Communicating climate change uncertainty

The U.S. Climate Change Science Program finally managed to release its Final Report of Synthesis and Assessment a couple of weeks back. There's not a lot of new material in the first four of five volumes which deal with the state of the science, mostly because the report was supposed to be released a while back. Must have been some delay at the top of the food change. But the fifth volume came with an intriguing title: "Best practice approaches for characterizing, communicating, and incorporating scientific uncertainty in decision-making"

The 156-page report tries to give scientists a few pointers on translating their work into something the public can swallow, and in doing so provides an excellent backgrounder for anyone not already intimately familiar with the subject. Here's a typically cogent excerpt, describing the "three key facts that it is important to understand to be an informed participant in policy discussions about climate change:"

• When coal, oil and natural gas (i.e., fossil fuels) are burned or land is cleared or burned, carbon dioxide (CO2) is created and released into the atmosphere. There is no uncertainty about this.

• Because CO2 (and other greenhouse gases) trap heat, if more is added to the atmosphere, warming will result that can lead to climate change. Many of the details about how much warming, how fast, and similar issues are uncertain.

• CO2 (and other greenhouse gases) are not like conventional air pollution such as SO2, NOx or fine particles. Much of the CO2 that enters the atmosphere remains there for more than 100 years. In order to reduce concentration (which is what causes climate change), emissions must be dramatically reduced. There is no uncertainty about this basic fact, although there is uncertainty about how fast and by how much emissions must be reduced to achieve a specific stable concentration. Most experts would suggest that a reduction of CO2 emissions of between 70 and 90% from today's levels is needed. This implies the need for dramatic changes in energy and other industrial systems all around the globe.

A fair portion of the report is devoting to the various kinds of uncertainty involved in climatology. Some of it's a bit technical, but the authors have a little bit of fun nevertheless:

While Donald Rumsfeld (2002) was widely lampooned in the popular press, he was absolutely correct when he noted that "...there are known unknowns. That is to say, we know there are some things we do not know. But there are also unknown unknowns, the ones we don't know we don't know." But perhaps the ever folksy but profound Mark Twain put it best when he noted, "It ain't what you don't know that gets you in trouble. It's what you know for sure that just ain't so."

There's also what amounts to an implicit criticism of the latest IPCC assessment's use of qualitative language to describe the likelihood of climate change scenarios.

Think about betting odds. Suppose that to one person "unlikely" means that they think there is only 1 chance in 10 that something will happen, while to another person the same word means they think there is only one chance in a thousand that that same thing will happen. In some cases, this difference could be very important. For example, in the second case, you might be willing to make a big investment in a company if your financial advisor tells you they are "unlikely" to go bankrupt - that is, the odds are only 1 in 1000 that it will happen. On the other hand, if by unlikely the advisor actually means a chance of 1 in 10, you might not want to put your money at risk.

The same problem can arise in scientific communication. For example, some years ago members of the EPA Science Advisory Board were asked to attach odds to the statement that a chemical was "likely" to cause cancer in humans or "not likely" to cause cancer in humans. Fourteen experts answered these questions. The odds for the word "likely" ranged from less than 1 in 10 down to about 1 in 1000! The range was even wider for the odds given on the word "not likely." There was even an overlap...where a few experts used the word "likely" to describe the same odds that other experts described as "not likely."

Because of results like this, it is important to insist that when scientists and analysts talk about uncertainty in climate science and its impacts, they tell us in quantitative terms what they mean by the uncertainty words they use.

The authors eventually get around to suggesting climatologists use graphical representations of probability. This, I expect, will receive a variety of responses. Here's a proposed visualization technique:

Hmmm. Still, if you have the time, give the thing a read.