Just because a paper is published in Nature doesn’t automatically mean it’s wrong. Nobody’s perfect. But if it’s a climate paper in that journal it’s a pretty safe bet that it is. The most recent example is a 2025 item that claimed that drought is accelerating globally. All the usual suspects gave it all the usual hype. Sadly, rainfall data don’t show a pattern of global decline that would indicate accelerating drought. But the authors found a way around that detail. Drought risk, they argue, is based not just on availability of rain but on atmospheric “thirst” or evaporative demand. And how do we measure Atmospheric Evaporative Demand, or AED? With a model of course. And their model said that even where rainfall isn’t declining AED is rising, so drought severity is getting worse due to global warming: AED means QED. But other experts looked at the AED model, and the study’s methodology, and poured cold water all over it. The result? Well, for once at least the authors conceded their analysis was garbage and retracted the paper. Now a new paper by an international team of scientists has appeared arguing that their AED model is fundamentally biased, and a more sensible one shows no evidence of accelerating global drought, indeed no evidence that it is even increasing. So did the usual suspects withdraw their hype or claims of worsening drought? Hoo hah.

It’s a telling example partly because rainfall is a straightforward concept. Water comes from sky. And it stands to reason that if water does not come from sky, the risk of drought goes up, where “drought” refers to availability of freshwater including soil moisture. But there is another piece of the puzzle, the potential for surface water to evaporate away, and that is where AED comes in. Because hotter weather implies higher AED, as does drier air.

For all the chatter about “simple physics”, the actual potential for evaporation is the result of some complex physics, so scientists use models to estimate it, which they can compare to measures taken from changes in streamflow, reservoir storage levels and pan evaporation gauges. The most common models use something called the Penman equation. But the new paper explains that this equation has, of all things, been found to overestimate evaporative demand. An alternative method, the so-called energy-based AED models, have proven to be more accurate.

When the authors of the new paper use Penman-based models they too find evidence not only of increasing drought severity but accelerating changes in severity. But when they correct the AED model the results change. Now they find no significant trend and no acceleration. And this finding holds up for shorter recent data sets or long data sets back to 1900.

So kudos to the science community for correcting a flawed study and publishing a newer and better method even though it didn’t support the alarmist narrative. It would be nice if it became a habit. And while we’d prefer if it didn’t seem like the mistakes always go in the direction of alarmism, which creates the impression of bias rather than simple error, reality being what it is, we’ll reserve the right to continue being suspicious of the validity of alarmist science papers, especially when they appear in Nature, unless and until it does become a habit. Or better yet, they developed a habit of not publishing flawed studies that confirm their preconceptions without doing due diligence.