Life After People: The Preservation Experiment

People asking whether an earlier civilization could have vanished without trace are really asking about preservation. Running the experiment on ourselves is the cleanest way to calibrate it, because we know exactly what is there to be lost. Most of us badly overestimate how long objects last, and badly underestimate how long chemistry lasts.

• Steel-reinforced concrete fails on century timescales as rebar corrosion spalls the surrounding concrete.
• Deep excavations, tunnels and mine systems flood within months to years once pumping stops.
• Ceramics, glass, gold and platinum-group metals are among the most chemically durable things we make.
• Voids cut into bedrock — mines, tunnels, quarries — are structurally more durable than surface buildings.
• Sediments of the mid-20th century onward carry globally correlated markers: fly-ash spherules, artificial radionuclides, elevated heavy metals, fixed nitrogen and plastic particles.
• Plutonium-239 has a half-life of roughly 24,100 years; caesium-137 roughly 30 years, so the radionuclide signature changes composition over time rather than simply fading.
• Burning fossil carbon produces a measurable shift in the carbon isotope ratio recorded in sediments.
• Current species extinction and redistribution rates, and the global abundance of a few domesticated vertebrates, constitute a distinctive palaeontological signature.

• How long specific markers remain diagnostic after diagenesis, subduction, erosion and burial — estimates vary widely by depositional setting.
• Whether the Anthropocene should be formally ratified as a geological epoch and where its boundary sits; the stratigraphic markers are documented, the formal classification is contested.
• How much of the plastic signal will survive as identifiable material versus as chemically altered residue.
• How confidently a future investigator could distinguish an intelligent cause from an unusual natural excursion without additional structural evidence.

• ~100 years

  Likely lost

  Paper, film, magnetic and optical media; wooden structures in wet climates; every powered system and digital archive; flat roofs, then floors.

  May survive

  Most masonry and steel frames still standing but corroding; road and rail alignments; the overall street plan of every city.

  Planetary signature

  Atmospheric and sediment markers still accumulating; no geological perspective yet — this is a ruin, not a record.

• ~1,000 years

  Likely lost

  Almost all reinforced concrete, as rusting rebar splits it apart; bridges; most exposed metals in humid and coastal settings.

  May survive

  Stone foundations, brick rubble, ceramics, glass, gold and platinum-group metals; anoxic landfill contents; canals, embankments and drainage patterns.

  Planetary signature

  A distinct urban chemical horizon: concentrated metals, fired clay, fly-ash particles, disturbed ground.

• ~10,000 years

  Likely lost

  Nearly all surface structures as recognisable buildings; anything in eroding uplands or glaciated regions may be removed entirely.

  May survive

  Buried ceramics and glass fragments; deep foundations; mine and tunnel voids; waterlogged and desert-preserved organics.

  Planetary signature

  Cities read as low mounds with anomalous mineralogy; the radionuclide and isotope layers are still clearly measurable.

• ~100,000 years

  Likely lost

  Objects, as a useful category. Sea level has moved tens of metres, coastlines migrated, metres of surface eroded.

  May survive

  Voids cut into bedrock; deeply buried deposits in subsiding basins and deltas; the most chemically stable ceramics.

  Planetary signature

  A thin globally correlated layer: heavy metals, fixed nitrogen, fly-ash spherules, plutonium-239 and its decay products, plastic residues.

• Millions of years

  Likely lost

  Any expectation of recognisable artefacts; most sediment of our age is eroded, subducted or metamorphosed.

  May survive

  Sections preserved in the right depositional settings — a small fraction of the surface, but not zero.

  Planetary signature

  A carbon-isotope excursion, synthetic and transuranic traces, an extinction and redistribution pulse, and a fossil assemblage dominated worldwide by a handful of domesticated species.

The traces this subject should have left, listed before we ask what we actually find.

• Geochemical anomalies in sediment: lead, mercury, copper, cadmium, fixed nitrogen, fly-ash spherules.
• Artificial radionuclides with long half-lives, notably plutonium-239 at roughly 24,100 years.
• A negative carbon-isotope excursion from burning isotopically light fossil carbon.
• Novel synthetic materials and their breakdown products, including plastic-sediment composites.
• A biological signature: extinction pulse, global species redistribution, and domesticated animals appearing on every continent at once.
• Mines, tunnels and quarries — voids in bedrock rather than objects on it.

Preservation bias: the processes that erase, degrade or relocate the evidence.

• Corrosion and freeze-thaw dismantle metal and concrete within centuries.
• Glaciation can scrape a region's entire surface record away.
• Erosion in uplands removes metres of section; subduction removes it permanently.
• Sea-level change drowns or exposes coastal evidence, where most of our population lives.
• Short-lived radionuclides such as caesium-137 decay to nothing within centuries, changing the signature's composition over time.
• Diagenesis and burial alter chemistry, blurring sharp markers into gradients.

Concrete findings that would strengthen, weaken or change this assessment.

• Not applicable in the usual sense — this is our own record. The useful test is the reverse: which of these markers has anyone actually searched for in older strata?
• Higher-resolution dating of candidate horizons, since event duration is one of the few things separating an industrial spike from a volcanic one.
• Synthetic organic molecules or transuranic isotopes in deep-time sediments, which have no known natural source.

This is a thought experiment, not a prediction and not a claim about the past. Suppose modern industrial civilization ended today — no maintenance, no pumping, no repair, no one left to curate anything. What would a geologist find at 100 years, 1,000, 10,000, 100,000 and a few million? The exercise is useful because it calibrates intuition. Most people badly overestimate how long objects last and badly underestimate how long chemistry lasts. At roughly a century, the built environment is failing but obvious. Steel-framed buildings corrode where water enters; flat roofs fail first, then floors. Subway systems and deep mines flood within months once pumps stop. Wooden structures are largely gone in wet climates. Vegetation reclaims paved surfaces through frost cracking and root wedging. Paper, film, magnetic media and optical discs are substantially degraded, and every digital archive is unreadable for the simpler reason that nothing has power. Cities remain unmistakable, but their information content is already collapsing faster than their masonry. By around a thousand years, almost all steel-reinforced concrete has spalled and slumped, because rusting rebar expands and splits the concrete that protects it. Glass survives well but is buried or shattered. Most metals in humid or coastal settings are corroded or reduced to stains, while gold, platinum-group metals and some stainless alloys persist. Bridges are down. What survives best are things that were massive, chemically stable and already buried: stone foundations, ceramics, brick rubble, landfill contents in anoxic conditions, quarry and mine voids, canal cuts, road embankments, land drainage patterns. At ten thousand years, the object-scale record is thin and depends almost entirely on burial environment. Ceramics and glass persist as fragments because they are already close to geologically stable materials. Buried anoxic deposits — bogs, landfills, deep sediment, some caves — preserve organics that would have vanished at the surface. Above ground, in a temperate climate, a city is a low mound with an anomalous mineral and metal chemistry. In a glaciated region, it may be scraped away entirely. In a subsiding delta, it may be tens of metres down and superbly preserved. At a hundred thousand years and beyond, discussing objects is mostly the wrong frame. Sea level has moved tens of metres more than once, coastlines have migrated, ice sheets have advanced and retreated, and continental erosion has removed metres of surface. What survives is not artefacts but signatures: the shape and chemistry of what we did to the planet. Mines and tunnels are among the most durable single structures we have made, because they are voids in bedrock rather than objects on it. Anomalously altered sediments — sudden increases in lead, mercury, copper, cadmium, spherical carbonaceous fly-ash particles from fossil fuel combustion, nitrogen from industrial fertiliser — form a thin but globally correlated layer. Some markers reach into millions of years. Artificial radionuclides from mid-twentieth-century weapons testing decay on very different schedules: caesium-137 is effectively gone within centuries, but plutonium-239 has a half-life of about 24,000 years, and traces of long-lived isotopes plus their decay products persist much longer. The carbon-isotope excursion from burning isotopically light fossil carbon is a global geochemical shift of exactly the kind used to identify events deep in the record. Plastics degrade, but their fragments, breakdown products and the distinctive way they fossilise into sediment and rock-plastic conglomerates are novel materials with no natural precedent. Then there is biology: a mass extinction pulse, a worldwide redistribution of species, and a bizarre fossil assemblage dominated by a handful of domesticated animals whose skeletons appear simultaneously on every continent. Two conclusions matter, and they pull in opposite directions. First, we would not become undetectable. A future investigator with our own toolkit examining sediments of our age would find a stack of correlated anomalies — chemical, isotopic, biological and structural — that no known natural process produces together. Second, detecting an anomalous episode is not the same as reconstructing the civilization that caused it. From the signature alone, you could infer intense global combustion, novel synthetic materials, artificial nuclear reactions and abrupt ecological reorganisation. You could not recover our languages, our politics, our religions, our cities' names, or even reliably our body plan. The record would say something extraordinary happened. It would say very little about who we were. That asymmetry is the point of the exercise, and it cuts both ways for any claim about a lost civilization. It means a prior technological episode need not have left recognisable ruins to have left something detectable. It equally means that "we would have vanished without trace, therefore they could have too" is false as stated: the geochemical record is far more persistent than the architectural one, and it is the geochemical record that has been searched.

Modern synthesis drawing on taphonomy, Anthropocene stratigraphy and popular-science treatments such as Weisman's The World Without Us (2007)

Sources are labelled by type and access. Arcanum preserves claimant material where it documents the claim, while using academic, institutional and primary records to evaluate it.

• Waters et al. — The Anthropocene is functionally and stratigraphically distinct from the Holocene (Science, 2016)(opens in a new tab)

  Peer-reviewed synthesis of the physical, chemical and biological markers modern civilization is laying into the sedimentary record.

  Peer-reviewed

• Anthropocene Working Group — Subcommission on Quaternary Stratigraphy(opens in a new tab)

  Open official working-group record — no login. Documents both the proposed markers and the ongoing formal dispute over classification.

  Reference work

• Zalasiewicz et al. — The geological cycle of plastics and their use as a stratigraphic indicator (Anthropocene, 2016)(opens in a new tab)

  Peer-reviewed assessment of how plastics enter, move through and persist in the sedimentary record.

  Peer-reviewed

• Kidwell & Holland — The Quality of the Fossil Record (Annual Review of Ecology and Systematics, 2002)(opens in a new tab)

  Peer-reviewed review of preservation bias: what gets fossilised, what does not, and how that skews reconstruction.

  Peer-reviewed

• IAEA — Radioactive decay data and half-lives(opens in a new tab)

  Open reference database — no login. Source for the decay schedules that determine how long artificial radionuclide markers stay detectable.

  Reference work

• Alan Weisman — The World Without Us(opens in a new tab)

  Popular-science treatment of the same thought experiment. Useful as a structured scenario, not as a technical source.

  Interpretive

Each question searches Arcanum’s own index — nothing is sent to an external search provider. Use “Copy question” to save the full wording anywhere you like.

• Am I reasoning about objects or about planetary signatures? They have completely different survival curves. — search the Arcanum index for this questionCopy question
• In which depositional environment is this trace supposed to have survived — eroding upland, subsiding delta, glaciated shield, deep marine? — search the Arcanum index for this questionCopy question
• Which of the markers I am relying on would still be diagnostic after a million years, and which decay or homogenise away? — search the Arcanum index for this questionCopy question
• Does the proposed civilization's technology produce any of these signatures at all, or a different set entirely? — search the Arcanum index for this questionCopy question
• What is the difference between detecting an anomaly and reconstructing its cause, and which one does this claim actually require? — search the Arcanum index for this questionCopy question

• The Silurian Hypothesis
• The Museum Paradox: Archaeology After Archaeology
• How Archaeology Dates Stone Sites
• Apply the Survival Test

Notes are stored locally on this device only in this phase.