Would It Be a Bad Thing If I Made the Econ 2 Freshmen Learn Python?
Noted for Your Morning Procrastination for September 27, 2013

Richard Muller and the New York Times on Estimating the State of the Climate: Department of Whiskey-Tango-Foxtrot-Bang-Query-Bang-Query Weblogging

I know Richard Muller is trying as hard as I am to be helpful and informative, but this made me wince a bit:

Fig A gif 656×446 pixels 2

Richard Muller draws a straight line from the 1998 NASA GISS temperature datapoint through the 2012 datapoint and writes:

Richard Muller: A Pause, Not an End, to Warming:

For all of its warnings, and despite a steady escalation of greenhouse gas emissions into the atmosphere, the planet’s average surface temperature has remained pretty much the same for the last 15 years…

But that is not how you draw trends through noisy data that is estimated with error and uncertainty.

What NASA GISS does--which is a very reasonable smoothing procedure--is to estimate the state of the climate by taking a five-year moving average of its annual observations. Thus its estimate of the temperature state of the climate anomaly as of 1988 is not the noisy 0.82C measurement, but instead the 0.69C centered five-year moving average. And the forecast of its estimate of the state of the climate in 2012 is not the noisy 0.77C measurement, but rather the 0.83C average of what the 2010, 2011, and 2012 measurements are as a forecast of what the estimate for 2012 will turn out to be when we can calculate it in 2015.

Whether you think 0.69C is "about the same" as 0.82C can be seen as a matter of taste, I suppose. But when Muller writes:

I didn’t actually predict a pause in the warming but a possible period of cooling. But that’s close enough. We are now in that pause…

I think he gets it somewhat wrong. What he ought to say is not that global warming has paused, but rather that we expect one year in about every 50 to show us an annual observation as far as 0.2C above the actual climate trend at that date--and that in the aftermath of such a high outlier we should expect to see annual observations fail to reliably break that high-water mark for a couple of decades.

And, of course, for still better estimates we would like to lean heavily on our intuition that the true state of the climate is something that doesn't jiggle up or down but evolves smoothly, and fit splines…