Back in 2016, Dr Friederike Otto, Associate Director of the Environmental Change Institute at the University of Oxford, put her name to a paper in which the following words can be found:
“The answer to such an open question as have CO2 emissions caused the 2003 European heatwave is thus dramatically affected by (i) how one defines the event 2003 European heatwave and (ii) whether causality is understood in a necessary or sufficient sense. Precise causal answers about climate events thus require precise causal questions.”
Given that the words ‘dramatically’ and ‘precise’ play such an important role in the above statement, one would have thought that Dr Otto would thenceforth be fastidiously careful when publically communicating the causation of heatwaves. Well, maybe she has been or maybe she hasn’t. This is what she recently said regarding the causation of the Pacific Northwest heatwave of 2021:
“The Pacific Northwest 2021 heatwave is still rare in today’s climate, yet would be virtually impossible without human-caused climate change. As warming continues, it will become a lot less rare.”
Subject to the usual caveats regarding the uncertainties in climate models and the statistically dubious manner in which uncertainty is treated when such models are handled as ensembles, there is nothing wrong with the above statement – but only as far as it goes. The two problems with it are that:
i) It doesn’t adequately define the event.
ii) It understands the heatwave’s causation in a purely necessary sense. Nothing is said regarding sufficiency.
Does this matter? Well it seemed to matter a great deal to Dr Otto when she was talking to other scientists back in 2016, and she had seemed very keen then to emphasise to her colleagues how much it should matter when communicating results to a wider audience. But when it came to doing so herself, it didn’t seem to matter to Friederike all that much.
There may be a number of reasons for this apparent ambivalence regarding the importance of providing precise and complete causal answers. Some seem quite reasonable to me, whilst others reflect less well upon the scientific community. I think it depends upon one’s view regarding the essence of causation and the importance of downplaying uncertainty.
The Humpty-Dumpty syndrome
Firstly, it is indeed very easy to influence the narrative when one enjoys the freedom to define an event as one wishes. For example, it didn’t seem to bother any of the scientists involved in the Pacific Northwest 2021 heatwave attribution study that the duration of the exceptionally high temperatures failed to meet the World Meteorological Organization’s criterion for the definition of a heatwave – at least not for the area covered by the study. Such detail didn’t matter to the attribution team because the actual event they were interested in was the recurrence of such extreme temperatures within a stated period – no matter how transient such a recurrence may be. Furthermore, one can also be very choosey regarding the geographical extent of the ‘heatwave’. The scientists knew where the newsworthiness lay, and so didn’t think it relevant to reference the unusually cold temperatures contemporaneously visiting other areas of the North American content. Fair enough, but it would still be interesting to see attribution studies that demonstrate just how unlikely certain cold snaps are in the light of global warming. Unfortunately, the message doesn’t communicate as well when focusing upon events that are less likely with human-caused climate change, rather than events that are ‘virtually impossible’ without it.
Secondly, what about that failure to mention the sufficiency of human-caused climate change? Dr Otto was very keen to draw attention to the necessity of human-induced climate change in order to understand the heatwave, but where was the equally valid statement to the effect that human-caused climate change was almost certainly insufficient to explain it? Was this fact downplayed because it was giving the public too much to think about? Was it deemed too much off-message? The omission seems even more troubling when one considers the extent to which the records were broken. The freakishness of the event was so obvious that one has to say (from an anthropogenic climate change perspective) that the levels of insufficiency were high enough for everyone to start panicking that the climate models must be wrong. But does such high insufficiency matter when it comes to making a statement of causation?
Well not to the majority of climate scientists. They are more than happy to emphasise the necessity of anthropogenic climate change and leave it at that. However, despite this, they are not lying by omission. Rather, it is a case of most of them simply failing to understand that a causation statement is incomplete without a reference to sufficiency. And even those that do understand this still believe that (when push comes to shove) it is only necessary causation that truly matters. In the very same paper in which Dr Otto and her co-authors bemoan the common failure to offer full disclosure in attribution statements, you will find a simplistic example of how probability of necessity (PN) and probability of sufficiency (PS) work in law, followed by this assertion:
“The probability of sufficient causation PS is thus close to zero here, but this is not important in a legal context, in which it is only PN that matters, while PS does not.”
Frankly, particularly given the facile nature of the example the authors use, I think they are well out of their depth here in discussing the legal angle. Many natural disasters are man-made in the sense that the necessary weather conditions were far from sufficient in causing the damage or loss of life. Poor forest management, delinquent flood protection, poor drainage, arson, bad building design and inappropriate siting of buildings are all acts and omissions that bear upon matters of sufficiency from the AGW perspective, and yet they all add to the culpability; it is certainly not a case of being able to ignore low PS values. But that does not matter here. All that matters is that Dr Otto, along with the remainder of the climate scientists and activists, are fully signed up to their view and so cannot be accused of hiding a factor (low sufficiency) when they secretly know it to be very important. They are not hiding it – they are either ignorant of the issue or just convinced it can be ignored. Furthermore, there is once again the relevance of event definition. According to Dr Otto and her fellow authors back in 2016:
“Even in the few such cases where evidence supporting necessary causation is strong, assertive causal statements appear to have been shied away from, possibly by the perception that sufficiency was lacking. A statement such as “CO2 emissions have not caused the particular event Y: they have only caused the probability of occurrence of Y-like events to increase” may actually often be too conservative and even wrong; as in the above example, it may indeed be the case that CO2 emissions did cause event Y, although in a restrictively necessary causation sense. Further, by defining the event to mean not just occurrence in a particular year but during the entire industrial era, it may be possible to establish that event Y was in fact caused by increased CO2 emissions—this time wrt both necessity and sufficiency.”
So it seems that a lack of sufficiency need not be a problem. One can always convince oneself it isn’t really relevant if one takes a judiciously long term view. That certainly seems to be the position taken by Dr Otto in the Northwest Pacific heatwave attribution study. When she said in 2016 that a different definition of ‘event’ can dramatically affect the answers to attribution questions, I took that as a warning. I didn’t appreciate that she meant it to be taken as a glorious opportunity.
Staying on message
However, there is another way of looking at this issue, in which the lack of attention to matters of sufficiency appears far less benign. The problem, it seems, is a concern that any allusion to sufficiency could be construed as an exaggeration of the uncertainties – and we can’t have that. As Dr Peter Stott of the Hadley Centre puts it:
“If the public hears that a particular weather event is consistent with climate change they may conclude that it is further proof of the immediate consequences of human-induced global warming. On the other hand, if the public hears that it is not possible to attribute an individual event, they may conclude that the uncertainties are such that nothing can be said authoritatively about the effects of climate change as actually experienced.”
In other words, a leading climate scientist is saying that one cannot overestimate the importance of ensuring confirmation bias when it comes to extreme weather event attribution, and this objective is reason enough to expunge any talk of sufficiency from the debate. Furthermore, there are plenty of others who are willing to take this cause seriously. For some time now, the BBC has used the following standard statement whenever reporting upon weather event attribution:
“Experts say climate change is expected to increase the frequency of extreme weather events, such as heatwaves. However, linking any single event to global warming is complicated.”
Be it ever so glib, there is, at least, an obscure hint at insufficiency in there. But climate scientist Ed Hawkins is not happy with such ‘excess caution’, suggesting that the journalists replace that statement with:
“Experts say that climate change is already increasing the frequency of extreme weather events, and many single events have been shown to have been made worse by global warming.”
So that’s fine. Keep it simple, keep it on message, and don’t give the public a chance to think about anything that might cast the seeds of doubt.
In science we can trust – just keep your eye on the scientists
Censorship takes many forms. One of them is when well-intended but uncomprehending climate scientists step in to tell journalists how to do their job. It is well known that state control of the media sounds the death knell for objective journalism, but one does not expect that sort of jackboot intolerance to be coming from the scientific community — of all places. And one certainly doesn’t expect it to come in the form of a seemingly innocent attempt to ‘correct’ an inexactitude.
I’d like to make a confident statement as to whether Dr Friederike Otto is playing the same game with her own handling of the press, but I can’t be sure. With so much hocus-pocus going on, what is a poor denier supposed to think? She tells her colleagues one thing, then she acts as if she doesn’t care about her own advice. This may be a justified ambivalence, but it’s difficult to say. And so I’ll finish as a I started, by simply asking the question: Friederike Otto, just what is your game?