Russell's Teapot and the Climate Unicorn

Why RCP8.5 is not a proxy for anything

Dall-E dreams of the climate unicorn

Physicist Richard Feynman once described a key difference between religion and science as follows: “Religion is a culture of faith; science is a culture of doubt.” Scientific reasoning and provisional truths that result are based on evidence, not simply belief.

As debate has grown over the continued misuse of outdated climate scenarios in research and policy I’ve noted a tendency for defenders of RCP8.5 to rely on decidedly unscientific justifications for its continued misuse. Among defenders of the indefensible are leading scientists and even the Intergovernmental Panel on Climate Change (IPCC).

For instance, the recent IPCC AR6 report acknowledged our recent work on scenario plausibility but reasserted the value of RCP8.5:

In the scenario literature, the plausibility of some scenarios with high CO2 emissions, such as RCP8.5 or SSP5-8.5, has been debated in light of recent developments in the energy sector. However, climate projections from these scenarios can still be valuable because the concentration levels reached in RCP8.5 or SSP5-8.5 and corresponding simulated climate futures cannot be ruled out.

The IPCC provides no guidance on the criteria for “ruling out” a scenario. The logic here is that RCP8.5 is “in” unless it can be “ruled out.”¹

This shift of the burden of proof to critics of RCP8.5 to rule it out reminds me of Russell’s teapot. In 1952 philosopher Bertrand Russell proposed the following analogy:

Many orthodox people speak as though it were the business of skeptics to disprove received dogmas rather than of dogmatists to prove them. This is, of course, a mistake. If I were to suggest that between the Earth and Mars there is a china teapot revolving about the sun in an elliptical orbit, nobody would be able to disprove my assertion provided I were careful to add that the teapot is too small to be revealed even by our most powerful telescopes. But if I were to go on to say that, since my assertion cannot be disproved, it is intolerable presumption on the part of human reason to doubt it, I should rightly be thought to be talking nonsense.

Similarly, when it comes to climate scenarios designed to help guide us into the future, under typical scientific reasoning the champions of those scenarios have the burden of proof to demonstrate to their peers that the proposed scenarios represent plausible futures — before they are prioritized in research and used in policy. The burden of proof is continuous as time moves on because scenarios become outdated, often very quickly. That is normal and to be expected.

Discussion of climate scenarios is not the only place in climate research where non-scientific reasoning has taken hold. Writing on the science of detection of trends in extreme weather and their attribution to human causes, Naomi Oreskes and Elizabeth Lloyd criticize the IPCC’s approach to this issue and argue that “there is no “right” or “wrong” approach to [detection and attribution] in any absolute sense.”

Instead, they suggest that we evaluate scientific claims — including the risks of making false claims — based on whose political interests are served by the claims, not their ultimate accuracy.

Lloyd and Oreskes explain the practical importance of sometimes claiming that there is an undetectable teapot orbiting the sun:²

. . . the traditional [IPCC] risk-based approach to extreme event [detection and attribution] may lead to a challenge in communication, and to the impression that climate science is less epistemically secure than it actually is. Climate scientists have been repeatedly asked in real-time interviews to deliver expert opinions on whether a particular extreme event—a drought, a flood, a hurricane—was attributable to [anthropogenic climate change, ACC]... but at present, without doing an attribution study, scientists are essentially obliged to say that they cannot answer this question. . . This may create the impression that there was no relationship between climate change and the event, even where there may have been one, thus leading to harmful under-appreciation of the extent to which ACC is already occurring, as well as to under-preparation for the future. Because no event can be attributed to climate change without an attribution study, this effectively means that scientists following community norms will nearly always convey the message that individual events are not related to climate change—or at least, that we cannot say if they are. In short, it conveys the impression that we just do not know, which feeds into both contrarian claims that climate science is in a state of high uncertainty, doubt, or incompleteness, and the general tendency of humans to discount threats that are not imminent.

The fact of the matter is that the IPCC has concluded that connections of carbon dioxide emissions and most types of extreme weather are “in a state of high uncertainty, doubt, or incompleteness.”³ The proposal here to change the scientific community’s norms is to elevate political expediency over empirical understandings.

On climate scenarios the IPCC AR6 Working Group 1 justified the continuing value of RCP8.5 based on an empirical claim:

That is because of uncertainty in carbon-cycle feedbacks which, in nominally lower emissions trajectories, can result in projected concentrations that are higher than the central concentration levels typically used to drive model projections.

And IPCC AR6 Working Group 3 presented a similar hypothesis:

[H]igh emissions cannot be ruled out for many reasons, including political factors and, for instance, higher than anticipated population and economic growth. Climate projections of RCP8.5 can also result from strong feedbacks of climate change on (natural) emission sources and high climate sensitivity (AR6 WGI Chapter 7). Therefore, their median climate impacts might also materialise while following a lower emission path . . .

If someone claims that there is an undetectable teapot orbiting the sun, I — and I suspect you as well — will need proof before taking that claim seriously. A hypothesis is not a conclusion. Extreme climate scenarios from low emissions trajectories are like orbital teapots.

The IPCC, among others, has hypothesized that a scenario with lower emissions, a level of cumulative emissions to 2100 at or below the RCP4.5 scenario can produce a radiative forcing at RCP8.5 levels. I am calling this hypothetical world a climate unicorn — I don’t think it exists out there in model land.

Across the 5,000+ official scenarios of the IPCC since 2000, not one has cumulative emissions of RCP4.5 levels or lower and a radiative forcing of 8.5 watts-per-meter-squared or greater. I am also unaware of any literature using earth system models that produces such a result from emissions scenarios using plausible inputs.⁴

It is increasingly common to hear justifications for the continued misuse of RCP8.5 based on an assertion that it serves as a meaningful proxy for these hypothetical worlds where low emissions lead to dramatic levels of climate change.⁵

OK, it is an interesting hypothesis — Show us the unicorn.

Note: Comments on this post are open to everyone, as I welcome input especially from those who defend RCP8.5 as a valuable proxy, claim to have seen the climate unicorn or whose work may actually have produced one.

Thanks for reading! THB is reader-supported and reader-engaged. If you are a subscriber, thanks. If you are not, please join in at whatever level makes sense for you.

1. Of course, we have already “ruled out” RCP8.5 — see our recent paper and Justin Ritchie’s classic demolition.

2. Oreskes has recently made a similar argument with respect to masks in COVID-19.

3. These views are not isolated — An assessment of tropical cyclones and climate change by the World Meteorological Organization adopted explicitly the approach recommended by Lloyd and Oreskes.

4. I am actually unaware of any such a result using implausible inputs.

5. Of course, we should never be using outdated scenarios as a proxy for anything. Instead, we should use updated scenarios and the results that emerge from their use. That we are even talking about using RCP8.5 as a proxy it itself an indication of problems in scenario development and application.