Uncertainty makes tackling climate change more urgent

Scientific uncertainty has been described as a ‘monster’ that defies our best efforts to understand the Earth’s climate system. Commentators and politicians routinely cite uncertainty about the severity of climate change impacts to justify their opposition to mitigation measures such as a price on carbon.

What is the appropriate response to uncertainty about the future of the Earth’s climate? Is there too much uncertainty to warrant action? Should we wait for more certainty?

On the face of it, complacency in the light of uncertainty might appear tolerable or even advisable.

However, a mathematical analysis of the implications of uncertainty about future temperature increases shows otherwise.

My colleagues and I recently published two articles in the journal Climatic Change – “Scientific Uncertainty and Climate Change"Part I and Part II – which strongly suggest that greater uncertainty about future climate change increases the likelihood that the real-world effects will be even more severe than anticipated.

Uncertainty is not harmless

In a nutshell, greater uncertainty should make us worry more rather than less. Greater uncertainty should compel us to invest greater rather than lesser effort into mitigation.

Far from being a ‘friend’, scientific uncertainty is an angry beast that may come to haunt us.


For starters, we need to realise that uncertainty cuts both ways.

If a doubling of atmospheric CO2 levels from pre-industrial times is expected to raise global temperatures by between 1.5ºC and 4.5ºC, with a best estimate of around 3ºC, then it is a logical fallacy to consider only the lower estimate. Just because the ultimate temperature increase is uncertain doesn’t mean that it’ll be 1.5ºC – it could equally be 4.5ºC.

This elementary mathematical fact is often forgotten or overlooked by commentators who equate ‘uncertain’ with ‘harmless’.

Uncertainty does cut both ways, however much we may wish to ignore or downplay the discomforting upper estimates.

What’s more, the implications of uncertainty are being amplified in such a manner that greater uncertainty maps into greater expected damages, all other factors being equal.

This mathematical fact is difficult to explain but easy to illustrate. Suppose that our best estimate of a future temperature increase is constant at 3ºC, then the damages that result from this will be greater, on average, if the uncertainty range is wide – for example, 1ºC to 5ºC – than if it is narrow, for instance 1.5ºC to 4.5ºC.

The reason for this is that all economic models of climate change agree that the damage costs from climate change increase at an accelerating rate as temperature increases. By a 100-year-old mathematical theorem known as Jensen’s inequality, it follows inevitably that greater uncertainty is associated with greater expected cost.

In a nutshell, if we expect X°C warming, the expected damage cost associated with that degree of warming is a function not only of X but also of the uncertainty associated with our estimate of X – and the greater that uncertainty, the greater the expected damage.

Greater loss

Not only is uncertainty not your friend, but greater uncertainty translates into greater expected loss.

This result is perhaps counter-intuitive – see here for a more detailed explanation – but it is mathematically virtually inescapable.

The same implications of uncertainty arise in connection with sea-level rise. Sea levels have been rising inexorably over the last few decades as the climate has been warming, and rising sea levels undoubtedly exacerbated the impact of Hurricane Sandy in 2012.

To illustrate, if we could predict with absolute certainty that sea levels will rise by exactly 50 centimetres – no more and no less – by 2100, then all we need to do is to build a 50 centimetre tall levee that can cope with this rise.

Add any uncertainty to the estimate and a 50 centimetre levee will no longer be sufficient – in fact, even a relatively small amount of uncertainty in estimates of sea level rise means building a levee almost twice as high as the expected increase in sea level.

This result again follows from some fairly basic mathematics and is therefore difficult to circumvent. Greater uncertainty means greater effort is required to deal with sea level rise.

In light of those results, there is a further aspect of uncertainty that should give us particular cause for concern. Whatever uncertainty there is about future temperature increases, it is ‘leveraged’ by the total amount of greenhouse gases we emit.

That is, the adverse effects of uncertainty we just discovered increase with increasing emissions: not only is uncertainty not your friend, but it is an angry beast that we poke harder and harder as we continue to emit greenhouse gases.

For the past 8,000 years, civilisations have sought to reduce uncertainty about the environment. From the invention of agriculture to the construction of dams and bridges, humans have endeavoured to reduce uncertainty and thus the risks that arise from the environment.

Continuing with unabated carbon emissions reverses those efforts of the past millennia by introducing greater uncertainty and greater risk into our future.

* Stephan Lewandowsky is a professor of cognitive psychology at the University of Bristol in the UK. He was awarded a Discovery Outstanding Researcher Award by the Australian Research Council in 2011. He also received a Wolfson Research Merit Award from the Royal Society in 2013. On Twitter he is @STWorg. His research examines the distinction between scepticism and denial and the role of uncertainty in people’s thinking about climate change.