On cue, to prove our point that climate orthodoxy has abandoned data for dogma including asserting a level of wildfires in 2026 quite incompatible with the fact that they’re down on every continent, Scientific American (in one of those newsletters inexplicably not posted online in the year of grace 2026) blares “Europe is on fire.” Not once, but twice, including in its August 20 “Today in What Used to be Science” email newsletter and then in the actual story. And the newsletter goes on to report a thing that hasn’t happened based on processes that can’t: “But the area of Europe affected by fires will essentially triple by 2100, assuming humans keep emitting high amounts of carbon, according to a new calculation.” As with RCP8.5, the fatal flaw, or one of them anyway, is that if Europe is on triple fire by 2100, we won’t keep emitting high amounts of carbon because we’ll all be charred skeletons or else hiding under the simmering waters of the Med and North Sea. Unless of course it’s all ridiculous pseudo-science.

In the related story, which also starts “Europe is on fire” although Europe is not on fire so already they’re jeopardizing their credibility, they talk about a study that got our spidey senses tingling about what scenario or scenarios, exactly, underlay this new “calculation”, a word here meaning not calculation in the olde-tyme sense of calculating whether sacrificing the bishop on h2 will lead to checkmate in four moves or what you get if you multiply six by nine. Here it meant what a computer model will tell you if you tell it to tell you it.

As they admit, the models aren’t worth much of anything:

“The mechanics of simulating wildfire behaviour in fire-enabled DGVMs vary widely (Hantson et al. 2020; Forrest et al. 2024). In particular, the influence of human activities is approached in different ways and has evolved substantially over time (Le Page et al. 2015; Hantson et al. 2016; Teckentrup et al. 2019).”

Evolved, sure. They’re bigger, more complicated and slicker. But since they still don’t model reality or even try to, relying instead of all kinds of estimates rather than measurements of various relationships between various variables, what you get out depends on what you put in. Literally in fact:

“Burned area=𝑐0+∝1·FAPAR12+∝2·Slope+∝3·TPI +∝4·Popdens+∝5·HDI+∝6·Treecover +∝7·Treecover2+∝8·MEPI+∝9·ln(FWI) +∝10·MEPI·ln(FWI)”.

So the science is settled. Unless it isn’t. (Trivially, surely burned area feeds back on Popdens.) But anyway, you tweak the inputs until you get the outputs you want and call it future history.

In this case they’re messing around with SSPs, specifically SSP1-2.6 and SSP3-7.0 which, they say, are associated respectively with “the low-emission climate scenario RCP2.6” and, yes, almost 8.5, “the higher-emission climate pathway RCP7.0”. And it’s all a load of rubbish anyway because it’s based on guesswork that isn’t made better by assigning a variable to it, guesswork about everything from land use to population growth, government policy and energy use. If they could predict that stuff, they’d get rich in the stock market. They can’t.

Indeed, we defy them to use their vaunted models to plug in 1900 and get out 2000. And if they can’t we say “Pfui” to claims they can plug in 2000 and get out 2100. Or this claim:

“For each of the five climate models, we performed a spin-up of the natural vegetation by starting the simulation from a bare, non-vegetated state. The simulations were run for 2500 years, looping over input forcing data for the years from 1960 to 1989. Thereafter, we conducted a 390-year land-use spin-up to account for land-use change legacy. After those two spin-up phases, we run the simulations from 1960 to 2014 (historical period of each climate model), which was the starting point for the future simulations to diverge into the two SSPs (SSP1-2.6 and SSP3-7.0) scenarios.”

What bare, non-vegetated state ever existed in Europe? Except when it was under ice, we mean. And who dares run simulations for 2500 years and claim vindication unless they match what actually happened? Babbling about 390 years of “land-use change legacy” doesn’t make it better. Especially as they’re working with “9 x 9 km grids” not individual fields; is this Minecraft?

They got what they were looking for:

“By the end of the century, intensified fire weather alone could increase annual burned area by approximately +39% under a low emission scenario (SSP1-2.6) and by nearly +192% under a high emission scenario (SSP3-7.0). Continued improvements in fire management capacity could substantially moderate these increases, reducing burned area by 72%–92% compared to scenarios without these improvements. However, under strong climate change, fire activity still rises in about 55% of Europe’s fire prone regions, even if investments in fire management capacity continue at current levels.”

But of course you only get such a result if you build in the assumption that fires are getting worse and doing so because climate change whatever. And fires are not getting worse; they’re a lot less severe than they were in cooler pre-industrial times, at least where we think we know enough to make a reasonable estimate.

Still, all climate change all the time and as a causal force not a statistical description. The Scientific American story gives the final word to a scientist who says, who says:

“Fuels management can mitigate some of the worst effects near homes and critical resources. And proactive fire management is crucial for restoring fire-adapted systems and supporting a range of social and ecological objectives. But it will not fully solve the looming fire-weather dangers caused by climate change.”

We know because a computer said so when told to.