Guide

Communicating uncertainty in climate science

Download the pdf

When the media talk about cli­mate change scep­ti­cism, they usu­ally mean that people are uncer­tain in some way about the reality or ser­i­ous­ness of cli­mate change.

A number of polls of public atti­tudes towards cli­mate change have doc­u­mented an increase in the degree of per­ceived uncer­tainty about cli­mate change over the past three years (BBC, 2010; Pew Research Centre, 2009; Spence, Venables, Pidgeon, Poortinga & Demski, 2010). These data on public opinion about cli­mate change can be con­trasted with a recent survey of active and pub­lishing cli­mate sci­ent­ists. Among this group, Doran and Zimmerman (2009) found 97% agree that human activity is con­trib­uting to cli­mate change. On the basic ques­tion of whether human activity is influ­en­cing the global cli­mate there is very little uncer­tainty among cli­mate scientists.

But des­pite the over­whelming amount of sci­entific evid­ence that shows human activ­ities are causing the earth to warm, debates about cli­mate change are still char­ac­ter­ised by an enormous amount of uncer­tainty (Zehr, 2000). The everyday meaning of uncer­tainty is neg­ative – and so when it comes to cli­mate change, people tend to infer that sci­ent­ists do not know any­thing about a topic, just because they do not know everything about it. Uncertainty about cli­mate change is a major bar­rier to public engagement.

Perhaps the most important task for com­mu­nic­ating cli­mate change effect­ively is being clear that uncer­tainty is not an enemy of sci­ence that holds it back – it is actu­ally a stim­ulus that drives it for­ward. Just like any area of com­plex sci­ence, uncer­tainty is a fea­ture of cli­mate sci­ence that will never go away. Science doesn’t deal in cer­tain­ties – it weighs up the evid­ence and tells you which of sev­eral pos­sible answers has the most sup­port. Where uncer­tain­ties remain, sci­ent­ists know where to put their efforts into what to invest­igate next. Getting the mes­sage across that uncer­tainty is not a bad thing, and that we make decisions every day based on less than cer­tain inform­a­tion is an important place to start for cli­mate change communicators.

So how much uncer­tainty is there about cli­mate change?

On the basic ques­tion of whether the earth is warming, and whether human activity is causing this to happen, there is very little uncer­tainty – but on more spe­cific ques­tions about when, where and how soon the impacts of cli­mate change will take place, there is still a great deal that is unknown.

The British Royal Society has pro­duced a guide to the sci­ence of cli­mate change that describes the areas where the sci­ence is well estab­lished, where there is still some debate, and where genuine uncer­tain­ties remain. Although it is clearly written, it is aimed at people who have some under­standing of the sci­ence. But the basic mes­sage is very clear: there is strong and reli­able evid­ence that human activity has caused the earth to warm over the past 50 years – on this crit­ical point, there is next to no uncer­tainty. Emphasising that uncer­tainty in sci­ence is normal will help to combat the belief that uncer­tainty should equal inaction.

As a com­mu­nic­ator, it is crit­ical to explain to people the dif­fer­ence between sci­entific uncer­tainty (that is, the extent to which sci­ent­ists agree about the answer to a par­tic­ular ques­tion), and uncer­tainty that comes from deciding how to respond to what the cli­mate sci­ence tells us. In his book, Why we dis­agree about cli­mate change, the cli­mate sci­entist Mike Hulme shows how easy it is to con­fuse the dif­ferent types of uncer­tainty that sur­round cli­mate change (Hulme, 2009).

No matter how much sci­ence is con­ducted, it will never tell us which of a set of dif­ferent policies to choose. A sci­entist can tell you how much carbon is in the atmo­sphere, and cal­cu­late the effect­ive­ness of dif­ferent methods of redu­cing it (e.g. redu­cing the number of cars on the road vs. reg­u­lating emis­sions from cement factories). But a sci­entist cannot tell you which of these is the right choice to make – this is a decision for cit­izens and politi­cians. Being as clear as pos­sible about the dif­fer­ence between cli­mate sci­ence, and the choices we make based on that sci­ence, is essen­tial for effective com­mu­nic­a­tion – oth­er­wise, these dif­ferent types of uncer­tainty become confused.

For people inter­ested in com­mu­nic­ating more spe­cific aspects of uncer­tainty in cli­mate sci­ence – for example, how con­fident sci­ent­ists are that a par­tic­ular impact will occur during the next 20–30 years – there is now some guid­ance avail­able. The Intergovernmental Panel on Climate Change (IPCC) has attempted (not always suc­cess­fully) to quantify and com­mu­nicate the uncer­tain­ties around cli­mate sci­ence in their Assessment Reports for policy makers. In the most recent report, in 2007, they spe­cified numer­ical prob­ab­ility ranges (e.g. 90% or more) at the outset of the report, linked them to spe­cific terms (e.g. “very likely”) and then used these terms throughout the report. Based on the most recent assess­ment of the cli­mate sci­ence evid­ence, the IPCC con­cluded that it was ‘very likely’ that human activity was causing cli­mate change.

However, when psy­cho­lo­gists from the University of Illinois (Budescu, Broomell & Por, 2010) ran an exper­i­ment to test how people inter­preted these written terms, they found a big dis­crep­ancy between the meaning the IPCC had intended, and the con­clu­sion people actu­ally drew. The numer­ical ranges that people actu­ally assigned to the verbal expres­sions were often not the ones the IPCC had used. The authors of the study con­cluded that written expres­sions such as ‘very likely’ need to be accom­panied by the prob­ab­ility range they refer to if they are to be effective: oth­er­wise people will inter­pret the same term in very dif­ferent ways. For com­mu­nic­ators wishing to talk about spe­cific events, using numer­ical inform­a­tion as well as a verbal term is likely to convey more accurate information.

Media-generated con­tro­versy is often cited as a reason for scep­ti­cism about cli­mate change – because media reports exag­gerate the uncer­tainty about the basic ques­tions of cli­mate change. Radio, tele­vi­sion and news­paper reports have been cri­ti­cised for inter­preting too simplist­ic­ally the notion of providing a ‘bal­anced’ set of views, which can lead to com­peting points of view on a sci­entific issue being presented as equally sup­ported when in fact they are not (Boykoff, 2007; Ward, 2008; Zehr, 2000). This cre­ates the impres­sion that the causes of cli­mate change are more con­tro­ver­sial than they actu­ally are.

Climate change com­mu­nic­ators can try to combat this problem by using examples of other sub­jects where a ‘bal­anced’ news report would not mean giving equal space to opposing points of view (e.g. the health impacts of smoking). Emphasising that ‘bal­ance’ doesn’t mean ‘equal’ but ‘rep­res­ent­ative of the evid­ence’ is one way of explaining to people why media cov­erage of cli­mate change can often paint an unreal­istic pic­ture of the under­lying science.

In a paper that was pub­lished in the journal Nature Climate Change, the social sci­ent­ists Nick Pidgeon and Baruch Fischhoff sug­gest that the best way to deal with uncer­tainty is to talk about cli­mate change as a risk. Although this might not seem a major dif­fer­ence, framing the issue as being about risk (rather than a set of pre­dic­tions about the future) turns the problem into some­thing that most people are used to dealing with (and that social sci­ent­ists know a fair amount about): per­ceiving and man­aging risks. It is also the lan­guage of the insur­ance, health and national security sec­tors. Pidgeon and Fischhoff sug­gest that the more the risks of cli­mate change can be brought to life through vivid ‘mental models’, the better (using clear prac­tical examples of the risk of sea level rise, or the risk of melting glaciers).

An aca­demic paper by psy­cho­lo­gists from the University of Exeter in the UK showed that uncer­tainty is not neces­sarily a bar­rier to com­mu­nic­a­tion if the right ‘framing’ of the problem is used (Morton et al, 2011). They gave people short mes­sages to read that con­tained uncer­tain inform­a­tion, but framed in either a pos­itive or neg­ative way. So, half the par­ti­cipants in the exper­i­ment read a mes­sage describing an 80% risk of abrupt changes to Monsoon pat­terns (neg­ative framing), while the other half of the par­ti­cipants read a mes­sage describing a 20% chance of avoiding abrupt changes to Monsoon pat­terns (pos­itive framing). The researchers found that when uncer­tainty was used to indicate that losses might not happen if cau­tious action is taken to pre­vent them (i.e. the pos­itive frame), then people were more likely to indicate stronger inten­tions to act in a pro-environmental way. The authors con­cluded that uncer­tainty is not an inev­it­able bar­rier to action, provided com­mu­nic­ators frame cli­mate change mes­sages in ways that trigger cau­tion in the face of uncertainty .

A good prac­tical example of an attempt to com­mu­nicate uncer­tainty in cli­mate sci­ence is the UK Climate Impacts Programme. It con­tains a great deal of inform­a­tion about the prob­able out­comes of cli­mate change for dif­ferent regions in the UK, and is a tool for national and local gov­ern­ments. The last major assess­ment, in 2009, used visual images and prob­ab­ility inform­a­tion to com­mu­nicate com­plex uncer­tainty inform­a­tion. Resources and reports can be down­loaded from their web­site.

Beyond uncer­tainty in cli­mate sci­ence, there is also a problem with public per­cep­tions of dis­agree­ment among cli­mate sci­ent­ists. One study in Nature Climate Change (Ding et al, 2011) dir­ectly studied the link between people’s sense of whether there was a ‘con­sensus’ on cli­mate change among sci­ent­ists, and their sup­port for cli­mate policies. Using survey data from the US, the study found that per­cep­tions of con­sensus were dir­ectly linked to a range of key beliefs (including whether humans were causing cli­mate change), which in turn determ­ined whether people sup­ported policies for redu­cing carbon. The mes­sage from this study is important – repeatedly stating that the vast majority of sci­ent­ists agree that humans are causing cli­mate change is one way of over­coming per­cep­tions of uncertainty.