{
  "id": "cretaceous-paleogene-extinction",
  "title": "Cretaceous–Paleogene Extinction",
  "lang": "en",
  "source_lang": "en",
  "summary": "A mass extinction about 66 million years ago, driven principally by an asteroid impact, that eliminated non-avian dinosaurs and many other organisms.",
  "category": "science",
  "subcategory": "earth-science",
  "type": "event",
  "aliases": [
    "K–Pg extinction",
    "Cretaceous–Paleogene extinction event",
    "Cretaceous–Tertiary extinction",
    "K–T extinction",
    "end-Cretaceous mass extinction"
  ],
  "generated_at": "2026-10-25",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/cretaceous-paleogene-extinction/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/cretaceous-paleogene-extinction/",
      "json": "https://aiwiki.page/zh/cretaceous-paleogene-extinction.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The Cretaceous–Paleogene extinction was a global mass extinction approximately 66 million years ago, marking the boundary between the Cretaceous and Paleogene periods and between the Mesozoic and Cenozoic eras. An estimated three-quarters of Earth's species disappeared, including all non-avian dinosaurs, ammonites, pterosaurs, and several groups of marine reptiles. Birds survived, as did representatives of mammals, crocodilians, turtles, and many other lineages, although survival varied greatly among groups and regions.\n\nThe principal trigger was the impact of a large asteroid at Chicxulub, on what is now Mexico's Yucatán Peninsula. The impact dispersed material through the atmosphere, causing rapid environmental disruption, prolonged darkness and cooling, and a severe reduction in photosynthesis. These changes destabilized terrestrial and marine food webs. Massive eruptions in the Deccan Traps of present-day India also altered environmental conditions around this time; their contribution to the extinction and subsequent recovery remains an active research question.\n\nThe event is one of the best-documented major crises in the history of biodiversity. Its worldwide geological signature allows researchers to connect an abrupt physical disturbance with biological losses and the subsequent rebuilding of ecosystems. Its aftermath transformed the course of evolution, creating ecological opportunities for surviving mammals, birds, and other organisms while leaving a recovery record that varied across habitats and timescales."
    },
    {
      "heading": "Geological age, terminology, and boundary definition",
      "text": "The Cretaceous–Paleogene extinction occurred approximately 66 million years ago, at the boundary between the Maastrichtian Age, the last subdivision of the Cretaceous Period, and the Danian Age, the first subdivision of the Paleogene Period. This same boundary separates the Mesozoic and Cenozoic eras. These divisions describe both geological time and the corresponding rock record: Maastrichtian and Danian are ages when referring to time, but stages when referring to bodies of strata. The numerical age assigned to the boundary can be refined as dating methods improve without changing the physical reference point that formally defines it.\n\nThe abbreviation K–Pg combines K, the conventional geological symbol for the Cretaceous, derived from the German Kreide, meaning chalk, with Pg for Paleogene. The older abbreviation K–T means Cretaceous–Tertiary and remains common in historical publications. “Tertiary” formerly designated the interval after the Cretaceous and before the Quaternary, encompassing what are now the Paleogene and Neogene periods. Modern international classification recognizes those periods separately rather than retaining Tertiary as a formal period. Consequently, Cretaceous–Paleogene is the preferred formal boundary terminology; the change in name does not identify a different extinction event.\n\nThe boundary’s Global Boundary Stratotype Section and Point—the internationally agreed physical reference, often called a “golden spike”—is at El Kef in northwestern Tunisia. Ratified in 1991, it defines the base of the Danian Stage and, simultaneously, the Paleogene System and Cenozoic Erathem. The reference point lies at the base of a boundary clay bed overlying Maastrichtian marls in a marine sedimentary succession. The section preserves an abrupt change in microscopic fossil assemblages together with an iridium anomaly associated with impact fallout. Other sections are correlated with this reference through several independent markers; they need not reproduce its sediment thickness or rock type. In stratigraphy, a boundary is a specified position in a rock succession, not a prescribed thickness of “extinction sediment.”\n\nA formally defined boundary must also be distinguished from the duration of biological extinction. Placing a boundary at a particular surface does not establish that every affected population disappeared at precisely that instant. Fossil occurrences record organisms through the filters of preservation, sediment accumulation, and sampling. Conversely, an individual fossil found above the boundary may have been eroded from older sediment and redeposited. Establishing the timing of losses therefore requires comparing fossil distributions with independently recognized boundary markers, rather than treating every last fossil occurrence as an exact time of death.\n\nSeveral complementary methods establish the boundary’s age and correlate it across regions. Radiometric dating measures radioactive decay in suitable minerals, especially those in volcanic ash layers associated with boundary-bearing successions. Argon–argon dating of feldspar and uranium–lead dating of zircon provide numerical ages; ash beds above and below a boundary can bracket its age even where the boundary sediment itself cannot be dated directly. Magnetostratigraphy identifies changes in Earth’s magnetic polarity preserved in rocks: the K–Pg boundary falls within reversed-polarity chron C29r, so a polarity record supplies context rather than a unique boundary marker. Fossil correlation, particularly using planktonic foraminifera and calcareous nannoplankton, provides additional resolution in marine deposits. Numerical ages retain measurement uncertainties arising from analytical precision, decay constants, calibration standards, and the geological relationship between a dated mineral and the event being timed. Thus, “66 million years ago” is a rounded age, not a claim of exact numerical knowledge."
    },
    {
      "heading": "Earth and its ecosystems before the extinction",
      "text": "During the late Maastrichtian, approximately 68–66 million years ago, the continents were recognizable but their connections and coastlines differed substantially from those of today. The Atlantic Ocean was narrower, South America and Africa were separated, and India was an island continent moving north toward Asia. Europe consisted largely of islands separated by shallow seas, while Australia and Antarctica remained connected. High sea levels flooded continental interiors, although these habitats had contracted from their greater mid-Cretaceous extent. In North America, the Western Interior Seaway, which had previously connected Arctic waters with the Gulf of Mexico, had retreated substantially, exposing new coastal plains. The future Chicxulub impact site lay on a shallow tropical carbonate platform at the northern edge of the Yucatán Peninsula, partly beneath the present Gulf of Mexico. The Deccan volcanic province occupied part of the northward-moving Indian landmass, with its preserved lava fields concentrated in present-day western and central India.\n\nThe late Cretaceous had a greenhouse climate, with warmer high latitudes and a smaller equator-to-pole temperature contrast than today. Forests grew at polar latitudes despite prolonged winter darkness, and there is no evidence for persistent continental ice sheets comparable to those of modern Antarctica and Greenland. Nevertheless, conditions varied strongly with latitude, elevation, ocean circulation, and distance from the coast. Tropical seas were warm, whereas high-latitude environments remained distinctly cooler and seasonal. The Maastrichtian was generally cooler than the exceptional warmth of the mid-Cretaceous, but it was not an interval of uninterrupted cooling: marine geochemical records document warming and cooling episodes, including warming late in the stage. Sea-level changes repeatedly shifted shorelines, altering the extent of coastal wetlands, inland lowlands, and shallow marine shelves. Such changes operated over longer intervals than the abrupt disruption at the boundary.\n\nOn land, non-avian dinosaurs occupied many large-herbivore and large-predator roles, but communities differed among continents. The Hell Creek Formation of western North America, deposited approximately 67–66 million years ago, preserves river channels, floodplains, and forests inhabited by Triceratops, Edmontosaurus, and Tyrannosaurus. Southern continents supported different assemblages, often including abundant titanosaur sauropods and abelisaurid predators. Flowering plants, which had diversified substantially earlier in the Cretaceous, were prominent in many lowland floras; conifers, ferns, and other plant groups remained important. These landscapes were not modern grassland ecosystems. Leaves, pollen, and fossil wood record mosaics of vegetation shaped by drainage, disturbance, and regional climate.\n\nBirds already included several evolutionary lineages, among them enantiornithines and toothed diving birds such as hesperornithines. Mammals were generally small relative to the largest dinosaurs but were neither uniformly insectivorous nor ecologically identical: multituberculates possessed specialized chewing teeth, while other groups exploited insect-eating, predatory, and omnivorous diets. Crocodile relatives, turtles, lizards, snakes, amphibians, and freshwater fishes added substantial vertebrate diversity. Insects acted as herbivores, predators, decomposers, and pollinators. Fossil leaves preserve mines, galls, and characteristic feeding damage, documenting specialized interactions between insects and plants.\n\nMarine ecosystems likewise combined familiar groups with organisms that have no living representatives. Ammonites occupied diverse marine settings, while mosasaurs, plesiosaurs, sharks, and large fishes formed parts of upper-level predator communities. Rudist bivalves built extensive carbonate accumulations in some warm, shallow seas, although their distribution and abundance had already changed regionally before the boundary. Planktonic foraminifera and calcareous nannoplankton were important components of open-ocean communities; their shells contributed to widespread carbonate sediments. Other microscopic producers, including dinoflagellates and diatoms, also supported marine food webs. Regional changes in these assemblages record shifting temperatures, circulation, nutrient supply, and sea level. They do not establish a single worldwide ecological decline immediately before the extinction: long-term environmental change and the condition of communities in the final pre-boundary interval must be distinguished."
    },
    {
      "heading": "Development of the impact explanation",
      "text": "Nineteenth- and early twentieth-century fossil studies established that characteristic Cretaceous animals, including ammonites and non-avian dinosaurs, disappeared before the better-known Cenozoic assemblages appeared. Explaining these losses was harder than recognizing them: incomplete fossil sequences could make abrupt extinction appear gradual, and correlations between marine and terrestrial records remained uncertain. Proposed explanations included climatic deterioration, retreat of shallow continental seas, and volcanic activity. Extraterrestrial explanations also preceded the eventual impact hypothesis. In 1956, Otto Schindewolf proposed that radiation from a nearby supernova could cause mass extinctions; other researchers subsequently considered similar astronomical disturbances. Such proposals lacked a uniquely identifying geological signal at the boundary.\n\nThe decisive change came from measurements of the boundary clay near Gubbio, Italy. Geologist Walter Alvarez and physicist Luis Alvarez initially investigated whether its iridium content could reveal how slowly it had accumulated, using the continual arrival of extraterrestrial dust as a potential clock. Measurements by nuclear chemists Frank Asaro and Helen Michel instead revealed an unexpectedly large enrichment of iridium, an element scarce in Earth’s crust but relatively abundant in meteorites. In their 1980 Science paper, the four researchers reported anomalous boundary concentrations from Italy, Denmark, and New Zealand and proposed the impact of an asteroid approximately 10 kilometres across. Their explanation connected a measurable global chemical anomaly with a physical mechanism: impact-generated material dispersed through the atmosphere could obstruct sunlight and suppress photosynthesis, disrupting food supplies.\n\nThe proposal initially lacked an identified crater of suitable age and size. Its iridium evidence also required testing against alternatives, including volcanic sources and unusual sedimentary concentration processes. During the 1980s, discoveries of shocked quartz, impact-related spherules, and additional iridium-rich boundary layers strengthened the impact interpretation. Quartz containing characteristic planar deformation features was particularly important because it indicated extreme shock pressures, rather than ordinary sedimentation or volcanic ash deposition. Meanwhile, paleontologists investigated whether extinctions were sufficiently abrupt and closely associated with the boundary deposits. These investigations linked the chemical anomaly to a short geological event rather than merely to a prolonged interval of environmental change.\n\nIdentification of Chicxulub developed partly independently of this work. In 1978, geophysicists Glen Penfield and Antonio Camargo, working with data collected for Mexico’s petroleum company Pemex, recognized a large circular structure beneath the northern Yucatán Peninsula and adjacent sea. Magnetic and gravity anomalies suggested an impact structure, and they presented their interpretation at a geophysical meeting in 1981. The connection with the extinction hypothesis nevertheless remained unestablished for several years. In 1991, Alan Hildebrand and colleagues, including Penfield and Camargo, published evidence identifying Chicxulub as a candidate boundary-age impact crater. Its location also helped explain the unusually thick impact-related deposits around the Gulf of Mexico and Caribbean. Studies of material recovered from petroleum exploration wells provided evidence of shock metamorphism and impact melting.\n\nAcceptance rested on convergence among independent observations rather than on iridium alone: a large crater, diagnostic shock features, geographically widespread ejecta, abrupt biological losses, and increasingly precise age determinations. In 1992, argon-isotope dating linked Chicxulub melt rocks with boundary-associated impact glass; subsequent improvements in radiometric dating tightened the chronological relationship. A 2013 study by Paul Renne and colleagues placed the impact and extinction within approximately 32,000 years of each other, within the resolution of its dating framework. A 2010 synthesis led by Peter Schulte, involving 41 researchers, concluded that the Chicxulub impact triggered the mass extinction. This conclusion does not establish every proposed consequence: the relative roles of dust, sulfur aerosols, soot, heating, and oceanic disruption remain distinct research questions, as does the contribution of contemporaneous Deccan volcanism."
    },
    {
      "heading": "Chicxulub crater and the geological evidence for impact",
      "text": "The Chicxulub structure lies beneath the northern Yucatán Peninsula and the adjoining Gulf of Mexico, buried by younger sediments rather than exposed as an open depression. Approximately 180 kilometres across, it is a complex impact crater with concentric faults and a prominent inner peak ring, about 80–90 kilometres in diameter. Gravity surveys reveal a broad, roughly circular anomaly caused by contrasts in subsurface rock density; seismic reflection profiles image displaced sedimentary layers, the crater basin, and its ring structures. These geophysical observations establish its geometry, while drilling supplies direct evidence of impact-deformed and melted rocks. The crater’s dimensions should not be confused with those of the smaller, short-lived excavation cavity that preceded collapse.\n\nIn 2016, International Ocean Discovery Program–International Continental Scientific Drilling Program Expedition 364 drilled into the offshore peak ring, reaching approximately 1,335 metres below the seafloor. Beneath post-impact sediments and impact-generated deposits, the core encountered fractured granitic basement containing shock deformation. Its position supports models in which deeply buried crust rose during crater excavation and then collapsed outward to form the peak ring. The core also recovered impact melt rock and suevite, a breccia containing rock fragments and impact-derived melt particles. These materials record melting, fragmentation, transport, and deposition during crater formation, rather than ordinary sedimentation or volcanic construction.\n\nThe connection between Chicxulub and the extinction boundary rests on geographically distributed, mutually reinforcing evidence. Boundary sediments contain an anomalous concentration of iridium, an element scarce in Earth’s crust but comparatively abundant in many meteorites. They also contain shocked quartz, whose microscopic planar deformation features record pressures far beyond those generated by normal sedimentary processes. Small spherules represent droplets of impact melt, commonly altered to clay minerals after burial. Together with other impact ejecta, these occur in deposits that generally become thicker and coarser toward the Gulf region. Iridium enrichment alone would not locate a crater, and rounded grains alone would not demonstrate impact; their stratigraphic association with diagnostic shock features and impact-derived materials makes the combined case much stronger. An iridium-enriched interval identified in the Expedition 364 core further connects the crater’s sedimentary record with the worldwide boundary layer.\n\nEstimates commonly place the impactor’s diameter at roughly 10–15 kilometres, but this is a reconstruction rather than a direct measurement. Calculations use crater dimensions, target properties, and impact-scaling relationships; different assumptions about velocity, density, and trajectory produce different combinations of projectile size and kinetic energy. A speed near 20 kilometres per second is frequently used in simulations, not independently measured. Three-dimensional simulations published in 2020 found that trajectories approximately 45–60 degrees above the horizontal best reproduced several observed crater asymmetries, although the inferred angle remains model-dependent. Geochemical evidence favors a carbonaceous-chondrite-like projectile, including ruthenium-isotope results published in 2024; such evidence constrains the impactor’s broad compositional affinity rather than identifying a unique parent asteroid.\n\nAround the Gulf of Mexico, the event produced deposits much thicker and more complicated than the thin distal fallout layer. Crater-interior sequences include impact melt and breccias, overlain by materials deposited as water returned to the basin. Coastal and deep-water sections contain combinations of ejecta, eroded sediment, graded beds, and sediment-gravity-flow deposits. Tsunami waves, earthquake-induced slope failure, and repeated remobilization could all contribute, so not every thick boundary bed is uniquely diagnostic of a tsunami. Reworked fossils or ejecta can also occur above their original depositional level. Distinguishing these locally transported deposits from globally dispersed atmospheric fallout requires sedimentary structures, grain-size trends, stratigraphic context, and geochemical correlation."
    },
    {
      "heading": "From impact to global environmental disruption",
      "text": "Within seconds of impact, shock waves compressed and heated the projectile and target rocks, melting or vaporizing part of both. An expanding plume carried vapor, molten droplets, and fragmented rock above the atmosphere, while excavation opened a transient cavity that subsequently collapsed into the Chicxulub structure. Over minutes, ejecta spread across the surrounding region; seismic waves disturbed sediments and slopes, and displacement of seawater generated tsunamis. Returning water transported debris into the crater, where drilling has recovered roughly 130 metres of impact-related deposits emplaced during the first day. These processes devastated the Gulf of Mexico region, but blast damage and tsunami inundation alone cannot explain extinction on distant continents and throughout the oceans. Globally effective disruption depended especially on material injected into the atmosphere.\n\nThe impact site’s geology strongly influenced that atmospheric burden. Above crystalline basement lay a thick sedimentary succession containing carbonate rocks and sulfate-rich evaporites, including anhydrite. Shock heating released sulfur-bearing gases from sulfates and carbon dioxide from carbonates, although the amounts reaching the atmosphere depended on vaporization, chemical reactions, and subsequent condensation or recombination. Pulverized rock supplied mineral dust, while combustion of organic material supplied soot. Estimates of these emissions differ substantially because they require reconstructions of the target’s composition, impact conditions, and the proportion of material lofted high enough to avoid rapid removal. The environmental consequences therefore depended not simply on impact energy, but also on what the asteroid struck.\n\nAn impact winter could develop when suspended particles reduced incoming sunlight. Dust, sulfate aerosols, and soot behaved differently rather than forming a single uniform atmospheric veil. Coarse dust settled relatively quickly, whereas fine particles could remain aloft much longer. Sulfur gases underwent chemical conversion into sulfate aerosols, which scattered sunlight; their climatic effect depended on particle growth, atmospheric circulation, and removal. Soot absorbed solar radiation, heating the atmosphere while reducing illumination at the surface. Injection into the stratosphere prolonged these effects because particles there were less exposed to removal by rain than particles in the lower atmosphere.\n\nReduced illumination simultaneously cooled the surface and constrained photosynthesis, disrupting terrestrial vegetation and marine primary production. Food shortages could then propagate through food webs, even where organisms escaped the initial physical disturbance. A 2023 modeling study led by Cem Berk Senel used measurements of fine silicate dust from a boundary deposit in North Dakota, including particles approximately 0.8–8 micrometres across. Its simulations indicated that dust could remain in the atmosphere for about 15 years, produce global-average surface cooling of as much as approximately 15 °C, and contribute to nearly two years of photosynthetic shutdown. These are model-dependent estimates, not direct measurements of worldwide post-impact conditions; they demonstrate how particle-size assumptions can change the inferred importance of dust relative to sulfur and soot.\n\nAdditional stresses operated on different geographical and temporal scales. Re-entering ejecta radiated heat toward the ground, potentially igniting vegetation, but the intensity of this thermal pulse and the extent of simultaneous wildfire remain debated. Charcoal and soot document combustion without establishing that every forest burned; some soot may also have originated from organic material in the target rocks. Sulfur- and nitrogen-bearing compounds could produce acid deposition. Rapid changes in surface-ocean chemistry, supported by boron-isotope evidence for a fall in pH, added stress to marine organisms, although ocean acidification varied with circulation, buffering, and biological activity.\n\nAtmospheric darkening was temporary: settling, chemical transformation, and precipitation progressively removed particles. Recovery of sunlight, however, did not immediately restore ecosystems or ocean biogeochemistry. Altered photosynthesis, decomposition, and marine carbon export changed the carbon cycle beyond the initial atmospheric disturbance. Carbon dioxide also persisted longer than most light-blocking particles, allowing its warming influence to outlast their cooling effects; its contribution must be distinguished from contemporaneous volcanic emissions. Local heating, regional inundation, prolonged darkness, cooling, and chemical disruption thus overlapped without acting everywhere at equal strength or for equal durations."
    },
    {
      "heading": "Deccan volcanism and other environmental influences",
      "text": "The Deccan Traps are a large igneous province formed principally by repeated flood-basalt eruptions in what is now western and central India. The surviving volcanic rocks cover approximately 500,000 square kilometres; their original extent was larger because erosion and burial have removed or concealed parts of the province. Most of the main lava succession accumulated approximately 66.4–65.5 million years ago, spanning the extinction boundary at about 66 million years ago. Individual eruptions produced extensive lava sheets, separated in places by weathered surfaces and sedimentary beds. This episodic construction matters environmentally: the effects of eruptions depended on their spacing and duration, not simply on the final thickness of basalt.\n\nVolcanic carbon dioxide and sulfur dioxide exert different, potentially opposing climatic effects. Carbon dioxide strengthens the greenhouse effect, with perturbations potentially persisting for thousands to hundreds of thousands of years as the ocean and geological carbon cycle redistribute and remove the added carbon. Sulfur dioxide can form sulfate aerosols that scatter sunlight and cool the surface. Individual stratospheric aerosol injections generally act over years rather than geological intervals, although closely spaced eruptions could renew the cooling. The atmospheric consequences depend on gas content, eruption intensity, injection altitude, and latitude. Lava volume is therefore an imperfect measure of climatic forcing: magma can release gases before eruption or while remaining underground, and different magma batches can contain different volatile concentrations.\n\nHigh-resolution radiometric dating confirms that major Deccan activity occurred both before and after the boundary, but published chronologies differ over its detailed tempo. In 2019, Blair Schoene and colleagues used uranium–lead dates from zircon-bearing horizons between lava flows to infer four high-volume eruptive intervals, including one beginning tens of thousands of years before the extinction. Courtney Sprain and colleagues, using argon–argon dating of volcanic minerals, reconstructed a change in eruption tempo near the boundary and estimated that roughly three-quarters of the sampled lava volume was emplaced afterward. These approaches date different materials and require different assumptions when interpolating between sampled horizons. Neither directly measures the timing or quantity of gas release, so an apparent peak in lava emplacement need not coincide with the strongest atmospheric disturbance.\n\nLate Maastrichtian warming provides evidence of environmental change before the impact. Marine temperature records indicate warming of roughly 2–4 °C during an interval beginning several hundred thousand years before the boundary, commonly linked to Deccan carbon emissions; its magnitude and timing vary among records. Volcanism could consequently have altered habitats or imposed additional physiological stress before the extinction, while later emissions may have influenced early Paleocene climate and recovery. Nevertheless, the abrupt losses concentrated at the impact-marked boundary do not match a simple model of extinction tracking cumulative lava production. Pincelli Hull and colleagues’ 2020 synthesis of temperature records and carbon-cycle modelling distinguished pre-boundary volcanic warming from the impact-associated extinction. Proposals that impact-generated seismic waves changed Deccan magma transport or eruption rates remain unresolved: temporal proximity does not independently establish a physical trigger.\n\nSea-level fluctuations supplied another environmental influence. Retreating shallow seas changed the area and connectivity of continental-shelf habitats, altered sediment delivery, and could affect nutrient cycling and regional circulation. Such changes were geographically uneven, however, and some apparent disappearances in coastal fossil sequences reflect erosion, non-deposition, or migration of sedimentary environments rather than biological extinction. Testing a causal contribution therefore requires matching independently dated environmental changes to biological losses across multiple settings. Sea-level regression, regional climatic trends, and volcanism provide context for the boundary crisis, but their coincidence with it does not by itself demonstrate that they produced the globally abrupt extinction pattern."
    },
    {
      "heading": "Extinction in the oceans",
      "text": "The boundary approximately 66 million years ago marks the final disappearance of ammonites, mosasaurs, plesiosaurs, and the remaining rudist bivalves. These were extinctions of entire evolutionary groups, not simply reductions in abundance or geographic range. Ammonites had been widespread marine cephalopods; mosasaurs and plesiosaurs included large predators occupying different feeding niches; and rudists were specialized bivalves that built extensive carbonate structures in warm, shallow seas. Their disappearance did not mean that all comparable marine animals vanished: nautiloid cephalopods, many other bivalves, marine turtles, sharks, and ray-finned fishes crossed the boundary, although surviving groups also suffered losses. Some rudist communities had already declined before the boundary, making their local disappearance less consistently abrupt than the global loss of ammonites.\n\nAmong microscopic organisms, the turnover was particularly severe in planktonic foraminifera, single-celled organisms commonly bearing calcium-carbonate shells. More than 90% of late Cretaceous planktonic foraminiferal species disappeared in many global compilations, although the precise percentage depends on classification and on whether apparent earliest Paleogene occurrences represent survivors or reworked fossils. Large, elaborately ornamented forms were replaced by assemblages dominated by small, comparatively simple species. Calcareous nannoplankton, including coccolithophores, likewise lost roughly 90% or more of their species. These organisms occupied different ecological roles: coccolithophores contributed directly to primary production through photosynthesis, whereas planktonic foraminifera consumed other organisms, with some also hosting photosynthetic symbionts. Both supplied carbonate particles to seafloor sediments, so their losses changed not only marine food webs but also the production and accumulation of chalk and pelagic limestone.\n\nExtinction severity varied strongly among habitats. Shallow tropical carbonate communities lost rudists and experienced major disruption of reef-associated ecosystems, while losses among other molluscs differed between regions and ecological groups. In the open ocean, the destruction of plankton communities affected food supplies available to larger consumers, although particular feeding pathways are harder to reconstruct than the disappearance of shelled species. Deep-sea benthic foraminifera, living on or within seafloor sediment, provide a marked contrast: they show substantial changes in abundance and community composition but no extinction comparable in magnitude to that of planktonic foraminifera. Their persistence indicates that organic matter continued to reach the deep ocean, though its quantity, nutritional quality, and geographic distribution changed. Seafloor environments therefore were not insulated from the crisis, but neither did they experience uniform starvation.\n\nCarbon-isotope records provide evidence for disruption of the biological pump, the processes transferring biologically fixed carbon from surface waters into the ocean interior. Before the extinction, surface-dwelling and deep-dwelling organisms commonly recorded different carbon-isotope compositions because photosynthesis preferentially removes the lighter carbon isotope from surface waters. The sharp reduction of this surface-to-deep gradient after the boundary was initially interpreted as evidence for a nearly lifeless ocean. However, a weakened gradient does not uniquely demonstrate that photosynthesis ceased. It can also reflect changes in plankton size, food-web structure, organic-matter recycling, and the depth at which sinking material decomposes. Geographic differences among sediment records further argue against a single, globally uniform productivity collapse.\n\nChanges in carbonate chemistry added another stress. In 2019, Michael Henehan and colleagues used boron-isotope measurements in foraminiferal shells, including material from Geulhemmerberg in the Netherlands, to reconstruct rapid surface-ocean acidification associated with the boundary. Such a fall in pH could disadvantage calcifying plankton, although responses depended on physiology and local conditions. Their proxy-based modeling also showed that reduced global productivity, rather than its complete elimination, could explain important features of the post-extinction carbon-isotope record. The evidence thus distinguishes severe losses of major carbonate producers and impaired carbon export from a complete shutdown of ocean life."
    },
    {
      "heading": "Extinction on land and in freshwater habitats",
      "text": "All non-avian dinosaurs disappeared at the boundary, including the latest Maastrichtian populations of Triceratops, Tyrannosaurus, hadrosaurs, and smaller theropods. No securely established Paleocene population demonstrates their survival; isolated dinosaur remains above the boundary can result from erosion and redeposition of older fossils. The remaining pterosaurs, including large azhdarchids, also became extinct. Birds survived only through a subset of their evolutionary diversity. Enantiornithines, a widespread and anatomically diverse group of Cretaceous birds, left no known Paleogene descendants, as did toothed forms such as ichthyornithines and hesperornithines. Their disappearance shows that the survival of birds as a lineage did not amount to general survival among flying vertebrates.\n\nMammals also underwent severe losses rather than passing through the crisis largely unaffected. A 2016 analysis by Nicholas Longrich and colleagues estimated that approximately 93% of mammalian species and 59% of genera represented in its North American dataset disappeared across the boundary. These estimates apply to the sampled record, not to mammals worldwide. Studies of northeastern Montana likewise document substantial turnover, including losses among metatherians—the broader group containing marsupials—and multituberculates, an extinct group with distinctive grinding teeth. Fossil teeth provide much of this evidence, allowing changes in taxonomic composition and feeding adaptations to be tracked through successive deposits. Squamates, including lizards and snakes, also experienced substantial losses in western North America. However, differences in collecting intensity, sedimentary environments, and taxonomic identification complicate comparisons between regions; the exceptionally detailed Western Interior record cannot simply be treated as a global census.\n\nPlant fossils document both abrupt vegetation disturbance and lasting changes in composition. Pollen records register the disappearance or sharp decline of numerous Cretaceous flowering-plant taxa, while leaves reveal losses that pollen alone may not distinguish. In North Dakota, Peter Wilf and Kirk Johnson’s 2004 analysis estimated extinction of about 57% of the sampled plant species. At several boundary localities in western North America and New Zealand, a fern spike—a pronounced increase in fern spores relative to other terrestrial palynomorphs—records the spread of ferns over disturbed landscapes. These deposits indicate regional replacement of established vegetation by pioneer communities, not a globally simultaneous fern-dominated landscape. Fossil wood, leaves, pollen, and spores provide complementary evidence of forest disruption: wood identifies woody vegetation, whereas pollen and spores can register community changes even where recognizable trunks and leaves are scarce.\n\nThe disruption extended to relationships between plants and insects. Fossil leaves preserve mines, galls, punctures, and characteristic feeding traces, enabling researchers to compare ecological interactions before and after the boundary. Conrad Labandeira, Kirk Johnson, and Peter Wilf’s 2002 study of North Dakota floras documented a marked loss of insect damage diversity, particularly specialized associations such as leaf mining. Some generalized external feeding persisted even where specialized damage types disappeared. Because many fossil insects are known only indirectly through their traces, such records demonstrate losses of interactions more directly than they establish the extinction of named insect species. Turnover of host plants also complicates interpretation: disappearance of a particular mine or gall can reflect loss of the insect, its host, or their local association.\n\nRiver and lake assemblages contrast with the severe terrestrial losses. Across the Hell Creek–Fort Union succession of the northern Great Plains, many freshwater fishes, turtles, crocodyliforms, and amphibians persisted, although individual taxa disappeared and community composition changed. Their relative continuity is consistent with freshwater food webs receiving organic material from sediments and surrounding catchments, reducing immediate dependence on living foliage or fresh aquatic primary production. This contrast is neither evidence that freshwater environments escaped disturbance nor a universal survival rule. River channels, floodplain ponds, and lakes differed in hydrology and food supply, and their fossil assemblages sampled different communities; persistence of broad groups can therefore coexist with substantial local extinction."
    },
    {
      "heading": "Selectivity of extinction and survival",
      "text": "Survival across the boundary was selective, but no single trait consistently separated survivors from victims. Body size, diet, habitat, geographic distribution, and reproductive biology interacted with the particular stresses organisms experienced. Large terrestrial animals disappeared disproportionately, consistent with their substantial food requirements and generally slower population replacement. Yet small non-avian dinosaurs also became extinct, alongside many small birds and mammals. Size therefore cannot by itself explain the losses. Statistical associations also depend on the organisms and geographic regions sampled: a relationship established among North American mammals need not apply to marine invertebrates. A plausible survival mechanism requires more than a correlation between a trait and occurrence above the boundary.\n\nFood supply provides a mechanistic explanation for some contrasts between habitats. Reduced photosynthesis would rapidly diminish living plant material and the prey dependent on it, whereas dead organic matter could continue supporting decomposition and detritus-based food webs. Peter Sheehan and David Fastovsky developed this explanation in 1992 to account for the comparatively high survival of freshwater vertebrates in western North America. Rivers and lakes received organic debris from surrounding land, potentially sustaining invertebrates, fishes, and their predators during a shortage of fresh primary production. The persistence of many crocodilian and turtle lineages is consistent with such buffering, combined with relatively low metabolic demands and an ability to withstand prolonged food shortages. It does not establish that freshwater systems were unaffected or that every survivor depended on detritus.\n\nShelter and life history could modify exposure without eliminating dependence on functioning ecosystems. Burrows, submerged habitats, and sediment could protect organisms against brief surface disturbances and temperature extremes. Dormant seeds, resistant eggs, and other resting stages could bridge an unfavorable interval without continuous feeding, provided that conditions improved before stored reserves or viability were exhausted. These mechanisms are biologically plausible, but direct evidence that a particular boundary-crossing animal survived specifically through burrowing or dormancy is usually lacking. Flexible diets likewise offered potential advantages: an omnivore could switch among seeds, carrion, invertebrates, and other available resources. Nevertheless, the extinction of numerous small mammals shows that neither small size nor presumed dietary flexibility guaranteed persistence.\n\nBird survival illustrates the distinction between testable hypotheses and demonstrated ancestral traits. In 2016, Derek Larson, Caleb Brown, and David Evans proposed that seed eating helped some ancestors of modern birds survive the collapse of vegetation and associated food chains. Seeds can remain edible after the parent plant dies, providing a resource less immediately dependent on continuing photosynthesis. Their analysis of fossil tooth disparity suggested that toothed bird-like dinosaurs retained diverse feeding adaptations until near the boundary, rather than undergoing a simple prolonged decline. In 2018, Daniel Field and colleagues combined evidence of forest destruction with evolutionary reconstructions to argue that predominantly ground-dwelling ancestors were favored over tree-dependent birds. Seed eating and non-arboreal habits are compatible explanations, but neither the diets nor the habitat preferences of all surviving lineages are securely established.\n\nGeographic range could spread risk across regions, although its protective effect depended on whether those regions retained suitable conditions. Marine invertebrate studies have identified geographic breadth as a correlate of survival at some taxonomic levels; a widespread genus, however, may contain many narrowly distributed species. Survival of that genus can therefore coexist with severe species losses. The same distinction applies to birds, mammals, turtles, and other surviving groups: their presence after the boundary records the persistence of particular evolutionary branches, not the survival of every constituent species. Reproductive rate, age at maturity, and population size may also have influenced extinction risk, but these characteristics are difficult to reconstruct for most fossil organisms and should not be treated as measured universal predictors."
    },
    {
      "heading": "Ecological recovery and evolutionary consequences",
      "text": "Recovery after the extinction proceeded at several distinct rates. The resumption of primary production or the appearance of burrowing animals demonstrates biological activity, but not the restoration of pre-extinction biodiversity, complex food webs, or mature forests. Communities initially dominated by a few abundant survivors could remain species-poor even as their total productivity increased. Likewise, renewed surface-ocean photosynthesis did not necessarily restore the export of organic carbon to deep water: that process depended on the composition, size, and interactions of plankton communities. Local recolonization could therefore precede the recovery of regional diversity and global carbon cycling by hundreds of thousands of years or more.\n\nOn land, pioneer vegetation occupied landscapes where forests had been destroyed. Fern spores dominate some earliest Paleocene pollen assemblages, recording rapid colonization of disturbed ground rather than the immediate return of a forest canopy. The Corral Bluffs succession in Colorado provides an unusually detailed regional record of subsequent recovery. A 2019 study led by Tyler Lyson integrated vertebrate fossils, fossil plants, and geological dating to reconstruct approximately the first million years after the boundary. Within about 100,000 years, mammalian species richness had doubled relative to the earliest post-extinction assemblages, and maximum mammalian body size approached pre-extinction values. Around 300,000 years after the extinction, larger mammals appeared alongside increasing plant diversity. By approximately 700,000 years, the local record included legumes and another increase in mammalian size and dietary specialization. Several changes coincided with warming intervals, suggesting links among climate, vegetation, and mammalian ecology, although temporal association alone does not establish which factor drove each transition.\n\nThe disappearance of non-avian dinosaurs and numerous other organisms altered access to food, space, and ecological roles. Surviving mammals subsequently expanded their range of body sizes and diets, with increasingly specialized herbivores, predators, and arboreal forms appearing during the Paleocene and Eocene. This adaptive radiation was not the origin of mammals: mammals and several major mammalian branches already existed in the Cretaceous. Similarly, birds originated long before the boundary, and fossil and molecular evidence places at least some divergences among living bird lineages before the extinction. Post-boundary diversification nevertheless contributed substantially to the assembly of modern avian diversity. Estimates of the timing of individual divergences differ between fossil-based and molecular analyses; the ecological expansion of a group should not be equated automatically with the date of its evolutionary origin.\n\nMarine recovery was also geographically uneven. A 2018 study led by Christopher Lowery found evidence of burrowing organisms in the Chicxulub crater’s earliest post-impact deposits and a productive marine community established within roughly 30,000 years. This comparatively rapid recolonization near the impact site contrasts with prolonged recovery at some distant sites, showing that distance from the crater alone did not determine recovery time. Surviving planktonic foraminifera and calcareous nannoplankton initially formed low-diversity assemblages; subsequent speciation rebuilt diversity and produced communities unlike those of the Maastrichtian. Across ocean basins, changes in nutrient supply, water-column structure, and plankton ecology influenced the pace of recovery. Geochemical records indicate that disrupted organic-carbon export persisted much longer than the initial interruption of photosynthesis, with some reconstructions placing major reorganization of the biological pump on million-year timescales.\n\nEcological release did not simply refill empty positions with equivalent replacements. The loss of ammonites, large marine reptiles, and non-avian dinosaurs permanently changed the organisms participating in marine and terrestrial food webs. Surviving lineages diversified within environments already transformed by extinction, vegetation change, and altered competition and predation. Their subsequent histories also depended on dispersal, later climatic changes, and newly evolving interactions. Cenozoic ecosystems thus emerged through contingent sequences of recolonization and diversification, rather than through an inevitable progression toward modern communities."
    },
    {
      "heading": "Fossil-record limitations and outstanding research questions",
      "text": "The fossil record samples organisms unevenly: preservation depends on burial, sediment chemistry, skeletal durability, and subsequent erosion, while collecting is concentrated in accessible, well-studied regions. Many boundary sections also contain intervals of non-deposition or erosion, so an apparently instantaneous biological change may conceal an unrecorded interval. The Signor–Lipps effect, formalized by Philip Signor and Jere Lipps in 1982, describes how incomplete sampling places a species’ last recovered fossil before its actual extinction. Several species disappearing simultaneously can therefore appear to decline successively. Conversely, erosion can redeposit older fossils in younger sediment, producing apparent post-boundary survivors. Establishing survival requires evidence that fossils are contemporaneous with their enclosing deposits, rather than merely occurring above a mapped boundary.\n\nThese problems are central to debate over dinosaur diversity before the impact. The intensively sampled Hell Creek and Lance formations of western North America provide detailed late Maastrichtian records, but do not represent every continent or habitat. Local abundance—the number of recovered individuals—is not equivalent to regional species richness, and regional replacement of one fauna by another need not indicate a global decline. Fossil counts also vary with exposed rock area, collecting effort, and the environments represented by sedimentary formations. Analyses based on species occurrences and those estimating diversification from evolutionary trees have consequently reached different conclusions about long-term pre-boundary trends. Statistical corrections and phylogenetic models help evaluate these alternatives, but depend on assumptions about sampling and evolutionary relationships. Uncertainty about earlier diversity trajectories does not negate the disappearance of all non-avian dinosaurs at the boundary.\n\nThe principal physical uncertainties concern the amounts, particle sizes, atmospheric distribution, and persistence of impact-generated dust, sulfate aerosols, and soot. These properties govern how strongly sunlight was attenuated and how long cooling and disruption of photosynthesis lasted. A 2023 modeling study led by Cem Berk Senel, using fine silicate dust measured in the Tanis deposit in North Dakota, simulated dust persistence for approximately 15 years, global mean surface cooling of as much as about \\(15\\,^\\circ\\mathrm{C}\\), and nearly two years of severely inhibited photosynthetic activity. These are model-dependent estimates, not direct measurements of global boundary conditions; the representativeness of the input sediment and assumptions about atmospheric transport matter. Sulfur release, soot production, particle aggregation, and removal by precipitation remain additional sources of uncertainty. Distinguishing their effects is difficult because several mechanisms can produce overlapping fossil and geochemical signals.\n\nIndependent observations can narrow these uncertainties. The 2016 International Ocean Discovery Program–International Continental Scientific Drilling Program Expedition 364 recovered cores from Chicxulub’s peak ring, providing material for reconstructing crater formation, deposition, and biological recolonization. Additional marine and continental cores can test whether apparently different recovery histories reflect real ecological differences or missing sediment. Isotopic measurements, sediment composition, fossil assemblages, and organic biomarkers provide complementary evidence for temperature, productivity, and environmental change, although each proxy requires calibration and can be altered after burial. Improved radiometric dating helps compare impact deposits and volcanic episodes, but analytical precision must be distinguished from uncertainty in correlating separate sedimentary sections. Quantitative occurrence databases and coupled climate–ecosystem models can test whether proposed atmospheric disturbances reproduce observed geographic patterns and extinction selectivity, rather than only global average conditions.\n\nThe event is conventionally included among the “Big Five” mass extinctions, a grouping established through analyses of marine fossil diversity, notably work by David Raup and Jack Sepkoski in 1982. Comparisons with the end-Permian and other crises help identify recurring vulnerabilities involving warming or cooling, oxygen availability, ocean chemistry, and disrupted food webs. They do not establish identical causes: the unusually abrupt impact trigger distinguishes the end-Cretaceous event from crises associated with prolonged volcanism and carbon-cycle disturbance. Comparisons with present-day biodiversity loss likewise illuminate ecological selectivity and the potentially prolonged consequences of extinction, but require different temporal scales and sampling methods. Modern observations resolve populations and decades; most ancient records resolve much longer intervals, making direct comparisons of extinction rates sensitive to the time windows used."
    }
  ],
  "links": [
    "extinction",
    "atmosphere-of-earth",
    "photosynthesis",
    "food-web",
    "biodiversity",
    "ecosystem",
    "evolution",
    "radiometric-dating",
    "measurement-uncertainty",
    "india",
    "habitat",
    "north-america",
    "greenhouse-effect",
    "carbon-dioxide",
    "primary-production",
    "ph",
    "ocean-acidification",
    "carbon-cycle",
    "forest",
    "species",
    "decomposition",
    "climate",
    "speciation",
    "phylogenetics",
    "isotope",
    "climate-model"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}