As a result, it urges the panel to always give the percentage chance of an event occurring whenever it uses such risk phrases.
The confusion arises with phrases such as ‘likely’, ‘very likely’, ‘about as likely as not’ or ‘virtually certain’, which the IPCC uses in its reports to denote the probability of extreme events such as floods arising. The numerical definition of the phrases typically appears in a separate appendix (see table for a list of the phrases’ definitions).
|Virtually certain||More than 99|
|Very likely||More than 90|
|Likely||More than 66|
|About as likely as not||Between 33 and 66|
|Unlikely||Less than 33|
|Very unlikely||Less than 10|
|Exceptionally unlikely||Less than 1|
Participants were randomly divided into two groups. One was given the IPCC’s risk phrases, translated into their language. The other saw the same phrases, but the numerical risk was also presented in brackets after the words.
The second format would, for example, state: “It is very likely (more than 90 per cent) that hot extremes, heat waves and heavy precipitation events will continue to become more frequent.”
The authors, from Fordham University and Carnegie Mellon University in the United States, and the Australian National University, asked all participants to judge the upper and lower limits of probability represented by the statements.
They found that 27 per cent of those in the first group correctly understood the IPCC’s meaning, but this rose to 40 per cent in the second group. Most people tended to underestimate what was meant by high probabilities of extreme events occurring and overestimate low probabilities.
It might seem surprising that only 40 per cent of people gave percentages matching the IPCC’s definitions when presented with the figures. David Budescu, the study’s lead author, tells SciDev.Net that the respondents may have allowed their own judgment of the events’ probability to creep in, and this could explain the discrepancy. “Although we went out of our way to stress that we don’t ask people for their opinions, it is possible that some can’t totally supress their own views and beliefs — and these colour their estimates,” he says.
“Since most readers interpret the terms conservatively, the danger is that the message of the IPCC would be diluted.”
The authors propose that the IPCC adopts the second format — which they call the ‘verbal-numerical’ format — for describing risk.
“Since most readers interpret the terms conservatively, the danger is that the message of the IPCC would be diluted since the likelihood of some of the events and the severity of some of the consequences would be underestimated,” says Budescu. “We found that one can improve communication in all countries by adopting the proposed method,” he says.
He also tells SciDev.Net that the paper has been sent to the IPCC. “I hope they will take the results into account when planning the next report,” he says.
But two Indian authors of the IPCC reports disagree over the suggestion.
Shreekant Gupta, coordinating lead author for the chapter on risks and uncertainties in the latest IPCC report on climate change mitigation, says: “The IPCC would do well to consider the study carefully, if it wishes to communicate a certain message. Such studies should be seen as constructive, and not as criticism.”
But N. H. Ravindranath, a drafting author for the summary report for policymakers on climate change mitigation, which was released last month, sees no reason for change.
He says the policymakers from 195 countries who were at the launch understood the current descriptions and none “raised any serious objection”. So the current methods “are unlikely to be changed”, he says.
> Link to full paper in Nature Climate Change.