It’s official. 2020 is among the hottest years ever. And no silly complaining that it hasn’t technically happened yet. Gavin Schmidt has spoken, and when Gavin Schmidt speaks, computers speak. See, “With today’s update to GISTEMP, 2020 is now virtually certain to be a top 3 year (in timeseries back to 1880), the 6th year in a row more than 1ºC above the late 19th C, and a good chance of breaking the record (~70%).” So never mind Toronto breaking an 81-year-old record for cold the day before. Forget April 2020 being colder than April 1998 on land globally. Never mind waiting to see what we can actually measure, on the ground or from space. That’s just real-world stuff. We’re talking models here.
It will not be easy for people like Schmidt to admit that computer simulations of climate did not work nearly as well as expected in the heady early days. It’s what he does and he has long been very proud of it. For instance in an enthusiastic, even evangelical TED talk in 2014 he showed us how well his models had captured 20th-century temperature. Or rather, how well he thought they did.
It’s intriguing to watch him dismiss their failure to match reality throughout the first half of the century as mere “noise”, then praise their “predicting” rising temperatures in the face of rising CO2 from about 1970s on. But there’s nothing very magical about tweaking a complicated set of equations to match a limited set of known inputs to a limited set of known outputs.
Especially when what should always have bothered him, and many others, is that these same “tweaked” models were rubbish at predicting other outputs. And not just the unknown future. They could not cope with the first half of the 20th century without being retweaked in ways that made them unable to cope with the second half. They couldn’t do the 19th century. Or the 14th. Or any other century. (Unless of course you did something funny to the data so temperature flatlined and you could then match it to flatlined CO2 if not to, say, real-world temperature readings.) They couldn’t even cope with the last few decades in the part of the atmosphere where the models were supposed to have the clearest view on greenhouse warming.
In the end, the argument will not be settled by Gavin Schmidt’s computer telling us 2020 was hotter than an oven before it even happens. Nor by computer simulations of what temperature ought to have been. It will be settled by what temperature is. And in that spirit, and without exaggerating the precision of measurements of temperature around the world even if the approximations seem to support our own point of view (as Schmidt did in 2014, saying flatly “We know what happened over the 20th century, right? We know that it’s got warmer, we know where it’s got warmer”) we want to wait and see what it was like before crying “Gotcha” or indeed anything else.