The significance of the “pause”

I wrote a post a while ago about there being no “pause” in global warming. It was based on a paper by Risbey et al. called [a] fluctuation in surface temperature in historical context: reassessment and retrospective on the evidence, which essentially argues that

The results show that there is little or no statistical evidence for a lack of trend or slower trend in GMST using either the historical data or the current data.

John Kennedy has a post that seems rather critical of this paper. We’ve also had some brief discussions on Twitter.

I’m still not entirely sure that I understand John’s criticism (hence this post) but I think it relates to a particular analysis that is carried out in the paper. They used a Monte Carlo approach to generate 1000 synthetic realisations of global surface temperatures. They did this by assuming the same linear trend as in the actual data (between 1970 and 1998) and then superimposing white noise with the same standard deviation as the residuals in the actual data (and they did this for each temperature dataset). They then compared trends over the pause period with trends of the same duration, sampled across the full period of their synthetic data.

What they found is that rarely was the trend during the “pause” period some kind of outlier (i.e., rarely were fewer than 5% of the synthetic trends smaller than during the “pause” period). In other words, there was little to suggest that the trend during the “pause” period differed, statistically, from what might have been expected for trends of the same duration over the whole period considered.

I think John’s crtiticism is that this does not mean that there aren’t interesting differences between these different periods. I, of course, agree; just because two periods seem statistically consistent does not mean that the underlying physical processes associated with these periods are the same. However, I don’t think that the authors of this paper are really suggesting that these different periods are the same in all respects. I think they are simply demonstrating that there is little to indicate that the period, often referred to as the “pause”, is somehow statistically distinct from other periods of a similar duration. It is essentially an attempt to address statistical arguments that have suggested that this period is indeed unusual.

Of course, maybe I misunderstand John’s argument, and I’m not suggesting that there couldn’t have been better ways to address this issue. I don’t think, however, that this paper is really making any claim about the underlying processes associated with the “pause” period; I think it is mainly demonstrating that there is little to indicate (based on a statistical analysis) that there is anything particularly unusual about this period. This doesn’t mean that there aren’t interesting things about this period. It just suggests that there wasn’t really anything about this period that implied something unexpected had happened to challenge our basic understanding of anthropogenically-driven climate change.

Addendum
I seem to have trouble keeping my posts short, but wanted to add something to this. I do wonder if some of the issue about this paper is related to context. I think it is partly motivated by extensive public discussions of the “pause”, which was clearly used by some to argue that it challenged our understanding of AGW. The large number of scientific studies that considered this period, also played a role in suggesting that it was indeed a period of particular interest. As someone involved in discussing this publicly, I found some of this quite frustrating. It seemed quite clear that this period was not really interesting in some way that differed from what might make other similar periods interesting, or that it somehow challenged our basic understanding. Despite this, it still seemed to dominate much of the public discussion.

From a scientific perspective, you could argue that this is all irrelevant; science is about trying to gain/improve understanding and if something seems interesting, then people will study it. I largely agree with this, but I don’t think this means that there is no merit in reflecting on a research topic. Even if we wouldn’t have done anything differently at the time, that doesn’t mean we can’t learn something so that we might do something different in the future.

Even though we shouldn’t let societal factors influence what we conclude from some research, this doesn’t mean that we can’t think about how to present some research, what terminology to use, or how to promote some research results. I think it’s important that the public draws conclusions that are a reasonable representation of our scientific understanding, and scientists have an important role to play in trying to ensure this.