Climate Change’s Stubborn Fat Tail
Climate change uncertainty. We have it.
But that’s not a reason for inaction [pdf]. Why? Because climate change uncertainty cuts both ways. Sure, climate change might not be as bad as we predict, but it also could be worse. It’s a point definitely worth serious consideration — but that point doesn’t quite capture the whole climate-change-uncertainty conundrum. Turns out that while that uncertainty cuts both ways, one side cuts a lot deeper than the other.
First of all, when we say there is uncertainty, that doesn’t mean we don’t know anything. We know a lot. For example, while we don’t know exactly how much temperatures will rise from greenhouse gases, we do have a pretty good handle on the range of possible temperatures increases.
The Climate Sensitivity
This range of temperature outcomes is often expressed in terms of the uncertainty in the so-called climate sensitivity — which is defined as the change in global average temperature that would result from a doubling in the concentration of carbon dioxide (CO2) once the climate system comes to equilibrium.
This climate sensitivity — let’s call it T2x for short, as climate scientists are wont to do — is estimated from observations (e.g., past climate changes) and/or model simulations. Current estimates place T2x’s central (or most likely) value at about 3 degrees Celsius (or 4.5 degrees Fahrenheit). But because of uncertainties, the value could be smaller (which would be good) or larger (which would not be good).
How Much Smaller or Larger?
In the literature, you will often see estimates of the uncertainty in T2x expressed as a range between a low value and a high value. For example, the most recent assessment of the International Panel on Climate Change states that the “likely” range is from 2 to 4.5 degrees Celsius. But here’s where things get more complicated and more worrisome: That range is itself not well-defined. More extreme results (outside this range) cannot be ruled out and therefore just focusing on those two numbers can be misleading.
A more comprehensive way of expressing the uncertainty range in T2x is in terms of its so-called probability function — a plot of the probability of a specific value of T2x over the full range of possible T2x’s. The figure below provides a rough illustration of what that plot typically looks like — the larger the value of the plotted probability density for a given value of T2x, the more likely that T2x is. (Consistent with the discussion above, you will note that T2x of about 3 degrees Celsius is the most probable. And all values that have probability densities above about 0.1 lie between 2 and 4.5 degrees Celsius.)
The problem that keeps many climate scientists awake at night is that there is a non-zero probability of T2x’s that lie at the extreme edges of the distribution — in other words, not very probable but not entirely improbable. And even more worrisome is the fact that the distribution of these extreme values is not symmetric — it is skewed to the right. In the jargon of a statistician, there is a long or fat tail extending all the way out to 10 degrees Celsius.
What that means is that we cannot rule out the possibility that climate’s response to increasing greenhouse gas concentrations will be huge and catastrophic.
There’s been a fairly healthy debate in the scientific community of whether that fat tail is real — intrinsic to the way the climate works and the data we have, or just an artifact of the way we do the analysis. Into this fray come Gerard Roe and Kyle Armour of the University of Washington. In a paper published last week in the journal Geophysical Research Letters, they ask the questions: How robust is that fat tail, and how likely is it that through better observations or improved modeling the fat tail could be eliminated? As a result of their modeling analysis, their replies are: pretty robust, and not very likely.
In other words, the spectre of catastrophic climate change will continue to hang over the debate about climate change and what to do about it. The one saving grace is the fact that, as Roe and Armour point out, “very high temperature responses, if they develop, are associated with the very longest time scales” (i.e., thousands of years). That long view provides some comfort, but large climate sensitivities likely also come with larger short-term changes as well, as the system moves to a more profoundly perturbed state.
Roe and Armour’s paper is certainly not the last word on the fat tail of climate change. But it does suggest that it’s going to be around for a while. And that suggests that climate scientists will continue to lose some sleep over it. It should be keeping non-scientists up at night as well.
Correction 8/5/2011, 5:44 a.m:
Post was corrected to fix a typographical error (Tx to T2x) in the sentence explaining the graph.