March 11, 2013

# Suppressio variation, suggestio falsi

The global mean land-surface temperature reconstruction from the Berkeley Earth Surface Temperature project (other reconstructions are very similar), looks like this:

I’ve scaled the axes using Bill Cleveland’s method of “banking to 45 degrees“, that is, so the median of the slope is 45 degrees.  Based on his research, this seems to give close to optimal perception of patterns.

This is a graph of temperature anomaly, that is, departure from an average.  The use of temperature anomaly makes sense, because there isn’t an intrinsic reference value for global mean land temperature.

If you knew almost (but not quite) nothing about graphs, you might think that it was important to add back the estimated average and include 0 celsius on the y-axis — after all, zero should be zero, as tobacco-company shill Darrell Huff taught us. Of course, this ignores the fact that for this purpose zero Celsius is just another number, like zero Fahrenheit. Temperatures below zero Celsius do exist, as I can testify personally.  The graph would then look like this

Now, you might want to focus on the important years of the graph, like, say, 1997-2012. I’m sure no true New Zealander would be cynical enough to ask why 1997 was the appropriate starting point — after all, it was the year the first successfully cloned sheep was announced. The graph looks like this:

That’s basically what the Telegraph, to its shame, printed on Saturday (they used sea and land temperatures, which I didn’t have as easily available, and which added another 5 degrees to the y-axis).

Of course, they missed a trick with the y-axis. As every scientist knows, if you really want an absolute temperature scale, you start at zero. Absolute zero.

There’s a graph that would make Exxon and Shell happy.

For the rest of us, as the legal maxim almost has it:  to suppress the variation (deliberately or otherwise) is to suggest a falsehood.

Update: from Twitter this morning it appears that Christopher Monckton is visiting NZ next month to tell us that there isn’t any climate change. He certainly picked a good year for it. For those who haven’t encountered him: Monckton is also famous for claiming to be a member of the House of Lords, and pretending to be the Burma delegate to the Doha climate change talks.

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. See all posts by Thomas Lumley »