Climate change uncertainty is no reason for inaction since we can't rule out risk

We don't have to believe that our house will burn down to take out insurance. So why delay taking action to reduce emissions?

Tim Palmer guardian.co.uk, Monday 22 March 2010 07.30 GMT



A man walks in the rain along the Albert Embankment in London. Photograph: Shaun Curry/AFP/Getty Images

Climate change is sometimes presented in simple black and white terms. You either believe it or you don't. Perhaps after the recent controversies over email leaks and melting Himalayan glaciers, some may have decided to change camp.

But this is a false dichotomy. Indeed the notion of "belief" plays no role at all in science, whether about climate change or anything else. The Royal Society, the UK's national academy of science, was founded 350 years ago on this very basis, with the motto Nullius in verba, or "take nobody's word for it". The founders took nothing for granted and chose to investigate observations and search for the conclusions that best fit them. The notion that these conclusions can never be considered certain and immutable, underpinned both their actions and those that came after them. As James Gleick wrote about the great 20th century theoretical physicist Richard Feynman; he believed in the primacy of doubt, not as a blemish on our ability to know, but as the essence of knowing.

Modern day weather prediction is inherently uncertain. Every day, weather forecast centres will generate an ensemble, typically of 50 individual weather predictions, in order to assess uncertainty in the weather up to a week or more ahead. The individual predictions have very slightly different starting conditions, reflecting the fact that the weather observations which generate a forecast's initial state are neither complete nor wholly accurate. When the atmosphere is in a predictable state, all 50 predictions are more or less identical for the coming week and the forecaster can say with great confidence what the weather will be like. On the other hand, when the atmosphere is in a chaotic state, the best the forecaster can talk about are probabilities of different outcomes.

Some, perhaps those without scientific training, may see probabilistic predictions as an evasion of responsibility. However, in reality, probabilistic predictions embody the scientific method. In any case, what is better for decision making, a forecast with some realistic measure of uncertainty, or some grossly overconfident prediction with no hint of uncertainty? Worldwide, probabilistic weather forecasts are now used by those making decisions to evacuate people exposed to river flooding, or to intense storms. For better or for worse, they are also central to who trade on future energy prices and other weather-sensitive commodities.

Similar "ensemble" methods are used to predict climate change, except here it is also critical to vary uncertain parameters in the climate models as well as uncertain estimates of how greenhouse gas emissions will change over the course of the century. As with the weather forecasts, contemporary climate predictions are essentially probabilistic. Hence for example, based on the ensemble of the world's climate models, we can estimate that over the Asian monsoon region, a season that was so wet it would only have a 1 in 20 chance of occurring in the 20th century, could have a 1 in 3 chance of occurring by the end of the 21st century.

We don't have to believe that our house will burn down in the coming year to take out insurance. Similarly we don't have to believe that dangerous climate change will occur to take action to cut emissions. A key question that everyone concerned by the climate change issue should ask, particularly those who are sceptical, is this. How large does the probability of serious climate change have to be before we should start cutting emissions? To be specific, how large does the probability have to be that by the end of this century, large parts of Bangladesh will under water because of sea level rise and a substantially more intense monsoon system? Or that the Amazonian rainforest will die because of shifts in rainfall patterns over South America? Or that the type of drought that plagued sub-Saharan Africa in the 1980s will become a quasi-permanent feature? 0.1%, 1%, 10%, 50%? Considered this way, it's clear that the dichotomy between the "climate believers" vs "climate sceptics" is indeed a false one.

The scientific method is sometimes described as "organised scepticism", and this, rather than some logical progression from one certainty to the next, characterises the inherently uncertain path of scientific progress. As one leading climate scientist put it: "In truth, we are all climate sceptics." However, despite the climate scientists' best efforts at scepticism, it simply has not been possible to rule out the risk of the sort of climate changes discussed above.

Handling uncertainty is key to the scientific method, but, conversely, the existence of uncertainty is not itself cause for inaction.

• Tim Palmer is a Royal Society 2010 Anniversary Research Professor at the Unversity of Oxford, and is organiser of a two-day discussion meeting at the Royal Society (22-23 March) on Handling Uncertainty in Science. The meeting includes speakers as diverse as Professor Sir Roger Penrose, eminent cosmologist and theoretical physicist, and Mervyn King, governor of the Bank of England.