One paradox of climate change science is that ability to dazzle, or intimidate, matters more than ability to predict. Even that tired “you’re not a climate scientist” is invoked if, for instance, some scientist predicts the end of snow and you point out that it’s still snowing. And now we get the vexed problem of ocean heat in which experts can wave equations about and program computers until their chips ache. But to the limited extent we can measure sea surface temperature, they can’t explain why it’s still rising. And if you’re thinking well duh, global warming, you dense non-scientist, the puzzle Paul Homewood zeroes in on here is that it’s been rising since 1900 and for the most part at a very steady rate. Leaving atmospheric CO2, well, out of the picture. And the equation.

Homewood’s starting point is the superficially scary headlines about how 2020 tied 2016 for the hottest year ever. Which he notes can also be expressed rather less scarily as “there has been no global warming in the last 5 years” and which, in consequence of this possible implication, has led to a lot of tsk-tsking about the fact that obviously all the heat went into the oceans which are starting to give off steam or something.

Homewood quotes one tweet critical of the GWPF’s own observation about no warming over five years that includes a very steep rising graph of ocean heat since 1958. But then he asks: Why start in 1958? As he notes, one possible reason is that the data is pretty shaky before then. But unfortunately it’s also pretty shaky afterward, at least until the deployment of “ARGO” buoys in 2004. And as he also observes, it’s pretty shaky in another sense to do a high-rise chart of “Zettajoules” soaring upward when if you instead graphed some variable dimly related to warming such as, say temperature, you’d find the ARGO system showing a profoundly unscary warming of one one-hundredth of a degree since 2004.

By contrast, the Hadley Centre’s HADSST3 series shows a rise since 1958 of as much as half a degree. Despite cooling from 1940 to 1980. But also after rising from 1900 on. Which brings us back to the central issue. Any suitably trained non-fool can write equations that “predict” the warming after 1980 and pin it on CO2. But they cannot write equations that also predict the cooling from 1940 to 1980 and pin it on CO2 or, crucially, predict the warming from 1900 to 1940 and pin it on CO2.

It doesn’t matter how clever or intimidating the equations are. What matters is how well they explain what we can actually see. And so Homewood’s chart (generated from the HADSST3 series by WoodForTrees.org) is a picture worth a thousand equations, because it tells you what the equations must explain to be worth anything.