Getting climate forecasts in line
The climate forecasting business is always couched in uncertainty. One model shows this, another shows that. One federal scientist says one thing, another from XYZ subagency PDQ (Federal Building J, subbasement G-5c) says something else. Blah blah blah. The only thing they all seem to agree on is that, if you think the climate’s bad now, just wait 50 or 100 years. Yet they rarely get specific.
In fact, not all climatologists feel that way, as readers of Environment & Climate News well know. In their wittily titled 2000 book, The Satanic Gases, contributing editors Patrick Michaels and Robert C. Balling Jr. did something rare for climatologists: They made a few predictions about future climate. Here’s an excerpt:
Prediction 1: “[Within the next 50 years] scientists will confirm that although the functional form of the climate models is correct, the amount of warming is already dictated by nature. . . . In surface temperatures for the last third of this century [remember, the year 2000 is technically still the twentieth century], a trend has been established, and it is near or below the low limit of the model calculations, but it is a straight line, as the models predicted. There is no reason to suspect this is going to suddenly stop.”
Prediction 2: “The earth’s average surface temperature will warm 0.65C to 0.75C by 2050.”
Now that is controversial stuff. Michaels and Balling are saying the planet’s warming is linear, and therefore predictable. This is no doubt upsetting to the climate modeling community (and their funding agencies) since—despite the alleged sophistication of the model’s various cloud and surface parameterization schemes, the integration of the biosphere, and the development of dynamic, multilayered circulating oceans—those models’ prediction of future climate is the same one you or I could make using a straight edge to connect the dots (cost: zero).
In Prediction 2, given that we now know the slope, the temperature forecast for the future is easy. We’ve experienced about 1.0°C of global warming since pre-industrial times. So by the year 2050, the authors us, the total warming will be a nonastounding 1.6°C to 1.8°C.
Climate modelers claim there are far too many uncertainties to make so bold a forecast. It’s difficult to know the real climate sensitivity (how much warming you get, say, per slug of CO2), how much heat the oceans are storing, the actual cooling impact of sulphate aerosols, and so forth. Reality is certainly much more complicated.
Enter a new paper in Nature by England’s Myles Allen, cowritten by modelers from Hadley Center (the UK Meteorological Office), the Max-Planck Institut in Germany, and the Geophysical Fluid Dynamics Lab at Princeton. “Quantifying the Uncertainty in Forecasts of Anthropogenic Climate Change” is what they call it, and well, they’ve done just that, and darn if the numbers don’t match Michaels’ forecast!
According to Allen, “We expect global mean temperatures in the decade 2036-46 to be 1-2.5[°C] warmer than in pre-industrial times. . . . This range is relatively robust to errors in the models’ climate sensitivity, rate of oceanic heat uptake or global response to sulphate aerosols as long as these errors are persistent over time.”
Furthermore, Allen and his colleagues are 90 percent confident of this prediction (in other words, the real warming would fall within their projected range nine out of 10 times if the “experiment” of adding greenhouse gases at the projected rate were repeated). Michaels and Balling’s more refined numbers are dead smack in the middle of this range.
That kind of confident estimate is not possible using current climate model forecasts. Rather, Allen and colleagues arrived at it using simple statistics, comparing observed temperature patterns for each decade from 1946 to 1996 with the decadal projections of various climate models.
Now, we all know the model projections of current climate are not correct. So, for each model and for each decade, Allen and colleagues calculated a “scaling factor” that best brings the projections in line with the data, using basic multiple regression analysis. And when that scaling factor is used to correct the model and then project the future climate change, it turns out the relationship is linear.
In Nature’s “News and Views” section (the layperson’s introduction to each article), Andrew Weaver and Francis Zwiers note that:
The beauty of this result is that it is independent of the “climate sensitivity” of the model and the rate at which that warming happens. In particular, it is independent of the rate of oceanic heat uptake, which varies between models with different representations of ocean physics. More complicated models have indeed exhibited a linear response to changes in radiative forcing caused by increases in atmospheric greenhouse gases and sulphate aerosols. For example, when separate model runs with twentieth-century levels of greenhouse gases and aerosols are combined, the result is similar to runs which included them together.
Of course, a lot of other factors could modify the model’s forecasts: Volcanic eruptions, variability in the sun’s output, and ozone depletion are all legitimate components of climate. But Allen’s results suggest it doesn’t really matter if some or all of these components are missing from the models, because any feedbacks they might have on temperature have already been factored into the projection using their approach.
It’s also clear that the various model projections of future temperatures are too high. So, if it’s true we are faced with a future in which we can actually predict the magnitude of warming, and it turns out to be a small number, can we stop fussing about our climatic future? Abandon trying to implement a Kyoto Protocol that would have the net effect of reducing global temperatures by 2050 by only 0.07°C while simultaneously wreaking havoc on the global economy?
The demise of the Kyoto Protocol is one forecast that would serve us all quite well.
Michaels, P.J., and Robert C. Balling Jr., 2000. The Satanic Gases, Cato Institute Press.
Allen, M.R. et al., 2000. Quantifying the uncertainty in forecasts of anthropogenic climate change, Nature, 407,617-620.
Weaver, A.J. and Zwiers, F.W., 2000. Uncertainty in climate change, Nature, 407, 571-572.