Today’s Combination of CO₂ and Temperature Is Unprecedented in 66 Million Years

How One Diagram Reveals the Climate Change Story

Dean Rovang · substack.com/@deanrovang · ORCID: 0009-0006-1351-5320

Abstract

Three independent climate records — a deep-time Cenozoic reconstruction, an 800,000-year ice-core archive, and the modern instrumental record — converge on the same CO₂–temperature relationship and meet at a single point in 1850, the moment the industrial era began. From that hinge, the modern trajectory departs in a direction with no precedent in 66 million years. This essay shows how that single diagram structurally defeats the most common climate myths, explains why current warming understates where the physics is pointing, and places the standard 2100 endpoint of climate projections in its proper paleoclimate context. The case does not rest on models, projections, or consensus. It is written in the paleoclimate record itself.

Author’s Note

Today’s combination of CO₂ and temperature is unprecedented in 66 million years. That is the central finding of the diagram at the heart of this essay — and it is not a model result. It emerges from three completely independent climate records, each developed by different scientific communities using different methods to answer different questions, plotted together on a single set of axes.

I am a retired research engineer from Sandia National Laboratories, where I developed experimental capabilities for fusion and high-energy density physics on the Z Machine. Since retiring in 2018, I have been drawn to the broader question of how science is communicated — and increasingly frustrated by how much of the climate debate takes place in a context too narrow to reveal what the physics actually shows.

I sit in the gap between scientist and science communicator. I am not a climate scientist, and I do not pretend to generate the research. But I am trained to read it, to evaluate independent lines of evidence, and to recognize when they converge on something real. That instinct — overlay independent diagnostics; if they converge, trust the result — is what produced the diagram at the center of this essay. What I found, and what I have not seen presented this way before, is that three completely independent climate records land on top of each other, and that the modern trajectory occupies a position with no precedent in 66 million years of Earth history.

This essay is written for two audiences. The first is the general reader who wants to understand the climate change story on its own terms, without appeals to models or consensus. The second is the technically trained reader who may find value in seeing the empirical record assembled this way — as a teaching tool and as a framework for contextualizing model projections. Both audiences will find, I hope, that the diagram does most of the work.

This is not a policy argument. It is an attempt to show what the physics says, grounded in the climate record itself.

I. The Climate Record Beyond Models

Much of the public debate about climate change centers on climate models — their accuracy, their assumptions, their uncertainty ranges. That debate is legitimate and ongoing. Models are tools for projecting how fast temperature responds to rising CO₂, which feedbacks accelerate or moderate that response, and what the regional and seasonal texture of future change will look like. They are indispensable for those questions.

But the most fundamental relationship in Earth’s climate system does not come from models at all. It comes from the climate record itself. Nature does not know anything about climate models. The relationship between atmospheric CO₂ and global temperature was being written into geological archives long before humans existed to read it, and long before anyone built a model to simulate it.

The diagram introduced in this essay does not project the future. It places the present moment within the context of Earth’s past. That context — 66 million years of empirical data — provides an independent constraint on where the physics is pointing, one that stands entirely apart from any model.

That is where this essay begins.

II. A Geological Relationship

When researchers reconstruct atmospheric CO₂ and global temperature across the Cenozoic era — the last 66 million years, following the extinction of the non-avian dinosaurs — a coherent relationship emerges. Higher CO₂ is consistently associated with warmer global temperatures. Lower CO₂ is consistently associated with cooler temperatures. The relationship is approximately logarithmic, reflecting the radiative physics of greenhouse gases and the Earth system’s long-term adjustment toward equilibrium.

This pattern is visible in Figure 1, drawn from the synthesis by Judd and Tierney and colleagues published in Science in 2024. That work combined hundreds of paleoclimate records with improved statistical and spatial methods to reconstruct Earth’s surface temperature over geological timescales. What is striking about the result is how tightly temperature tracks CO₂ over the Cenozoic — despite changes in continents, ocean circulation, ice sheets, and ecosystems. The tightness of the relationship is remarkable given the complexity of the Earth system and the uncertainties inherent in deep-time reconstructions.

Figure 1. Deep-time equilibrium between CO₂ and global temperature across the Cenozoic era. Source: Judd et al. (Science, 2024).

The slope of this relationship corresponds to an Apparent Earth System Sensitivity (AESS) of approximately 8.2°C per doubling of CO₂. The word “apparent” is an honest qualifier — this value is derived from regression over geological data, not from an energy-balance calculation. It reflects the combined influence of fast and slow feedbacks operating over long timescales: not just the rapid atmospheric responses of water vapor and sea ice, but the slower adjustments of ice sheets, vegetation, and the carbon cycle that take centuries to millennia to fully express.

This equilibrium curve will serve as the reference frame for everything that follows. It is the destination — not a projection, not a consensus position, but an empirical statement about where the physics points for a given CO₂ concentration, given enough time for the system to fully respond.

III. The Pleistocene Cloud

Figure 2 — the diagram at the center of this essay — overlays three independent datasets on the same CO₂–temperature axes. The first is the Judd et al. equilibrium curve just described. The second is the Pleistocene cloud: 800,000 years of glacial–interglacial variability from Antarctic ice cores.

Figure 2. Climate variability across three timescales plotted on common CO₂–temperature axes: (1) the deep-time equilibrium curve (black dashed) from Judd et al.; (2) glacial–interglacial variability over the past 800,000 years from the EPICA Dome C ice core record (purple); (3) the modern instrumental record since 1850 (green circles); and (4) a representative future trajectory under an intermediate Shared Socioeconomic Pathway, SSP2-4.5 (red dashed). Source: Rovang.

The purple cluster in Figure 2 records the rise and fall of the ice ages. These measurements come from Antarctic ice cores — primarily the EPICA Dome C record — which preserve ancient air bubbles containing past atmospheres alongside chemical proxies for local temperature. Antarctic temperature anomalies are scaled to global mean surface temperature and plotted at 1,000-year intervals stretching back 800,000 years.

When these data are placed on the same axes as the deep-time reconstructions, something important becomes visible. The purple Pleistocene cloud does not scatter randomly across the plot. It clusters along the equilibrium curve — a different dataset, from a different archive, developed by a different scientific community, converging on the same CO₂–temperature relationship.

The ice ages were substantial. Global temperatures swung by several degrees Celsius. CO₂ oscillated between roughly 180 and 280 parts per million across repeated glacial–interglacial cycles. But these were variations within a constraint — variability that remained broadly consistent with the underlying equilibrium relationship. The ice ages were not exceptions to the physics. They were the physics, expressed through orbital cycles that changed the distribution of solar energy reaching Earth, triggering changes in ice sheets and ocean circulation that amplified the temperature response.

It is worth noting that in many glacial transitions, temperature changes preceded shifts in CO₂. This is sometimes cited as evidence against CO₂ as a climate driver. The observation is correct, but the interpretation misses the mechanism. Small orbital variations initiated initial warming; as oceans warmed, they released dissolved CO₂, which then amplified the temperature increase. CO₂ functioned as a feedback in those cases, not as the original trigger. Today the sequence is reversed: human emissions are increasing atmospheric CO₂ directly, and temperature is responding to that forcing.

The slope of the Pleistocene cloud is if anything slightly steeper than the Judd et al. equilibrium curve. This likely reflects the influence of glacial–interglacial feedbacks — large ice sheets, surface albedo changes, and reorganizations of ocean circulation — that amplify temperature responses during ice-age cycles. In that sense the Pleistocene record captures a highly responsive climate regime. Both the deep-time curve and the Pleistocene cloud can be thought of as experimental measures of Earth System Sensitivity (ESS) — empirical, not modeled, derived from the actual behavior of the climate system over long timescales.

IV. The 1850 Hinge Point

The convergence between records becomes especially clear at the point where modern observations begin. Around 1850, atmospheric CO₂ was approximately 284 parts per million and global mean surface temperature was near its pre-industrial baseline. That point sits precisely at the intersection of all three independent datasets — the deep-time equilibrium curve, the Pleistocene cloud, and the start of the modern instrumental record.

That moment acts as a hinge. It connects the geological history of the Earth system, the natural variability record of the last 800,000 years, and the modern era of direct measurement — three lines of evidence, three different methods, one point in common. The system was at equilibrium right up until the industrial experiment began.

From this common starting point, the modern trajectory moves in a direction not seen in any of the earlier records. CO₂ rises rapidly as fossil fuel combustion releases carbon stored over millions of years. Temperature rises, but more slowly — the system cannot keep pace with the forcing being imposed on it.

This is not a model result. It is not a consensus position. It is three completely independent measurement systems — geological proxies, ice-core records, and direct instrumental observation — agreeing on where the climate system stood at the moment the modern era began.

Which brings us to the argument the diagram makes most powerfully.

V. A Coordinate Not Seen in 66 Million Years

Skeptics sometimes observe, correctly, that Earth has experienced today’s CO₂ levels before. They observe, also correctly, that global temperatures have been higher at various points in Earth’s history. Both statements are true. Neither is the point.

The diagram shows something those observations miss entirely. Nowhere in the past 66 million years — not once in the entire Cenozoic record — do we find the specific combination of CO₂ concentration and global temperature that exists today. Earlier climates with high CO₂ were also much warmer. The cooler climates of the ice ages had much lower CO₂. The modern climate occupies a different position entirely: CO₂ has risen rapidly while temperature has not yet caught up.

That is what is unprecedented. Not the individual values. The combination.

In terms of the diagram, the modern green circles sit below the equilibrium curve — in territory the climate system has not previously inhabited. The coordinate pair has no precedent in the paleoclimate record. It is the signature of a system being forced faster than it can respond, a disequilibrium with no analog in 66 million years of Earth history.

The argument this makes against cherry-picking is structural, not rhetorical. A skeptic can dispute the instrumental temperature record. Or the ice cores. Or the deep-time proxies. But not one at a time. All three datasets land on the same relationship. Three independent methods, developed by different communities without reference to each other, do not converge on the same answer by accident. Explaining that away requires all three to be wrong in exactly the same way. That is not skepticism. It is a conspiracy theory.

VI. Why Modern Warming Lags

The reason the modern observations sit below the long-term equilibrium curve is that the climate system has not yet reached equilibrium with the CO₂ already in the atmosphere. The system responds on multiple timescales, and not all of them are fast.

When greenhouse gases accumulate, the Earth’s energy balance shifts: more incoming solar energy is retained than is radiated back to space. That imbalance drives warming. But the climate system has enormous thermal inertia, particularly in the oceans, which absorb heat slowly and release it over decades to centuries. An ice cube does not melt the instant it is removed from the freezer; heat must first flow into it. In the same way, the Earth system takes time to warm as excess energy is redistributed through the oceans and atmosphere.

As the climate warms, fast feedbacks come into play over years to decades: water vapor increases, sea ice retreats, cloud patterns shift. These amplify the initial warming, but they too take time. Over longer timescales, slow feedbacks — changes in ice sheets, vegetation, and the carbon cycle — further amplify the temperature response. The Judd et al. AESS of approximately 8°C per doubling reflects all of these feedbacks fully expressed. The assessed Equilibrium Climate Sensitivity of approximately 3°C per doubling reflects only the fast feedbacks.

The slope of the modern instrumental record is neither. It is the transient climate response (TCR) — the warming observed while the system is still catching up to the CO₂ forcing. It understates where the system is headed even on decadal timescales, let alone centuries.

The gap between the modern trajectory and the equilibrium curve in Figure 2 is therefore not a discrepancy or an anomaly. It is the physical signature of a system still catching up to the forcing already imposed on it — the warming the system has yet to express in response to CO₂ already in the atmosphere.

The diagram makes this three-level hierarchy spatial and visible: the transient response at the bottom, the Equilibrium Climate Sensitivity (ECS) destination somewhere in between, and the AESS equilibrium curve at the top. Three different answers to the question “how sensitive is the climate?” — and the diagram shows why all three are correct for different timescales. Table 1 summarizes the three measures for reference.

Table 1. The three-level climate sensitivity hierarchy. Each measure answers the question “how sensitive is the climate?” correctly for a different timescale. Sources: TCR and ECS from IPCC AR6 Chapter 7; AESS from Judd et al. (Science, 2024).

VII. What Climate Models Are Trying to Predict

The three-level sensitivity hierarchy described in the previous section — TCR, ECS, and AESS — is not just a conceptual distinction. It defines the central problem that climate models have been built to address. The empirical record tells us the destination; models are our best attempt to describe the journey.

Figure 3 shows the same composite diagram with three additional lines added, each anchored at the 1850 preindustrial starting point. These lines represent Equilibrium Climate Sensitivity values of 1.5, 3.0, and 4.5 degrees Celsius per doubling of CO₂. The 1.5 and 4.5°C values defined the likely range in the IPCC Fifth Assessment Report; the Sixth Assessment Report narrowed that range to 2.5–4.0°C with a best estimate of 3.0°C. The three lines are retained here as a teaching tool — bracketing the full range of historical IPCC assessments and making the sensitivity question visually legible. Each line asks: if the fast feedbacks of the climate system settle at this sensitivity, where does a given CO₂ concentration take global temperature?

Figure 3. The same composite diagram as Figure 2 with ECS reference lines added. Blue dashed lines show equilibrium warming for three values of Equilibrium Climate Sensitivity (1.5, 3.0, and 4.5°C per doubling of CO₂), all anchored at the 1850 preindustrial baseline. Source: Rovang.

The SSP2-4.5 trajectory — SSP stands for Shared Socioeconomic Pathway, and the 2-4.5 variant represents a mid-range emissions scenario — tracks close to the 2.5°C ECS line through 2100, and the modern instrumental record since 1850 implies almost exactly the same value. This is consistent with the Transient Climate Response being somewhat lower than ECS: the ocean has not yet fully absorbed the heat being added, so the observed warming understates the equilibrium destination. A model with an ECS of 3°C or higher will show a transient response closer to 2.5°C — which is where both the modern record and the SSP trajectory sit. Notably, 2.5°C sits at the lower bound of the IPCC AR6 likely range of 2.5 to 4.0°C. The models are being honest about the distinction between transient and equilibrium response, not conservative about the outcome.

The hot model problem

When the IPCC’s Sixth Assessment Report was being assembled, something unexpected happened. A new generation of climate models — the CMIP6 ensemble — produced ECS values significantly higher than previous generations, with some models running at 5°C per doubling or above. These “hot models” ran substantially warmer than observational constraints suggested was physically plausible.

The scientific community did not dismiss this finding or accept it uncritically. Researchers cross-checked the high-ECS models against independent observational constraints — paleoclimate data, satellite energy imbalance measurements, and the historical temperature record. The hot models failed those tests. The AR6 assessment explicitly down-weighted them, narrowing the likely ECS range to 2.5–4.0°C and placing the best estimate at 3.0°C. The upper bound of the likely range had actually been tightened from the previous assessment.

This episode illustrates something important about how climate science works. The models are not the final word — they are hypotheses, tested against the empirical record. When models conflict with observations, the observations win. The process is self-correcting by design, and the AR6 ECS constraint is more confident, not less, because it was tested against multiple independent lines of evidence.

Uncharted territory

The CO₂ forcing now being applied to the climate system has no analog in the instrumental record and no close analog in the paleoclimate record. The rate of increase — roughly 2 to 3 parts per million per year — is unprecedented in the ice-core record and likely in the broader geological record over millions of years. The EPICA Dome C data in Figure 2 represent changes that unfolded over tens of thousands of years. We have accomplished the equivalent in roughly 175 years.

This means that climate scientists are doing something genuinely difficult: projecting the response of a complex system into a region of state space it has not previously occupied, at a rate of forcing with no precedent. The models are our best attempt to simulate a system that has never been in this state. The diagram shows why that is true. The ECS range provides the scientific community’s best assessment of where the fast-feedback equilibrium lies. The Judd curve shows where the physics ultimately points. The gap between them is not a model artifact — it is the slow feedbacks, waiting.

Scientists are doing the best they can to project the temperature response and its impacts under genuinely novel conditions. The uncertainty in that projection is real, and it runs in both directions. But the direction of the physics — toward the equilibrium curve, on timescales that extend well beyond 2100 — is not in doubt.

VIII. The Role of Climate Models

What the empirical record establishes independently

The equilibrium curve in Figure 2 sets the destination of the climate system for a given CO₂ concentration — and it does so without reference to any model. This is the diagram’s unique contribution. The physics is in the data itself, derived from 66 million years of the climate system’s actual behavior. Nature does not know anything about climate models, and the destination it encodes does not depend on any.

This matters because much of the public skepticism about climate projections is really skepticism about models — their parameterizations, their treatment of clouds, their regional accuracy. That skepticism may or may not be warranted in specific cases. But it is beside the point here. The destination is an empirical statement, not a model result.

What models do that the diagram cannot

Climate models are indispensable tools that the diagram cannot replace. They do things the empirical record alone cannot:

• Trajectory resolution — how fast temperature is rising, whether the rate is accelerating, and by how much. A recent analysis suggests the underlying warming trend has accelerated from roughly 0.18°C per decade to somewhere between 0.24 and 0.30°C per decade — a scientifically significant but actively debated finding. The diagram is deliberately silent on it. At the scale of 66 million years, the difference is invisible.

• Regional texture — monsoon shifts, Arctic sea ice loss, glacier retreat timelines, coral bleaching thresholds, extreme precipitation changes. These are the consequences that drive policy decisions on decadal timescales, and they require models to quantify.

• Scenario discrimination — the equilibrium curve shows where the system ends up for a given CO₂ level, but which CO₂ level depends on emissions choices. SSP1-2.6 and SSP3-7.0 land in very different places on the curve. Models translate emissions pathways into concentration trajectories, making them essential for connecting policy choices to physical outcomes.

Complementary, not competing

The empirical record and climate models answer different questions. The diagram provides what models alone cannot: an empirical grounding of the destination, independent of any assumptions baked into model parameterizations. Models provide what the diagram cannot: the rate of travel, the regional texture, and the consequences along the way.

Together they tell a more complete story than either can alone. A climate scientist showing that warming is accelerating needs the diagram to answer the follow-up question: accelerating toward what, and for how long? The diagram showing the destination needs models to show the path. The diagram is not a replacement for models. It is the context that makes model results matter.

2100 is a waypoint, not a destination

One of the most important things the diagram makes visible is that the year 2100 — the standard endpoint of most climate projections — is not the end of the story. Under SSP2-4.5, atmospheric CO₂ reaches roughly 600 parts per million by 2100. The red trajectory in Figure 2 ends there. But the equilibrium curve continues. The physics does not stop at 2100.

Research by Kirsten Zickfeld and colleagues, published in the Journal of Climate in 2013, examined multi-century model simulations of climate change commitment across a range of emissions scenarios. Their findings are sobering: after emissions reach net-zero, atmospheric CO₂ remains elevated for centuries — above 50 percent of its peak value at year 3000 — and temperature remains correspondingly high. The duration of elevated conditions grows longer the further the system travels before net-zero is reached. Sea level rise continues for centuries not because warming is locked in, but because the system remains displaced from equilibrium for so long that the slow feedbacks keep expressing.

Zeke Hausfather’s explainer at Carbon Brief provides an accessible treatment of what the science says happens to temperatures after net-zero is reached.

The diagram frames this directly. Reaching net-zero does not send the modern trajectory back toward the equilibrium curve. It stops the rightward movement. The system remains displaced below the curve and slow feedbacks continue to close the gap over centuries to millennia. The farther the system travels from its natural equilibrium range before net-zero is reached, the larger that gap — and the longer the journey back.

The impacts of decisions made today will be felt by people living in the 28th century and beyond. That is not a rhetorical point. It is a physical consequence of the climate system’s long memory, visible directly in the diagram.

IX. What This Diagram Does

Having walked through the science, it is worth being explicit about what the diagram accomplishes — both as a teaching tool and as a framework for engaging skeptics.

Myths it structurally discredits

• “Climate has always been changing.” Yes — and here are 66 million years of it on one plot. The relevant question is not whether climate changes naturally, but whether today’s combination of CO₂ and temperature has any precedent in that record. It does not.

• “It’s just a natural cycle.” The modern trajectory breaks from 800,000 years of glacial–interglacial cycling. The Pleistocene cloud shows the full range of natural variability. The green circles are not in it.

• “CO₂ and temperature aren’t linked.” Three independent datasets — deep-time proxies, ice cores, and direct measurement — developed by different communities using different methods, all show the same thing: when CO₂ goes up, temperature goes up. The relationship is consistent across 66 million years.

• “Models are driving the alarm.” The equilibrium curve is empirical, not modeled. The destination is written in the paleoclimate record, not in any simulation.

• “Current warming is within natural variability.” Today’s CO₂–temperature coordinate has no precedent in 66 million years. The modern system is rapidly being driven away from Earth’s natural equilibrium. It literally takes a hard right at the 1850 hinge point.

What it makes visible that other plots do not

• The disequilibrium gap — the distance between where the system is and where the physics is pointing — is visible as a spatial feature of the diagram, not an abstraction.

• The lag between CO₂ forcing and temperature response is intuitive: the system is chasing the curve.

• The three-level sensitivity hierarchy — TCR, ECS, and AESS — is made spatial. Three different answers to “how sensitive is the climate?” are all visible simultaneously, each correct for a different timescale.

• The 1850 convergence shows the system was at equilibrium at the moment the experiment began — making the departure from that equilibrium unambiguous.

As a teaching tool

• Introduces logarithmic CO₂–temperature scaling intuitively, without equations.

• Compresses 66 million years of climate history into a single coordinate system accessible to a non-specialist.

• Demonstrates convergence of independent evidence — a core principle of how scientific confidence is built.

• Naturally motivates the distinction between where we are, where we are headed by 2100, and where the physics ultimately points.

• Requires no models to interpret. The physics is in the data itself.

X. The Longer View

Over geological timescales, the Earth system’s tendency to return to equilibrium is one of the most consistent features of the climate record. The 66-million-year sweep of the Judd et al. reconstruction — spanning climates both much warmer and much cooler than today, through continental rearrangements, mass extinctions, and a sun that has grown roughly 4 percent brighter over that span — shows a system that has always found its way back toward the CO₂–temperature relationship.

That tendency is real. But “eventually” means something specific here. It means millennia to tens of millennia. It means that once the anthropogenic perturbation has finally worked its way through the slow feedbacks — the ice sheets, the deep ocean, the carbon cycle — the system will settle along the equilibrium curve again. That trajectory has been interrupted.

Some argue that because these feedbacks operate slowly, humanity will have time to adapt. That argument deserves scrutiny. The slow feedbacks are slow on geological timescales, not on human ones. Sea level rise driven by ice sheet loss does not wait for civilization to adjust its coastlines. The reorganization of climate zones does not pause while agriculture relocates. Sea level rise already committed by warming today will reshape coastlines within the lifetime of infrastructure being built right now and will continue for centuries. Slow, in this context, means centuries — not indefinitely deferred. The commitment is being made now. The consequences arrive later, on timescales that span many generations.

At its core, that is a moral question — about the obligations of one generation to all the ones that follow. The diagram does not answer it. But it makes the scale of the commitment visible in a way that models alone cannot, because it places the modern trajectory in the context of where the system has been for 66 million years, and where the physics says it is still pointed.

XI. What Would Have to Be Wrong

The diagram builds its case from multiple independent lines of evidence. Before concluding, it is worth making explicit what a skeptic would need to reject in order to dismiss that case.

To take exception with the conclusions of this essay, one must identify which of the following is wrong:

• The Judd et al. equilibrium curve. Sixty-six million years of proxy data from multiple independent methods showing a robust logarithmic relationship between CO₂ and temperature. If this is wrong, the error must be consistent across five different climate regimes, three ocean configurations, and two major ice age cycles, all studied by independent research groups.

• The Pleistocene ice core record. Eight hundred thousand years of glacial–interglacial data from Antarctic ice cores, independently corroborating the same CO₂–temperature relationship. If this is wrong, the error must align in exactly the same direction and magnitude as the error in the Judd curve — derived by entirely different methods from entirely different archives.

• The modern instrumental record. Berkeley Earth global mean surface temperature and Mauna Loa CO₂ measurements — the most directly verified of the three datasets, confirmed independently by eight other major research groups using different methodologies. If this is wrong, the error must somehow align perfectly with the errors in the other two.

• The 1850 convergence. All three datasets meeting at a single point — the moment the industrial experiment began. A coincidence of this precision across independent records requires explanation.

• Basic radiative physics. The greenhouse effect, established before modern climate science, confirmed in laboratory experiments, and observed directly in the atmosphere. This is the mechanism that connects CO₂ to temperature across all three records.

These are not one theory with five supporting pieces. They are five independent lines of evidence that happen to tell the same story. Dismissing the conclusion requires dismissing all of them simultaneously — and explaining why independent methods, developed by different scientific communities to answer different questions, all arrive at the same wrong answer.

Accepting that conclusion leads immediately to a further question: for how long? The Zickfeld commitment answers it. That the duration of elevated temperature and CO₂ extends 700 years or more after net-zero emissions is not a separate hypothesis to be proven — it is a physical consequence of the same carbon cycle and ocean heat capacity that explains why the modern trajectory lags the equilibrium curve in the first place.

The remaining uncertainty lies not in whether the relationship exists, but in how far the modern trajectory carries the climate system from the range that defined early human civilization — and what the passage back looks like.

Data Sources and Reproducibility

The figures in this essay synthesize multiple publicly available datasets spanning different timescales:

• Deep-time equilibrium: Cenozoic temperature and CO₂ reconstructions synthesized by Judd et al. (Science, 2024).

• Pleistocene variability: Antarctic ice-core CO₂ and temperature records from the EPICA Dome C reconstruction, with Antarctic temperature anomalies scaled by a factor of 0.55 and added to a preindustrial baseline of 13.8°C to estimate global mean surface temperature; plotted at 1,000-year intervals.

• Instrumental temperature: Berkeley Earth global mean surface temperature estimates.

• Modern and long-term CO₂: Long-term CO₂ compilations from Our World in Data.

• Future trajectory: SSP2-4.5 scenario projections assessed in the IPCC AR6.

• Long-term commitment: Zickfeld et al. (Journal of Climate, 2013).

All datasets are publicly accessible. The synthesis presented here reflects straightforward transformations and plotting intended to place observations across timescales into a common CO₂–temperature framework.