With the support of thePotsdam Institute for Climate Impact Research (PIK) An interactive visualization of up to 1.2 million years of global temperature — from deep Pleistocene ice ages through the narrow climatic window that made civilization possible, to the diverging futures ahead of us.
The Corridor of Life For most of human existence, Earth's climate has swung violently between extremes — ice ages that buried continents under kilometers of glaciers, and brief warm intervals between them. Then, around 11,700 years ago, something remarkable happened: the climate stabilized. Global temperatures settled into a narrow band of roughly −1°C to +1.5°C relative to preindustrial levels (1850–1900). This corridor of stability is the only climate our species has ever built anything in — agriculture, cities, writing, science, everything we call civilization.
The Safe and Just Climate Space (−1°C to +1°C) represents the core of this corridor, where ecosystems and societies can thrive. The Safe but Unjust zone (+1°C to +1.5°C) marks a range where global-scale risks remain manageable but impacts are already severe for the most vulnerable populations and ecosystems.
We are now leaving this corridor. As of 2025, global mean temperature has risen approximately +1.4°C above preindustrial levels.
How to navigate Zoom — scroll wheel, pinch, or the +/− buttonsPan — click-drag or touch-dragHover a point on the timeline to see temperature data for that yearClick any point to open a detail panel with full dataSearch — type any keyword to find historical events, figures, and periods from the Holocene Interactive databaseSurprise Me! — discover a random event, figure, or historical period Take the guided tour
What you see Temperature anomaly All temperatures are shown as anomalies — the difference in degrees Celsius (°C) from the preindustrial baseline (mean global surface temperature during 1850–1900). A value of +1.0°C means the planet is 1 degree warmer than that reference period. This is the standard used by the IPCC . Warming stripes In Stripes mode, each vertical bar represents a time period, colored from blue (cold) through green (near baseline) to red (warm). Inspired by climate scientist Ed Hawkins' warming stripes . Temperature spectrum In Spectrum mode, horizontal colored bands — one per degree step — create a full temperature spectrum from deep blue (cold) through green (baseline) to red (warm). A continuous line traces actual temperature through time. This mode better reveals the magnitude and pace of temperature changes. ERA5 daily overlay When you zoom into the modern era (post-1940), a high-frequency daily temperature trace appears behind the smooth annual curve, along with its own annual mean. This ERA5 reanalysis data (31,000+ daily measurements) reveals the day-to-day variability hidden by annual averaging — seasonal cycles, El Niño spikes, volcanic cooling events. The overlay fades in automatically as you zoom, and disappears when you zoom back to deep time.GMST : it blends land air temperature with sea surface temperature. ERA5 reports GSAT : air temperature measured 2 metres above all surfaces, including over oceans. Because air above the sea warms slightly faster than the sea surface itself, GSAT runs about 0.05–0.10°C warmer than GMST. Both are valid — the IPCC uses GMST for its headline figures, while GSAT is standard in reanalysis products. Corridor zones Two horizontal bands mark the safe climate space. The inner zone (−1°C to +1°C , solid green line) is the "Safe and Just Climate Space" per the Earth Commission (2023) . The outer zone (+1°C to +1.5°C , dashed line) marks the "Safe but Unjust" threshold from the IPCC Special Report on 1.5°C . Land Use overlay Toggle the Land Use button to reveal 12,000 years of human transformation of Earth's surface, drawn as trend lines in the lower portion of the canvas (below −1°C). The data comes from the HYDE 3.5 database, which classifies all of Earth's land surface into 20 categories grouped under 6 levels: Wildlands , Seminatural , Rangelands , Croplands , Villages , and Dense settlements .Thick lines show the 6 main levels; thin lines in matching color tones show the 20 sub-categories within each level. The dramatic story: in 10,000 BCE, Wildlands and Seminatural covered essentially all of Earth's land. By 2025, Rangelands have exploded from near zero to 27%, Croplands from 0% to 14%, and Dense settlements (including cities) have appeared from nothing. Seminatural land has collapsed from ~60% at 10,000 BCE to 24%.Why a square root scale? The raw percentages span a huge range — Urban areas cover less than 1% of Earth's land, while Seminatural peaks at 66%. If drawn on a linear scale, the categories that are growing most dramatically (cities, irrigated croplands, villages) would be invisible at the bottom of the chart. A square root scale compresses the large values and expands the small ones, making all 20 categories visible while preserving their relative ordering. This is the same mathematical principle used in scientific cartography for proportional symbol maps (Flannery, 1971) — where a circle’s radius (not area) is proportional to the square root of the value. The axis labels on the left show the real percentages at their non-linearly spaced positions. "Choose Your Future" fan Beyond the present year, the timeline splits into diverging bands representing different temperature futures. You can switch between three very different ways of thinking about what comes next:
Pathways (default) — Climate Action Tracker projections based on what governments are actually doing right now : enacted laws, stated targets, and net-zero pledges. This answers: "Where are we really headed?" What If — Shared Socioeconomic Pathways (SSPs) from the IPCC. Abstract, model-driven scenarios that explore “what if” worlds with different levels of global cooperation. This answers: "What could happen under very different futures?" Overshoot — Schleussner et al. (2024) scenarios showing what happens if warming temporarily exceeds 1.5°C before Carbon Dioxide Removal brings it back down. This answers: "Can we overshoot and come back? At what cost?"
Each band is colored by its projected temperature. The future is not a single trajectory; it is a choice.
Future projections — three ways of looking ahead
After 2025, the timeline fans out into possible futures. Three buttons let you switch between very different ways of thinking about what comes next. Each answers a different question, and each draws on different scientific sources.
Pathways — "Where are we actually headed?"
This is the reality check. The Climate Action Tracker (CAT) looks at what governments around the world are actually doing — the laws they have passed, the targets they have announced, and the pledges they have made — and calculates where all of that leads in terms of global warming. It is an independent project by Climate Analytics and NewClimate Institute , tracking climate action since 2009. The data shown here is from the November 2025 (COP30) update . Policies & action — Where current enacted laws lead, if no new legislation is passed. This is the most likely trajectory right now. 2.5–2.9°C by 2100. 2030 & 2035 targets — If countries actually meet the near-term targets they submitted to the UN (their “Nationally Determined Contributions”). ~2.6°C. Pledges & targets (dashed) — If every pledge is fully honored — including conditional commitments and long-term strategies. The best case under current political ambition. 2.0–2.2°C. Optimistic (net-zero) (dashed) — If all announced net-zero pledges are met in full and on time. ~1.9°C. The takeaway: The distance between these four lines reveals what is sometimes called the credibility gap — the difference between what politicians have promised and what they have actually put into law. As of COP30, even the most optimistic reading of current pledges leads to ~1.9°C — well above the 1.5°C limit that the Paris Agreement set out to protect.
What If — "What could happen under very different futures?"
Pathways tells you where current politics are going. But what if the world took a radically different course? What if nations cooperated fully on decarbonization? What if fossil fuels kept expanding? The IPCC built a framework for exploring these questions: Shared Socioeconomic Pathways (SSPs). Each one is a complete “what if” story about how global society could evolve — its demographics, economics, energy systems, and international cooperation — and the climate models then calculate the warming that would follow.not predictions . They are not tied to any real government’s plans. They are thought experiments — imagined worlds that scientists use to map the full range of possibility. Three are shown here:
SSP1-2.6 — Sustainability. A world of strong international cooperation, rapid decarbonization, and net-zero emissions by mid-century. Warming stays below +2°C. The “2.6” refers to how much extra energy is trapped in the atmosphere (measured in watts per square metre).SSP2-4.5 — Middle of the road. Moderate effort, uneven progress, some mitigation but no transformational change. +2.7°C by 2100.SSP3-7.0 (500 ka) / SSP5-8.5 (1.2 Ma) — High emissions. A world of regional rivalry or continued fossil fuel expansion with little mitigation. +3.6°C to +4.4°C by 2100. The takeaway: No single curve here is “the” future. The value is in the spread — the distance between the best and worst imagined worlds. That spread shows how much the choices of the next few decades matter. The shaded bands around each curve show a different kind of uncertainty: not which path society chooses, but how the Earth’s climate system responds to whichever path we are on (computed from 600 different model configurations).
Overshoot — "Can we exceed 1.5°C and come back? At what cost?"
The Paris Agreement set 1.5°C as its more ambitious warming limit. But in 2025, global temperature is already at +1.41°C and still rising. Realistically, the world is very likely to exceed 1.5°C — at least temporarily. So an urgent question emerges: can we overshoot the limit and then bring temperatures back down by pulling CO₂ out of the atmosphere?Schleussner et al. (2024) investigated for the PROVIDE project, published in Nature . Their paper’s title is a warning in itself: “Overconfidence in climate overshoot.” They modelled five futures, each relying on different levels of Carbon Dioxide Removal (CDR) — technology and methods that pull CO₂ back out of the air: Low Overshoot — The most aggressive near-term action: rapid electrification, demand reduction, societal transformation. Temperature barely touches 1.5°C (peak ~1.5°C in 2042) before declining. This is the safest path — but it requires an unprecedented pace of change within the next 15 years.Peak & Decline — Strong emission cuts plus moderate CDR deployment. Temperature peaks at ~1.7°C around 2059, then slowly comes back down. Even this modest overshoot requires sustained CO₂ removal for decades.Net-Zero Stabilization — The world reaches net-zero CO₂ emissions around 2060 — and then stops there, without going further into net-negative territory. The result is surprising: temperature initially dips, but then starts climbing again — reaching ~1.7°C by 2300. Why? Because the CO₂ already in the atmosphere continues to warm the planet as oceans release absorbed heat, permafrost keeps thawing, and the cooling effect of industrial aerosols (tiny particles from fossil fuel burning that once reflected sunlight) disappears. The lesson: reaching net-zero is necessary but not sufficient . To actually stabilize temperature, we need to actively remove CO₂ — net-zero is the beginning of the solution, not the end.High Overshoot — Moderate near-term action, followed by a massive bet on future CDR technology. Temperature peaks at ~2.7°C around 2122 — more than 1°C above the Paris limit for over a century — before CDR slowly pulls it back toward 1.5°C by 2300. This requires removing hundreds of billions of tonnes of CO₂, approaching the upper limit of what may be physically and sustainably possible.Extreme Overshoot — Current policies continue for decades, then emergency CDR. Temperature peaks at ~3.3°C around 2122 — nearly double the Paris target. The CDR requirement is beyond anything ever deployed, exceeding today’s removal capacity by over 10,000×. Two centuries of extreme warming would likely cause irreversible ecosystem collapse, dramatic sea-level rise, and trigger multiple Earth system tipping points. This is why Schleussner et al. titled their paper the way they did. Why overshoot is not “free”
It is tempting to think of overshoot as a temporary detour — that if temperatures eventually return to 1.5°C, the world that results is the same as one that never exceeded the limit. The science says otherwise. Sea-level rise: Every 100 years above 1.5°C adds roughly 40 cm of additional sea-level rise — on top of the ~80 cm already locked in from past emissions. Even after CO₂ is removed, ocean expansion and ice sheet disintegration continue for centuries. This rise is essentially permanent on any human timescale.Self-reinforcing feedbacks: Every 100 years above 1.5°C adds ~0.02°C of additional warming from thawing permafrost and decomposing peatlands. Once these frozen carbon stores start releasing CO₂ and methane, the process feeds on itself and cannot be reversed by removing CO₂ elsewhere.Tipping elements: Overshoot increases the probability of crossing tipping points — thresholds beyond which ice sheets, ocean currents, and forests undergo irreversible shifts. Once triggered, these systems continue changing on their own timescales (centuries to millennia), regardless of what happens to global temperature afterward.A different world, not just a warmer one: A planet at +1.5°C after a 3°C overshoot is fundamentally different from one that reached +1.5°C without exceeding it. Ocean circulation patterns, rainfall distribution, and ice sheet dynamics are all path-dependent — they remember where they have been.Lost options: Ecosystems and infrastructure damaged during peak warming cannot simply be restored when temperatures fall. Coral reefs bleached during overshoot, for example, may never recover — and with them go the coastal protection and food security they provided to hundreds of millions of people. The Carbon Dioxide Removal challenge
Every overshoot scenario that returns to 1.5°C depends on Carbon Dioxide Removal (CDR) — technology and methods that pull CO₂ out of the atmosphere and store it permanently. Current global CDR capacity (excluding natural reforestation) is approximately 2 million tonnes of CO₂ per year . The High Overshoot scenario would require several billion tonnes per year sustained for over a century — a thousand-fold increase.Warszawski et al. (2021) reviewed all 1.5°C-compatible scenarios from the IPCC and found that CDR is the most over-relied-upon lever — most scenarios assume removal at scales beyond what is considered sustainable. Their conclusion: “1.5°C cannot be attained with silver bullets — all-round portfolios are needed in which most available levers are used at challenging levels.”Fuhrman et al. (2024) showed there is a way to halve the CDR requirement: cut emissions aggressively at the source , sector by sector — shipping, aviation, buildings, industry, agriculture. The most effective way to avoid depending on speculative future technology is to act now.
How the three projections relate Pathways answers: where are current politics taking us? (~2.5–2.9°C under enacted laws.) What If answers: what is physically possible? (from below +2°C to over +9°C, depending on which imagined world we end up in.) Overshoot answers: what is the cost of betting on future technology instead of acting now? Ripple et al. (2025) warn, the longer we delay, the greater the risk of crossing tipping points that could lock the planet into a hothouse trajectory — a path of no return.
Controls Stripes / Spectrum Toggle between warming stripes and temperature spectrum visualization modes. Adaptive / Linear Adaptive uses a piecewise time scale that gives more visual space to recent centuries while still showing deep time. Zigzag break marks on the axis indicate where the scale changes density. Linear shows all years at equal width — use it to see the true scale of deep time, where modern civilization is a sliver.Theme (Moon / Sun icons) Switch theme. Dark theme is optimized for the warming stripes palette; light theme is useful for presentations and readability. 500 ka / 1.2 Ma Switch between two data profiles. 500 ka shows the last 500,000 years using EPICA ice cores + HadCRUT5 instrumental data. 1.2 Ma uses the extended dataset from Ripple et al. (2025) , stitching together Clark et al., Osman et al., and Trewin et al., with SSP projections extended to 2300. Details Show named geological and historical periods (glaciations, interglacials, key eras) as labels on the timeline. Corridor Show or hide the "Corridor of Life" and "Safe but Unjust" zone labels on the canvas. Generations Overlay Western world generational spans (Baby Boomers, Gen X, Millennials, Gen Z, Gen Alpha) on the modern era. Zooms into the 1880–present view. Birth year Enter your birth year to see your lifetime highlighted on the timeline — in the context of 500,000 years of climate history. Check “80 yr” to project an 80-year lifespan into future climate scenarios. Land Use Show trend lines for 20 categories of human land use over the past 12,000 years, overlaid in the lower portion of the temperature canvas. Six thick lines show the main levels; twenty thin lines show the sub-categories. Hover anywhere in the land-use zone to see a breakdown by category. Data from HYDE 3.5. Tipping Show 16 Earth system tipping elements as vertical bars at the right edge of the canvas. Each bar spans the element’s warming threshold range. Hover for details; click to navigate to the Tipping Interactive page. What If / Pathways / Overshoot Switch between three types of future climate projections.
Pathways (default) — Climate Action Tracker (CAT) policy-based temperature projections, updated November 2025 (COP30). Unlike abstract models, these are grounded in what governments have actually legislated, pledged, and committed to. Four scenarios:Policies & action — Where current enacted policies lead, without additional action. This is the most likely outcome if no new policies are adopted. Warming: 2.5–2.9°C by 2100.2030 & 2035 targets — If countries meet their near-term NDC (Nationally Determined Contribution) targets. These are the promises made at the UN climate talks. Warming: ~2.6°C.Pledges & targets (dashed) — Full implementation of all pledges including long-term strategies and conditional commitments. This is the best case under current political ambition. Warming: 2.0–2.2°C.Optimistic (net-zero) (dashed) — If all announced net-zero pledges are met in full and on time. Most ambitious realistic scenario. Warming: ~1.9°C.
What If — IPCC Shared Socioeconomic Pathways (SSPs). These are abstract “what if” scenarios used in climate modeling. They explore a wide range of possible futures — from aggressive mitigation to unchecked emissions — without tying to specific real-world policies. Useful for understanding the full range of possibility, but less grounded in current political reality.
Overshoot — What happens if warming temporarily exceeds 1.5°C before Carbon Dioxide Removal (CDR) brings temperatures back down? Five scenarios from the PROVIDE project (Schleussner et al. 2024) tell an escalating story about the risks and limits of betting on future CDR:Low Overshoot — Aggressive near-term action barely touches 1.5°C (peak 1.5°C in 2042). The most protective pathway.Peak & Decline — Strong action + moderate CDR brings temperatures back down from 1.7°C (peak 2059).Net-Zero Stabilization — Reaching net-zero CO₂ alone is not enough : temperature initially peaks, dips, then resumes rising to 1.7°C by 2300 as the climate slowly equilibrates.High Overshoot — Moderate near-term action + massive future CDR. Peaks at 2.7°C around 2122. Requires hundreds of GtCO₂ of removal.Extreme Overshoot — Current policies continued, then emergency CDR. Peaks at 3.4°C around 2122. The dashed 1.5°C line and red shading show the “danger zone.”
Data sources & methods Default mode (500 ka) Deep time: Jouzel et al. / EPICA Dome C ice core composite, 1,000-year averages, converted from years BP to CE. Supplemented by higher-resolution EPICA data for the last 25,000 years.Instrumental: HadCRUT5 v5.1.0.0 (Morice et al. 2021), global mean surface temperature, 1850–2025. Baseline: 1850–1900 mean.Projections: SSP1-2.6, SSP2-4.5, SSP3-7.0 from IPCC AR6 WG1 (2015–2099).
Extended mode (1.2 Ma) Deep time: Clark et al. (2024) , “Global and regional temperature change over the past 4.5 million years”, Science 386, eadi1908. GMST reconstruction with area-weighted global mean, converted from 0–5 ka baseline to 1850–1900 baseline.Holocene: Osman et al. (2021) , Last Glacial Maximum Reanalysis (LGMR), 200-year resolution, −22,000 to 1850 CE.Instrumental: Trewin et al. (2022) , consolidated global mean surface temperature, 1850–2020.Projections: MAGICC v7.5.0 (Nicholls et al. 2020), SSP1-2.6, SSP2-4.5, SSP5-8.5 (2015–2300). Percentiles (5th/50th/95th) from 600 ensemble members.
Why the two modes use different projection methods
Both modes draw their SSP projections from the same underlying source: the MAGICC v7.5.0 ensemble (Nicholls et al. 2022), which runs 600 model configurations for each Shared Socioeconomic Pathway. The difference lies in how those 600 runs are filtered.Default mode (500 ka) — observation-constrained: The IPCC AR6 applied observational constraints to the MAGICC ensemble — ruling out runs whose implied climate sensitivity fell outside the range supported by instrumental records, paleoclimate evidence, and process-based models. The result is a narrower, more rigorously assessed distribution. However, the IPCC only performed this assessment out to 2100 . No observation-constrained projections exist beyond that year, which is why the default mode ends at 2100.Extended mode (1.2 Ma) — unconstrained: Following Ripple et al. (2025) , this mode uses the full unconstrained 600-member ensemble, which extends to 2300 . Because it does not filter out high-sensitivity runs, the unconstrained median for SSP5-8.5 at 2100 is approximately +5.0°C — about 0.6°C above the IPCC assessed best estimate of +4.4°C. For SSP2-4.5 and SSP1-2.6, the difference is negligible (<0.1°C). ERA5 daily overlay ERA5 (Hersbach et al. 2020), global mean 2m air temperature (GSAT), daily resolution, 1940–present. Provided by the Copernicus Climate Change Service (C3S) via the Climate Pulse dashboard. The same ERA5 dataset used by The Climate Brink .Baseline conversion: the raw ERA5 anomalies are relative to 1991–2020. We add +0.88°C (the C3S official offset , adopted Autumn 2021 per IPCC AR6 ) to convert to the 1850–1900 preindustrial baseline. Cross-check: Copernicus reported 2024 as +0.72°C (1991–2020) = +1.60°C (preindustrial); 0.72 + 0.88 = 1.60. ✓GSAT vs GMST: ERA5 measures air temperature 2 m above all surfaces globally (GSAT). The primary curve (HadCRUT5) blends land air temperature with sea surface temperature (GMST). Air over the ocean warms faster than the ocean surface, so GSAT > GMST by roughly 0.05–0.10°C. This is a genuine physical difference, not a calibration error. The IPCC headline figures use GMST; reanalysis products like ERA5 use GSAT.
Historical events Search and "Surprise Me!" draw from a curated database of 865 events , 238 historical figures , and 173 multi-year spans , shared with the Holocene Interactive timeline. Land use (anthromes)
Classification from Ellis, Beusen & Klein Goldewijk (2020) , “Anthropogenic Biomes: 10,000 BCE to 2015 CE,” Land 9(5):129. Their Anthromes 2.1 system defines 20 land-use classes in 6 levels, based on population density, agriculture intensity, and land cover. Here, this classification is applied to updated population and land-use data from the HYDE 3.5 database (the original 2020 paper used HYDE 3.2).HYDE 3.5 (History Database of the Global Environment): Klein Goldewijk, Beusen, Vander Linden & Doelman (2025) , hosted at Utrecht University YODA . Licensed CC BY 3.0 .128 time points spanning 10,000 BCE to 2025 CE. Time resolution increases toward the present: 1,000-year steps from 10,000 to 1000 BCE, 100-year steps through most of recorded history, 10-year steps from 1700, and annual data from 1950 to 2025.Square root (√) scale: The vertical axis uses a square root transformation instead of a linear scale. This is a standard technique in scientific visualization (used in cartographic proportional symbols since Flannery, 1971 ) that compresses large values and expands small ones. Without it, Urban land (0.8% of Earth’s surface) would be invisible next to Seminatural land (24%). The axis labels show the actual percentages; their non-uniform spacing reflects the square root mapping. This preserves the relative ordering and direction of all trends while making every category visible.
Corridor boundaries
Safe and Just Climate Space: Rockström et al. / Earth Commission (2023) , “Safe and just Earth system boundaries”, Nature 619, 102–111.IPCC SR1.5 (2018) , Special Report on Global Warming of 1.5°C .
Climate Action Tracker pathways Climate Action Tracker (CAT), an independent scientific project by Climate Analytics and NewClimate Institute . Since 2009, CAT has tracked government climate action against the Paris Agreement goals. The November 2025 (COP30) update provides global mean temperature projections for four policy-based scenarios. Original decadal data (1990–2100) interpolated to annual resolution using monotone cubic Hermite interpolation (Fritsch-Carlson). Bridge point at 2025 = +1.41°C from HadCRUT5. See CAT methodology for full details.
Overshoot scenarios Schleussner et al. (2024) , “Overconfidence in climate overshoot,” Nature 634, 366–373. Five scenarios from the PROVIDE EU Horizon 2020 project. Temperature modeled with FaIR v1.6.2 (Finite Amplitude Impulse Response), 2,237 probabilistic ensemble members calibrated against IPCC AR6 assessed ranges. Uncertainty bands show the 10th–90th percentile range, reflecting geophysical climate response uncertainty.Bridge to observations: FaIR models Global Surface Air Temperature (GSAT), which differs slightly from observed surface temperature (HadCRUT5). A per-scenario offset aligns each curve to the observed 2025 value of +1.41°C, consistent with how existing SSP and Pathways data are treated.Key finding: Overshoot is not a free lunch. Even if temperatures eventually return to 1.5°C, irreversible consequences accumulate: every 100 years above 1.5°C adds ~40 cm of sea-level rise and ~0.02°C from permafrost and peatland feedbacks. Regional climate patterns are permanently altered, tipping elements may be triggered, and adaptation options are lost during the overshoot period.Data: Scenario data from Zenodo (CC BY 4.0) . Curves 1–2 (Low Overshoot, Peak & Decline) extend to 2100; curves 3–5 (Net-Zero Stabilization, High Overshoot, Extreme Overshoot) extend to 2300 in the 1.2 Ma data mode.
Key terms Preindustrial baseline The mean global surface temperature during 1850–1900, used as the reference point for all temperature anomalies. This is the standard adopted by the IPCC in its Special Report on Global Warming of 1.5°C and the Paris Agreement . Shared Socioeconomic Pathways (SSPs) Model-driven climate scenarios developed for the IPCC Sixth Assessment Report . Unlike the policy-based Pathways , SSPs explore hypothetical worlds that differ in demographics, economics, and technology. They are not predictions — they are “what if” stories that bracket the range of possibility. Select What If in the toggle to view these. This visualization includes:SSP1-2.6 — Sustainable development, strong mitigation, net-zero by mid-century. Warming likely stays below +2°C.SSP2-4.5 — "Middle of the road" — intermediate pathway, moderate mitigation efforts.SSP3-7.0 (500 ka mode) / SSP5-8.5 (1.2 Ma mode) — High emissions pathways with limited mitigation, representing worst-case trajectories.
Pleistocene The geological epoch spanning ~2.6 million to ~11,700 years ago, characterized by repeated glacial-interglacial cycles. The visualization shows the most recent 500,000 years (or 1.2 million in extended mode) of this epoch. Holocene The current geological epoch, beginning ~11,700 years ago. The remarkably stable, warm interglacial period during which all human civilizations arose. The Corridor of Life concept is fundamentally about this stability. Frequently asked questions Why does the Holocene look smooth in 1.2 Ma mode but spiky in 500 ka mode?
It comes down to temporal resolution. The 500 ka mode uses EPICA ice-core data, which preserves detail down to roughly 100-year intervals for the most recent millennia. You can see individual cold snaps (the 8.2 ka event, the 4.2 ka drought) and warm peaks. The 1.2 Ma mode stitches together deep-ocean sediment records (Clark et al.) and the Osman et al. reanalysis, both of which smooth the signal over longer windows — roughly 200–1,000 years . At that resolution, the Holocene’s small wiggles are averaged away, leaving a steady plateau. Neither view is wrong — one is a close-up, the other a wide-angle lens.
Why do interglacial peaks (like the Eemian) look so different between the two modes?
Around 125,000 years ago (the Eemian, or Last Interglacial — MIS 5e), the 500 ka mode shows a sharp warm peak (+1.9°C to +2.7°C), while the 1.2 Ma mode shows a much more muted one (+0.9°C). This is not a bug — it reflects a well-known discrepancy between proxy types.500 ka mode uses the EPICA Dome C ice core (Jouzel et al.), which measures deuterium (δD) in Antarctic ice — a proxy for local temperature that is then scaled to a global estimate. Ice cores preserve sharp peaks because they record conditions at the time snow fell.1.2 Ma mode uses Clark et al. (2024) , which derives temperature from a global benthic δ18 O stack — oxygen isotopes measured in deep-ocean sediment cores. This signal integrates both deep-water temperature and global ice volume, both of which change slowly. It dampens short-lived warm excursions that last only a few thousand years.all interglacial peaks: ice-core proxies consistently show warmer, sharper maxima than ocean-sediment records. Clark et al. themselves note that their reconstruction may underestimate peak interglacial warmth. Neither method is “wrong” — they measure different aspects of the climate system at different timescales.
How do scientists know what the temperature was thousands or millions of years ago?
No one was holding a thermometer 100,000 years ago, so climate scientists rely on proxies — physical or chemical traces left in natural archives that vary with temperature. The main methods, from most recent to most ancient:Tree rings (dendrochronology) — Width and density of annual growth rings record temperature and moisture year by year, reaching back about 12,000 years in the longest records. Best for regional climate, seasonal detail.Ice cores — Bubbles of ancient atmosphere and the isotopic composition of water molecules (δD, δ18 O) trapped in polar ice sheets. The EPICA Dome C core (used in 500 ka mode) reaches back 800,000 years . Ice cores are prized for their time resolution and the fact that they directly sample ancient air.Ocean sediment cores (benthic foraminifera) — The shells of tiny bottom-dwelling organisms (“forams”) accumulate on the seafloor over millions of years. Their oxygen isotope ratio (δ18 O) reflects both deep-ocean temperature and global ice volume. This is the backbone of Clark et al.’s 1.2-million-year record.Speleothems (cave formations) — Stalagmites and stalactites grow layer by layer; their isotopic chemistry records local temperature and rainfall. Useful for regional detail, especially in the tropics.Corals — Tropical coral skeletons record sea-surface temperature in their chemistry, with seasonal to annual resolution, going back hundreds of thousands of years.Instrumental records — Thermometers since the mid-1800s. The HadCRUT5 dataset (Morice et al. 2021) used here compiles thousands of weather stations and ship-based ocean measurements into a global mean from 1850 to the present . How do scientists project future temperatures?
Future projections are not forecasts in the weather-report sense — they are “if…then” calculations . Scientists feed a storyline about future human choices (how much fossil fuel we burn, how fast we decarbonize) into Earth system models — mathematical representations of the atmosphere, oceans, ice sheets, land surface, and carbon cycle. The models compute the climate response.Shared Socioeconomic Pathways (SSPs) are five such storylines, developed for the IPCC, ranging from aggressive decarbonization (SSP1) to continued fossil fuel expansion (SSP5). Because no one knows exactly how sensitive Earth’s climate is to a given amount of CO₂, each SSP is run through dozens of different models (or, in the case of MAGICC, hundreds of parameter configurations). The spread between those runs is the shaded uncertainty band you see around each curve.Pathways projections (Climate Action Tracker) take a different approach: instead of hypothetical storylines, they start from actual laws, targets, and pledges that governments have filed, and calculate where that political reality leads.Overshoot scenarios (Schleussner et al.) focus on a specific question: what happens if we temporarily exceed 1.5°C and then rely on Carbon Dioxide Removal to bring temperatures back down? These use the FaIR model with over 2,000 ensemble members to map the risks.
Why does the temperature line sometimes go above the colored stripes?
In Stripes mode, each vertical bar represents the average temperature over a time bin (whose width depends on zoom level). In Spectrum mode, the white line connects individual data points at their actual values. When a brief warm spike falls inside a time bin that also includes cooler years, the bin’s average color will be cooler than the spike. Switching to Spectrum mode reveals these extremes; switching to Stripes mode emphasizes the long-term trend.
What is the “Adaptive” time scale?
Half a million years is hard to fit on one screen and show the last 200 years in useful detail. The adaptive scale solves this by allocating more horizontal space to recent centuries and less to deep time, with smooth transitions between segments. Zigzag break marks on the axis show where the scale changes density. If you want to see the true, undistorted proportion of deep time to recent history — where civilization is a nearly invisible sliver — switch to Linear .
Is the current warming unusual compared to past warm periods?
Past interglacials (like the Eemian, ~125,000 years ago) reached similar or even slightly higher global temperatures — but they took thousands of years to get there, driven by slow shifts in Earth’s orbit. The current warming is happening over decades , driven by greenhouse gas emissions. It is not the destination that is unprecedented — it is the speed . Ecosystems and ice sheets that had millennia to adjust in the past now have a century or less.
Keyboard shortcuts + / − — Zoom in / out0 — Reset zoom← / → — Pan left / rightEsc — Close panels, clear search Suggested citation GLOBAÏA (2026). The Corridor of Life — Up to 1.2 Million Years of Climate [Interactive visualization]. Available at: globaia.org/corridor-of-life
Created by GLOBAÏA
Cultivating planetary awareness through science and art
globaia.org
