Dr Friederike Otto’s work – which considers whether human activities are increasing the likelihood of extreme weather events – has wide-ranging implications, from environmental damage claims to post-disaster reconstruction.
3 minute read
How has the science of climate attribution changed since you began working in the field?
It’s changed from something people had suggested was theoretically possible to something that is now being done at scale. This is mainly down to the change in computing power; you can run large ensemble simulations of climate models to look at weather, not just once or twice but several hundred times.
To get something sensible out of climate models, we also needed to develop methodologies to establish the relationships between an attributable weather event, the damages and the event people experience. We have developed a protocol where we use different models and methodologies, based on climate and statistical modelling. These can be combined with observational data and synthesised for an assessment of what the role of man-induced climate change is in an event people have just experienced.
Is climate attribution possible for all extreme weather events?
No. It is possible for assessments for heatwaves, large-scale rainfall events, flood events and droughts. We are beginning to develop methodologies for windstorms, but that’s quite difficult at this stage. We cannot carry them out for hail, or things occurring on small spatial scales: we just don’t have climate models that can reliably simulate these events.
In the developed world the network of weather stations and observed data is quite dense
In the developed world, particularly Europe and the US, the network of weather stations and observed data is quite dense, and the quality of the observations is high almost everywhere. There are weather observations for about the last 100 years. That is not the case in other parts of the world. Without observational data, you cannot evaluate whether the climate model is up to the job, because you don’t know what the model should be showing. This makes confidence in attribution assessments in developing countries much lower than in Europe and the US.
Can you give examples where you are certain human activities have contributed to more extreme weather?
One of the strongest examples is in the occurrence of European summer heatwaves. We have conducted a few studies in different parts of Europe, using different methodologies. In many places, climate change is making the heatwaves we have experienced recently orders of magnitude more likely.1 Climate change is a real game changer.
We have seen between a doubling and three-fold increase in the likelihood of extreme rainfall events
We have also seen the impact in the rainfall brought by hurricanes in the Gulf of Mexico. We have seen between a doubling and three-fold increase in the likelihood of extreme-rainfall events. We have looked at the droughts in Southern Africa and found climate change played a large role in changing the likelihood of these events.2 In other droughts in Brazil or East Africa, our analysis suggests climate change did not play a major role.
One possible extension of this work is in calculating damages. Is there much interest from clients?
Using the word “clients” implies the users will pay. We do not have clients in that sense. We work a lot for the Red Cross; many of the studies were initiated because the Red Cross contacted us and said: “We have huge damages from a storm or a drought or a heatwave. The question is whether to relocate or re-build.” That hinges on whether it was a freak event or the harbinger of more to come in a changing climate.
When we do studies quickly, we do them without peer review
We have had conversations with commercial companies. But because we do research, our models and our results are publicly available. This is important: when we do studies quickly, we do them without peer review. We need to be super transparent, so everyone can redo what we do to ensure scientific integrity. Commercial clients don’t really like that aspect, which has made working directly on projects with – say – the insurance industry difficult, because it wants to keep that information closed.
You have been producing curves exploring “non-linear” relationships between a climate event and potential damages. Can you explain this work?
What we can do with these attribution methodologies is not just stop at the meteorology, but go down the subsequent pathways as well. For flooding, we have looked at the rainfall, river discharge and the inundation that followed, and found the relationships are not linear. In principle, you could carry out these studies for all sorts of damages.
We look at the overall effect of man-made climate change, including all the greenhouse-gas and aerosol emissions from the beginning of the Industrial Revolution
We have also looked at the other end of the spectrum – mostly we look at the overall effect of man-made climate change, including all the greenhouse-gas and aerosol emissions from the beginning of the Industrial Revolution. But we have done analyses at a country and regional level. For example, how much did the actions of the US or Europe change the likelihood of a specific extreme event? And we have even looked at a company level. How much might one company be responsible for?
Of course, there are a lot of political, social and legal questions that influence the outcome. When we did the country study, you could use all the emissions since the beginning of the Industrial Revolution. Or you might say: “Well, you can’t say that people really knew about climate change at the time. Are they really responsible for it? Maybe we only take the emissions from 1990, when the first Intergovernmental Panel on Climate Change report was published.” You can do that as well, and of course you will get a different number. There is a large difference between the two.
Scientifically you can do both, and both make sense, but which is the one you might want to use in court, or for any other purpose?
These are sensitive areas, geopolitically and commercially. Are people actively working on them?
Yes. I’ve just been in Dundee at a conference of lawyers who all have live climate-litigation cases. I wouldn’t say it’s massive yet, but it is growing rapidly.
What are the limiting factors, from your perspective?
Now that we have the scientific methodology, we need to work with people in adaptation planning – that’s where there is a huge gap in knowledge transfer. We will need to work with lawyers too. How will we transfer this into legal language, so it’s most useful? How do we do the translation backwards and forwards between real intervention points in the real world? We need funding and manpower to work on these interdisciplinary questions.
You have been working collaboratively to deliver event attribution in almost real time. You need to work fast – how do you do it?
Every time we do a rapid study, we ask colleagues around the world: “Do you want to join in? It would be fantastic if you would provide your data.” They also know society is asking pressing questions on climate change and it is important to answer them. We drop everything else for a short time, don’t sleep much – and just do it.
