Settling some Science

Once again Margaret Wente is hyperventilating over hot new talking points in climate-change denial circles. During an interview, climatologist Dr. Phil Jones said that CRU temperature data shows the increase in global temperature over the last fifteen years is ‘less than statistically significant’.

To quote Homer Simpson in a similar state of mind, whoo hoo! A real climate scientist is saying that global warming might not even be happening! It’s all over, this is denialism’s wet dream on a plate (ugh)! Right? Well, no. Dr. Jones said, truthfully, that warming over the last fifteen years falls just short of being statistically significant. Ms. Wente is over the moon about this: “Hello? When other people say that, they’re called deniers”, she proclaims. Not so. This is a trivial result, a product of the short period and the wide year-to-year variation in global temperatures. The warming over this period, as measured by CRU, falls short of being statistically significant.

This means that the possibility of the observed warming being due to chance variation in year to year temperatures is not ruled out with 95% confidence. The rate of warming, at about .12 ^oC per decade, is slow enough that a chance pattern of year-to-year variation over the last fifteen years that just happened to finish as much warmer as that period did has a probability higher than 5% (in fact, as Dr. Jones pointed out in his interview, this probability is just barely over 5%).

This 5% probability is the standard threshold for ‘statistical significance’—but it’s not a magical line handed down from on high; it’s just a choice about how much evidence we standardly require before rejecting the hypothesis that an observed correlation is due to mere chance. If we suspect there’s a real effect (and we certainly do in this case) then we go looking for more data to check. The larger the sample size, the better your chance of significant results. And, lo and behold, over the last twenty years observed warming in the CRU data is statistically significant! (It’s also worth noting that we are now in a deep solar minimum, but 2009 was still the second warmest year on record—it’s funny how facts about solar output are important to deniers when they might help to explain observed warming, and completely neglected when they don’t.)

Also, if we use the GISS temperature measurements instead (which include wider coverage of the arctic based on correlations between measured temperatures over fairly large regions), the increase over the last fifteen years is statistically significant—that is, the probability of the observed increase being due to chance turns out to be a bit less than 5%. No doubt, this difference between the two main measures of global temperature will soon be a new topic of intense discussion in the denialsphere: how could our two best measures of earth’s temperature disagree on this basic question?

So if all the evidence we had for global warming was the warming over the last fifteen years in CRU data, we would be uncertain (by the arbitrary standard that rules out chance only when the probability of an observation being due to chance is less than 1 in 20). But the balance of probability would strongly favour warming. The odds of the observed increase being due to chance variation are less than 1 in 10, but not less than 1 in 20. Of course we have much more evidence than that, and the odds that the observed warming is due to chance when we look at more data and longer periods are far less than 1 in 20. Finally, on longer scales the warming grows more and more significant, that is, our confidence that it is not a matter of chance becomes higher and higher.

Ms. Wente’s excitement about Dr. Jones’s statement reveals nothing about climate science, but a lot about her ignorance of statistics and probability. Would she also be excited to know that if we consider only a few days in February and a few days in May, the temperature difference between the two months can easily fail to be statistically significant?

Along the way Ms. Wente once more tosses out a brief concession, admitting that anthropogenic global warming is settled. The ‘third position’ she now identifies herself with accepts “the underlying science, which says that climate change is happening and human activity is a factor”. Surely this implies that she does admit the science is settled? Of course not! Having conceded the conclusion she has resisted for so long, she continues to dismiss and belittle the problem we face. The science is still not settled for her, because her ‘third position” (endorsed, she claims, by unnamed experts) says that “threats of imminent catastrophe have been wildly exaggerated,” and “(w)e don’t know much about what might happen in the future, especially when it comes to specifics such as rising sea levels or regional droughts.”

There are indeed still questions about the details of just how much warming will result, and just how bad the consequences will be—and no serious defender of the science of global warming, from Al Gore to David Suzuki to James Hansen, Michael Mann and Phil Jones himself has denied this. But, pace Ms. Wente, this doesn’t mean we don’t know a lot about what might happen: the paleoclimatological record shows that the last time temperatures were just a few degrees warmer than they now are, sea level was meters higher. How quickly that rise in sea level occurred is not known, but nothing we know rules out a rapid rise in the near future: sea levels fall slowly as the earth cools, but they rise much faster during warming, and the resolution of our data doesn’t show a delay. Climate shifts can be dramatic in other ways too, altering rainfall patterns, disrupting life cycles for plants and animals and increasing destructive large-scale weather events. Both models and the past climatological record show that the climate can be very sensitive, and any major changes will be difficult if not impossible for many people and countries to cope with. Further, though the worst consequences may be delayed, our failure to act now can lock in drastic changes later: our descendants will suffer from our reluctance to make changes in our energy systems.

So Ms. Wente is pulling another fast one here—she is moving the goalposts. It wasn’t long ago that she disputed that anthropogenic global warming was happening, eagerly joining in the attack on Michael Mann’s ‘hockey stick’ reconstruction of past temperatures and questioning whether present warming is in any way exceptional. Now she accepts that it’s happening, and that it’s at least partly due to our greenhouse gas emissions. But she demands (before she’ll admit the science is ‘settled’) that we know just how much warming will happen, how soon, and what the consequences will be. But this is ridiculous: there will always be some level of uncertainty about these things: Ms. Wente has picked a reading of what it takes to ‘settle the science’ that implies the science can never be settled. This is a nice rhetorical trick, but pretty silly when you think about it.

Part of moving the goalposts is forgetting the past. Several years ago, Ms. Wente was similarly excited over the ‘Hockey stick’ debate, a short-term obsession of the denialsphere triggered by the prominence of Michael Mann’s reconstruction of past temperatures over the last 1,000 years. Ms. Wente gleefully announced that MacIntyre and McKitrick had shown Mann’s work was wrong and that current warming is not exceptional. But subsequent work by multiple climate scientists and a detailed review by the American Academy of Science have confirmed Mann’s pioneering work and extended it using different proxies and more refined statistical methods. (Worse for Ms. Wente, the Republican-led House ‘investigation’ of Mann’s work that led to the Wegman report attacking Mann’s work has been exposed as a set-up job, manipulated by Representative Joe Barton and others: see DeSmog and Deepclimate.) But it seems Ms. Wente has forgotten all about this episode. Past manipulations and the failures of denialist arguments simply disappear from her story, as she spins the latest ‘defeat’ for global warming science while pronouncing her own views to be moderate, sensible, balanced and wise. Isn’t it funny that her views also keep turning out to be wrong?