I’ve been reading a paper by Shannon Osaka and Rob Bellamy called Weather in the Anthropocene: Extreme event attribution and a modelled nature–culture divide. I’ve written about event attribution before, and I’m largely in favour of the storyline approach; given that an event has occured, how might climate change have influenced this event?
This new paper is somewhat critical of extreme event attribution. There are a number of aspects that are considered, some of which I broadly agree with (although, some I agree with because they’re true, not because they’re all that relevant – yes, event attributions relies on models, so what?). However, there was one in particular that I found very confusing. The claim is that in trying to separate the human influence from the natural variability of weather, extreme event attribution creates a new nature-culture divide.
Okay, but this is kind of the whole point of event attribution; we’ve gone from living in a world where atmospheric CO2 had been at around 280ppm for a very long time, to a world where it’s at 410ppm, and rising. We’d like to understand how this is impacting, and will continue to impact, climatic events. If you don’t like the terminology, you could change the framing, but the reason that atmospheric CO2 has gone from 280ppm to 410ppm is because we’ve been dumping CO2, and other greenhouse gases, into the atmosphere through our use of fossil fuels.
The paper suggests that
The danger of such approaches is that they might obscure, elide, or distract from the many other forms of causality: for example, human influence that cannot be modelled as greenhouse gas emissions, but which is instead enacted along axes of vulnerability, inequality, and other socio-political dimensions.
The problem here is that extreme event attribution typically tries to understand how the event might be different because of anthropogenic-driven climate change, rather than trying to understand how the actual impact of that event might be different. The above does complicate some events, such as flooding, but – as far as I’m aware – this is acknowledged. I’ll add that anthropogenically-driven climate change isn’t only due to greenhouse gas emissions, but this is the dominant factor.
The paper goes on to suggest that
Modellers could attempt to incorporate other aspects of risk, to make visible those aspects of causality that are currently elided.
I think it would be very good to consider how these other factors influence the impact of extreme weather events, but putting them into models (at least the ones used for event attribution) is extremely difficult. I also think that this is a slightly different question. Event attribution is typically trying to understand how the properties of a climatic event has been influenced by anthropogenically-driven global warming. How this then impacts communities, how we might respond, and the influence of other socio-political factors is a related, but different, issue.
These other factors are clearly important, but it’s not clear why they should really be considered by those doing event attribution studies. I will agree that those who do these studies should think a little about the implications of what they present, but – at the same time – we should be cautious of suggesting that scientists should take socio-political considerations into account. There is a difference between being careful about how you present your results because of socio-political sensitivities, and letting these socio-political sensitivities influence how you do your research.
One obvious concern with what is suggested in this paper is that if we don’t distinguish between natural and anthropogenic influences, how do you then avoid people simply concluding that it’s natural, or using this to argue that it’s natural? It’s bad enough that some already use the complexity of attribution studies to suggest that we still don’t know if climate change is influencing extreme weather events, without also then blurring the distinction between natural and anthropogenic.
Of course, maybe I really just misunderstand what’s being suggested in this paper. If so, I’d be more than happy to have it clarified.