SF-0023Myth or Evidence?
Yellowstone is not overdue, and the recurrence arithmetic gives it another 100,000 years
The USGS computes the average interval between Yellowstone's eruptions — 725,000 years — calls it meaningless because it rests on two intervals, and notes that even if you trust it, there are about 100,000 years to go.
A thermal pool that did not exist in 2024. The US Geological Survey records that it formed in a series of mildly explosive events between late December 2024 and early February 2025, and that the rocks and white silica mud around it were probably thrown out as it formed. It is about four metres across — the ordinary scale of what actually happens at Yellowstone.
View looking northwest at a new thermal pool in the Porcelain Basin area of Norris Geyser Basin, Yellowstone National Park, that probably formed in a series of mildly explosive events between late December 2024 and early February 2025. The rocks and white material (silica mud) surrounding the pool were probably ejected as the feature formed. The pool is about 4 meters (13 feet) across.
- Image
- Mike Poland
- Taken
- 2025-05-12
- Location
- Porcelain Basin, Norris Geyser Basin, Yellowstone National Park, Wyoming
- Archive ref.
- USGS media library — PXL_20250512_174659809
- Licence
- Public domain as a work of the US federal government, corroborated by the item's Sources/Usage statement (licence record)
- The image itself
- Authentic, unaltered
New thermal pool, Porcelain Basin, Norris Geyser Basin, Yellowstone National Park. Mike Poland, US Geological Survey, 12 May 2025. Public domain. Cropped.
Verdict
"Yellowstone goes off every 600,000 years and the last one was 640,000 years ago. We're overdue."
The claim is demonstrably untrue, and the evidence against it is clear.
The claim is not a distortion of the science; it is contradicted by the agency that does the science, in a sentence written to answer exactly this question.
ESTABLISHED, and not in dispute here: Yellowstone is an active volcanic system, it has produced three caldera-forming eruptions, and those eruptions are dated to about 2.08 million, 1.3 million and 0.631 million years ago. The USGS states all of this plainly.
THE FIRST FAILURE — on its own terms. Those three dates give two intervals, averaging about 725,000 years. The most recent eruption was about 631,000 years ago. Subtract, and the popular arithmetic does not produce "overdue" at all: the USGS writes that on this basis "there is still about 100,000 years to go." The claim gets the sign wrong on its own calculation. The circulating version usually rounds the cycle down to 600,000 years to close that gap.
THE SECOND FAILURE — on the terms it should have used. The USGS immediately disowns the average it just computed: it "is based on the average of just two time intervals between the eruptions, which is meaningless." Michael Poland, the scientist-in-charge of the Yellowstone Volcano Observatory, gives the mechanism: volcanoes are not like faults, and "with rare exceptions, volcanoes do not accumulate magma at a constant rate." Magma arrives "only in discontinuous batches". A system that is not filling steadily cannot be due.
So the verdict is False rather than Misleading. Misleading would be the right call if the claim were a true fact wrapped in a bad frame. Here the load-bearing assertion — that Yellowstone is overdue — is one the responsible agency denies in those words.
False requires 1 primary or authoritative source. This file cites 5.
In short
Yellowstone is said to erupt every 600,000 to 700,000 years, with the last 640,000 years ago, so it is overdue. The US Geological Survey answers directly: "Yellowstone is not overdue for an eruption." The claim fails twice. It fails on its own method: average the real intervals, subtract the real elapsed time, and the next eruption is about 100,000 years away rather than in the past. And it fails on the arithmetic it should have used: volcanoes do not accumulate magma at a constant rate, and a mean drawn from two intervals is not a schedule.
The claim does not survive its own arithmetic
Three caldera-forming eruptions at Yellowstone are real, and dated: about 2.08 million, 1.3 million, and 0.631 million years ago. Two intervals. They average about 725,000 years.
The last one was about 631,000 years ago.
Take the average seriously as a schedule — which is exactly what the claim does — and the next eruption is not late. The USGS does the subtraction on its own page and reports the result: there is still about 100,000 years to go.
The circulating version gets to “overdue” by rounding the cycle down to 600,000 years and the elapsed time up to 640,000. Two small roundings in opposite directions, and a 100,000-year surplus becomes a 40,000-year debt.
And then the arithmetic itself is withdrawn
Having computed the average, the USGS immediately says what it is worth:
…this is based on the average of just two time intervals between the eruptions, which is meaningless.
Two intervals cannot establish a period. And when supereruptions do recur in a volcanic system, the agency notes, “they are not evenly spaced in time.”
Why “overdue” is the wrong kind of word
The clearest statement of the mechanism comes from Michael Poland, the scientist-in-charge of the Yellowstone Volcano Observatory, in a 2019 article whose title does most of the work:
“Overdue” can apply to library books, bills, and oil changes, but it does not apply to Yellowstone!
The reason is that volcanoes are not like faults, which do load at a roughly steady rate:
With rare exceptions, volcanoes do not accumulate magma at a constant rate.
Yellowstone’s system receives magma “only in discontinuous batches, causing several eruptions in a short period of time with long periods of quiet in between”. Eruption depends on there being enough liquid magma and enough pressure to drive it upward — supply and state, not elapsed time. Nothing is accruing. Nothing can be owed.
What is actually likely
Smaller and stranger than the story. The most likely eruption at Yellowstone is not volcanic at all:
The most likely type of eruption would not be volcanic but, rather, hydrothermal.
Steam and hot water rather than molten rock — the same shallow reservoirs that feed the geysers. These happen every few years and leave craters a few metres across; every few thousand years, one leaves a crater a few hundred metres across. If a volcanic eruption does occur, lava flows of rhyolite or basalt are the likely form.
The magma reservoir beneath the caldera is currently about 5–15% molten, and YVO reports no monitoring signal suggesting an eruption is imminent.
What this file does not claim
That Yellowstone is dormant, harmless, or unworthy of a dedicated observatory. It has one, staffed, because the system is active and capable of large eruptions.
That the odds are precisely 1 in 730,000 per year. That figure is published by the USGS and is quoted here for scale, but it is calculated from the same three-eruption record the agency calls too sparse to support a recurrence interval. Treating it as a measurement would repeat the claim’s own mistake while appearing to correct it.
And it does not claim that another caldera-forming eruption will never happen. The USGS is careful in both directions: scientists are not convinced there will be another one. “Not overdue” is a statement about a schedule that does not exist — not a forecast of quiet.
The evidence
The USGS answers the question in its own words: "Yellowstone is not overdue for an eruption."
Cited sources:S1
The premise the claim rests on is denied directly: "Volcanoes do not work in predictable ways and their eruptions do not follow predictable schedules."
Cited sources:S1
The three caldera-forming eruptions are real and dated — about 2.08 million, 1.3 million and 0.631 million years ago — giving an average of about 725,000 years between them.
Cited sources:S1
On that same average, the claim's own arithmetic points the other way. The USGS writes that "there is still about 100,000 years to go."
Cited sources:S1
The USGS then disowns the average it has just computed: it "is based on the average of just two time intervals between the eruptions, which is meaningless." It adds that when supereruptions do occur more than once in a system, "they are not evenly spaced in time."
Cited sources:S1
The mechanism, from the Yellowstone Volcano Observatory's scientist-in-charge: volcanoes are unlike faults, and "with rare exceptions, volcanoes do not accumulate magma at a constant rate." Yellowstone's system receives new material "only in discontinuous batches, causing several eruptions in a short period of time with long periods of quiet in between."
Cited sources:S2
Eruption requires supply and pressure, not elapsed time: volcanoes "erupt when there is a sufficient supply of liquid magma in the subsurface and sufficient pressure to cause that magma to ascend to the surface."
Cited sources:S2
What is actually most likely is not a volcanic eruption at all. "The most likely type of eruption would not be volcanic but, rather, hydrothermal" — from shallow reservoirs of steam or hot water rather than molten rock.
Cited sources:S3
Those events are small and routine: hydrothermal explosions "occur in Yellowstone National Park every few years and form a crater a few meters across", with craters up to a few hundred metres across every few thousand years.
Cited sources:S3
A worked example, documented by the USGS in 2025: a new thermal pool in Porcelain Basin "probably formed in a series of mildly explosive events between late December 2024 and early February 2025", throwing out the rocks and silica mud around it. It is "about 4 meters (13 feet) across" — the ordinary scale of a Yellowstone event, and it happened while the "overdue" claim was circulating.
Cited sources:S5
The yearly probability of another caldera-forming eruption is given as about 1 in 730,000, or 0.00014% — described as roughly similar to the chance of a one-kilometre asteroid striking Earth.
Cited sources:S4
The subsurface does not currently look like a system about to erupt: the magma reservoir is only about 5–15% molten, and YVO reports no monitoring signs suggesting an eruption is imminent.
Evidence that cuts the other way
Findings that complicate or argue against the verdict above. They are listed because leaving them out would make the file look stronger than it is.
Yellowstone is an active volcano and this file does not suggest otherwise. It is monitored continuously by a dedicated USGS observatory precisely because it is capable of large eruptions, and the same sources that reject "overdue" describe lava flows and ash hazards as real possibilities.
The arithmetic in the claim is not fabricated. The eruption dates, the roughly 725,000-year average and the framing of it as a recurrence interval all appear on USGS pages. What the agency adds — that two intervals cannot establish a period — is the part that does not survive the retelling.
Cited sources:S1
The 1-in-730,000 annual figure inherits the same weakness as the claim it is used to answer. It is derived from the same three-eruption record whose sparseness the USGS calls meaningless, so it is best read as an order-of-magnitude statement about rarity rather than a precise odds quote. This file cites it as framing and does not treat it as a measurement.
"Not overdue" is not "will not happen". The USGS notes that scientists are not convinced another caldera-forming eruption will occur at all, which is a statement of uncertainty in both directions rather than reassurance.
Cited sources:S1
What we still don’t know
- The annual probability figure is not a measurement. It is derived from the same three-eruption record the USGS elsewhere calls too sparse to support a recurrence interval, so quoting it to four significant figures gives it a precision the underlying data does not have. It belongs in this file as an indication of rarity and nothing more.
- Whether a fourth caldera-forming eruption will happen at all is open. The USGS notes that scientists are not convinced one will, and this file does not treat "not overdue" as a prediction that nothing will ever occur.
- Melt fraction figures vary between studies and imaging methods. The 5–15% range is what the USGS states; this file uses that number and does not attempt to adjudicate between seismic tomography results that have produced different estimates.
- The date of the most recent caldera-forming eruption is given here as about 631,000 years ago, following the USGS. Popular retellings usually say 640,000. The difference is immaterial to the argument and is noted only because the rounder number is the one that circulates, and rounding the interval down while rounding the elapsed time up is precisely how "100,000 years to go" becomes "overdue".
- Hydrothermal explosion frequency is stated qualitatively by the USGS — "every few years" for metre-scale craters — and this file does not have a catalogue of dated events behind that phrase.
Sources
Every source below was read in full before it was cited. Archive snapshots are recorded at research time, because a file whose sources have evaporated is indistinguishable from one that never had any.
S1Tier 1Read in full
Is Yellowstone overdue for an eruption? When will Yellowstone erupt?
US Geological Survey
The spine of this file. The agency's direct answer to the exact question, establishing that Yellowstone is not overdue; that eruptions do not follow predictable schedules; the three eruption dates and the ~725,000-year average; the "about 100,000 years to go" consequence; that the average rests on two intervals and is meaningless; that supereruptions are not evenly spaced; and the 5–15% melt fraction.
S2Tier 1Read in full
"Overdue" can apply to library books, bills, and oil changes, but it does not apply to Yellowstone!
Michael Poland · US Geological Survey — Yellowstone Volcano Observatory · 2019-03-25
The mechanism behind the refusal, written by YVO's scientist-in-charge: that volcanoes are unlike faults, do not accumulate magma at a constant rate, receive magma in discontinuous batches, and erupt on supply and pressure rather than elapsed time. Cited for why "overdue" is the wrong category of word, not merely the wrong number.
S3Tier 1Read in full
Questions About Future Volcanic Activity at Yellowstone
US Geological Survey
What is actually expected instead: that the most likely eruption is hydrothermal rather than volcanic, from shallow steam or hot-water reservoirs; the size and frequency of those events; that rhyolite or basalt lava flows are the most likely volcanic case; and that the probability of a caldera-forming eruption in any century or millennium is "exceedingly low".
S4Tier 1Read in full
Questions About Supervolcanoes
US Geological Survey
The annual probability framing — approximately 1 in 730,000, or 0.00014%, compared to a one-kilometre asteroid impact. Cited for scale, and explicitly flagged in the contradicting evidence as inheriting the same sparse-record weakness as the claim it answers.
S5Tier 1Read in full
New hydrothermal feature near "Tree Island" in Norris Geyser Basin, Yellowstone National Park
Mike Poland · US Geological Survey · 2025-05-12
This file's photograph and its rights record — the credit line naming the agency, and the item-level public-domain statement. Cited separately as evidence for its caption, which documents a specific recent hydrothermal event: a pool about 4 metres across formed by mildly explosive activity between late December 2024 and early February 2025. Not independent of S2: the photographer is the same Michael Poland who wrote that article. It is treated as one voice, not two.
Checked for circularity: S1 and S2; S3 and S1; S3 and S2; S4 and S2; S5 and S1; S5 and S3; S5 and S4 were confirmed not to derive from one another. Two sources that trace back to the same origin are one source, and counting them twice is the most common way a false claim looks well-evidenced.
How we checked this
What we went looking for, and what we found
Three checks, one of which changed the verdict and one of which added a source.
First, against the verdict being Misleading. Misleading was the natural reflex, since the claim is built from real dates and a real average. It was rejected because the load-bearing assertion is not a true fact in a bad frame — it is a statement the responsible agency denies in the same words the claim uses. "Yellowstone is not overdue for an eruption" leaves no room for the claim to be technically correct.
Second, and this is the finding that makes the file: the claim was tested on its own arithmetic rather than only against the agency's objection to it. It fails there too. The average the claim depends on puts the next eruption roughly 100,000 years away, not in the past. That is a stronger result than "the statistics are invalid", because it does not require the reader to accept anything about volcanology — only subtraction.
Third, against the correction overshooting into reassurance. Searching for what the USGS says IS likely produced the hydrothermal material and the annual-probability figure, and also the caution that scientists are not convinced another caldera-forming eruption will occur at all. The 1-in-730,000 figure was then deliberately demoted in the file: it derives from the same sparse record the agency calls meaningless, so citing it as precise odds would have repeated the claim's own error while appearing to refute it. It is recorded in the contradicting evidence for that reason.
- We searched for evidence against our own conclusion
- We traced where the claim started
- The claim is stated in its strongest form, not a weak version
- The photograph was verified independently of the story
- We checked whether this was settled after the popular version froze
- Image rights were traced to a stated licence
- We checked what we were leaving out
- The verdict sat before publication, and was re-read cold
- Confidence in the verdict
- high
- Verdict assigned
- Most recent authoritative source
- Researched by
- The Strange File
RESEARCH AND IMAGE COMPLETE. TWO GATES OUTSTANDING, AND NEITHER MAY BE ASSUMED.
The text, evidence, sources, verdict, social copy and image are finished. Five tier-1 USGS pages were read in full, and the photograph carries a us-federal-work rights basis established by reading the credit block for literal text: "U.S. Geological Survey photo by Mike Poland, May 12, 2025."
The image was chosen to argue the file's own point rather than to decorate it. The obvious picture for a Yellowstone piece is a caldera panorama or a doom shot; this is a four-metre pool that formed in early 2025 from mildly explosive activity — the ordinary scale of what actually happens there, photographed fifteen months ago. Its USGS caption is cited as evidence in its own right, as source S5.
One coincidence is recorded rather than leaned on: the photographer is the same Michael Poland who wrote source S2. S5 is therefore marked as not independent of S2 and the two are treated as one voice.
COOLING-OFF HAS NOW RUN, by an independent session on 2026-08-05. It re-fetched every USGS quote rather than trusting this file's transcription, found them all verbatim, and found one defect in the headline's arithmetic. Recorded in full under coolingOffNotes.
OUTSTANDING — the owner has not approved publication, and DRAFT is nowhere near permission to publish. The calendar side of the cooling-off gate also does not clear until the day after the verdict was assigned.
ARCHIVE SNAPSHOTS — all sources carry an archiveUrl. Corrected 2026-08-05: an earlier pass recorded some of them as archiveCheck: none-exists, on the strength of the Internet Archive's availability API, which had answered that no capture existed. It was wrong, and so was the note that called that "a fact about the archive's coverage". The CDX API lists captures for every one of them. The availability endpoint returns an empty answer for URLs the archive demonstrably holds, and an empty answer is indistinguishable from an authoritative no. `npm run archive` now queries CDX for exactly this reason.
No sensitivity flag. The subject is a statistical claim about a volcano. The only individual named is the USGS scientist cited as a source and credited as the photographer, in both cases because the agency names him.
Part of a collection
The Correction OvershootsDebunking has a second error, and almost nobody guards against it.
The Qualifier Is The StoryThe interesting content is in the clause everyone skims.
Also related
Nobody has settled how Devils Tower formed, and the Park Service says so
The rock is phonolite porphyry and the columns are cooling cracks — that much is agreed. How the magma got there is not. The NPS lists four origin hypotheses and says discrepancies remain for all of them.
A 30-metre rock made a 1.2-kilometre hole, and a mineral made impact craters recognisable
The USGS puts the impactor at about 30 metres across and 100,000 tons, arriving at 12–20 km/sec with the force of over two million tons of TNT. The hole it left is 1.2 km wide and 180 m deep.