Fact-Checked Science

Paper check · Geomagnetism

The core field accelerates most near the equator, in patterns that persist, not collapse

Three geophysicists in Madrid mapped the secular acceleration of Earth’s magnetic field, how fast its rate of change itself changes, at the top of the core over 125 years, and related it to geomagnetic jerks [1]. S0 News read their maps as “a full equatorial collapse”, with concentration points at the South Atlantic Anomaly and two folklore triangles [2 00:03:13–00:03:29]. We could not read the paper itself: its publisher’s page sits behind a CAPTCHA that the site’s rules do not let us pass, and we found no other copy. So we read its record and an index copy of its abstract [3][4] and re-ran the public field models it uses: the acceleration does gather near the equator, as the abstract says, but not at the anomaly or the Dragon’s Triangle, and its strongest patch lay inside the Bermuda Triangle only in 1950–1975; the century’s apparent growth lies in scales the early data cannot resolve; and the equatorial field fell 0.99% in 25 years, the same as worldwide [6][8].

By Fact-Checked Science · Published

Our analysis marks our own reasoning and calculations, including our re-run of the CHAOS-8.5 and COV-OBS.x2 field models, and the verdicts are our assessment. Everything else comes from the numbered sources. For the paper [1] we could read only its record and an index copy of its abstract [4], so nothing on this page quotes its text; elsewhere § gives a section, p. a page, Fig. a figure and Abstract the part of a source we quote.

Repeat our re-runDownload as PDFSourcesHow this was checkedCorrections

The acceleration gathers near the equator in lasting patterns, and nothing there is collapsing

Our assessment

54%

of the secular acceleration’s power at the top of the core lies within 20° of the equator, a band covering 34% of it, in the model the paper uses for 2000–2025, by our re-run [6]

0.8–1.1×

the global level: the acceleration beneath the South Atlantic Anomaly in the same model, depending on where the anomaly’s edge is drawn, against 2.8× at the strongest patch, by our re-run; the anomaly is not where it gathers [6]

−0.99%

the change in the surface field’s strength along the equator from 2000 to 2025, the same as worldwide, by our re-run [6]

The claims

Misread

A new data point on a worsening pole shift

The presenter: “a new data point on the worsening situation of Earth’s ongoing magnetic pole shift” [2 00:00:18–00:00:22]

The paper maps how fast the core field’s rate of change itself changes, and relates that to geomagnetic jerks [1]. Its record and an index copy of its abstract say nothing about a pole shift, and what the abstract reports is persistence, a few long-lived regional structures, not a worsening [3][4]. In the model the paper uses for 2000–2025 the acceleration was strongest in 2010–2015 and has eased since, by our re-run [6]. “A pole shift is under way” is rated Misleading on our review of October 20. See section 6.

Misleading

From sparse events to growing power and coverage

The presenter: “What they found is that we’ve gone from sparse moderate accelerative events 120 years ago to increasing power and coverage of the variation anomaly.” [2 00:02:57–00:03:04]

Maps of the models the paper uses do darken toward the present, by our re-run. In COV-OBS.x2, one of the two it uses for the 20th century, the acceleration at the top of the core is 70% stronger in 2000–2020 than in 1900–1925, and the share of the core surface above 250 nT a year squared rises from 30% to 82% [8]. But almost all of the rise is at fine scales the early data cannot resolve [9 Results], a limit the abstract itself notes as larger uncertainties before the satellite era [4]; the large scales stay flat for 120 years, and in the sharp satellite era the acceleration peaked in 2010–2015, by our re-run [6]. What the paper’s text says about a trend, we could not read [1]. See section 4.

False

Earth has lost the plot, in a full equatorial collapse

The presenter: “earth has completely lost the plot. The wheels are coming off almost coming completely off the rails in a full equatorial collapse.” [2 00:03:13–00:03:20]

Nothing at the equator is collapsing. The surface field along it weakened 0.99% from 2000 to 2025, as much as worldwide, and 3.8% from 1900 to 2020, less than the worldwide 6.1%, by our re-run of the models the paper uses [6][8]. What gathers near the equator is the acceleration, an expected signature of fast waves and flows at the top of the core [12 Abstract]; the CHAOS-7 team called recent Pacific events “an integral part of the expected spectrum of rapidly changing SV behaviour that takes place at low latitudes” [10 Results]. See section 6.

Misleading

The concentration points are the South Atlantic Anomaly, a West Pacific anomaly and two Triangles

The presenter: “Concentration points are the South Atlantic anomaly in Bermuda Triangle and then on the West Pacific anomaly and the Dragons Triangle exactly as we expect them to be.” [2 00:03:20–00:03:29]

One of the four places held a hot spot decades ago: in COV-OBS.x2, the strongest acceleration on the whole core in 1950–1975 lay inside the Bermuda Triangle, at 29°N 67°W [8]. In 2000–2025 the triangle is 0.8 to 1.5 times the global level, and the strongest patch lies 23° to its south, under Colombia, at 2.8 times, by our re-run [6]. The rest fails now: beneath the South Atlantic Anomaly the acceleration is 0.8–1.1 times the global level, depending on where its edge is drawn [6]; the published West Pacific Anomaly disappeared after 1800 [24 Introduction]; the Dragon’s Triangle lies over an ordinary part of the core; and the Indian sector, which the abstract names, is left out. “Exactly as we expect” cites no prediction, and the abstract makes none [4]. See section 5.

True

Secular acceleration gathers in a few low-latitude regions

The paper’s abstract, paraphrased from an index copy [4]

For 2000–2025, the abstract says, the acceleration is concentrated in a few low-latitude regions, paired with opposite signs, chiefly in the Atlantic, Pacific, Caribbean and Indian sectors [4]. Our re-run of the same model agrees on the low-latitude concentration: 54% of the power within 20° of the equator, a band covering 34% of the core’s surface, and six of the seven strongest patches within 5° of it [6]; the pairing, the opposite signs and the component analysis we could not test. Other teams’ maps of the same era also put the strongest acceleration mostly at low latitudes [11 Abstract][23 Abstract]. See section 3.

1 · A paper we could not read

Open access, but behind a CAPTCHA, and we found no other copy; we read its record and an index copy of its abstract

The paper is a peer-reviewed article in Physics of the Earth and Planetary Interiors, online since October 5, 2026, in the volume dated December 2026, and open access under a CC BY 4.0 licence [1][3]. Its three authors work at the Complutense University of Madrid, two of them also at the geosciences institute it runs with the Spanish National Research Council [4]. Its text was out of our reach.

What the record shows

Title: “Persistent regional patterns of secular acceleration at the core–mantle boundary and their relation to geomagnetic jerks over the past 125 years”, by Mario Serrano, Saioa A. Campuzano and F. Javier Pavón-Carrasco; volume 381, article 107615 [1][3]. 49 references; funding from the Community of Madrid and Spain’s Ministry of Science and Innovation [3].

No abstract is deposited with Crossref [3]. OpenAlex holds one as an index copy, which we paraphrase on this page and do not quote [4].

Our analysis

What that means for this check

We rate what the paper’s record, its abstract and its public data support, and we say so in each verdict. The publisher’s page asked our tools, and a real browser, to prove we are human; the site’s rules do not let us pass a CAPTCHA or any other bot check, so we stopped there, and we found no other copy. Everything that needs the text, listed below, stays unverified.

The three field models the paper uses are public, though. So we did what its text would have let us check, and re-ran two of them ourselves: sections 3 to 5.

The abstract, in brief (Abstract only, from an index copy)

The authors study the secular acceleration of Earth’s field, and its changes, at the boundary between the core and the mantle over the past 125 years. For 2000–2025, with the CHAOS-8.5 model, they find it concentrated in a few low-latitude regions, in pairs of opposite sign, chiefly in the Atlantic, Pacific, Caribbean and Indian sectors; these behave in persistently opposite ways and carry most of the rapid changes tied to recent jerks [4].

A regional principal component analysis finds one dominant pattern in each of the Atlantic, Pacific and Caribbean sectors and two in the Indian sector, and ties each region to the 21st century’s sequence of jerks. For the 20th century, with the Kalmag and COV-OBS.x2 models, the authors note larger uncertainties before the satellite era but find persistent, recurring structures: a dominant Caribbean–North America pair through much of the first half of the century, and the Pacific and Indian sectors growing in importance after the 1970s. They suggest that the acceleration’s variability is organised around a few long-lived regional structures whose relative strengths change over time [4].

What we could not read, and so do not rate

QuestionStatus
Which figure the episode showed (“The darker spots are the increase. Oldest is at the bottom. newest top left”)Unverified [2 00:03:06–00:03:09]
The jerk epochs the paper usesUnverified; section 2 gives the published record
The regions’ coordinates and boundariesUnverified; section 3 gives the patches in the model the paper uses, by our re-run
Any numbers for a change in the acceleration’s strength or extent over timeUnverified; section 4 gives our re-run
The authors’ caveats, beyond the abstract’s note on uncertainty before the satellite eraUnverified [4]
How the acceleration and its changes are computed: truncation, time step, the component analysisUnverified
The pairing of opposite polarity, the anti-correlated behaviour and the component analysis’s modesUnverified; our re-run tests only where the acceleration concentrates
Competing interests, data availability, received and accepted dates; whether the full text mentions a pole shift, the South Atlantic Anomaly or the trianglesUnverified
Journal, publication dates, licence, funding, affiliations, referencesKnown [3][4]
Every route to the paper we tried, and what each returned
RouteResult on October 25, 2026
The episode’s link to ScienceDirect, and its PDF link, fetched as a browser wouldHTTP 403 and a CAPTCHA page: “Are you a robot? Please confirm you are a human by completing the captcha challenge below.” Not attempted
The DOI, through a page-reading toolA redirect page, then HTTP 403
The article page in a real browserThe same CAPTCHA, still there after waiting. Not attempted
Crossref’s metadata recordFull record, with references and funders; no abstract [3]
OpenAlexOpen access at the publisher only, no repository copy; an abstract as an index copy [4]
Semantic ScholarNo abstract and no other copy
Elsevier’s text-mining interfaceMetadata only without an API key, which we do not seek; it confirms the CC BY 4.0 licence
The Complutense University’s repository, DoctaThe group’s earlier jerk papers, not this one
Europe PMCNo record
arXiv, and web searches for the exact titlearXiv limited our requests; no preprint, conference abstract or repository copy found

Each route was tried on October 25, 2026; the log, with file sizes and checksums, is in our verification record for this page.

2 · Jerks and the models

Jerks have broken the field’s slow change for at least a century, and the three models the paper uses see them with very different sharpness

Earth’s internally generated field changes slowly, the secular variation, and the rate of that change itself changes, the secular acceleration. “The secular variation of the core field is generally characterized by smooth variations, sometimes interrupted by abrupt changes, named geomagnetic jerks” [14 Abstract]. Mapped down to the top of the core, about 2,900 km below the surface, the acceleration shows where the field’s rate of change is itself changing fastest.

A century of jerks

A wavelet search of 97 observatories’ records found seven jerks in the 20th century: two of global extent, in 1969 and 1978, three possibly so, in 1901, 1913 and 1925, and two seen only in places, in 1932 and 1949 (an abstract read through an index, paraphrased) [15 Abstract]. Of the jerks after 1969, “Those with the widest consensus have turned out to correspond to epochs around 1978, 1991, 1999, and 2003” [16 §1]; and in observatory data from 1957 to 2008, changes that fit the definition of a jerk are frequent and not simultaneous worldwide (also paraphrased from an index) [17 Abstract].

Since 1999, jerks have come “at a regular rate of one every 3–4 years”, in 2003, 2007, 2011 and 2014 [18 Introduction], then over the Pacific in 2017 [20 Abstract], worldwide between mid-2019 and early 2020 [19 Abstract], and in 2024, after a pulse of acceleration centred in 2022 [21 Abstract].

Why jerks matter to field models

A field model forecasts five years of secular variation from the trend it last saw, and a jerk breaks the trend. The 2024 jerk “was not captured by our candidate SV model or the IGRF-14 final SV in Western Europe”, wrote the authors of one candidate model for the International Geomagnetic Reference Field, and the resulting errors “are expected to contribute, at least in part, to the degradation in the performance” of both forecasts; “A similar situation occurred shortly after the development of IGRF-13, due to the 2020 geomagnetic jerk and subsequent 2022 SA pulse” [21 §5].

Drift of that kind is ordinary: it is why the reference models are reissued every five years, as our review of October 20 explains. It is not the failure our video check examined, where the 2025 World Magnetic Model, by NOAA’s annual report, proved accurate in its first year (“The 2025 model already failing?”, rated False).

The three models the paper uses, as their makers describe them

Scroll sideways to see all four columns →

ModelUsed in the paper forBuilt fromWhat its makers say about its sharpness
CHAOS-8.52000–2025 [4]Satellite and observatory data; “valid from 1999 to 2025” [5 Abstract]. Release of February 2, 2026, with Swarm data to December 21, 2025 [6]Users of the acceleration at the core surface should truncate it at degree 15 [5 §8]. Its prior “allows stronger accelerations of the sectoral terms compared with the zonal terms, due to the low-latitude focus of acceleration patterns associated with the hydromagnetic waves in the dynamo simulation” [5 §3.2.2]
KalmagThe 20th century [4]A Kalman filter and smoother over satellite, observatory and survey data, with uncertainties, for the 122 years from 1900 [9 Abstract]“At early times, between 1900 and 1920, only some limited amount of temporal scales of the SV up to SH degree ℓ=4 are resolved”; up to degree 6 in 1920–1960; in the satellite eras, degree 15 partly resolved over about five years and degree 10 well resolved over two [9 Results]
COV-OBS.x2The 20th century [4]Observatories, satellites and older surveys, 1840–2020 [7 Abstract]“the use of a cubic B-splines basis with knots separated by 2 years does not permit the capture of short time-scales features” [7 Results and discussion]

SV is secular variation and SH spherical harmonic; the degree measures the scale of a pattern, larger degrees for smaller patterns. Which model the paper uses for which years is from its abstract; how it combines them, and at what truncation, we could not read [1].

3 · Where it gathers

In the model the paper uses for 2000–2025, the acceleration gathers within a few degrees of the equator

We computed the secular acceleration of the radial field at the top of the core from CHAOS-8.5, the model the paper uses for 2000–2025, every quarter of a year, and took its root-mean-square strength in each 2° cell [5][6]. The map is our own re-run, not a figure from the paper, which we could not see.

How strongly the field accelerated at the top of the core, 2000–2025

Our analysis
  • under 250
  • 250–500
  • 500–750
  • 750–1,000
  • 1,000 or more nT a year squared

Scroll sideways to see the whole map →

Bermuda TriangleDragon’s Triangleold WPA, 1650SAA minimaABCDEFG30°N0°30°S
  1. A3°N 73°W, under Colombia: 1,254
  2. B5°N 43°W, the equatorial Atlantic off Brazil: 1,170
  3. C3°S 103°E, under Sumatra: 1,014
  4. D1°N 147°E, the western Pacific north of New Guinea: 1,007
  5. E27°N 87°E, under the Himalaya north of the Bay of Bengal: 957
  6. F5°N 99°W, the eastern Pacific off Central America: 957
  7. G3°N 125°W, the eastern Pacific: 829

In blue: × the South Atlantic Anomaly’s two surface minima in 2025, by our re-run; the Bermuda Triangle, drawn between Miami, Bermuda and San Juan; the Dragon’s Triangle, a dashed box from about 24° to 34°N and 139° to 145°E, since it has no fixed bounds; the old West Pacific Anomaly’s core about 1650, dashed.

Our analysis of CHAOS-8.5 [6] (DTU Space, CC BY 4.0): the second time derivative of the radial field, degrees 1–9, at the core–mantle boundary, radius 3,485 km, every quarter of a year from 2000 to 2025, as a root mean square over time; 2° cells merged into 4° cells for drawing; patch values in nT a year squared. Degree 9 is the truncation the CHAOS team advises for the acceleration before 2014 [10 Conclusions]; at degree 13 the map is busier and two high-latitude patches join the strongest, but the same low-latitude patches lead. Coastlines: Natural Earth [33].

Share of the acceleration’s power, and of the surface, in each band of latitude

Our analysis
  • Share of the acceleration’s power, 2000–2025
  • Share of the core’s surface

Scroll sideways to see the whole chart →

0%10%20%30%40%0–10° N or S33%17%10–20° N or S21%17%20–30° N or S14%16%30–45° N or S12%19%45–60° N or S10%17%60–90° N or S11%13%
Our analysis of CHAOS-8.5, degrees 1–9 [6]. Each band counts both hemispheres; the power is the area-weighted square of the root-mean-square acceleration. A band with more power than area holds more than its share.

Within 20° of the equator

Our analysis

54%

of the acceleration’s power, on 34% of the core’s surface; 33% of it within 10°, on 17% [6]

The strongest patches

Our analysis

6 of 7

lie within 5° of the equator; the seventh, under the Himalaya, at 27°N [6]

Finer scales kept, to degree 13

Our analysis

48%

of the power still within 20° of the equator: the belt does not depend on the truncation [6]

Other maps of the same era

Mapping the CHAOS-6 model in 2015, its makers found the acceleration at the core surface “largest at low latitudes under the India–South-East Asia region, under the region of northern South America, and at high northern latitudes under Alaska and Siberia” [11 Abstract]: our patches A, C and E, and the map’s Alaskan and Siberian highs. In CHAOS-7 since 2000, the maxima of the acceleration’s pulses are “mainly concentrated near the equator with a north-south span from 14°S to 18°N in the region from ∼10°E west to 150°W”; there are none from about 10°E to 60°E, and from 60°E to 150°E they lie farther from the equator, between 35°S and 45°N [23 §3.2]. The same authors find the pulses “the strongest near the equator (2°N) and more robust in the high-latitude region (68°S) of the Southern Hemisphere” [23 Abstract]. In CHAOS-8’s own maps, “The CMB SA is generally strongest at latitudes below ±30°” [5 §4.5].

In data from the CHAMP satellite, the pulses of 2006 and 2009 appeared as intense patches at the core surface in the Atlantic sector at low latitudes, and in the Indian Ocean sector, read as parts of a standing wave with a period of about six years (an abstract read through an index, paraphrased) [22 Abstract].

Why there

The leading explanation is waves in the liquid core. In geodynamo simulations, “jerks are caused by the arrival of localized Alfvén wave packets radiated from sudden buoyancy releases inside the core. As they reach the core surface, the waves focus their energy towards the equatorial plane and along lines of strong magnetic flux” [12 Abstract].

In satellite data, the field’s year-to-year changes follow suit: “Downward continued to the core surface, these variations primarily show up in the equatorial belt” [13 Abstract]. The CHAOS-8 team reports “further evidence of traveling hydromagnetic waves at the core–mantle boundary” [5 Abstract].

Our analysis

One caveat, and why the verdict stands

CHAOS-8 is built with a prior drawn from geodynamo simulations that “allows stronger accelerations of the sectoral terms compared with the zonal terms, due to the low-latitude focus of acceleration patterns associated with the hydromagnetic waves in the dynamo simulation” [5 §3.2.2]: the model leans toward finding acceleration at low latitudes. But maps of the older CHAOS-6 and CHAOS-7 models, built without that prior, also put the strongest acceleration mostly at low latitudes [11 Abstract][23 Abstract], and so does COV-OBS.x2 over 2000–2020, where 27% of the power lies within 10° of the equator, on 17% of the surface, by our re-run [8]. The paper’s finding of a low-latitude concentration stands, and we rate it True.

4 · A century of maps

Maps of the models the paper uses darken toward the present, at the scales the old data cannot see

The episode described the paper’s figure as maps across time: “The darker spots are the increase. Oldest is at the bottom. newest top left” [2 00:03:06–00:03:09]. Which figure it was, we could not see [1]. Here is the same measure in COV-OBS.x2, one of the two models the paper uses for the 20th century, at the start, the middle and the end of its span, on one scale.

The same measure at the start, the middle and the end of a century, in one model

Our analysis
  • under 250
  • 250–500
  • 500–750
  • 750–1,000
  • 1,000 or more nT a year squared

1900–1925

30°N0°30°S

1950–1975

30°N0°30°S

2000–2020

30°N0°30°S

In blue: the Bermuda Triangle, drawn between Miami, Bermuda and San Juan.

Our analysis of COV-OBS.x2, the model of Huder and colleagues [8], drawn as in section 3: degrees 1–9, root mean square over each period. The darker recent map is what sharper data produce, as the chart below shows; its patches also sit where CHAOS-8.5 puts them. In 1950–1975 the strongest patch on the whole core lies inside the Bermuda Triangle, at 29°N 67°W (section 5).

How strong the acceleration is, by scale, decade by decade

Our analysis
  • Degrees 5–9, smaller scales: COV-OBS.x2, by decade
  • Degrees 1–4, large scales: COV-OBS.x2, by decade
  • The same in CHAOS-8.5, by five years
  • The satellite era

Scroll sideways to see the whole chart →

satellite era01002003004005001900192019401960198020002020nT a year squareddegrees 5–9: smaller scalesdegrees 1–4: large scales

Up: the root-mean-square acceleration of the radial field at the core surface. Across: the year

Our analysis of COV-OBS.x2, the model of Huder and colleagues [8], and of CHAOS-8.5, DTU Space’s (CC BY 4.0) [6]: the global root mean square over each decade or five years, plotted at its middle.

1900–1925 to 2000–2020

Our analysis

+70%

in the acceleration’s overall strength in COV-OBS.x2, 241 to 409 nT a year squared; the share of the core surface above 250 rises from 30% to 82% [8]

Large scales, degrees 1–4

Our analysis

79–112

nT a year squared, decade by decade since 1900: no trend [8]

The satellite era, CHAOS-8.5

Our analysis

Peak 2010–15

514 nT a year squared, against 384 in 2000–2005 and 405 in 2020–2025 [6]

What the models’ makers say about the early decades

Kalmag’s makers: “At early times, between 1900 and 1920, only some limited amount of temporal scales of the SV up to SH degree ℓ=4 are resolved. The situation slightly improves between 1920 and 1960 where some signal up to SH ℓ=6 can be accurately recovered”, while in the satellite eras “spatial scale up ℓ=15 can be partially resolved down to periods of approximately 5 years, and 2-year fluctuations can be very well captured up to ℓ=10” [9 Results]. COV-OBS.x2’s makers: its splines, knots two years apart, do “not permit the capture of short time-scales features” [7 Results and discussion]. Even within the satellite era, the CHAOS-7 team advises truncating “the secular acceleration, which is more challenging to extract, at degree 9 (degree 15 after 2014)” [10 Conclusions].

The choice of model moves the numbers too. Over 2000–2020, CHAOS-8.5 shows 11% more acceleration than COV-OBS.x2 at degrees 1–9 and 73% more at degrees 10–14, by our re-run [6][8]; the CHAOS-8 team finds the same between its own versions: “These differences showcase the differing regularization methodologies adopted in CHAOS-8.1 and CHAOS-7.18” [5 §4.3].

Our analysis

Why “growing power and coverage” misleads

The true part: in the models the paper uses, by our re-run, the recent maps are darker and their dark patches more widespread. The false impression: that this is a finding about the core. Almost all of the century’s rise sits at degrees 5–14, fine scales the early data cannot resolve; the large scales, which they can, stay flat; and in the sharp satellite era the acceleration rose to 2010–2015 and fell back. COV-OBS.x2’s mean model is pulled toward zero acceleration where its data are thin, because its prior has every coefficient but the axial dipole varying about zero [7 Methods]; so its early maps are fainter by construction.

What does change, in the abstract and in our re-run, is where the acceleration gathers: under the Caribbean and North America early in the century, with its strongest patches at 9°N 75°W and 41°N 67°W in 1900–1925, and at 29°N 67°W, inside the Bermuda Triangle, in 1950–1975; in the Pacific and Indian sectors after the 1970s, with the strongest patch of 1975–2000 at 7°N 83°E, off Sri Lanka; along the equator since 2000 [4][8]. Jerks were recorded throughout the century (section 2). Whether the paper reports a growth in power, and how it treats the early uncertainty beyond its abstract’s note, we could not read [1].

5 · The anomaly and the triangles

Not at the South Atlantic Anomaly, the old West Pacific Anomaly or the Dragon’s Triangle; inside the Bermuda Triangle in 1950–1975, south of it now

The episode named four concentration points [2 00:03:20–00:03:29]; the abstract names none of them [4]. We measured the acceleration at the top of the core beneath each, in the models the paper uses.

How much of the core surface has weaker acceleration than directly beneath each place

Scroll sideways to see all seven columns →

PlaceCHAOS-8.5, 2000–2025The same, to degree 13COV-OBS.x2, 1900–19251925–19501950–19751975–20002000–2020
Bermuda Triangle, centre81.6%74.4%81.3%98.8%99.7%84.7%91.7%
San Juan, one of its corners89.8%73.5%94.5%99.8%98.1%85.1%95.3%
Dragon’s Triangle, about 30°N 141°E15.3%68.6%58.0%39.4%48.9%48.1%17.9%
Old West Pacific Anomaly core, about 13°N 115°E76.5%87.9%80.0%87.8%33.7%92.0%84.7%

Our analysis of CHAOS-8.5 and COV-OBS.x2, degrees 1–9 unless stated [6][8]: the share of the core surface, by area, where the root-mean-square acceleration is weaker than directly beneath the place, 2,900 km down. 50% is the median; the strongest patches score close to 100%. The folklore regions have no fixed bounds, so we use a central point for each.

The South Atlantic Anomaly

The anomaly is the surface region where the field is weakest. In CHAOS-8.5 in 2025 it has two minima, 22,095 nT at 26.5°S 59.5°W, over northern Argentina near the border with Paraguay, and 23,452 nT at 40.5°S 0.5°W, south-west of Africa, by our re-run [6]; the CHAOS-7 team saw that second minimum develop “around 40° South on the Greenwich meridian” [10 Results]. Beneath the anomaly, at the core surface, lie patches of reversed flux “coinciding with the location of the SAA” [5 §4.5].

The anomaly has no single edge, and beneath it the acceleration is about average or weaker, depending on where the edge is drawn: 0.82 times the global level beneath the region where the surface field is below 26,000 nT, 0.96 times below 28,000 nT, and 1.08 times below 32,000 nT, the outline two of the paper’s authors used in 2021, with 28,000 nT as their alternative [18 Methods and Results]; 0.79, 0.91 and 1.00 times at degree 13; and 0.70 times in the box the same authors drew around the anomaly’s roots, by our re-run [6][18 Fig. 4]. Under no outline is it where the acceleration gathers: the strongest patch, under Colombia, is 2.8 times the global level, and the strong Atlantic and South American patches lie 32–35° of arc north of the anomaly’s main minimum.

The authors’ own earlier link

Two of the authors did tie the anomaly to jerks in 2021: minima in the acceleration of its area coincided with the jerks of the satellite era, and they concluded that “the SAA area is a region especially sensitive for the identification of geomagnetic jerks” [18 Discussion]. That is a link in timing, through the anomaly’s growth, not a finding that the acceleration concentrates there; the new abstract puts the concentration at low latitudes [4]. Whether the new paper discusses the anomaly, we could not read [1].

The West Pacific Anomaly

The name exists in the literature, for a weak-field region of the past. “The WPA separated from the main anomaly between 1540 and 1550 AD and its core, located above the South China Sea between the Philippine Archipelago and current-day Vietnam, reached its minimal intensity around 1650 AD”; then the field’s evolution caused “the WPA to shrink until disappearance after 1800 AD” [24 Introduction]. A conference abstract based on the gufm1 model dates it to 1620–1810 [25].

A 2026 conference abstract speaks of a less pronounced minimum “below the West Pacific”, with no date in its text [26]. In CHAOS-8.5 in 2025 we find none: across 30°S–40°N and 100°E–180° the surface field is at least 32,919 nT, against 22,095 nT in the South Atlantic, by our re-run [6]. The old core lies over a moderately active part of the core today (76.5% in the table), and the strongest western Pacific patch is north of New Guinea, D on the map.

If the presenter meant today’s active western Pacific rather than the published anomaly, that part comes closer to the paper: patch D is the fourth strongest on the core [6], the abstract names a Pacific sector [4], and the 2017 jerk was a Pacific event [20 Abstract].

The two Triangles

Both are folklore names with no official bounds. NOAA: “The Bermuda Triangle is a region in the western part of the North Atlantic Ocean in which ships, planes, and people are alleged to have mysteriously vanished”; “The U.S. Navy and U.S. Coast Guard contend that there are no supernatural explanations for disasters at sea”, and “no official maps exist that delineate the boundaries of the Bermuda Triangle” [27]. In 1964 Vincent Gaddis gave its corners as Miami, San Juan and Bermuda, the outline we draw, though later writers did not always follow him [28]. The Dragon’s Triangle, or Devil’s Sea, is “a region of the Pacific, south of Tokyo”, in a 1955 newspaper account from the Izu to the Ogasawara Islands [29].

In the models the paper uses, the acceleration beneath the Bermuda Triangle’s centre is stronger than over at least four-fifths of the core surface in every period we mapped, at degree 9, and among the strongest 1–2% in 1925–1975 [6][8]. In COV-OBS.x2 in 1950–1975, before the satellite era and its sharper data [4], the strongest acceleration on the whole core lay inside the triangle, at 29°N 67°W: 736 nT a year squared, with the whole triangle at 1.7 to 2.4 times the global level [8]. The CHAOS-7 team noted that “an even stronger episode of acceleration change was seen at San Juan observatory around 1970” than the recent one in the Pacific [10 Results]. In 2000–2025, in CHAOS-8.5, the triangle ranges from 0.8 to 1.5 times the global level, and the strongest patch lies 23° to its south, under Colombia [6]. The Dragon’s Triangle lies over an average-to-quiet part of the core at degree 9 in every period, and a little above average at degree 13.

Our analysis

Why “Misleading”

Part of the claim holds, for one of the four places and one era: in COV-OBS.x2 the strongest acceleration on the whole core lay inside the Bermuda Triangle in 1950–1975, and the triangle’s centre still lies over a more active part of the core than most [8][6]. The abstract, too, places the concentration in a few sectors, the Atlantic and the Pacific among them [4]. But the rest fails now. The South Atlantic Anomaly is not a concentration point: the acceleration beneath it is 0.8–1.1 times the global level, against 2.8 times at the strongest patch. The published West Pacific Anomaly disappeared after 1800 [24 Introduction]; read as today’s active western Pacific, that part comes closer, as patch D shows. The Dragon’s Triangle is ordinary; since 2000 the strongest patch lies 23° south of the Bermuda Triangle; and the Indian sector, which the abstract names and which probably holds two of the seven strongest patches, goes unmentioned. “Exactly as we expect them to be” cites no prediction, and the abstract makes none [4]. Acceleration 2,900 km down has no known bearing on ships or aircraft.

6 · No collapse, no pole shift

The equatorial field is changing by 1% in 25 years, and the paper’s record says nothing of a pole shift

The episode introduced the paper as “a new data point on the worsening situation of Earth’s ongoing magnetic pole shift” [2 00:00:18–00:00:22] and read its maps as “a full equatorial collapse” [2 00:03:13–00:03:20]. Neither is in the paper’s record or abstract [3][4], and the models the paper uses contradict the second.

Along the equator, 2000–2025

Our analysis

−0.99%

34,153 to 33,814 nT at the surface, the same as the worldwide average, 42,623 to 42,200, in CHAOS-8.5 [6]

Along the equator, 1900–2020

Our analysis

−3.8%

35,243 to 33,897 nT, less than the worldwide 6.1%, 45,032 to 42,283, in COV-OBS.x2 [8]

The dipole, 1900–2020

Our analysis

−7.2%

in COV-OBS.x2, in line with the axial dipole’s loss of about 9% since 1840 [30 p. 1]

In the episode’s words

Quoted from the automatically generated transcript; “Dragons Triangle” is its spelling.

“today we’ve got big [clears throat] stories on the ozone, solar storm deaths, and a new data point on the worsening situation of Earth’s ongoing magnetic pole shift.”

“And now our top story. We are back at the ongoing magnetic pole shift of Earth. And this team had the goal of tracking every last little bit of geomagnetic jerk evidence there is and relating it to secular variation of the field.”

“zooming in on there we can see that earth has completely lost the plot. The wheels are coming off almost coming completely off the rails in a full equatorial collapse.”

Our analysis

“A data point on a worsening shift”: Misread

The paper studies where and how the core field’s acceleration is organised. By its record and abstract it does not mention a pole shift, a reversal or an excursion, and what it reports is persistence, a few long-lived structures whose relative strengths change [3][4]. Its 2000–2025 model shows no worsening: the acceleration peaked in 2010–2015 [6]. The premise itself, “A pole shift is under way”, is rated Misleading on our review of October 20: the axial dipole is about 9% weaker than in 1840 [30 p. 1], the field is “probably not reversing” [31], and no excursion is expected for centuries (our review of the forecasts; the field’s decline). A real paper cited for what its record and abstract do not say is a misreading.

Our analysis

“A full equatorial collapse”: False

What the paper maps near the equator is acceleration, the change in the field’s rate of change, gathered in structures that persist for decades [4], the signature of waves and fast flows at the top of the core (section 3). The field itself is not collapsing there: along the equator it weakened 0.99% in 25 years, as much as everywhere else, and over the century less than the world as a whole [6][8]. The evidence contradicts the claim.

The episode’s statements beside the paper

Scroll sideways to see our verdicts →

The episode saidThe paper’s record, abstract and modelsOur verdict
“a new data point on the worsening situation of Earth’s ongoing magnetic pole shift” [2 00:00:18–00:00:22]No pole shift in the record or abstract; persistence, not worsening [4]; the acceleration peaked in 2010–2015 [6]Misread
“this team had the goal of tracking every last little bit of geomagnetic jerk evidence there is and relating it to secular variation of the field” [2 00:02:50–00:02:57]An analysis of three global field models, relating the acceleration’s regions to jerks [4]Context
“we’ve gone from sparse moderate accelerative events 120 years ago to increasing power and coverage of the variation anomaly” [2 00:02:57–00:03:04]Darker recent maps, from sharper data at fine scales; large scales flat; a satellite-era peak in 2010–2015 [8][6]Misleading
“earth has completely lost the plot … in a full equatorial collapse” [2 00:03:13–00:03:20]The equatorial field fell 0.99% in 2000–2025, as much as worldwide [6]False
“Concentration points are the South Atlantic anomaly in Bermuda Triangle and then on the West Pacific anomaly and the Dragons Triangle exactly as we expect them to be.” [2 00:03:20–00:03:29]0.8–1.1 times the global level beneath the anomaly, by its outline [6]; no West Pacific Anomaly since 1800 [24 Introduction]; the Dragon’s Triangle ordinary; the strongest patch on the core inside the Bermuda Triangle in 1950–1975, 23° south of it now [8][6]; the Indian sector left outMisleading
“the extinction risk of magnetic pole ships like the one we’re in now” [2 00:03:33–00:03:35]Not in the abstract [4]; rated on our review of the episodeSee that page

Quotations from the episode are from its automatically generated transcript, with the times at which each begins and ends [2]; “pole ships” is its rendering of pole shifts. The episode’s own framing is rated on our review of the episode.

7 · Repeat our re-run

One script, two public model files, about half a minute

Every number from our re-run in the text comes from what one script prints, a few by simple arithmetic on it (percentage changes, ratios and distances); the script also writes the grids behind the maps and the latitude chart. It needs only Python 3.

Run it

Download CHAOS-8.5_spline-coefficients.dat from the CHAOS-8.5 section of DTU Space’s CHAOS-8 archive [6] and COV-OBS.x2-int from DTU Space’s COV-OBS.x2 page [8], put them beside the script, and run python3 sa_cmb.py. It checks both files against the SHA-256 checksums of the copies we used, then prints its results in five parts, in about half a minute on a laptop. Run as python3 sa_cmb.py --grids grids, it also writes the grids behind the maps and the latitude chart, as CSV files, to a folder named grids.

How we checked it

Our code reproduces the International Geomagnetic Reference Field: CHAOS-8.5 at 2020.0 matches its definitive 2020 coefficients to within 0.5 nT in the six we compared, and our evaluation of the reference field at 0°N 120°E on August 25, 2025 matches the British Geological Survey’s calculator to the nanotesla [32]. The global pulses of acceleration it finds in CHAOS-8.5, peaking in 2006, 2009, 2012–2013, 2016 and 2021–2022, match those reported from other models [22 Abstract][16 §1] and [21 Abstract], the 2016 one by two of the paper’s authors [18 Discussion]. Kalmag, the third model the paper uses, could not be downloaded: its site did not answer.

8 · Limits

Where this check could be wrong

What would change these verdicts

  1. 01

    The paper’s text: its figures, epochs and regions, any numbers for a change in the acceleration’s power or extent, and its own caveats.

  2. 02

    A rise in the large-scale acceleration over the century, or a sustained rise through the satellite era, in independent models.

  3. 03

    A forecast, published before the paper, that the acceleration would concentrate at the South Atlantic Anomaly or under the two Triangles.

Prepared by Fact-Checked Science from the paper’s record and an index copy of its abstract, the episode’s transcript, the literature cited and our own re-run of two public field models. No figure or text comes from the paper, which we could not read. The assessment is our own and has not been reviewed by the authors.

Sources

Bracketed numbers refer to this list. For the paper [1] we could read only its bibliographic record [3] and an index copy of its abstract [4], so no locator points into its text and nothing from it is quoted. For the episode [2], times are positions in the video (hours:minutes:seconds), from the first to the last transcript line of each quotation. For the other sources, § gives a numbered section and Fig. a figure, p. a page, and Abstract, Introduction, Methods, Methods and Results, Results, Results and discussion, Discussion and Conclusions name those parts of a source; where only an abstract was read, the entry says so, and abstracts read only through an index are paraphrased, not quoted. Years are those of the printed issue where there is one, otherwise of online publication. Links checked October 25, 2026.

  1. 1

    Serrano, M., Campuzano, S. A., and Pavón-Carrasco, F. J. (2026). Persistent regional patterns of secular acceleration at the core–mantle boundary and their relation to geomagnetic jerks over the past 125 years. Physics of the Earth and Planetary Interiors 381, 107615. Online October 5, 2026; volume dated December 2026. Peer reviewed; open access, CC BY 4.0. We could not read it: ScienceDirect met our tools and a real browser with a CAPTCHA, which the site’s rules do not let us pass, and we found no other copy. Abstract only, from an index copy [4], paraphrased on this page and not quoted; no figure is redrawn.

  2. 2

    SpaceWeatherNews (S0) (October 24, 2026). Earth Field Chaos, Solar Storm Deaths | S0 News Oct.24.2026. YouTube video, https://www.youtube.com/watch?v=e526hqAnTwg. Quoted from the automatically generated transcript; times are positions in the video. The label “Growing/Spreading Jerk Variation” is the episode’s own, on its link to the paper in the video’s description.

  3. 3

    Crossref. Metadata record for doi:10.1016/j.pepi.2026.107615. Retrieved October 25, 2026: record created October 5, 2026; volume 381, article 107615, issued December 2026; 49 references; funders the Community of Madrid (PIPF-2022/TEC-25005) and Spain’s Ministry of Science and Innovation (RYC2023-044408-I and PID2024-159020NB-I00); CC BY 4.0 for the version of record; no abstract deposited.

  4. 4

    OpenAlex. Work W7220411698. Retrieved October 25, 2026: publication date October 5, 2026; open access at the publisher, no repository copy; the authors’ affiliations; and the paper’s abstract, held as an index copy, which we paraphrase.

  5. 5

    Kloss, C., Finlay, C. C., Olsen, N., Tøffner-Clausen, L., Gillet, N., and Grayver, A. (2026). The CHAOS-8 geomagnetic field model. Earth, Planets and Space 78, 21. CC BY 4.0. We read the full text.

  6. 6

    Kloss, C., Finlay, C. C., Olsen, N., Tøffner-Clausen, L., Gillet, N., and Grayver, A. (2026). CHAOS-8 geomagnetic field model, version 8.5 (data set). Zenodo; distributed by DTU Space: release of February 2, 2026, with satellite data to December 2025, the file CHAOS-8.5_spline-coefficients.dat from the CHAOS-8 archive. CC BY 4.0. The model the paper uses for 2000–2025, re-run by us.

  7. 7

    Huder, L., Gillet, N., Finlay, C. C., Hammer, M. D., and Tchoungui, H. (2020). COV-OBS.x2: 180 years of geomagnetic field evolution from ground-based and satellite observations. Earth, Planets and Space 72, 160. CC BY 4.0. We read the full text (Europe PMC).

  8. 8

    Huder, L., Gillet, N., Finlay, C. C., Hammer, M. D., and Tchoungui, H. (2020). COV-OBS.x2 geomagnetic field model: the mean internal model, file COV-OBS.x2-int, distributed by DTU Space; the download page states no licence. One of the two models the paper uses for the 20th century, re-run by us.

  9. 9

    Baerenzung, J., Holschneider, M., Saynisch-Wagner, J., and Thomas, M. (2022). Kalmag: a high spatio-temporal model of the geomagnetic field. Earth, Planets and Space 74, 139. CC BY 4.0. We read the full text; the model’s own files could not be downloaded.

  10. 10

    Finlay, C. C., Kloss, C., Olsen, N., Hammer, M. D., Tøffner-Clausen, L., Grayver, A., and Kuvshinov, A. (2020). The CHAOS-7 geomagnetic field model and observed changes in the South Atlantic Anomaly. Earth, Planets and Space 72, 156. CC BY 4.0. We read the full text (Europe PMC).

  11. 11

    Finlay, C. C., Olsen, N., Kotsiaros, S., Gillet, N., and Tøffner-Clausen, L. (2016). Recent geomagnetic secular variation from Swarm and ground observatories as estimated in the CHAOS-6 geomagnetic field model. Earth, Planets and Space 68, 112. CC BY 4.0. We read the full text.

  12. 12

    Aubert, J., and Finlay, C. C. (2019). Geomagnetic jerks and rapid hydromagnetic waves focusing at Earth’s core surface. Nature Geoscience 12, 393–398. We read the abstract on the publisher’s page.

  13. 13

    Gillet, N., Gerick, F., Angappan, R., and Jault, D. (2022). A dynamical prospective on interannual geomagnetic field changes. Surveys in Geophysics 43, 71–105. CC BY 4.0. We quote its abstract from the publisher’s page.

  14. 14

    Mandea, M., Holme, R., Pais, A., Pinheiro, K., Jackson, A., and Verbanac, G. (2010). Geomagnetic jerks: rapid core field variations and core dynamics. Space Science Reviews 155, 147–175. We read the abstract on the publisher’s page.

  15. 15

    Alexandrescu, M., Gibert, D., Hulot, G., Le Mouël, J.-L., and Saracco, G. (1996). Worldwide wavelet analysis of geomagnetic jerks. Journal of Geophysical Research: Solid Earth 101(B10), 21975–21994. Abstract only, read through an index (OpenAlex) and paraphrased.

  16. 16

    Torta, J. M., Pavón-Carrasco, F. J., Marsal, S., and Finlay, C. C. (2015). Evidence for a new geomagnetic jerk in 2014. Geophysical Research Letters 42(19), 7933–7940. We read the full text in an author’s copy; § 1 is its introduction.

  17. 17

    Brown, W. J., Mound, J. E., and Livermore, P. W. (2013). Jerks abound: an analysis of geomagnetic observatory data from 1957 to 2008. Physics of the Earth and Planetary Interiors 223, 62–76. CC BY 3.0. Abstract only, read through an index (OpenAlex) and paraphrased.

  18. 18

    Campuzano, S. A., Pavón-Carrasco, F. J., De Santis, A., González-López, A., and Qamili, E. (2021). South Atlantic Anomaly areal extent as a possible indicator of geomagnetic jerks in the satellite era. Frontiers in Earth Science 8, 607049. CC BY 4.0. By two of the paper’s authors; we read the full text.

  19. 19

    Pavón-Carrasco, F. J., Marsal, S., Campuzano, S. A., and Torta, J. M. (2021). Signs of a new geomagnetic jerk between 2019 and 2020 from Swarm and observatory data. Earth, Planets and Space 73, 175. CC BY 4.0. We read the full text.

  20. 20

    Whaler, K. A., Hammer, M. D., Finlay, C. C., and Olsen, N. (2022). Core surface flow changes associated with the 2017 Pacific geomagnetic jerk. Geophysical Research Letters 49(15), e2022GL098616. CC BY 4.0. We read the full text (Europe PMC).

  21. 21

    Chulliat, A., Nair, M., and Califf, S. (2026). Candidate geomagnetic field models for IGRF-14 and secular acceleration since 2020. Earth, Planets and Space 78, 26. CC BY 4.0. We read the full text.

  22. 22

    Chulliat, A., and Maus, S. (2014). Geomagnetic secular acceleration, jerks, and a localized standing wave at the core surface from 2000 to 2010. Journal of Geophysical Research: Solid Earth 119(3), 1531–1543. Abstract only, read through an index (OpenAlex) and paraphrased.

  23. 23

    Bai, C., Gao, G., Wen, L., and Kang, G. (2024). Dynamic evolution of amplitude and position of geomagnetic secular acceleration pulses since 2000. Frontiers in Earth Science 12, 1383149. CC BY 4.0. We read the full text.

  24. 24

    He, F., Wei, Y., Maffei, S., Livermore, P. W., Davies, C. J., Mound, J., Xu, K., Cai, S., and Zhu, R. (2021). Equatorial auroral records reveal dynamics of the paleo-West Pacific geomagnetic anomaly. Proceedings of the National Academy of Sciences 118(20), e2026080118. CC BY 4.0. We read the full text (Europe PMC) and quote its introduction, an unheaded opening section.

  25. 25

    Yue, Y., Gao, J., Wei, Y., He, F., Cai, S., Wang, H., Wang, Y., and Rong, Z. (2024). Evolution and disappearance of the paleo-West Pacific Anomaly: implications to the South Atlantic Anomaly. EGU General Assembly 2024, abstract EGU24-13711. Conference abstract, not peer reviewed. CC BY 4.0.

  26. 26

    Yue, Y., Amit, H., Terra-Nova, F., Marum, V., Wang, Y., and Wei, Y. (2026). Time dependence of multiple geomagnetic surface intensity minima. EGU General Assembly 2026, abstract EGU26-12741. Conference abstract, not peer reviewed. CC BY 4.0.

  27. 27

    NOAA Ocean Service. What is the Bermuda Triangle? National Ocean Service fact page. Retrieved October 25, 2026.

  28. 28

    Wikipedia. Bermuda Triangle, revision of October 25, 2026. CC BY-SA 4.0. Used only for the region’s usual bounds.

  29. 29

    Wikipedia. Devil’s Sea, revision of July 10, 2026. CC BY-SA 4.0. Used only for the region’s location.

  30. 30

    Finlay, C. C., Aubert, J., and Gillet, N. (2016). Gyre-driven decay of the Earth’s magnetic dipole. Nature Communications 7, 10422. CC BY 4.0. Cited as on our page When Earth’s magnetic field falters.

  31. 31

    Brown, M., Korte, M., Holme, R., Wardinski, I., and Gunnarson, S. (2018). Earth’s magnetic field is probably not reversing. Proceedings of the National Academy of Sciences 115(20), 5111–5116. Cited as on our page When is the next magnetic excursion due?.

  32. 32

    IAGA Working Group V-MOD (2024). International Geomagnetic Reference Field, 14th generation (IGRF-14): Gauss coefficients 1900–2025 and secular variation 2025–2030. Data file, NOAA National Centers for Environmental Information; checked against the British Geological Survey’s online geomagnetic calculator. Used only to test our code.

  33. 33

    Natural Earth. Coastline, 1:110 million scale. Public domain. Used for the outlines on our maps.

Method

How this was checked

On October 25, 2026, we tried every route the site allows to the paper’s text: the episode’s link, the publisher’s PDF link and the DOI, through our tools and in a real browser; Crossref, OpenAlex and Semantic Scholar; Elsevier’s text-mining interface; the Complutense University’s repository; Europe PMC; arXiv; and web searches for the title. The publisher’s site met our tools and our browser with a CAPTCHA, which we did not attempt, and we found no other copy. So the page rests on the paper’s bibliographic record and an index copy of its abstract, the episode’s automatically generated transcript, the papers behind the three field models it uses (read in full), 19 other papers and conference abstracts on jerks, the core’s acceleration, the field’s decline, the South Atlantic Anomaly and the West Pacific Anomaly (11 read in full, 6 as abstracts and 2 cited as on our earlier pages), NOAA’s page on the Bermuda Triangle, two Wikipedia entries for the folklore regions, and our own re-run of two of those three models: about 250 numbers, quotations and statements in all, gathered on October 25 and 26, 2026. Every quotation was checked word for word against its saved source; abstracts that reached us only through an index, the paper’s among them, are paraphrased, not quoted.

The verdicts and the items tagged “Our analysis” are our own assessment, built on the cited sources. They include every number from our re-run: where the acceleration at the top of the core is strongest, its share near the equator, its strength beneath the South Atlantic Anomaly, however its edge is drawn, the West Pacific Anomaly’s old core and the two folklore triangles, its global strength by scale and decade, and the surface field along the equator and worldwide; the distances between those places; and our reading of what the paper’s record and abstract do and do not say.

We re-ran CHAOS-8.5, the model the paper uses for 2000–2025, and COV-OBS.x2, one of the two it uses for the 20th century, from the files their makers publish, with our own code, which reproduces the International Geomagnetic Reference Field’s published coefficients and the British Geological Survey’s calculator; the script is on this page. Kalmag, the third model, could not be downloaded. Nothing from the paper itself could be recomputed, because its text could not be read: its figures, its component analysis, its jerk epochs and its regions’ coordinates are unverified, and no figure of it is redrawn.

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.

The verdicts on this page

False
The evidence contradicts the claim.
Misleading
Contains something true but leaves a false impression, usually by leaving out key context.
Misread
Cites a real source but gets wrong what it says.
True
The evidence supports the claim.
Context
A question or aside rather than a claim. We add the facts it leaves out.

All ten verdicts, and how we check

Corrections

Corrections and changes

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.

Change log

  1. Published on factchecked.science after a full check against the sources. Built from the paper’s bibliographic record and an index copy of its abstract, because its text could not be read, together with the episode’s transcript, the literature on geomagnetic jerks and the core’s acceleration, and our own re-run of two of the paper’s three public field models; no figure or text from the paper is reproduced.

Suggest a correction

Spotted an error, or know a better source? Tell us here. Submissions aren’t published.