainty-quantified cycle amplitudes during grand minima provide empirical targets for solar dynamo theorists probing the minimum operating point of the solar cycle engine. The authors note that their forward model templates were built from modern, regular solar cycles, so reconstructions of grand minimum cycles should be treated as indicative rather than definitive, and that the additive cross-calibration between the two modulation potential datasets is itself a simplification. Even so, the framework offers what the team describes as a probabilistic ensemble of physically admissible solar histories rather than a single deterministic answer, and it opens the door to pushing the same technique further back in time as longer and older cosmogenic isotope records become available. The Sun, it turns out, kept meticulous records all along; the trick was learning to read them without breaking the laws of physics.

Subject of Research: Physics-constrained reconstruction of annually resolved sunspot numbers from millennial-scale heliospheric modulation potential records.

Article Title: Physics-Constrained Reconstructions of Sunspot Number from Millennial-Scale Annual Heliospheric Modulation Potential

Article References: Saha, C., Owens, M., Lockwood, M., Barnard, L., Brehm, N., Dalla, S., Herbst, K., Muscheler, R., & Wang, J. (2026). Physics-Constrained Reconstructions of Sunspot Number from Millennial-Scale Annual Heliospheric Modulation Potential. Solar Physics, 301(9), Article 144. https://doi.org/10.1007/s11207-026-02731-0

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02731-0

Keywords: sunspot number, solar cycle, heliospheric modulation potential, cosmogenic isotopes, radiocarbon, open solar flux, Maunder minimum, Spörer minimum, approximate Bayesian computation, space climate, solar dynamo, galactic cosmic rays