Monday, May 05, 2008

More global cooling lunacy

I recently cautioned patience in the reporting of each month's climate data or every new study, as if they were proof or disproof of climate change. Sadly, we didn't have to wait long for an example of the problem. The media bungling and flagrant online abuse of a study on decadal climate prediction published in last week's Nature is a textbook case of pouncing on - and twisting - every new scrap of data related to climate change. This reporting and blogging train wreck is so textbook that I already plan to use it as a case study in my climate class next year (which limits what I can write here...)

The study by Keenlyside et al. tested whether a simplified climate model could make decadal (short-term by climate standards) forecasts. At issue is realistically simulating the many internal variations in the climate system - oscillations like the Atlantic Multi-decadal Oscillation (why people claim Atlantic hurricane activity 'naturally' varies over time) and the Pacific Decadal Oscillation - that make some decades warmer or cooler than others. The authors developed a modeling approach that exhibited some 'skill', to use the meteorologists term, in hindcasting previous ten-year periods. All in all, it is a reasonable effort; I'll spare a detailed analysis of the limits of their methodology here (but see William Connolley for a critical reading).

At the end of the manuscript, it was used to forecast temperatures over the next ten years (figure). The model forecasts a leveling off in the global temperature rise over the next decade (green / black lines), due to these internal variations in the climate system, followed by a continuation of the temperature rise.

Nowhere does it say climate change is over. This is a paper about whether models can make decadal predictions. The results confirm what we know, that internal variability is superimposed on any long-term warming trend. The global warming projections you see often look smooth or monotonic because they are averages of many individual model simulations or "realizations" of the climate (averaging them together reduces the "noise" in the data). So the paper doesn't contradict the existing consensus on a long-term rise in temperatures.

Their model forecasts the possibility that temperatures will not increase in some decades like the next decade - which is NOT the same as saying that temperatures will decrease - but will increase more in other decades.
Look at the figure: the next decade is WARMER than any of the previous decades. Nothing about the end of global warming. Yet the story that hit the press was "scientists predict global warming may stop" (Telegraph).

As with the Darfur lake fiasco last year, the authors of the press release may shoulder some of the blame. The opening "Global warming may take a short break" played up an angle sure to draw interest from the media. That does not absolve the authors of the global warming may stop silliness, both in the traditional media and that grand echo-chamber we call the blogosphere. It is not clear many of them read the press release, let alone the paper, (Pielke's flubbing the numbers seems proof of that) as the release very clearly states that the goal of the study was to look at decadal variability, and included this important quote from the authors:

"Just to make things clear: we are not stating that anthropogenic climate change won’t be as bad as previously thought”, explains Prof. Mojib Latif from IFM-GEOMAR. “What we are saying is that on top of the warming trend there is a long-periodic oscillation that will probably lead to a to a lower temperature increase than we would expect from the current trend during the next years”, adds Latif. “That is like driving from the coast to a mountainous area and crossing some hills and valleys before you reach the top”, explains Dr. Johann Jungclaus from the MPI for Meteorology. “In some years trends of both phenomena, the anthropogenic climate change and the natural decadal variation will add leading to a much stronger temperature rise."

The good news is that there are people like Andy Revkin of the NY Times (Dot Earth) out there, who tends to avoid the cheap stunt headlines and stories. Revkin's NY Times piece was like a pre-empitive strike against all the silliness to come.

[Update: For more, try this post from the prolific Joe Romm.]


John Fleck said...

Simon -

Thanks for a helpful explanation of what this paper means, rather than what people think they want it to mean.

I think this paper is a textbook example of a huge problem caused by Nature magazine, and by the misunderstanding of many in my business (the news media) regarding what the "Nature stamp of approval" means.

I know there's no way of conducting this experiment, but it's fascinating to imagine what might have happened in the news ecosystem had the paper been published in GRL instead.

Simon Donner said...

I suspect you are right. But I think the "Nature stamp of approval" affects scientists as much it affects the media. To get a paper reviewed in a top journal like Nature, it often is not enough for a paper to be important scientifically. The paper also has to have a good "headline". So scientist are pushed -- often willingly, we're not innocent here -- to play up the headline, rather than the research. In this case, the decadal prediction method made the research significant, but the specific conclusion that temperatures may level off for a few years was the headline.

Sun Tzu said...

Whenever I encounter blogs about Global Warming and Global Climate change, I inevitably must read something about Stopping Global Climate Change. That certainly may be a laudable goal, but it begs a greater question. How will we know when we've stopped Global Climate change? Whenever I ask that question, I rarely get a rational answer. As for rational answers, this recent post from John A. Warden III entitled: Thinking Strategically about Global Climate Change actually does a pretty decent job of thinking about the endstate of Global Climate versus all the tactical things people are trying to do. I found it very enlightening.

Simon Donner said...

The goal is really to slow human-induced climate change and return the climate system to an equilibrium (natural variations still going on, of course) with a minimum risk of dangerous impacts on the economy and the environment. Stabilizing atmospheric GHG concentrations at an acceptably low level (e.g 450 ppm, maybe lower) would achieve that. The climate itself would continue to warm slowly and change for some time (decades) afterwards because of lags in the climate system. Warden's post would be useful it is wasn't so confused about the facts. In the one post, he manages to bungle the causes of glaciation, the climate warming affect on crops in the subtropics and tropics, the definition of "usable" land, and rates of biodiversity loss, and fails entirely to take into account the rising human population.

Climate Chaos said...

Why does your graph start at 1960? Why not start at 1900 and show the peak period in the 1930's including the peak year 1934?

Simon Donner said...

No need for the conspiracy theory. The graph is from the published paper I discuss. The authors probably used the period since 1960 because it features more complete climate data: remember, the point of the study was to test a new model formulation.

(Regardless, the globe is warmer now than in the 1930s.)

Anonymous said...

Unfortunately no one is asking the question of why important influences on climate change such as solar activity and shifting of warm currents in the oceans are not given as much emphasis as carbon emmissions. Those two items could counteract the effects of greenhouse gases. There is no way to predict how solar and oceanic changes as well as possible volcanic activity will change things. Why are we giving one possible influence more emphasis over all the others?

Simon Donner said...

We are asking and answering those question. Solar variability, density-driven ocean circulations, even volcanic emissions are considered in climate models and discussed at length in the IPCC reports.

Anonymous said...

But everyone is conducting themselves under the assumption that catastrphic global warming is a foregone conclusion. It is entirely possible that the other, equally strong influences, could overide any greenhouse gas effect. And those influences are in no way predictable and make any assumptions just that...assumptions.

Simon Donner said...

First, "catastrophic" is not a word from the science. That's being applied by commentators. Second, the things you mention, solar variability, ocean circulation, etc., is exactly what climate scientists have been studying all these years. Your doubts are fully addressed in the published literature and in the IPCC report. For example, we have very good records of solar variability that show it is NOT an equally strong influence, that is can only be responsible for a small change in radiative forcing compared to that caused by the change (past century and future) in greenhouse gas emissions. I recommend reading the IPCC working group 1's technical summary (