A New Reconstruction of Phanerozoic Temperature and CO2

In January, I blogged about a lecture I watched from Jessica Tierney, a geologist who has done some fascinating work in paleoclimate, in which she described the research behind a new paper that was at the time still undergoing peer review. You can watch the video and see my previous thoughts about this here. Ever since I watched this lecture, I have been anticipating the publication of the paper and hoping that the text wouldn't be behind a pay wall so I could learn more about what Tierney shared in her lecture. Well, late last week, both of my hopes became a reality. The paper is published and the full version is available. The lead author is Emily Judd[1], and this appears to be a remarkable paper.

The tl;dr for this paper is that Emily Judd and her colleagues put together a data analysis (PhanDA) reconstruction of global temperatures and CO2 for the last 485 million years (most of the Phanerozoic), and they found that GMST varied greatly on geologic time scales, ranging from 11°C to 36°C. They also did a simple regression analysis between log(CO2) and GMST and found an “apparent” earth system sensitivity of GMST of 7.7 ± 0.3°C for 2xCO2 (1σ).

PhanDA Reconstruction

In my previous post about Tierney's lecture, I shared a little bit about how the PhanDA reconstruction works. It's most analogous to the types of reanalysis products used in weather forecasting, where a model "prior" can be updated with proxy evidence in Bayesian fashion. This type of DA work has been shown to be skillful, even when working with comparatively data poor time periods in the Phanerozoic. What I appreciate about this paper is not that it provides any sort of final word on paleo climate or climate sensitivity. It certainly isn't. But the PhanDA project can be updated as more proxy evidence is collected, and the the analysis can even be run with a higher resolution, more state of the art model "prior" when computing capacity and time make that more feasible. But for right now, this seems to be a very large step forward compared to previous attempts I've seen, particularly those attempting to reconstruct global temperatures and CO2 in the Paleozoic and Mesozoic.

There's a lot that could be shared here, but I'd like to focus on two aspects of this paper that are likely to be relevant to current discussions about climate change, and in particular AGW. First, I'd like to cover what can be inferred about sensitivity in this paper, and second, I'd like to point out a couple common objections to AGW that this paper (as well as others previously published) should put to rest for good.

Apparent Earth System Sensitivity

This paper did a simple regression analysis of GMST with log(CO2) and found a correlation with a slope of 7.7 ± 0.6°C for 2xCO2 (2σ) and with r = 0.72. The paper calls this “apparent” earth system sensitivity (I'll call this aESS). This is intriguing and plausible finding, but I think it needs context, much of which is described in the paper but may not be reported as frequently. Any sensitivity estimate technically needs to account for external forcings distinguished from feedbacks, where forcings in W/m^2 add, and sensitivity can then be calculated as the slope of GMST/forcings and then converted to temperature by multiplying by the forcing for 2xCO2 (usually in the range of 3.7 to 4 W/m^2). This paper makes no attempt to account for all known external forcings; instead, it does a simple regression of log(CO2) and GMST; this would be a true sensitivity estimate for ESS if all non-CO2 forcings canceled out or were negligible. I think this aESS value is very useful to demonstrate that the Earth's climate system is highly sensitive to forcings on geologic time scales, but without further analysis of other external forcings, I think it's premature to say that ESS is significantly higher than ~6°C for 2xCO2 on geologic time scales. I do think it is fair to say that overall 2xCO2 correlates with almost 8°C on geologic time scales, but those statements are not the same.

I think it's also important to acknowledge that the observed correlation is strongest during the Cenozoic and Paleozoic, but it's not as apparent during the Mesozoic. The correlation observed for the Cenozoic was r = 0.97 (P < 0.01), where the data density was the highest for both CO2 and GMST. The correlation was less strong during but still well-established during the comparatively data-poor Paleozoic, where the correlation was r = 0.73 (P < 0.01). But the relationship is not clear at all during the Mesozoic, where the correlation is r = −0.37 (P = 0.18). The authors suggest that the reason for this lack of correlation may be due in part to the fact that there is not nearly as much of a range of of climate states represented in the Mesozoic as there are in the Paleozoic and Cenozoic; "the compressed CO2 range (and large uncertainties) makes it difficult to detect a CO2-GMST relationship." The source of the discrepancy appears to be that their CO2 reconstruction does not increase with mid-Cretaceous hothouse, and others have noted this disconnect in the past. It could be that there are other forcings operable during the Cretaceous that account for this apparent discrepancy, but we simply don't know. The authors refer to this as the "Mesozoic Conundrum," and it will be interesting to see if new research on CO2 and GMST during the Cretaceous can resolve this. But even so, using all the data we have from the Phanerozoic, the correlation is clear, and it shows that a climate response to CO2 on geologic time scales is a robust feature of the climate system. It just so happens that the "outliers" seem to bunch up in the Mesozoic, where perhaps either proxy evidence for CO2 or GMST need improving, or perhaps where other external forcings play a larger role.

Putting Objections to Rest

There are two claims I commonly hear from those who argue against mainstream climate science, one is that it's the Sun, not CO2 that drives climate, and the second is that CO2 is currently saturated and any further increase in CO2 will not cause significantly more warming. Both of these claims have been shown to be false using multiple lines of evidence, but this paper should be seen as a kind of nail in the coffin for these objections.

False Claim #1: The Sun Drives Warming

I've noted before that the Sun has increases in luminosity over time on geologic time scales. When the Earth was formed, the Sun was only about 70% as bright as it is now, and TSI has been increasing in roughly linear fashion throughout geologic history. This means that we know that solar forcings were almost 10 W/m^2 less 485 million years ago than they are now. This means necessarily that if the Sun was the primary driver of global temperatures, the Earth would have had to have been significantly colder at the beginning of the Phanerozoic than it is today. But this is not what we observe at all. there were times in the Ordovician and Silurian during which temperatures were in "hothouse" conditions. This occurred when CO2 was high but solar forcings were low.

This has been observed by others[2] in the past, but it is reinforced here. The reconstruction of GMST for the last 485 million years is not reconcilable to what we know about solar forcings across the Phanerozoic without the greenhouse gas drivers for global temperatures.

False Claim #2: CO2 Saturation Limits Future Warming

You sometimes hear that any additional CO2 added to the atmosphere by human activity cannot warm the planet any more than it has, because CO2 "saturates" at ~400 ppm; any future increases will have no discernable impact on GMST. I've responded to this in previous posts as well, but this paper observes that the the relationship between log(CO2) and GMST continues from coldhouse and into hothouse conditions: "The CO2–GMST relationship also indicates a relatively consistent aESS of ~8°C across the Phanerozoic." In other words, on geologic time scales, there is no evidence that CO2 "saturates" even when CO2 levels are extremely high. It can be clearly shown that this saturation claim is a myth without the evidence from this paper, but this paper should add an additional nail in the coffin for those who have not been convinced by the evidence available elsewhere.

Conclusion

I'm not the only one that appreciates this new work. Gavin Schmidt at RealClimate also has some very helpful things to say about it. I consider this paper to be a major step forward in paleoclimate studies, not just because it provides a coherent picture of past climates during the Phanerozoic but because it was done in such a way that future paleoclimate work can improve these results. It also shows a more -or-less consistent relationship between CO2 and GMST, though with areas of further research needed to resolve the so-called "Mesozoic Conundrum." I strongly suspect  that this paper will not cause people to stop objecting that the Sun drives climate or that CO2 is "saturated," but at the very least, this paper provides conclusive evidence that these objections are not founded upon any real-world data or observations. The evidence flatly contradicts their claims.

References:

[1] Emily J. Judd et al., A 485-million-year history of Earth’s surface temperature. Science 385,eadk3705 (2024).DOI:10.1126/science.adk3705

[2] Gavin L. Foster, Dana L. Royer & Daniel J. Lunt, “Future climate forcing potentially without precedent in the last 420 million years,” Nature Communications 8.14845 (2017).
https://www.nature.com/articles/ncomms14845