Pekka
This kind of errors are typical for those, whose method of study is the search of signals in data or data mining, but who don’t have full understanding of the caveats of those methods (and of course also those who do understand, but want to mislead others).
That is not fair.
I am trying to correct IPCC’s mistake in the following graph.
IPCC compared trends that have different periods 150, 100, 50 & 25 years. In contrast I used a constant 30-years period (1850-1880, 1851-1881, …, and 1981-2011)
IPCC arbitrarily selected 2005 as the end point of all the trends. In contrast, I used every year starting from 1880 as end point.
IPCC used only four trend periods to make its “accelerated warming” conclusion. I used 132 trend periods.
Here is what I got => http://bit.ly/yni1Ug