Not exact matches
The standard deviation
of the residuals from a
linear regression to annual averages 1975 - 2007 is 0.472, so we expect a
range of variation
of roughly + / - 0.94 deg.C from the long - term
trend.
12 - month running averages are shown as well as
linear trend lines, and compared to the scenarios
of the IPCC (blue
range and lines from the 2001 report, green from the 2007 report).
In our 2007 paper we found
trends in the upper part
of the TAR
range (both GISS and Hadley give 0.22 ºC per decade
linear trends for 1990 - 2006) and proposed that «the first candidate reason for this is intrinsic variability in the climate system».
Without a validated model there is no justification for fitting a
linear trend and extrapolation way outside the
range of the data.
Figure 2.2 a presents both - LOD and dT
ranges, with the
linear trends removed by fitting
of linear regression and using a «detrending» procedure (Statgraphics 1988).
But I do know the difference between a simple
linear interpolation and principal component analysis, and I can calculate the two standard deviations
range of uncertainty on a white noise
linear trend.