We did not start from scratch»
The IPCC (Intergovernmental Panel on Climate Change) published the first special report on the risks of extreme events and disasters (SREX). Together with the Australian climate scientist Neville Nicholls, ETH-Zurich professor Sonia Seneviratne was coordinating lead author of one of the nine chapters. She talks about her work and the report in an interview.
The new IPCC SREX special reportevaluates and summarises scientific literature on changes in weather and climate extremes, and assesses the consequences of such changes in the context of disaster risk reduction and climate change adaptation. The chapter that Sonia Seneviratne coordinated provides the physical scientific basis regarding the observed and projected changes in climate extremes. It is the first IPCC report exclusively devoted to extreme events.
Ms Seneviratne, as coordinating lead author you supervised a chapter of this special report. The full title of the report is Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. Writing a chapter within such a report sounds like a challenging task.
Although the topic of our chapter is related to my research, it was indeed a considerable additional work load. It may have partly been a disadvantage while I was working on it as I had less time for my own research. But in the long run, it was very useful to me as a researcher because I now have a much better overview of the entire field and can see where there are gaps and which areas are most in need of further research. Of course, it is also an honour to have been given this task.
How long did the project take?
The project began in 2009. In all, we worked on this special report for two and half years.
The chapter you supervised is entitled Changes in Climate Extremes and their Impact on the Natural Physical Environment. What does it mean?
The whole report was coordinated by IPCC Working Group 1, which addresses the physical science basis on climate change, and IPCC Working Group 2, which addresses climate change impacts on ecosystems and society, as well as climate change adaptation. Our chapter was the only one by Working Group 1 in the report. The term “natural physical environment” highlights that our chapter addresses the purely physical aspects, including physical impacts, but not impacts on ecosystems and society.
What are the purely physical aspects of climate extremes?
We distinguished three separate categories of events relevant to disasters: first, weather and climate extremes in temperature, precipitation and wind; then, phenomena that are linked to these occurrences: tropical and extratropical cyclones, as well as El Niño and other large-scale modes of variability; finally, impacts on the physical environment, such as droughts, floods, coastal impacts, or impacts in mountainous environment.
You and the other authors of the chapter considered around 1,100 publications on extreme events. How did you know where to start?
We did not start from scratch. There were previous reports by the IPCC that had addressed aspects of this topic. But this is the first IPCC report exclusively devoted to extreme events. Previously, the topic was divided into different reports and, in the Fourth Assessment Report, the AR4, into different chapters. We began with this material. Of the 1,100 publications, however, over seventy percent were published after the AR4, in other words since mid-2006. Even while we were compiling the report, new relevant literature kept on being published. We compared the new publications with the AR4 assessment and that of an IPCC technical paper from 2008 on the topic “climate change and water”. We checked what was new or what was no longer correct. We had for instance a significant amount of new literature documenting more extensive analyses on the topics of droughts and tropical cyclones.
How did you approach the evaluation of publications on the specific extreme events?
The evaluation process became very systematic compared to the AR4, and we used the new IPCC uncertainty guidance for the assessments. As a first step, we determined the confidence we had in the data basis available for the considered extreme events. For instance, if for a given extreme we had only few publications based on a limited number of datasets or if the process understanding still showed considerable gaps, the level of confidence was assessed as being low. If the quality of the underlying data basis was good enough for us to be able to evaluate the sign of the change in a given extreme but not sufficient to make quantitative assessments regarding the probability of change, then the level of confidence was assessed as medium. As an example, drought is an important topic, but we assessed the level of confidence in the projections as only medium given that there are not many relevant observations and that the involved processes are complex. It is also difficult to assess how good the models are in cases where there are strong natural fluctuations and few events for the analyses, such as for El Niño. Quantitative statements on the probability of a change are only provided in cases where we assessed that we had high confidence in the underlying evidence.
Can you give an example?
Based on the available evidence, we can state that a future increase in heat extremes on the global scale is virtually certain, that is that it will occur with a degree of certainty of ninety-nine to 100 percent.
You even provided assessments of observed and projected changes in extremes at the regional scale.
Yes, we provided such assessments for temperature extremes, heavy precipitation events and droughts for twenty-six regions. It is the first time that such detailed regional information on extreme events has been compiled. We also provided global-scale assessments for all considered types of extremes, which are summarized in another table.
How did you deal with contradicting studies?
We outlined the different views on the given topics and provided an objective synthesis on the available literature.
What are the underlying messages of your chapter?
The main message is that there are already changes in extremes that we can observe. How sure we are with regard to these changes depends strongly on the considered extremes and regions. For instance, we assess it as very likely that increases in hot extremes have been observed on the global scale. On the other hand, cold extremes have become less frequent. In the case of heavy precipitation events, we identify more regions with increases than decreases globally, hence an overall tendency for increases in these events at the global scale, but at the level of single regions it is often more difficult to identify a signal.
Why is that?
We have to determine the presence of a signal from statistical analyses. The identification of the signal depends on the signal to noise ratio. If the noise is large, for instance because there is a strong natural variability, it is harder to see a signal. The signal-to-noise ratio is generally larger at global than regional scale, because the regions present different types of background noise that cancel out when all the data is averaged on the global scale. Finally, there are also extreme events for which the data is still limited and for which we are thus still unable to assess whether they are influenced by the increasing greenhouse gas concentrations.
Were there some surprises?
No, with a few exceptions. Globally speaking, it is clear that there are changes for some extremes that are caused by humans, especially for temperature and heavy precipitation, and that we expect an increase of these trends for the future. But there were also some changes compared to the AR4 assessments. On the one hand, we have now in the SREX more detailed assessment for aspects that were not previously treated, for instance more detailed regional assessments as well as partly quantitative assessments on the magnitude of changes for temperature and precipitation extremes. These analyses reveal large regional differences in the data basis, the observed trends and the projected changes in extremes. On the other hand, the latest literature has also revealed a greater level of uncertainty for certain extremes than was discernable at the time of the AR4. This is especially the case for droughts and tropical cyclones. Nevertheless, we can also provide assessments for these extremes – for instance, with medium confidence, that some large regions, including Central Europe and the Mediterranean, will be at a greater risk of droughts in future.
Apart from the two coordinating lead authors, twelve lead authors and twenty-eight contributing authors worked on the chapter. Was there a clear division of responsibilities?
I was extremely lucky that we were two coordinating lead authors and that Neville Nicholls already had experience as lead author on previous IPCC reports. The main structure and title of the chapter were provided by the IPCC based on a scoping meeting for the report. We could provide some suggestions for the choice of some lead authors given their scientific expertise and country of origin. In particular, we had experts on specific extremes or specific methods in climate research.
The chapters of the special report were reviewed by other scientists and government representatives for quality check. How did that work?
We had first an informal review of the zero-order draft of the report. The first-order draft was then officially reviewed by scientific experts. Following this, the second-order draft was reviewed by scientists and government representatives. There were also separate reviews of the Summary for Policymakers, in which material from our executive summary was included. We received a total of ca. 5,000 comments, not all on content, as a few of them also pointed out typos. However, we had to answer them all in writing. All the authors from our team had to do this for their respective contributions. But in the end Neville Nicholls and I had to read through and check them all to make sure everything had been answered correctly. This represented a large amount of work, but it was a valuable process to ensure the quality of our chapter..
Is quality assurance new within the IPCC and a response to errors that were found in the AR4?
No, these were also the quality standards at the time of the AR4. But we were particularly careful in this context to make sure that there were no mistakes.