Most of us have seen graphs of global temperature anomaly, like this one using data from NASA:
The temperature increase is, not to put too fine a point on it, obvious.
If we were climate deniers, how could we graph some temperature data that would “hide the incline”? Let’s start by graphing data, not for the whole world, but for a very small piece of it. After all, the variability of temperature in small regions is greater than the variability of global temperature so this will increase the amount of fluctuation and make the trend far less obvious — even if it’s still there.
So let’s pick a record that is well-known and highly reputable: CET, the central England temperature record. It even covers a longer time span than the global estimates! Here’s yearly average temperature (not temperature anomaly) from the CET record:
We’re part way there; the incline is less obvious, but it is still plain to see. In fact the total rise is a bit greater than the global total rise, so if we want to “hide the incline” we’ll have to do better.
How about using less time? To prevent people from thinking we’re cheating (even though that’s exactly what we’re doing), let’s compute seasonal averages rather than annual. The variability in seasonal averages is greater than annual averages so we’ll pump up the fluctuation level even more! But we’ll still have a problem, because for the CET data the autumn season shows even more warming than the annual averages, so despite the greater variability the rise for autumn will still be a bit too obvious:
What next? Let’s stretch out the time axis to make the trend seem more flat. We can do that by graphing, not all the data, but just the last century. I’ll do it like this:
A little better — but the rise is still just too big to be ignored. So let’s stretch things again, or rather shrink things. Instead of plotting the autumn averages from roughly 8 to 13 °C let’s make the y-axis go from about -2 to +18 °C, a range of fully 20 °C, in spite of the fact that the data only cover a range of 5.13 °C.
Now we’re getting somewhere! The rise is still there, of course, but it sure doesn’t look like much.
But how will we justify that axis change? Here’s a great way: plot data for all four seasons on the same graph so it will be crowded and the individual seasons will obscure each others’ detail. That’s why we wanted raw temperature rather than anomaly — so when we put different seasons on the same graph we’ll have to use a y-axis covering 20 °C. We end up with this:
Compare that to the graph of yearly average CET. The fact that CET has risen more than 1 °C no longer stands out, so we’ve managed to “hide the incline” in plain sight. “Mission accomplished.”
If you think this whole approach is just too misleading to be honest, too ridiculous to be used even by the most staunch and/or deluded climate denier, think again. It’s exacty what was done in this post at the WUWT blog.
It seems to be part of Anthony Watts’ attempt to discredit using anomalies in general, in order to discredit a great deal of climate data (which so often uses anomalies). But you don’t really have to use anomalies anyway — look at the graph of annual (rather than seasonal) average CET, which doesn’t use anomalies at all. Perhaps we shouldn’t be too surprised that he so dislikes anomalies; it seems to me that there’s plenty of other evidence he has a serious problem understanding how anomalies really work.
This blog is made possible by readers like you; join others by donating at My Wee Dragon.