Arguments
Human-caused climate change is unmistakably distinct from Earth’s natural climate variability
Posted on 14 April 2026 by Guest Author
This is a re-post from Staying Curious by Dean Rovang
This post presents two figures that are the culmination of an extended effort to build the strongest possible empirical case for what the paleoclimate record shows about CO? and temperature. They draw on five independent regression fits across four independent archives and 66 million years of geological evidence. The argument stands on its own merits.
Earth’s natural climate relationship
Three completely independent temperature reconstructions — the EPICA (European Project for Ice Coring in Antarctica) Dome C ice core, the Snyder (2016) Bayesian stack of 61 marine sediment cores, and the Clark et al. (2024) area-weighted reconstruction of 111 marine sediment cores — are plotted against the same CO? record on common axes. All three share Bereiter et al. (2015) as their CO? source, the best available composite for this period. The geological stages from Judd et al. (2024) extend the record to 66 million years.
Five regression fits are shown. Four use York/ODR regression — the method that accounts for measurement uncertainty in both CO? and temperature simultaneously, and the method used by Judd et al. themselves. The fifth is the canonical EPICA OLS fit using the original EDC CO? source, included because it was the first fit introduced relative to the Judd curve and remains the most conservative estimate in the set. All five land between 8.24 and 9.91 K/doubling of CO? — the same tier regardless of dataset, CO? source, or regression method.
The figure spans Earth’s climate from deep ice ages — CO? near 175 ppm, global temperatures near 8°C — to the warm Cenozoic periods when there were no significant continental ice sheets and sea levels were 60–70 meters higher. A natural question is whether the ice-albedo feedback drives this slope — and whether it therefore only applies to glaciated climates. The figure addresses this directly: most of Judd’s data comes from greenhouse climates with little or no continental ice — the Eocene, Oligocene, early Miocene — yet the same slope holds across the full range. If ice-albedo were the dominant driver, the relationship should look different in ice-free regimes. It does not, which Judd et al. themselves describe as surprising. The relationship appears to be a fundamental property of the Earth system across a wide range of boundary conditions, with and without ice sheets.
The 1850 hinge point: where modern climate change departs
Figure 2. The same natural CO?–temperature relationship as Figure 1, with the modern instrumental record (green, Berkeley Earth + OWID/NOAA, 1850–2025) and SSP2-4.5 projection (red, CMIP6/AR6, 2025–∼2100) added. Open circles mark 1850, 2025, and ∼2100. The gap between the modern trajectory and the natural equilibrium fits reflects thermal lag and slow feedback mechanisms — not a change in the underlying physics.
Figure 2 adds the modern instrumental record and the SSP2-4.5 projection to the same axes. At 1850, the modern record begins exactly where the natural relationship sits — the hinge point where Earth’s climate history meets direct measurement. From there, CO? rises rapidly while temperature lags behind, driving the system into a region of CO?–temperature space with no analog in the natural record.
The gap between the modern trajectory and the natural equilibrium fits is labeled directly on the figure: thermal lag and slow feedback mechanisms. The fits in Figure 1 reflect the apparent Earth System Sensitivity (AESS) — Judd et al.’s term for the empirical CO?–temperature relationship across states where all feedbacks, including slow ones like ice sheets, vegetation changes, and carbon cycle responses, have had time to fully operate. Those slow feedbacks operate on timescales of millennia, not centuries. The gap is not a simple warming commitment. It is the distance between where the system currently sits and where the physics ultimately points, given enough time.
The figure also raises the question of reversibility — one of the most important questions in climate science. Will the system eventually return to the equilibrium curve? If so, where on the curve? How long after net-zero emissions is reached? What is the path, and what are the impacts along the way? The system will tend toward the equilibrium curve as long as CO? remains elevated, but the journey operates on geological timescales. The destination on the curve is determined by the CO? concentration at which the system eventually stabilizes — which is precisely why the timing of net-zero matters so profoundly. Sea level rise, driven by slow ice sheet dynamics, will continue for centuries regardless of when emissions reach zero. These are not hypothetical futures — they are physical consequences already set in motion.
The modern trajectory does not move toward the equilibrium curve. At any given CO? concentration, the gap keeps growing as CO? rises faster than temperature can follow. The only way to stop that gap from growing is to reach net-zero CO? emissions. The longer it takes to get there, the higher the CO? concentration at which the system stabilizes — and the higher the equilibrium temperature the physics ultimately points to. Clark et al. (2016) showed that decisions made in the next few decades will lock in sea level rise, ice sheet loss, and temperature change that persist for thousands of years — well beyond any planning horizon human civilization has ever contemplated. The equilibrium curve does not forget where we sent it. The physics does not negotiate on timeline. It only responds to what is in the atmosphere.
The bottom line
Five independent fits of Earth’s natural CO?–temperature relationship — derived from different archives, different methods, and different timescales — all land in the same tier. That relationship holds from the depths of the ice ages to warm greenhouse climates with no continental ice, across 66 million years of Earth history. The modern trajectory departs from it at 1850 and has not looked back. The departure is not subtle, not a matter of interpretation, and not an artifact of any single dataset or method. It is unmistakable.
Sources
Full sources in the preceding post: justdean.substack.com/p/unprecedented-in-66-million-years
Clark et al. (2016). Nature Climate Change, 6, 360–369. https://doi.org/10.1038/nclimate2923
But other data shows less of a link to CO2. https://pureadmin.qub.ac.uk/ws/portalfiles/portal/676577731/revisit_2.pdf, https://www.nature.com/articles/s41586-025-10032-y.
[BL] You fail to explain what it is you expect people to see at those links. The Comments Policy is explicit:
No link or picture only. Any link or picture should be accompanied by text summarizing both the content of the link or picture, and showing how it is relevant to the topic of discussion. Failure to do both of these things will result in the comment being considered off topic
Your first link is to an unpublished manuscript, written by a lecturer in a university business school. This person appears to have no background in climatology or carbon cycles. There is no reason to think that he has anything useful to add to the science.
The second link leads to a paywalled paper. The visible abstract discusses ice core records of CO2 and other gases, but gives absolutely no indication that they make any detailed examination of links between CO2 and global temperature.
Moderator Response @1,
The amateur analysis of 'global average' temperature linked @1 by rkcannon is entirely naive in its method and in its reporting of conclusions.
It concludes "...the notion that CO2 is the primary driver of global warming. If this were the case, periods with higher CO2 emissions would exhibit a faster rate of warming than periods with lower emissions," pointing to what the amateur calls his finding that "...long-term temperature rise was steeper in earlier periods when CO2 emissions were modest compared to current levels. These results hold despite changes in how time periods are defined ... and how weather stations are selected ... .")
The grand analysis supporting such a bold assertion looked at 100, 500 and then 992 selected weather stations (so all land sites), selected for the level of data available and then calculates the temperature trends for 42, 35, 30 & 21 year periods. The 100 station results presented show the temperature trends for the latest periods are by far the steepest in two centuries under analysis, 1815-2024. (42y +0.24ºC/decade, 35y +0.25ºC/dec, 30y +0.33ºC/dec, 21y +0.41ºC/dec) which of course entirely contradicts the conclusions presented in the analysis.
So that's worse than "amateur"!!
The other link @1 by rkcannon is to Marks-Peterson et al (2026) which is paywalled but an associated paper Shackleton et al (2026) 'Global ocean heat content over the past 3 million years' is not. These two papers drew coverage at RealClimate. Both papers examine very very old ice which provides data with less accurate age such that ice age cycles are fuzzed out.
The two papers are pointing to a more complex cooling 3My-0,5My bp. From the press release:-
Somehow there are crazy folk gleaning straws from the science to present misguided support for their crackpot version of reality. The account of Marks-Peterson et al (2026) nailed-up on the rogue planetoid Wattsuppia was headlined 'Shock New Evidence Showing No Link Between CO2 and Temperature Over Last Three Million Years Stumps Net Zero Activists' and such coverage prompted a few grownups to explain the true implications fo the two papers.
[BL] Thanks for taking the time to dig further than I had the patience for. Just the abstract of the first paper was enough to convince me that the author had no idea what he was talking about. From your description, it seems that he has no idea about time scales and can't differentiate between short-term variation and long-term trends. Of course, that error is common in the contrarian literature (e.g., Salby).
As for the second paper, I'm shocked to find out that there is gambling going on here a contrarian is taking a paper out of context. In the past, rkcannon has used NoTricksZone as a source of information. That source rarely gets anything right.
Rebuttal: Systems Analysis vs. Circular Reasoning
Subject: Response to Moderator Comments regarding Bhatta (2024) and Marks-Peterson (2026)
While the ad hominem labels—"amateur" and "naive"—provide a look into the moderator’s temperament, they do not address the physical and statistical discrepancies presented. As a Professional Engineer (PE), I prefer to evaluate the Transfer Functions of a system rather than the consensus of the "grown-ups."
1. On Circular Reasoning and System Gain
The moderator admits that the Nature paper (Marks-Peterson et al., 2026) requires "important contributions" from albedo and ocean circulation to explain a 2.5°C cooling while CO2 remained stable.
The Logic: To claim CO2 is the "key" control knob, only to demote it to a "passenger" whenever the data shows the planet cooling without its help, is circular reasoning.
The Math: Since the early 1900s, human CO2 emissions have increased by over 1,700%. If a seventeen-fold increase in the supposed "driver" results in a warming rate statistically similar to 1910, a rational systems analysis concludes the system is insensitive to that input.
2. The Failure of "Aerosol Masking"
The argument that mid-century cooling was "masked" by aerosols fails the spatial and modern test.
The Discrepancy: If industrial aerosols were a primary "cooling shield," China—with the world’s highest coal-related aerosol loading—should have been a global cool spot. Instead, China has warmed faster than the global average.
The Conclusion: You cannot invoke a "masking shield" to explain the 1940s cooling while ignoring its failure to stop warming in modern Asia. This is curve-fitting, not physics.
3. The Measured Driver: Albedo and the CERES Data
The moderator’s focus on 21-year surface trends ignores the most robust data set we have: the CERES satellite record.
The Data: Since 2000, CERES has measured a 0.8% drop in Earth’s albedo. This change in reflectivity has added roughly 2.7 W/m2 to the Earth's energy budget—effectively 100% of the warming forcing that the IPCC attributes to CO2 over the last 250 years.
4. The Missing "Fingerprint" and UHI Bias
If CO2 were the driver, the laws of physics dictate a "Tropical Hot Spot" in the upper troposphere. Decades of radiosonde and satellite data show this fingerprint is missing. The warming we do see is surface-based and highly correlated with Urban Heat Island (UHI) contamination. When you "homogenize" data by forcing rural stations to match urban trends, you aren't measuring global climate; you're measuring the encroachment of asphalt on thermometers.
Conclusion
Rational skepticism demands that models reconcile with empirical history. If the planet cooled 2.5°C with no change in CO2 in the Pliocene, and cooled for 40 years during a CO2 surge in the 20th century, the "Control Knob" theory is functionally dead. It is fascinating to watch the "Immune System" of this forum react; the Killer T-cells are working overtime to neutralize empirical data that looks like a "foreign invader" to the dogma. Nature doesn't care about your PhD or your moderation policy if your math is wrong.