Once again IPCC’s math doesn’t check out

The high and rising costs of climate policy—now including the inability of jurisdictions that bet big on renewables to guarantee enough energy for their citizens to survive the coming winter—not only entitles us to question the basis of these decisions but demands we do so.

Ultimately the justification is the view that carbon dioxide emissions have a big effect on the climate, which will cause devastating harm at some point in the future. Scientists measure the effect using a concept called “Equilibrium Climate Sensitivity” or ECS, which estimates how much long-run average warming will occur from doubling the amount of carbon dioxide in the atmosphere. Some important new evidence pointing to a low ECS value just emerged in the scientific literature.

ECS has long been shrouded in uncertainty. In 1979 the U.S. National Academy of Sciences estimated it to be between 1.5 and 4.5 degrees Celsius, with a best estimate of 3.0 degrees C. That range, which encompasses the range from no-big-deal to very bad outcomes, was accepted by the UN Intergovernmental Panel on Climate Change (IPCC) in its first report in 1990 and thereafter until 2007 when, citing greater warming projections in newer models, it raised the bottom end to 2.0 degrees C.

But over the next few years a literature developed using observed warming rates since the late-1800s, rather than model simulations, to estimate ECS and those results were typically centred around 2.0 C or less. So in 2013 the IPCC reduced the bottom end of the range back to 1.5 C and declined to offer a best estimate. In other words, after three decades climate science hadn’t narrowed the uncertainty at all.

The economic implications of ECS being 2 C rather than 3 C are enormous. Economic models used by the U.S. Environmental Protection Agency and others assume ECS is 3 C when computing the social cost of carbon. Some coauthors and I have shown that if the ECS parameter is instead centered around 2 C the estimated social cost of carbon plummets and becomes very small at least through the middle of this century. The justification for costly climate policy essentially disappears.

Given the discrepancy between models and observations, the IPCC changed the way it handled the ECS issue in its latest (2021) report. It no longer relied on model estimates, but it didn’t go with the existing estimates in the empirical literature either. Instead it turned to a 2020 paper by Australian climate scientist Steven Sherwood and 10 coauthors, which used a new technique to combine data from modern climate change with that from the end of the last ice age and even further back in the Earth’s history. They concluded the likely sensitivity range was from 2.6 to 3.9 C. On this basis the IPCC revised its estimate of the likely ECS range to between 2.5 and 4.0 C with a best estimate of 3.1 C. And specifically ruled out ECS being less than 2.0 C.

But as so often happens when a new paper appears in the literature that solves a political problem for the IPCC, they pounce on it before experts in the field have had a chance to check the numbers.

A new paper in the peer-reviewed journal Climate Dynamics by U.K. mathematician Nicholas Lewis does just that. Lewis shows that the Sherwood paper made some mathematical errors, and also relied on outdated data. Interestingly, many of the data updates were done by the IPCC itself in other parts of its 2021 report, but weren’t applied in the Sherwood study. Other data updates were done in the wider peer-reviewed literature by specialists in the field.

Lewis showed that correcting the math actually increased the sensitivity estimate slightly. But updating the data did the opposite—the ECS best estimate dropped to 2.2 C with a likely range from 1.8 to 2.7 C. And if the analysis focuses only on the period after 1870 (recognizing that most of the world has little or no reliable temperature data prior to that) the best estimate drops even more to 1.8 C. In other words, the results in the Sherwood paper with updated data would largely confirm the empirical literature the IPCC had passed over.

This is a big deal. But amid the nonstop stream of climate news you won’t hear about it because the world’s climate journalists aren’t trained to follow important topics like this, though that never stops them from lecturing everybody about what they should think about climate science. It will be five years or more before the IPCC issues a new report, and if history is any guide, months before they publish it, a group of authors heavily involved in the report-writing process will rush a paper into print that drags ECS back up to the 3 C range just long enough for the new IPCC Summary for Policymakers to declare the same old best estimate. But reality keeps pointing to lower values.

Don’t let anyone tell you “the science” demands we simply accept the increasingly lethal climate policy agenda. It would fail a cost-benefit test even if ECS were 3 degrees C, but it’s that much more unjustified with an ECS of 2 degrees C, which the level the evidence seems to insist on.