Tom Mitchell, Theme Leader, Climate-related Disaster Risk Management, CDKN

Have you ever been drenched by heavy rain or sweltered in searing heat and wondered whether it might have been exacerbated by climate change? Until recently, the answer from scientists might have been: ‘today’s weather was consistent with the kind of extremes that we can expect in the future, but we can’t link individual weather events to climate change’. This picture is changing rapidly though, with climate scientists much more willing to tie weather events to climate change; an exercise known as ‘attribution’. Improvements in data and statistical techniques based on more refined models of the climate have allowed scientists to claim, for example, that the 2012 US drought has been made 20 times more likely by climate change. Such is the volume of new scientific papers on ‘attribution’ that the findings of the Intergovernmental Panel on Climate Change Special Report on Extreme Events and Disasters, released earlier this year, probably need updating already.

To find out more about these scientific advances and to understand the implications of attribution, I attended a gathering of the world’s foremost attribution scientists in Oxford last week. The meeting also involved a myriad of other groups keen to hear what the scientists had to say:  journalists, lawyers, insurance experts, government officials, NGOs, army engineers and water managers. The breadth and quality of the participants, combined with a room bursting at the seams, gave a clear indication that this was no standard scientific conference.

Being more technical for a moment - ‘attribution’ is the ability to decipher an anthropogenic climate change signal in the cause of a particular weather or climate event (e.g. extreme temperature, drought, heat wave, tropical cyclone) from other potential causes. These other causes could include El Nino/La Nina, natural longer-term climatic cycles, or just regular variability. Until recently, separating a climate change signal from the other causes has been hampered by the lack of long-term data on observed weather extremes. By definition, weather extremes are rare and you need a long data record to see if one event is a significant departure from the historic trend. This record does not exist in many places, and is certainly rare in developing countries.

A new technique however, based on the availability of improved climate models, allows scientists to run a set of statistical experiments in the model to see how often a particular weather or climate event is predicted. Based on analyses of many such experiments, they can then say how much more (or less) likely the event is given anthropogenic influence on the climate. Importantly, this even allows such judgements to be made in areas or countries without long data records or incomplete coverage of weather stations. The result is a set of statements about the enhanced probability of experiencing such an event given greenhouse gas emission volumes.

Of course, using models, rather than historic records introduces a higher degree of uncertainty, especially in places where the climate model lacks observational data to calibrate it. The greatest degree of certainty emerges where the model includes larger numbers of variables and the data records extend back over a century or more. Naturally, attribution scientists spend lots of time discussing and worrying about uncertainty and want to communicate it as clearly as possible. Nonetheless, using models does offer the potential to make predictions about the future of weather and climate extremes under different emissions scenarios – valuable information given that the historical record is an increasingly poor guide to the weather of 20 years time.

When asked to reflect on what attribution information meant for their work, interested groups said some uncomfortable things that made the scientists squirm in their seats:

• Lawyers, primarily from the US, are already engaged in law suits trying to get big oil companies to pay for a share of the damages caused by weather and climate extremes. They were clear that proving liability in such cases involves two elements – what role did climate change play in causing the extreme and what role have the oil companies played in causing climate change? Clear enough answers to both these questions based on high quality science have the potential to shape a generation of legal battles and lead to costly payouts for historical emitters. Whether the attribution scientists in the room were willing to be expert witnesses was a different matter. I was left wondering whether enough negative outcomes of such courtroom battles might eventually change the tide of US political opinion and the international climate negotiations too.

• Journalists talked about media coverage of the US drought (and other disasters) and how this was being portrayed on weather channels and in the newspaper. Public opinion about climate change is notoriously fickle, but recent articles have suggested that extreme events have created a reasonably significant shift in the US public’s belief in climate change. A clear message that says the more you emit, the worse the weather will be appears to have some currency at a time when individuals are experiencing damaging extremes.

• Insurers were more circumspect, saying that marginal changes in probabilities of extremes or their attribution was unlikely to have much impact given the way that they aggregate risks in their pricing of policies. They did however admit that any changes to tropical cyclone frequency or severity was likely to have more fundamental consequences (at the moment evidence of attribution in tropical cyclones is inconclusive). Other wider changes were also likely to result in increases in the pricing of policies, with the potential implication being that fewer and fewer people will be able to afford to take out insurance for extreme weather.

• Governments talked about the need to understand changing extreme event risk in the design of infrastructure, land-use planning and economic investment plans, with a representative of the UK government talking about the importance of public perceptions of climate change for formulating policy. For many developing country governments that CDKN works with, there is a certainly a need for more precise information. Learning that rainfall extremes or drought recurrence might exceed current design thresholds in the future is an important part of CDKN’s climate resilience work with the Ministry of Transport in Colombia or in post-flood reconstruction in Pakistan for example.

• I talked about the international climate change negotiations and the way that attribution questions were being discussed in the ‘loss and damage’ work programme, where CDKN is a key supporter of the Government of Bangladesh in its capacity as co-chair. ‘Loss and damage’ is a term used to describe what happens when adaptation thresholds have been exceeded. For poor countries, this is most acutely experienced when a disaster strikes where climate change can be identified as a contributing factor. The loss and damage work programme is considering how international mechanisms to reduce disaster risk, insure losses and boost humanitarian action might play be developed with the support of historic emitters. The science questions in the work programme are challenging and mirror some of those of the US courtroom described above – how do you distinguish the extent to which climate change has caused the ‘loss and damage’ and therefore what proportion should be addressed by a mechanism? Such questions resonate for a number of CDKN’s projects, especially in El Salvador, where CDKN is helping the country to deal with a radically changing pattern of extreme rainfall and economic losses. The government there wants to know who will help them cover these losses.

Undoubtedly the attribution sciences and the way society digests and uses the information are in their infancy but there are clear implications for how we tackle climate change. At times the meeting did give the impression that this corner of the climate sciences is shaped by a small, tight-knit community from the US, UK, Australia and a handful of other countries. As it evolves it will need to globalise by involving scientists and institutions from many more developing country regions. The challenge for CDKN, in its role in supporting the policy community in developing countries, is to optimise the way attribution information is used to create the best possible outcomes for poor and vulnerable people. The diversity of user groups, the depth of uncertainty and the number of communications angles makes this both a challenging and exciting frontier.