There is a new paper from G.J. van Oldenborgh et al. in Climate of the Past reporting that Western Europe has warmed much more from 1950 to 2007 than predicted by a 17 runs of the ECHAM5/MPI-OM climate model. The abstract tells us that “the warming trend of the last decades is now so strong that it is discernible in local temperature observations“. Hmmm! The observational data are from the HADCRUT3 database; the authors write that only mean global temperatures are taken from this dataset (if I understand this correctly), but they use their “own merged dataset” for the maps given in the paper (what is that dataset?). The essence of the paper is that the observed temperatures are higher than the predicted ones. One cause could be the reported warming of the eastern Atlantic, which is given as larger than what the models told. As West Europe’s temperature is under the big influence of the atmospheric circulation, the author write that changes in that circulation pattern may explain more than 50% of the variance in temperature observed (there is a circulation shift to more westerly winds). The AMO (Atlantic Multiannual Oscillation) has, according to the paper, only a very little effect on global mean ( = Western Europe mean?) temperature. I remain uneasy with many of these reportings, but the paper is well written and agreable to read. Where I jumped up, was a chapter on solar irradiance (shortwave radiation) increases from 1971 to 2007. Six Dutch stations are reported as showing an average increase of 14+/-2 Wm-2K-1 corresponding to about 0.7K of the observed global warming.
Here at meteoLCD my solar measurements since 1998 show a very different picture:
The annual or seasonal averages have been computed from **ALL** measurements ( i.e. two recordings per hour, each being the average of 30 1-minute-step measurements), not from monthly mean values as in the paper.
The yearly trend is negative, the spring trend nearly flat and the summer trend strong negative: just the opposite to the Dutch measurements. In a personal communication, G.J. Oldenborgh told me that the radiation data come from Wageningen University; they have been heavily “massaged”, i.e. corrected for circulation patterns. Without these corrections, no trend would have shown up. The meteoLCD data given in the graph are the raw data; the 11 year series obviously is to short to fully qualify for trending (should be at last 30 years, as G.J. Oldenborgh wrote in his very curtuous letter). I remain uneasy on the Dutch data handlings, that suggest a reality that in fact is only virtual. Changing wind circulations are a fact of life, and changing cloud patterns should be taken as is. Computing a hypothetical irradiance which would have existed in the abscence of the circulation patterns seems a dangerous procedure to look for real life trends!
The author conclude that the difference between the observed and predicted warming is very unlikely to be caused by decadal climate fluctuations; they remarkably refrain from citing anthropogenic emissions as the probable cause!