We’ve reviewed many studies that estimate the Earth’s Equilibrium Climate Sensitivity or ECS by observing actual conditions outside rather than staring at climate models in the lab. The data so far in our series have always been surface temperature observations, from urban airports, sea surface data gathered intermittently by ships, etc. So clearly there are lots of problems with that record (although even still, as we’ve seen, it doesn’t indicate nearly as much warming as models keep predicting). By contrast, this week’s study uses data from weather satellites that measure temperatures of the lower troposphere, i.e. the air space from about 1 km up to about 12 km up, and provide a complete sample for the whole world uncontaminated by urbanization and other such problems. The record is shorter since it only begins in 1979, and you can’t compare it directly to earlier thermometer readings taken very close to the ground. But as authors John Christy and Richard McNider point out, the models insist that the troposphere will respond more quickly than the surface to greenhouse warming, so 40 years is enough to draw conclusions. Namely that the warming trend due to greenhouse gases works out to only about 1° C per century, which implies short-term climate sensitivity to CO2 about half that of the average climate model, and a long-term figure under 2°C, again about half of the average climate model.

Climate is complicated, massive and slow-moving, which is why there’s a big gap between the response over a century and the response once everything has settled down, to the extent that it ever does. And Christy and McNider were estimating something called the Transient Climate Sensitivity or TCS, the short-run warming associated with doubling the amount of CO2 in the air, without waiting for all parts of the climate systems to respond which takes literally centuries.

They noted that there were some strong volcanoes in the early part of the satellite record, and strong El Niños in the latter part, and as the former tend to cool the planet temporarily, and the latter to warm it, this juxtaposition biases the temperature trend upward. But after removing those effects, they assumed whatever warming was left over was due to greenhouse gases, so they weren’t looking for every excuse to push them aside.

The result of their approach was a warming of about 0.09 degrees C per decade, the same as they had estimated in a similar study in 1994 when they had less than half as much data. So then they combined that result with the observed increase in atmospheric greenhouse gases and computed how much warming would happen right away if CO2 levels doubled, which gave them their TCS figure.

It came out at 1.1C, whereas climate models on average have a transient climate response of 2.3C. So the models exhibit too much warming in response to greenhouse gases--surprise surprise. The Equilibrium Climate Sensitivity is a multiple of the Transient Climate Sensitivity, so the satellite record would imply the observed ECS is under half that in the models too. Since the models have, on average, an ECS of about 3 degrees C, that puts the observed level at around 1.5C.

Which is one more data point telling us that the real climate system just isn’t as sensitive to greenhouse gases as the models insist it should be, and why alarmists get so annoyed at the actual data that they sometimes rough it up.