One retracted study doesn’t cancel climate science

It’s hard enough for most of us to predict what we’re having for dinner tonight, much less how the the world will look in 2100. By that time we might have cured cancer and started building giant space brains. Or maybe we’ll be nothing but batteries for our robot overlords. Or spraying crops with a sports drink, according to the Tribune News Service.

And yet without such predictions, we can’t envision alternative futures, which motivate us to make better ones possible. Consider climate change. People’s eyes glaze over when you tell them, “Climate change is going to be really bad.”

They need to know how bad, exactly. How specifically might it affect me, the protagonist of reality? So scientists and researchers keep delivering these forecasts to them, no matter how many reputational minefields they must traverse on the way.

The risks of prognostication were made grimly clear last week when the journal Nature retracted a 2024 study trying to estimate the damage global heating will do to economic growth in the decades ahead. The study, by researchers at the Potsdam Institute for Climate Impact Research, was striking because its long-term damage estimate was massive, at 60% of global GDP by 2100, or about three times bigger than previous forecasts.

Such an outlier naturally raised eyebrows, and within months Nature put a note on the study warning readers that “the reliability of data and methodology presented in this manuscript is currently in question.” This summer, a couple of studies pinpointed some significant flaws, and the retraction soon followed. The good news, if you’re the Potsdam researchers — but maybe not so good if you’re a person who enjoys living in a functioning economy — is that even with corrections their findings haven’t changed much. In a revised paper that’s still being peer-reviewed, they now estimate the global heating we’ve already experienced will slash income by 17% by 2050, down from 19%. The hit by 2100 will still be in the neighborhood of 60%. These are still very grim numbers.

There may also still be reasons to question them. A tweak to how the Potsdam scientists measure inflation cuts that 2050 income-hit estimate from 17% to something closer to 6%. With that tweak, the 2100 forecast plummets from 60% to roughly 15%. And the study’s forecasts for later decades are based on Representative Concentration Pathway (RCP) 8.5, the most extreme of the future carbon-emissions scenarios considered by the United Nations Intergovernmental Panel on Climate Change. This scenario assumes there will be no further efforts to curb emissions, driving global temperatures 4 or 5 degrees Celsius above preindustrial averages by the end of the century. This seems unlikely at this point, and thank goodness.

Then again, as noted above, predictions are hard. RCP 8.5 still has scientific uses when sussing out climate-change impacts. And the Potsdam paper is one of the few studies willing to take a stab at modeling how the effects of soaring heat and compounding disasters on productivity don’t just disappear but linger for years. In that way it’s at least trying to be more realistic than other efforts that use simpler variables.

Anyway, with forecasts this distant in systems this hopelessly complex, we might as well be talking about using a laser beam to split hairs on a pig riding a meteor through the Andromeda galaxy. The critical takeaway is that all the literature on climate and economics points in the same direction, even if it disagrees on the particulars: A hotter planet is one with less economic growth. And the hotter the planet, the lower the growth. Even at RCP 4.5, the current middle-of-the-road emissions scenario, we still risk heating of 3C to 4C. That will still be devastating to human health and well-being.