Compendium

Paleoclimate — Proxies, Records, Past Climate States, Tipping Points

19 min read

Paleoclimate — Proxies, Records, Past Climate States, Tipping Points

Paleoclimatology is the study of Earth’s climate prior to the instrumental record (which begins, depending on variable and region, ~1850 CE for global surface temperature; earlier locally — Central England Temperature since 1659, Stockholm 1756). Direct measurement gives roughly two centuries of context for a system whose memory and dominant variability operate on 10^3 to 10^8 yr timescales. Paleoclimate fills the gap by reading natural archives — ice cores, marine and lake sediments, speleothems, tree rings, corals, biomarkers — that recorded temperature, ice volume, atmospheric composition, ocean chemistry, hydrology, and circulation long before thermometers existed.

The field’s value is twofold: it benchmarks the sensitivity and tipping-point behaviour of the climate system across forcings far larger than the anthropogenic perturbation, and it provides analogues — imperfect but the only ones we have — for warmer and more acidic worlds.

1. Historical development of the field

The intellectual scaffolding of paleoclimate is older than most realize.

  • Louis Agassiz (1840) Études sur les glaciers — proposed that erratic boulders and striated bedrock across Europe and North America were evidence of a former continental ice sheet. The “Ice Age” concept entered geology.
  • James Croll (1864, 1875 Climate and Time) — first quantitative orbital theory of glaciation, invoking eccentricity and precession variations.
  • Alfred Wegener (1912, 1915 Die Entstehung der Kontinente und Ozeane) — continental drift; reconciled paleoclimate anomalies (Permian glaciation in India, equatorial coal in Antarctica) by repositioning continents. Plate tectonics not accepted until the 1960s seafloor-spreading evidence.
  • Milutin Milanković (Serbian astronomer; 1920 Théorie mathématique des phénomènes thermiques produits par la radiation solaire, completed 1941 Kanon der Erdbestrahlung) — computed insolation at all latitudes for the past ~600 kyr from orbital parameters; modern Milankovitch theory.
  • Cesare Emiliani (1955) — pioneered marine δ18O paleothermometry on foraminifera from Caribbean cores; identified glacial-interglacial cycles in oxygen-isotope stages.
  • Nick Shackleton (1967) — showed marine δ18O variation reflects mostly ice volume, not temperature, by analysing benthic forams.
  • Hays, Imbrie, Shackleton (1976) Science “Variations in the Earth’s Orbit: Pacemaker of the Ice Ages” — spectral analysis of marine cores matched orbital frequencies; Milankovitch confirmed quantitatively.
  • Willi Dansgaard (1969) — ice-core δ18O as paleothermometer (Camp Century, Greenland).
  • Wally Broecker (1987 Nature “Unpleasant surprises in the greenhouse”; 1991 Oceanography “The Great Ocean Conveyor”) — abrupt change, AMOC, deglacial freshwater hypothesis.
  • Lonnie Thompson (Ohio State, Byrd Polar) — tropical ice cores (Quelccaya 1983, Huascarán, Dunde, Kilimanjaro).

The “proxy revolution” of the 1950s onward — δ18O, then U/Th dating, then 14C calibration, then alkenones, Mg/Ca, TEX86, δ11B — turned paleoclimate from descriptive geology into a quantitative science.

2. Geological and stratigraphic timescales

Paleoclimate runs from yesterday to the Hadean. The relevant subdivisions:

  • Phanerozoic eon (541 Ma–present): visible life
    • Cenozoic era (66 Ma–present): “age of mammals”
      • Quaternary period (2.58 Ma–present): Pleistocene (2.58 Ma–11.7 ka) and Holocene (11.7 ka–present); Anthropocene proposed (Crutzen 2000) but the IUGS Subcommission on Quaternary Stratigraphy rejected its formalisation in March 2024.
      • Neogene (23.03–2.58 Ma): Miocene + Pliocene.
      • Paleogene (66–23.03 Ma): Paleocene + Eocene + Oligocene.
    • Mesozoic (252–66 Ma): Triassic, Jurassic, Cretaceous; “age of reptiles.”
    • Paleozoic (541–252 Ma): Cambrian → Permian.
  • Precambrian (4.6 Ga–541 Ma): Proterozoic, Archean, Hadean.

The Quaternary’s defining feature is repeated continental glaciation. Marine Isotope Stages (MIS) number these cycles backward from MIS 1 (Holocene): odd = interglacial, even = glacial. MIS 5e is the Last Interglacial (Eemian, ~125 ka); MIS 2 contains the Last Glacial Maximum (LGM, ~21 ka).

3. The major proxies

3.1 Ice cores

Polar ice preserves a continuous record of temperature (via δ18O and δD of the ice), trapped paleoatmosphere (CO2, CH4, N2O measured directly in air bubbles), dust, sea salt, volcanic sulfate, cosmogenic isotopes (10Be, 14C) and biomarkers.

  • Vostok (Russian; East Antarctica): 3623 m core; Petit et al. 1999 Nature “Climate and atmospheric history of the past 420,000 years from the Vostok ice core” — four glacial cycles, CO2 range 180–280 ppm, strong CO2/temperature covariation, CO2 lagged Antarctic temperature by ~800 yr at terminations.
  • EPICA Dome C (European Project for Ice Coring in Antarctica): 3270 m core; Lüthi et al. 2008 Nature extended CO2 record to 800 ka, covering 8 full glacial cycles; CO2 stayed within ~172–300 ppm pre-industrial.
  • EPICA Dronning Maud Land (EDML): 2774 m; bipolar seesaw via Dansgaard–Oeschger event correlation with Greenland.
  • Greenland: GISP2 + GRIP (1990s, Summit), NGRIP, NEEM — annually-layered back to ~110 ka (Eemian partially preserved); high-resolution view of Dansgaard–Oeschger cycles.
  • Beyond Epica – Oldest Ice (drilling at Little Dome C 2021–2026, target completed 2025–26): aiming for 1.5 Myr continuous ice — test of the Mid-Pleistocene Transition.

δ18O of ice empirically tracks site air temperature at ~0.7 ‰/°C (Dansgaard equation), though the slope varies with moisture-source geography. Air-bubble CH4 (interhemispherically well-mixed) lets one synchronise Antarctic and Greenland cores precisely (the “methane sync”).

3.2 Marine sediment cores

Foraminiferal CaCO3 in deep-sea sediment is the workhorse of pre-Quaternary and Quaternary paleoclimate.

  • δ18O of benthic foraminifera (e.g. Cibicidoides wuellerstorfi): integrates deep-ocean temperature and global ice volume. The Lisiecki–Raymo (2005) LR04 stack — composite of 57 globally distributed benthic δ18O records — is the canonical Plio-Pleistocene timescale.
  • δ18O of planktonic foraminifera (e.g. Globigerinoides ruber, G. sacculifer): surface temperature + δ18Osw + salinity.
  • Mg/Ca paleothermometry of foram tests: temperature-dependent partitioning, calibrations Nürnberg, Anand, Elderfield; cross-correlates with δ18O to deconvolve temperature from δ18Osw.
  • Alkenone UK37′ (di- vs tri-unsaturated long-chain ketones biosynthesised by haptophyte algae, principally Emiliania huxleyi and Gephyrocapsa oceanica): Prahl–Wakeham calibration, ~0–28 °C SST proxy.
  • TEX86: tetraether lipid distribution of marine Thaumarchaeota; Schouten et al. 2002 EPSL; useful into warm climates (>28 °C) where alkenones saturate.
  • δ11B: boric acid/borate equilibria pH-dependent; Hönisch & Hemming 2005, Foster 2008 — paleo-pH and (with [CO3 2−]) paleo-pCO2.
  • B/Ca, Cd/Ca, εNd, Pa/Th: deep-water mass tracing (carbonate ion, nutrient content, water-mass mixing, overturning strength).

Coring infrastructure: Deep Sea Drilling Project (DSDP) 1968–83 → Ocean Drilling Program (ODP) 1985–2003 → Integrated Ocean Drilling Program (IODP) 2003–13 → International Ocean Discovery Program (IODP) 2013–24 → post-IODP framework 2025+. Vessels: JOIDES Resolution (retired 2024), Chikyu (Japanese, riser-capable), MeBo seafloor drills.

3.3 Lake sediments

Lacustrine archives often deliver higher temporal resolution than marine cores (sedimentation rates 1–10 mm/yr vs 1–10 cm/kyr).

  • Varves — annually-laminated sediment (light/dark seasonal couplets): Lake Suigetsu (Japan; SG06 core; radiocarbon calibration anchor for 14C IntCal20 back to ~52.8 ka), Lake Petén-Itzá (Maya), Lago Grande di Monticchio (Italy).
  • Diatoms: silica-walled algae; assemblage-based transfer functions for pH, salinity, nutrients.
  • Pollen + plant macrofossils: vegetation reconstruction.
  • Biomarkers: brGDGTs (branched glycerol dialkyl glycerol tetraethers; soil/peat origin) — paleotemperature on land (Weijers 2007, De Jonge 2014 MBT’5me calibration).

3.4 Speleothems

Cave carbonates (stalagmites + flowstones) — U/Th-datable to ~600 ka, sub-annual sampling possible.

  • δ18O of calcite: rainfall δ18O (amount effect in tropics; temperature in mid-latitudes).
  • Chinese cave network (Wang/Cheng/Edwards): Hulu, Dongge, Sanbao, Yongxing — define the East Asian Summer Monsoon (EASM) record, beautifully tracking Northern Hemisphere insolation and ice-rafted-debris events in the North Atlantic via teleconnection.
  • Sofular (Turkey), Soreq (Israel), East African (Bunker, Anjohibe), Brazilian (Botuverá, Paixão) — extend the picture.

3.5 Tree rings (dendrochronology)

A. E. Douglass (1929) founded the field at the Laboratory of Tree-Ring Research, Tucson, originally to date archaeological sites.

  • Pinus longaeva (Great Basin bristlecone) — Methuselah (>4850 yr) and the Wheeler Peak chronology back ~5000 yr; consortia chronologies push past 9000 yr via cross-dated subfossil wood.
  • European oak (Hohenheim) and Irish bog oak: continuous ~12 ka.
  • Proxies extracted: ring width (TRW), maximum latewood density (MXD; Briffa et al.), stable isotopes (δ13C, δ18O, δD).
  • Mann, Bradley, Hughes (1998 / 1999 / 2008): the “hockey stick” Northern Hemisphere temperature reconstruction; M&M (McIntyre–McKitrick) statistical critique; subsequent reconstructions (PAGES 2k Consortium 2013, 2019 Nature Geoscience) broadly reproduced the result with refined methods.

3.6 Corals and other archives

Massive scleractinian corals (Porites, Diploria) deposit aragonite skeletons with annual density bands; Sr/Ca and Mg/Ca give SST, δ18O and Δ47 give SST + δ18Osw, B/Ca + δ11B give pH. U-Th datable back to ~600 ka. Living + fossil corals: Great Barrier Reef, Indo-Pacific, Caribbean, Red Sea, Mediterranean cold-water corals.

Other archives: loess (windblown silt; Chinese Loess Plateau Heller–Liu, magnetic susceptibility paleo-monsoon), paleosols, evaporites (fluid inclusions), tufa, ostracods (Sr/Ca + Mg/Ca + δ18O lake-temperature + salinity), sclerosponges (long-lived skeletons).

3.7 Atmospheric paleo-CO2 proxies (pre-ice-core)

Pre-800 ka, direct CO2 measurement is unavailable. Indirect:

  • Stomatal index/density of fossil leaves (Royer; inverse with CO2).
  • δ13C of paleosol carbonates and leaf-wax n-alkanes.
  • δ11B of planktic foraminifera (pH + alkalinity → pCO2): Foster et al. 2017 — 420 Myr record.
  • Alkenone δ13C (Pagani; carbon-isotope fractionation by phytoplankton scales with CO2).

The CenCO2PIP (Cenozoic CO2 Proxy Integration Project; Hönisch et al. 2023 Science) integrates these across the Cenozoic. The headline: today’s CO2 (424 ppm 2024 NOAA Mauna Loa annual mean) exceeds anything reliably reconstructed for the past ~3 Myr (mid-Pliocene CO2 was ~360–420 ppm).

4. Orbital forcing: Milankovitch theory

Earth’s orbit and rotation vary on quasi-periodic timescales, modulating insolation at high latitudes and altering monsoon and glacial dynamics.

  • Eccentricity (orbital ellipticity): dominant periods near ~100 kyr (short) and ~405 kyr (long; from Jupiter–Venus interactions). Modulates the amplitude of climatic precession.
  • Obliquity (axial tilt): currently 23.44°, varies 22.1–24.5° on a ~41 kyr period. Controls latitudinal insolation gradient.
  • Precession of the equinoxes: combination of axial wobble (~26 kyr) and apsidal precession; “climatic precession” parameter e·sinω has effective periods 19 and 23 kyr.

Northern-Hemisphere high-latitude summer insolation is the canonical driver of NH ice-sheet growth/decay (Milankovitch’s specific claim). Hays–Imbrie–Shackleton (1976) confirmed via spectral peaks in marine δ18O at orbital frequencies.

Mid-Pleistocene Transition (MPT): ~1.2–0.7 Ma — the dominant period of glacial cycles shifted from 41 kyr (obliquity) to ~100 kyr (eccentricity-paced, sawtooth). Mechanism is contested: Clark & Pollard (1998) — regolith stripping allowed thicker, more stable NH ice sheets; alternatives invoke CO2 thresholds, ice-sheet height–mass-balance feedbacks, ocean stratification.

5. Major paleoclimate events in deep time

5.1 Snowball Earth

The Cryogenian Period (720–635 Ma) saw at least two near-global glaciations: the Sturtian (717–660 Ma) and Marinoan (~650–635 Ma). Joseph Kirschvink (1992) coined “Snowball Earth”; Hoffman, Kaufman, Halverson, Schrag (1998) Science synthesised the evidence — banded iron formations (anoxic ocean under ice), cap carbonates (post-glacial alkalinity flush), diamictites at low palaeolatitudes, large δ13C excursions. Deglaciation likely proceeded via CO2 build-up from continued volcanism (no silicate weathering sink under ice) to ~10^5 ppm. The aftermath plausibly catalysed the rise of multicellular life and the Cambrian explosion (~538 Ma).

5.2 Late Paleozoic Ice Age (LPIA)

~360–260 Ma. Gondwanan ice sheets. Carboniferous atmospheric O2 reached ~30–35% (charcoal record, Berner GEOCARBSULF); CO2 plunged to ~300 ppm; vast coal-forming swamps; giant insects (Meganeura dragonflies 70 cm wingspan). Permian deglaciation as CO2 rose toward the End-Permian event.

5.3 End-Permian extinction (252.2 Ma)

The most severe mass extinction in Phanerozoic history: ~81% of marine species (Stanley 2016), ~70% of terrestrial vertebrate genera. Driver: Siberian Traps flood basalts (~4 Mkm³ erupted in <1 Myr; intrusion into Tunguska Basin coal/evaporite produced massive C release). Ocean hypercapnia, anoxia, euxinia, acidification, and ~10 °C surface warming over <60 kyr. Burgess, Bowring & Shen 2014 PNAS U-Pb geochronology pinned the extinction interval to ~60 kyr.

5.4 End-Triassic extinction (201.4 Ma)

Central Atlantic Magmatic Province (CAMP) — flood basalts as Pangaea began rifting. ~50% of genera lost; clears niches for the dinosaurs’ Jurassic radiation.

5.5 Mesozoic Ocean Anoxic Events (OAEs)

Black-shale intervals reflecting widespread anoxia driven by warm climates, sluggish circulation, and nutrient pulses from LIP volcanism.

  • Toarcian OAE (T-OAE) ~183 Ma: Karoo–Ferrar LIP; the Posidonia Shale and Jet Rock are source rocks for North Sea hydrocarbons.
  • OAE 1a (Selli) ~120 Ma; OAE 2 (Bonarelli) ~94 Ma (Cenomanian–Turonian boundary) — global δ13C excursion +2–6 ‰.

5.6 K-Pg boundary (66.04 Ma)

Luis & Walter Alvarez, Asaro, Michel (1980) Science — global iridium anomaly at the boundary clay (Gubbio, Stevns Klint). Hildebrand & Penfield (1991) identified the Chicxulub crater (Yucatán, ~180 km diameter, ~66.04 Ma) as the impactor. Non-avian dinosaurs, ~75% of species extinct; Deccan Traps volcanism (India) was contemporaneous and probably contributory. Mammals radiate in the Cenozoic.

5.7 PETM — Paleocene–Eocene Thermal Maximum (~56 Ma)

The closest deep-time analogue to anthropogenic carbon release.

  • Carbon-isotope excursion (CIE) of −2.5 to −4 ‰ in marine and terrestrial records over <20 kyr.
  • Estimated C release ~3000–7000 PgC (compare anthropogenic ~700 PgC 1850–2024 emitted, with ~600 PgC actually entering atm/ocean/biosphere).
  • Global surface warming +5 to +8 °C; deep-ocean acidification; benthic foraminiferal mass extinction (~30–50% of species).
  • Recovery via silicate weathering on ~100–200 kyr timescale (Penman 2014; Zachos 2005).
  • Caveat as an analogue: PETM CO2 release rate was ~0.6–1.1 PgC/yr (Zeebe, Ridgwell & Zachos 2016 Nature Geoscience), vs current ~10 PgC/yr — anthropogenic emissions are roughly an order of magnitude faster.

5.8 EOT — Eocene–Oligocene Transition (~34 Ma)

Onset of Antarctic continental glaciation. ~5–6 °C global cooling; benthic δ18O step (Oi-1 event); driven by CO2 drawdown crossing a threshold (~750 ppm in coupled GCM-ice-sheet models — DeConto & Pollard 2003). Drake Passage opening and the proto-ACC accelerated Antarctic isolation.

5.9 Miocene to Pliocene

  • Mid-Miocene Climatic Optimum (MMCO) 17–15 Ma: ~3–4 °C warmer than pre-industrial; CO2 ~400–600 ppm. Sphenoid/columnar deglacial events follow at ~13.8 Ma onward.
  • Messinian Salinity Crisis ~5.97–5.33 Ma: Mediterranean isolated, desiccated/refilled; massive evaporite deposits.
  • Mid-Pliocene Warm Period (mPWP, ~3.3–3.0 Ma): CO2 ~360–420 ppm (near present); global mean ~2–3 °C warmer; sea level +10 to +25 m (Dumitru 2019 — Mediterranean shoreline marker). PRISM (Pliocene Research, Interpretation and Synoptic Mapping; Dowsett et al.) provides boundary conditions for PlioMIP model intercomparison. The mPWP is the best modern analogue for a stabilised ~3 °C warmer world.

5.10 Plio-Pleistocene glaciation onset

~2.7–2.5 Ma: large-scale Northern Hemisphere glaciation began. Drivers debated: Panama Isthmus closure (Haug & Tiedemann 1998), tropical Pacific cooling, threshold CO2 crossing.

5.11 Pleistocene glacial cycles

LR04 stack reveals ~50 MIS stages over 5.3 Myr. Early Pleistocene: 41-kyr obliquity-paced cycles. Post-MPT (~700 ka): ~100-kyr eccentricity-paced sawtooth (slow build-up, rapid deglaciation across “Terminations” I through IX).

Last Glacial Maximum (LGM) ~26.5–19 ka (Clark et al. 2009 Science): global mean surface temperature ~5–6 °C cooler than pre-industrial (Tierney et al. 2020 Nature with data assimilation: −6.1 °C, 95 % CI −6.5 to −5.7), sea level −120 to −134 m, Laurentide and Cordilleran ice sheets covering N America, Fennoscandian sheet over N Europe. ECS inferred from LGM cooling: 2–4 °C, consistent with IPCC AR6 best estimate 3 °C (very likely range 2–5 °C).

5.12 Millennial-scale events of the last glacial

  • Dansgaard–Oeschger (DO) cycles: 25 numbered events in Greenland ice (NGRIP) between 115 and 14.7 ka; abrupt warming of 8–16 °C in Greenland in decades, gradual cooling over centuries to millennia. Atlantic Multidecadal/Centennial-scale, likely AMOC mode switches.
  • Heinrich events (Heinrich 1988): six layers of ice-rafted detritus in North Atlantic sediments (H1–H6); massive iceberg discharges from the Laurentide (Hudson Strait) ice stream during the cold phase preceding a DO warming.
  • Bipolar seesaw (Stocker & Johnsen 2003): Antarctic warming during Greenland Stadials; CH4 sync confirms.

5.13 Last deglaciation

  • Heinrich Stadial 1 ~17.5–14.7 ka — extreme N Atlantic cold.
  • Bølling–Allerød 14.7–12.9 ka — abrupt warming (~5–10 °C in Greenland in <100 yr at the BA onset).
  • Younger Dryas 12.9–11.7 ka — abrupt return to near-glacial cold (NH); Broecker (1989) freshwater-pulse-stops-AMOC hypothesis is the leading explanation; the Firestone et al. 2007 bolide-impact hypothesis is widely rejected (no impact crater, contested geochemical claims).
  • 8.2 ka event: ~160 yr of cooling triggered by final drainage of glacial Lake Agassiz/Ojibway via Hudson Bay (Barber et al. 1999 Nature).

5.14 Holocene

11.7 ka to present. Subdivisions (formalised by IUGS 2018): Greenlandian (11.7–8.2 ka), Northgrippian (8.2–4.2 ka), Meghalayan (4.2 ka–present; boundary set at the 4.2 ka aridification event recorded in a Mawmluh Cave (India) speleothem).

  • Holocene Climatic Optimum 9–5 ka: NH summer warmer than today (orbital forcing maximum), but recent reconstructions (Marcott et al. 2013 Science; Kaufman et al. 2020 Scientific Data; Osman et al. 2021 Nature) suggest global annual mean was probably comparable to or cooler than the 20th century — the “Holocene temperature conundrum.”
  • Medieval Climate Anomaly (MCA) ~950–1250 CE: warmth in the North Atlantic / Europe; not globally synchronous (PAGES 2k). IPCC AR6 concludes the MCA was not as warm as recent decades globally.
  • Little Ice Age (LIA) ~1300–1850 CE: NH cooling; volcanic clustering (Samalas 1257 — largest of the last 7000 yr; Kuwae ~1452; Huaynaputina 1600; Tambora 1815 “Year Without a Summer”); Maunder Minimum (1645–1715) and Dalton Minimum (1790–1830) solar irradiance lows.
  • Industrial era 1850 CE–present: anthropogenic warming, +1.45 °C above 1850–1900 in WMO 2024 reanalysis (calendar year 2024 was the first calendar year above +1.5 °C).

6. Dating methods

  • Radiocarbon (14C): half-life 5700 ± 30 yr; useful to ~50 ka. Calibration via IntCal20 (Reimer et al. 2020) for terrestrial, Marine20 for marine, SHCal20 for SH.
  • U-Th (U-series): 230Th/234U for carbonates (corals, speleothems) — useful to ~600 ka.
  • U-Pb: zircon dating; absolute geochronology in deep time (CA-ID-TIMS sub-permil precision).
  • 40Ar/39Ar: igneous and pyroclastic rocks; chronostratigraphy of LIPs and tephras.
  • Cosmogenic nuclides (10Be, 26Al, 36Cl, 21Ne, 3He): exposure dating of glacial moraines, fluvial terraces; burial dating (Granger).
  • OSL (optically stimulated luminescence) + TL: aeolian and fluvial quartz/feldspar; last 100s of kyr.
  • Tephrochronology: dated volcanic ash layers as stratigraphic markers (Icelandic tephra in N Atlantic + European cores).
  • Magnetostratigraphy: geomagnetic polarity reversals (Brunhes/Matuyama 0.78 Ma, Gauss/Matuyama 2.58 Ma) tie sediment sections to the GPTS (Geomagnetic Polarity Time Scale).
  • Biostratigraphy: planktonic foraminiferal zones, nannofossil zones (calcareous nannoplankton), dinoflagellate cysts, conodonts (Paleozoic), ammonites.
  • Astrochronology: orbital tuning of cyclostratigraphy (Meyers, Hinnov, Pälike).

7. Tipping points and paleoclimate evidence

The IPCC AR6 (WG1 Ch 5; WG2 Ch 16) and Armstrong McKay et al. 2022 Science identify ~16 candidate tipping elements. Paleoclimate constrains several:

  • AMOC shutdown: DO/Heinrich events demonstrate the system can switch states on decadal timescales for centuries; Eemian and Younger Dryas analogues. Modern monitoring (RAPID array since 2004; AMOC reconstructions: Caesar et al. 2018 Nature; Boers 2021 Nature Climate Change) suggests weakening; AR6 medium confidence weakening in 21st c, low likelihood of collapse before 2100.
  • Greenland Ice Sheet (GrIS) collapse: Eemian (MIS 5e) sea level was +6 to +9 m; ~1–2 m from Greenland (Dutton et al. 2015 Science; Helsen). Threshold ~1.5–3 °C above pre-industrial (Robinson et al. 2012).
  • West Antarctic Ice Sheet (WAIS): marine-based, susceptible to MISI (marine ice-sheet instability; Schoof 2007). Pliocene and LIG sea-level data imply WAIS lost or much-diminished. ANDRILL cores (Naish 2009 Nature) document WAIS collapses through the Plio-Pleistocene.
  • Amazon dieback: Last Glacial vegetation reconstructions (Mayle, Marchant; Bush + Silman) show savanna expansion during dry phases.
  • Permafrost C release: PETM is the analogue for cryosphere C destabilisation, though the source was contested (methane hydrates, peat oxidation, volcanic).
  • Tropical coral reefs: K-Pg, PETM, and OAEs all coincided with reef crises.

8. Climate sensitivity from paleo

Equilibrium Climate Sensitivity (ECS — response to doubled CO2 at equilibrium):

  • LGM constraint (Sherwood et al. 2020 Reviews of Geophysics): paleo evidence combined with process and historical-warming constraints — ECS 2.6–3.9 °C (66% range).
  • Pliocene constraint: Earth System Sensitivity (ESS, including slow feedbacks) ~4.5 °C/2×CO2.
  • IPCC AR6 best estimate ECS = 3 °C (very likely 2–5 °C; likely 2.5–4 °C). Paleo is consistent with the central estimate and rules out very high values (>5 °C) that would imply implausibly large LGM cooling, and very low values (<2 °C) that would imply implausibly little.

9. PMIP and paleoclimate modelling

The Paleoclimate Modelling Intercomparison Project (Joussaume & Taylor 1995; now PMIP4 within CMIP6) coordinates climate model simulations of standard time slices — Last Interglacial (lig127k), Mid-Holocene (midHolocene), LGM (lgm), Pliocene (midPliocene-eoi400), and Last Millennium (past1000). Models are benchmarked against proxy syntheses (e.g. SST data assimilation Tierney et al. 2020; Osman et al. 2021).

10. Modern context

NOAA Mauna Loa CO2 annual mean 2024: 422.7 ppm (preliminary; published 2025). Beyond the 800 kyr ice-core range; CenCO2PIP puts current CO2 above anything since the mid-Pliocene (~3 Ma). The rate of CO2 rise (~2.5 ppm/yr 2010–2024) is ~100× faster than any geologically reconstructed natural event including the PETM (Hönisch et al. 2012 Science “The geological record of ocean acidification”; Zeebe et al. 2016).

This is the paleoclimate-grounded statement of the anthropogenic anomaly: the current CO2 level is reasonably common in geological time, the rate of change is not.

11. Active research frontiers (2024–2026)

  • Beyond Epica – Oldest Ice: drilling at Little Dome C; target 1.5 Myr ice retrieved 2025–26.
  • PaleoCAR-style data-assimilation reconstructions (Osman et al. 2021; PALAEO-RA Franke 2017) — proxy-model fusion at unprecedented resolution.
  • Mid-Pleistocene Transition mechanism — competing CO2-threshold vs regolith-stripping hypotheses; Beyond-Epica ice will test directly.
  • PETM rate problem — improving age-model precision (orbital tuning, He-3 sedimentation rate) to constrain C-release rate, critical for analogue interpretation.
  • Holocene temperature conundrum — reconciling proxy-based mid-Holocene warmth (Marcott) with model simulations and Kaufman/Osman global cooler-than-now reconstructions.
  • Antarctic Quaternary: integrated drilling of marine + ice + sediment (e.g. SWAIS-2C — Sensitivity of the WAIS to 2 °C, ongoing 2023–2027).
  • CenCO2PIP continuation (Hönisch consortium): high-fidelity CO2 record across the Cenozoic.

12. Key figures

  • Wallace S. Broecker (1931–2019; LDEO Columbia): the great-ocean-conveyor, abrupt change, “the climate beast.”
  • Lonnie G. Thompson (Ohio State): tropical ice cores.
  • Maureen E. Raymo (LDEO): uplift–weathering hypothesis (BLAG); LR04 benthic stack with Lisiecki.
  • James E. Hansen (NASA GISS retired): climate sensitivity, paleoclimate constraints (Hansen et al. 2008 “Target atmospheric CO2: where should humanity aim?”).
  • Eric W. Wolff (Cambridge): EPICA, Beyond-Epica leadership.
  • Tas D. van Ommen (AAD): Australian Antarctic ice-core program.
  • Heinz Wanner (Bern): Holocene synthesis (PAGES).
  • Michael E. Mann (Penn → UPenn): hockey stick; PAGES 2k.
  • Raymond S. Bradley (UMass Amherst): paleoclimatology textbook author.
  • Gavin A. Schmidt (NASA GISS director 2014–): modelling + paleo; RealClimate.
  • Nicholas J. Shackleton (1937–2006, Cambridge): foundational marine δ18O work.
  • Bärbel Hönisch (LDEO): δ11B, CenCO2PIP.
  • Jess Tierney (Arizona): biomarker SSTs, LGM data assimilation.

13. Major data repositories

  • NOAA NCEI Paleoclimate (formerly WDC-Paleoclimate, Boulder): the canonical archive.
  • PANGAEA (AWI Bremerhaven): geoscience data.
  • NEOTOMA: Quaternary pollen + macrofossils + faunal remains.
  • PAGES 2k Network: 2000-yr proxy database.
  • LinkedEarth / LiPD: standardised paleo data format.
  • IntCal20 / Marine20 / SHCal20: radiocarbon calibration.
  • LR04 stack: Lisiecki–Raymo benthic δ18O composite.

Adjacent

Graph View

Backlinks