Natural Variability, Not CO2, Accounts for Late 20th Century Warming
DelSole, T., Tippett, M.K., Shukla, J. 2010. A significant component of unforced multidecadal variability in the recent acceleration of global warming. Journal of Climate doi: 10.1175/2010JCLI3659.1.
Specifically, a set of climate models run in "control" or unforced mode were used to develop a 300 year dataset of spatial ocean temperature data. It was found that an internal pattern, detectable using a spatial fingerprinting technique, could be identified in the simulated data. This spatial pattern of ocean temperature anomalies was labeled the Internal Multidecadal Pattern (IMP). It was found that this pattern is highly coherent with the Atlantic Multidecadal Oscillation (AMO) historical patterns and predicted the Pacific Decadal Oscillation (PDO), suggesting that the models were able to match the internal dynamics of the real Earth system.
Next, the authors extracted, also with discriminant fingerprinting, the forced component of the spatial patterns produced in the absence of the IMP as an orthogonal function, which they demonstrated has only a minor effect (less than 1/7 amplitude) on the IMP. They then used historical sea surface temperature data to evaluate the relative importance of the forced vs. IMP components of change from 1850.
In considering the latter portion of the record (1946-2008), results indicated that the internal variability component of climate change (the IMP) operated in a cooling mode between 1946 and 1977, but switched to a warming mode thereafter (between 1977 and 2008), suggesting that the IMP is strong enough to overwhelm any anthropogenic signal. Of this the authors state: "Specifically, the trend due to only the forced component is statistically the same in the two 32-year periods and in the 63-year period. That is, the forced part is not accelerating. Taken together, these results imply that the observed trend differs between the periods 1946-1977 and 1977-2008 not because the forced response accelerated, but because internal variability lead to relative cooling in the earlier period and relative warming in the later period" [italics added].
With respect to the entire record, the authors state that the 150 year-long trend of temperature is not explained by the IMP. In their Figure 4, it is seen that the forced component spatial fingerprint began to deviate from no trend sometime after 1920. But, this type of analysis does not distinguish between types of forcing (e.g., solar vs. anthropogenic). Nevertheless, the results in this paper suggest that simple extrapolations of rates of warming from 1980 onward overestimate the forced component of warming. Using this period without factoring out internal variability will likely lead to unrealistic values of climate sensitivity.