First and foremost, happy new year! I hope you enjoyed reading articles on the CSAG blog last year. Let’s see if we can keep the momentum going in 2013. Please do share the blog with interested friends and colleagues and feel free to post comments on current and past blog articles.
I hope everyone had an enjoyable Christmas and if you stayed (wisely) in South Africa, I trust you made the most of the warm and sunny weather. I spent the Christmas period with my family back in England and as much as I would like to say, “the weather was actually okay…the image of the UK being cold and rainy really isn’t fair”, unfortunately the reality is that it was a wet, as opposed to white, Christmas. 
Indeed in England it has been the wettest year since records began. Apart from a timely intermission during the London Olympics, for much of 2012 large parts of the UK rarely spent more than one or two consecutive days without rain. After beginning the year with drought concerns following one of the driest years on record, the year ended with rivers and reservoirs at record levels and widespread flooding affecting many thousands of people. Even the seals were impacted – the picture to the left shows a confused seal in a flooded nature reserve in Cambridgeshire, some 50 miles upstream of where they are usually found on the Norfolk coast. Inevitability, as creatures with an innate sense of curiosity, people (rather than seals) have been asking what caused this large departure from “normal” conditions (though seals may also be asking the same question – who’s to say?) and whether or not…yes you’ve guessed it…climate change might be to blame?
When the weather does something out of the ordinary or has an impact on people’s lives, as climate scientists we frequently get asked whether or not climate change has a role to play. When Hurricane Sandy hit the US mainland in October 2012, arguably the most notable “climate event” of 2012, many articles questioned whether or not this was a result of climate change. Here we enter the realm of attribution, a much loved research topic for those involved in climate science, including those working at CSAG. Of course a single weather event cannot provide any definitive signal of a change in climate, nor can a change in climate directly cause a particular weather event to occur. However we must be very careful before dismissing any link to climate change. Just as it is wrong to say “yes this is because of climate change” so too it is misleading to say “no this isn’t because of climate change”. The implication of the latter response is that a changing climate has no impact on the frequency or magnitude of weather events – of course, this is not the case.
The truth (or something resembling the truth) is that in a changing climate, irrespective of the causes of that change, the probability of experiencing a particular weather event may (rather than will if we are being precise) alter. So what we would really like to know is not whether climate change causes a specific weather event but whether or not climate change has substantially altered the probability of experiencing that event. Indeed, perhaps the track and magnitude of Hurricane Sandy and the record English rainfall could have occurred even if the global mean temperature had not increased. The question is not one of possibility but rather whether or not the chance (or probability or likelihood; choose your favourite noun) of the event occurring has increased or decreased as the climate has changed.
A number of scientists are developing methods to improve the probabilistic attribution of extreme events to climate change. Pall et al. (2011) use a ‘probabilistic event attribution’ framework to determine a percentage increase in the risk of specific extreme events (in this case flood risk in England and Wales) in a greenhouse gas forced world compared to a world in which greenhouse gas concentrations are kept at pre-industrial levels. Their approach, and the approach subsequently advocated by Stott et al. (2011), requires numerous climate model simulations with different forcing scenarios. Even if we were armed with these models and were able to run them quickly, we would still get imprecise answers about the relative change in the probability. This occurs for a number of reasons. Firstly, all climate models are wrong. This of course is a trivial point but it is especially important when we are trying to say something about probabilities of localised weather events rather than merely specifying global or hemispheric averages. Secondly, we would have to run these models many thousands (or at least hundreds) of times to get a statistically robust result; with only a handful of runs, we have limited reliability. Thirdly, but presumably not finally, there are many assumptions which we make along the way about the nature of the events in question, the impact of different system feedbacks, the role of different processes and the relevant boundary conditions for pre- and post-industrial worlds. Nonetheless, and I am trying to remain positive, in my view this kind of research (and it is research) is highly valuable given the potential relevance of the findings to scientists and society alike. Some of the issues noted are tractable and further work could provide more accurate and precise estimates of the role of climate change in changing the probabilities of weather events.
However, when you or I are asked such questions by the person sat next to us on a plane, train or automobile, rarely do we have the opportunity to engage in a lengthy analysis using climate models and statistical analytical techniques; if only! So can we create a generic model answer to say something more than “I don’t know” while avoiding pure speculation and/or taking a defensive approach to confuse our subject by introducing difficult-to-understand meteorological jargon (usually something about jet streams and quasi-geostrophic winds)? Well…let’s try!
Perhaps it is useful to adopt the notions of direct and systematic causation. As opposed to direct causation (e.g. I hit a drum and this creates a sound; odd example but that’s how my brain works), the notion of systematic causation refers to the controlling conditions or multiple causing factors that influence whether or not an event takes place (e.g. my distance away from the drum, my ability to hit a drum, the lack of anything between me and said drum; this example is getting odder). In such a way (perhaps forget the drum example if this confuses), the climate can therefore be said to systematically cause weather.
However, this new jargon might seem superfluous to communicate the concept to the non-scientist. So let us simplify by saying that the climate provides the constraints for the likely weather we experience. So our answer to the question, “is event X caused by climate change?” goes something like:
- No single weather event is a direct result of a change in climate.
- However, a changing climate can alter the chances of a weather event occurring.
- Event X can only occur if the climatic conditions are favourable.
- In the case of event X, this may (or may not if we have reliable evidence) have been possible even if the climate had not changed.
- (use with caution) Yet because of climate change, the (insert process) has become stronger/weaker. This has an influence on the conditions which control this type of weather event and therefore makes it more/less likely for event X to occur.
- (only if you are brave) What do you think?
Failing that, just show them this clip.
Oh, and before we embark on 2013, it is always interesting to start with a prediction. The results for the global mean temperature aren’t in for 2012 yet but early evidence suggests that it will probably rank in the top 10 warmest years on record. However, if the UK Met Office are correct there is a real chance that 2013 could be the warmest year on record. The upper end of their forecast is 0.71 degrees C above the 1961-1990 average while the current record is 0.54 degrees C above (2005 and 2010). Watch this space!