With climate change, uncertainty is no-one’s friend

Uncertainty cuts both ways highlights University of Bristol's Stephan Lewandowsky - if your preferred estimate is at the low end of a range, you're neglecting similarly likely high end estimates. Image credit: University of Bristol

Uncertainty cuts both ways highlights University of Bristol’s Stephan Lewandowsky – if your preferred estimate is at the low end of a range, you’re neglecting similarly likely high end estimates. Image credit: University of Bristol

Waiting longer to act on climate change will cost us more than taking immediate action. It’s a message that’s getting louder and louder, repeated from many sides. In March it was stressed by US Secretary of State John Kerry. In April it was highlighted by the UN Intergovernmental Panel on Climate Change (IPCC)’s latest report. Last month it was underlined by Hank Paulson, treasury secretary under George Bush, hedge fund manager Tom Steyer, and former New York Mayor Michael Bloomberg. This week the Council of Economic Advisors, the agency advising President Obama on economic policy, joined in.

These messages could hardly be any clearer, but still some of us remain uncertain on the need to act. The best argument for waiting until we’re more certain to act is that if climate change turns out to be harmless, our efforts to fight it will be wasted. Even simple things like current weather are enough to sway our opinions, and when uncertain it’s always tempting to feel like we don’t need to do anything. But that’s the wrong reaction to uncertainty on climate change, according to psychologist Stephan Lewandowsky from the University of Bristol, UK.

The researchers have found that greater uncertainty over how much Earth warms in response to our CO2 emissions would actually raise forecasts of average damage costs from climate change. Greater uncertainty also means projections see it as more likely that steps to cut emissions won’t keep the world below warming levels governments have agreed we must avoid. So, if we are unsure whether we can slow the climate juggernaut down before we smash into the wall, we’re better off hitting the brakes earlier. As Stephan explained, ‘uncertainty is no one’s friend’.

Evidence is piling up against the economic argument for waiting to see if climate sensitivity is less than 1C per doubling of CO2. Image copyright Springer, see reference below.

Evidence is piling up against the economic argument for waiting to see if climate sensitivity is less than 1°C per doubling of CO2. Image copyright Springer, see reference below.

Certain dissenting individuals raise risk forecasts

Warming caused by greenhouse gas emissions is often thought of in terms of climate sensitivity, the temperature increase from doubling CO2’s concentration in the air. There are many pieces of evidence for what we might expect that to be, including natural temperature records from the distant past, modern direct measurements and climate models. They typically come close to 3°C. According to the latest IPCC report: “Equilibrium climate sensitivity is likely in the range 1.5°C to 4.5°C (high confidence), extremely unlikely less than 1°C (high confidence), and very unlikely greater than 6°C (medium confidence).” The IPCC uses this range to build projections that governments can then use in forming their policies.

The fact that very high sensitivities are a bit more likely than very low ones becomes important if we become less certain. The estimates spread out and, if the average climare sensitivity is kept at 3°C, lower values actually make up a greater proportion of all sensitivity estimates. However, another key point is that economists agree that whatever the rate of warming, damage costs from climate change will grow even faster. For example, the cost of damage from the temperature rise from pre-industrial temperatures to 1°C above that level will be less than the cost of damage from the rise from 1°C to 2°C above that level. This amplifies the consequences of the raised chances of extreme warming, increasing average projected damage costs compared to current climate sensitivity estimates.

Projected chances of 5°C warming (red line) in response to distributions of estimates centred on 3°C warming,  but which Stephan's team progressively increased the uncertainty level on. The size of each bar shows the number, or frequency, of estimates of that value for climate sensitivity, which is shown in °C per doubling of CO2. The top left panel shows the least uncertainty, and least chance of exceeding 5°C, the bottom right shows most uncertainty and most chance of exceeding 5°C. This shows that far from delaying the need to act, building greater uncertainty into projections like the IPCC's would make action more urgent. Image credit: www.shapingtomorrowsworld.org

Projected chances of 5°C warming (red line) in response to distributions of estimates centred on 3°C warming, but which Stephan’s team progressively increased the uncertainty level on. The size of each bar shows the number, or frequency, of estimates of that value for climate sensitivity, which is shown in °C per doubling of CO2. The top left panel shows the least uncertainty, and least chance of exceeding 5°C, the bottom right shows most uncertainty and most chance of exceeding 5°C. This shows that far from delaying the need to act, building greater uncertainty into projections like the IPCC’s would make action more urgent. Image credit: http://www.shapingtomorrowsworld.org

Stephan and his team illustrate that this translates to real-life estimates, taking a study of experts’ subjective judgements from 1995. One of the 16 experts gave a very low but highly certain estimate for climate sensitivity, but in differing from the others the expert raises the overall uncertainty. That in turn increases the climate change risk levels in projections using the whole group’s estimates. “We show that removal of this outlier, which reduces ambiguity among experts, tends to reduce the likelihood of extreme risks of exceeding a global temperature threshold,” Stephan explained. “Conversely, we show that if this outlying expert is replicated and replaces some other observations at random, the likelihood of exceeding a tolerable risk of excessive warming increases in most cases.”

Closing in on the truth

In another study, the researchers show a similar impact of uncertainty on CO2 emission limits. To keep warming below the 2°C level governments have agreed to, CO2 levels in the air must be limited. One estimate of the limit is somewhere between 415-460 parts of CO2 per million (ppm) parts of other gas. In June 2014 the level was 401 ppm. Because CO2 stays in the air for centuries, we have a limited emission budget. The total budget is around 3.7 trillion tonnes, and although that sounds a lot, we’d used up about half by 2008. And the longer we carry on ‘business as usual’, the steeper the cuts we’ll need to make to squeeze under the 2°C threshold.

However, uncertainties about climate sensitivity turn into uncertainty about the emission budget, giving both more low and high limits. Again, the net effect is that this raises the average budget across all estimates, although “what is relevant to mitigation failure is the minimum, not the mean”, Stephan underlined. With more low budgets, greater uncertainty raises the number of estimates that suggest we’ve already passed the point where we’re locked in to 2°C warming. Using a figure close to the true uncertainty in climate sensitivity, 8% of the team’s carbon budget estimates gave this outcome. “It is unlikely that the public—or indeed governments—would tolerate nuclear power if it came with an 8% annual risk of a meltdown somewhere in the world,” they write.

How increased climate sensitivity uncertainty (top graphs) affects carbon budget uncertainty (bottom). σ is the level of uncertainty, which is lower in the left hand graphs (a) than the right hand graphs (b). Higher uncertainty raises the average carbon budget, but increases the number (frequency) of very low carbon budgets. Carbon budgets are in gigatonnes (Gt), or billions of tonnes, of carbon (C) emitted. 1 Gt C is equivalent to 3.7 Gt CO2. We've already emitted around 1000 Gt C. Image credit Springer, see references below.

How increased climate sensitivity uncertainty (top graphs) affects carbon budget uncertainty (bottom). σ is the level of uncertainty, which is lower in the left hand graphs (a) than the right hand graphs (b). Higher uncertainty raises the average carbon budget, but increases the number (frequency) of very low carbon budgets. Carbon budgets are in gigatonnes (Gt), or billions of tonnes, of carbon (C) emitted. 1 Gt C is equivalent to 3.7 Gt CO2. We’ve already emitted around 1000 Gt C. Image credit Springer, see references below.

That comment raises a philosophical question: Does 8% of a set of expert estimates actually convert into an 8% risk in the real world? Perhaps not, but it’s all we have, Stephan stressed. “The universe will unfold along its actual path, irrespective of the musings about risk by scientists,” he admitted. “However, given that we can never know this eternal truth, the best we can do is to provide an envelope of possibilities that we presume to enclose this truth ahead of time. And irrespective of what actually happens we must allow for a greater variety of possible outcomes if uncertainty is greater.”

  • You can read Stephan’s own explanation of his work at these links: part I, part II, part III. Stephan’s coauthors Ben Newell and Mike Smithson also share their explanation here.

Journal references:
Lewandowsky, S., Risbey, J., Smithson, M., Newell, B., & Hunter, J. (2014). Scientific uncertainty and climate change: Part I. Uncertainty and unabated emissions Climatic Change, 124 (1-2), 21-37 DOI: 10.1007/s10584-014-1082-7
Lewandowsky, S., Risbey, J., Smithson, M., & Newell, B. (2014). Scientific uncertainty and climate change: Part II. Uncertainty and mitigation Climatic Change, 124 (1-2), 39-52 DOI: 10.1007/s10584-014-1083-6

Advertisements

2 Responses to “With climate change, uncertainty is no-one’s friend”

  1. With climate change, uncertainty is no-one’s friend | Simple Climate Says:

    […] With climate change, uncertainty is no-one’s friend […]

  2. Richard Mallett Says:

    Global temperatures have been rising :-

    a) since 1880 at +0.65 C per century, according to NCDC and GISS.
    b) since 1850 at +0.47 C per century, according to HadCRUT4.
    c) since 1753 at +0.36 C per century, according to Berkeley Earth.
    d) since 1702 at +0.26 C per century, according to the 28 longest / most complete station records in CRUTem4.

    The correlation R^2 between CO2 and the NCDC / GISS / HadCRUT4 average from 1880-2004 is 0.789 (78.9% of the variation is related). During that time, CO2 levels have increased from 289.8 ppm to 374.6 ppm, an increase of 29.3%.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: