Paper check · Geomagnetic storms and ozone
A new preprint measured how four geomagnetic storms since May 2024 destroyed ozone high above the poles [1]. A video about it told viewers the losses were too big for Earth’s magnetic field, that climate models missed them, and that they point to a pole shift [V]. Here is what the paper actually reports.
Our analysis marks our own reasoning, and the verdicts are our assessment. Everything else comes from the numbered sources; § points to a section of the preprint [1].
“They measured the ozone loss and found between 30 and 80% depletion of those upper mesospheric layers.”
Video, 3:21 [V]
Matches the paper. The MLS instrument on the Aura satellite recorded peak losses of about 30% to 80% at roughly 60–80 km altitude after the four storms [1 Abstract, §3.1].
“Their models did not reproduce that, meaning this mega effect that they saw in reality is not in the climate models.”
Video, 3:29 [V]
Where those losses were measured, the model’s results “generally agree well” with the satellite’s, “both in spatial extent and in magnitude,” though often slightly smaller [1 §3.3]. The “not reproduced” in the paper’s abstract refers to something else: a dip higher up that isn’t statistically significant (section 2). And the model, WACCM6, is itself a climate model [1 §2.1].
“80% is way too much. That would never happen with Earth’s normal magnetic field unless it was like an X1000 solar flare.”
Video, 3:36 [V]
October 2024’s 80% loss followed an S3 “Strong” proton storm tied to an X1.8 flare [4], a level NOAA expects about 10 times per 11-year cycle [5]. A 2006 study found that “even moderate” proton events can deplete middle-atmosphere ozone by more than 70% [6]. See section 4.
“Pole shift right there in numbers.”
Video, 3:45 [V]
The paper never mentions a pole shift or a weakening magnetic field. Its model handles Earth’s magnetic shielding the standard way used for climate simulations, and its results still broadly match the measured losses in place and size [1 §2.2, §3.3].
1 · The mechanism
The basic pathways are “well established,” the authors write. Large uncertainties remain in how many particles arrive, how much air they ionize and how much the chemistry changes [1 §1].
Solar protons and electrons from Earth’s radiation belts pour into the polar atmosphere during major storms [1 §1].
Electrons above 130 keV ionize air at 60–80 km. Protons reach deeper, into the upper stratosphere [1 §1, §3.1].
The ionization produces odd nitrogen (NOx) and odd hydrogen (HOx) [1 §1].
NOx and HOx destroy ozone catalytically [1 §1]. Loss peaked about a day after each storm’s peak activity, a delay the authors say fits reactions that need sunlight to work best [1 §3.2].
2 · Where the model fell short
The satellite looked at three layers of the atmosphere. The model broadly matched the layer where the 30–80% losses were measured, though it often ran slightly low.
| Layer | What the satellite saw | What the model did | What the video said |
|---|---|---|---|
| Lower thermosphereUp to about 95 km, the top of the satellite’s range [1 §2.1] | A dip in ozone after all four storms, “generally not statistically significant” [1 §3.1] | Not reproducedThe authors think the model lacks the processes at work there [1 §3.3] | Not in the excerpt |
| MesosphereLosses measured at about 60–80 km [1 §3.1] | Peak losses of about 30% to 80% [1 §3.1] | Broadly reproducedResults “generally agree well… both in spatial extent and in magnitude,” though often slightly smaller [1 §3.3] | Misread “Their models did not reproduce that” [V] |
| StratosphereBelow about 50 km, home to about 90% of Earth’s ozone [3] | October 2024’s loss reached into the upper stratosphere [1 §3.1]. Otherwise the signal is “heavily masked” by background conditions [1 §3.3] | Sinking loss, model onlyPatches of loss sinking through winter, about 90% from electrons [1 §3.3] | Not in the excerpt |
In the abstract, “not reproduced by WACCM-D” refers to “these features,” the lower-thermosphere dips, five sentences after the 30–80% figure [1]. The video attached that miss to the wrong layer.
3 · The four storms
In the model, solar protons caused nearly all of the loss in October 2024, the 80% storm. Electrons caused most of it in the famous May 2024 superstorm [1 §3.3].
Peak ozone loss at about 60–80 km
May 2024
Peak loss May 12
October 2024
Peak loss October 10
November 2025
Peak loss November 13
January 2026
Peak loss January 21
| Storm | Peak loss day | Loss from the storm | Already gone before | Total | Particles, from the model | Reach |
|---|---|---|---|---|---|---|
| May 2024 | May 12 | ≈50% | ≈10% anomaly | Not stated | Electrons >80% | About 50°, at least 7 days |
| October 2024 | October 10 | ≈80% | Clean | ≈80% | Protons, nearly all | Pole to about 60°, about 3 days |
| November 2025 | November 13 | ≈30% | ≈30% | ≈60% | Protons ≈40%, electrons ≈60% | About 50° |
| January 2026 | January 21 | ≈50% | ≈30% | ≈80% | Protons ≈70%, electrons ≈30% at onset | Beyond 50° |
4 · Is 80% too much?
The video sized up the storm by its flare. NOAA rates flares on its radio blackout scale; the solar protons that drove October’s loss are rated on a separate scale for solar radiation storms, based on the flux of particles of 10 MeV and up [5].
October 2024
S3
“Strong” on NOAA’s five-level radiation storm scale [5]. Reached on October 9, 2024, after an X1.8 flare and its coronal mass ejection [4].
Done before
>70%
“Even moderate solar events (>10 MeV proton flux of >5000 pfu) can cause significant ozone depletion (>70%),” a study of the January 2005 storm found [6]. The October–November 2003 storms cut ozone by tens of percent at 36–60 km in the Northern Hemisphere’s polar winter, for several months [7].
| Level | Flux, pfu | Storms per cycle |
|---|---|---|
| S1 Minor | ≥10 | 50 |
| S2 Moderate | ≥100 | 25 |
| S3 Strong | ≥1,000 | 10October 2024 [4] |
| S4 Severe | ≥10,000 | 3 |
| S5 Extreme | ≥100,000 | Fewer than 1 |
The satellite saw more loss in October 2024 than the model produced. The authors see evidence that disturbed atmospheric dynamics also shaped the observed response, citing enhanced stratospheric ozone during the storm [1 §3.3].
About 10 S3 storms per 11-year cycle works out to roughly one a year on average [5]. On NOAA’s numbers, a storm like October 2024’s is a routine part of each solar cycle.
“…the magnetic field lines at the poles extend vertically downwards, allowing the particles to spiral down the field lines”
October 2024’s proton-driven loss reached only from the pole to about 60° latitude, the narrowest of the four storms. The other three reached about 50° latitude. During major storms, electrons can rain in down to about 50° geomagnetic latitude, and larger proton fluxes can push the boundary toward the equator [1 §1, §3.2].
Losses confined to high latitudes are what working magnetic shielding looks like. Protection depends on latitude and field strength [8], so a much weaker field would let solar protons in at lower latitudes. A failing field would show up as proton-driven losses spreading toward the equator.
5 · Caveats
The authors spell out the limits of their results. None of them involves Earth’s magnetic field.
Posted October 7, 2026, for open discussion. It is under review for Atmospheric Chemistry and Physics, with comments open until November 18 [2].
The lower-thermosphere anomalies “are generally not statistically significant” [1 §3.1].
Satellite anomalies in the stratosphere are “heavily masked” by background conditions [1 §3.3].
The electron contribution “may be systematically underestimated,” though “not necessarily” in extreme events [1 §3.3]. A May 2024 test found the electron input prescribed for CMIP7, the next round of coordinated climate-model experiments, too high in the upper mesosphere and too low at 60–65 km [1 §3.4].
Simulated mesospheric loss is “often slightly smaller than that observed by MLS” [1 §3.3].
The authors recommend focusing on the period from storm onset to 7 days after [1 §2.2].
If each storm’s top-layer dip were a coin flip, all four would come out negative 1 time in 16, about 6%. That is suggestive but not conclusive, which fits the authors’ view that these features “deserve closer examination” [1 §3.1].
The 30–80% figures describe 60–80 km. About 90% of Earth’s ozone lies in the stratosphere, below about 50 km, and most of the rest in the troposphere [3], so those percentages apply to a thin slice of the total. What reaches the stratosphere, such as October 2024’s deepest loss and the model’s slower sinking effect, matters more for the ozone layer [1 §3.1, §3.3].
Losses this large that neither the measured particles nor atmospheric dynamics can explain in a model using today’s magnetic field
Proton-driven ozone loss spreading far beyond the polar caps during a storm of ordinary strength
Magnetic-field measurements showing the shielding itself failing, not just strong storms passing through it
Bracketed numbers refer to this list; § marks a section of [1]. Links checked October 11, 2026.
Jia, J., Orsolini, Y., Szeląg, M. E., Kero, A., Verronen, P. T., Grandin, M., Espy, P. J., Käki, S., Thomas, N., van de Kamp, M., Koldobskiy, S., Zhang, J., and Partamies, N. (2026). Comparative analysis of middle atmospheric ozone response to four large geomagnetic storms of Solar Cycle 25. EGUsphere preprint, under review for Atmospheric Chemistry and Physics. doi:10.5194/egusphere-2026-5829. CC BY 4.0.
EGUsphere. Discussion page for [1]: received September 24, 2026; discussion started October 7, 2026; open until November 18, 2026.
NOAA Chemical Sciences Laboratory (2022). Twenty Questions and Answers About the Ozone Layer: 2022 Update, Q1. Part of the WMO Scientific Assessment of Ozone Depletion: 2022.
NOAA Space Weather Prediction Center (October 9, 2024). S3 (Strong) solar radiation storm in progress.
NOAA Space Weather Prediction Center. NOAA space weather scales: radio blackouts (R) and solar radiation storms (S).
Seppälä, A., Verronen, P. T., Sofieva, V. F., Tamminen, J., Kyrölä, E., Rodger, C. J., and Clilverd, M. A. (2006). Destruction of the tertiary ozone maximum during a solar proton event. Geophysical Research Letters 33, L07804. doi:10.1029/2005GL025571
Seppälä, A., Verronen, P. T., Kyrölä, E., Hassinen, S., Backman, L., Hauchecorne, A., Bertaux, J. L., and Fussen, D. (2004). Solar proton events of October–November 2003: ozone depletion in the Northern Hemisphere polar winter as seen by GOMOS/Envisat. Geophysical Research Letters 31, L19107. doi:10.1029/2004GL021042
NOAA National Weather Service. Types of space weather storms, section on solar radiation storms.
Video commentary on the preprint, transcript 3:07–3:45, as shared by a reader. We haven’t identified the channel it came from.
Method
On October 11, 2026, we checked every statement, number and quotation on this page against its source: the full preprint as posted October 7, 2026, the other sources listed here and the transcript of the video. That came to about 60 items. One section reference was wrong, and we’ve fixed it; the change log has the details.
The verdicts, and the items tagged “Our analysis,” are our own reasoning built on the cited figures. They aren’t findings of the sources. The magnetic-field drawing is a schematic.
The video is quoted from a transcript a reader shared with us [V].
This page was researched and written with Claude, an AI assistant made by Anthropic, and then re-checked against its sources in a separate verification pass.
Corrections
If we get something wrong, we fix it and record the change below with the date. If a correction changes a verdict, the log says what the old verdict was.