The most frequent false claim about the data is that there has been no warming since 1998. For a long time I’ve wanted to demonstrate explicitly the cherry-picking behind this claim. Now here it is (taking advantage of some free plotting software I recently downloaded, and generating an animated gif at a free online site).
Here is an example of the false claim, from Australia’s former Senator Fielding, using the analysis of Bob Carter, well-known denialist. It is purported to demonstrate that there has been no warming even though the carbon dioxide content of the atmosphere has been steadily rising.
The animated graph above shows a sequence of data segments, and for each segment a straight line that is the trend for that segment. The data are the so-called HadCrut mean annual temperature estimates, from the Hadley Centre of the UK Met Office in collaboration with the Climate Research Unit of the University of East Anglia. (Carter used monthly data above.)
Green: data window 1998-2011. The trend is nearly flat, which would mean there has been no warming.
Blue: 1996-2011. By extending the data window back only two more years you get a clear warming trend, because 1996 was unusually cool, just as 1998 was unusually warm.
Orange. Now let’s try cherry-picking the other way. 1992 was cool and 1998 was very warm, so the trend from 1992-1998 is very steep warming.
Teal. Finally, if you use the longer data set from 1970 to 2011 you can see a clear warming trend, without any analysis. You can see the other trends, based on shorter-term data, are misleading.
Climate scientists have stressed all along that to demonstrate global warming you have to look at longer-term trends, because there are short-term fluctuations that obscure the underlying trend. You need at least 5-10 years worth of data.
You might still argue, from the HadCrut data, that global warming seems to have paused. However these data do not include any allowance for polar regions, where measuring stations are scarce but warming is most rapid. NASA’s Goddard Institute for Space Studies does include an allowance for polar regions, and their data are included in the following graph (magenta), with the HadCruT data for comparison (blue). [If the animation isn’t working, try clicking on the graph.]
In the GISS data the warmest years are 2005 and 2010, not 1998, and there is a continuing upward trend. If there is an impression of slowing over the past decade, it is hardly any more pronounced that other “pauses” that you see, for example in 1978-1987, or 1987-1997. These “pauses” are the basis for the escalator of the previous post.
Finally, we know 1998 was an exceptional year, because the strongest El Niño condition for a century was operating, and El Niño tends to raise the air temperature for its own reasons. On the other hand for the past three years or so we’ve had persistent La Niña conditions, which tend to cool the air temperature. This does not mean the Earth is not holding extra heat, it just means not as much is being held in the air. The rest is presumably going into the oceans, which can hold enormous amounts of heat. If La Niña ends, the air temperature is likely to jump again.
Anyway the case for global warming does not rest just on these temperature data. Other direct measurements have now detected an imbalance between heat received at the Earth and heat radiated back into space. Other symptoms such as ice melting in the Arctic and mountain glaciers give clear evidence.
There is a great irony in the denialists’ choice of the Hadcrut data to claim warming has stopped, because the Climate Research Unit (the “CRU” part) is the very one that supposedly faked its data, according to the Climate-gate myth. (That’s the story, from emails illegally hacked from East Anglia University, that the scientists were trying to “hide the decline” and intimidate dissenters. The emails were totally misrepresented and this has to be one of the biggest beat-ups ever.) So not only do global-warming deniers choose the data set that suits them, it’s the data set they most vilify in a different context.