Richard Lindzen, the Alfred P. Sloan Professor of Meteorology at M.I.T. writes:
The notion of a static, unchanging climate is foreign to the history of the earth or any other planet with a fluid envelope. The fact that the developed world went into hysterics over changes in global mean temperature anomaly of a few tenths of a degree will astound future generations. Such hysteria simply represents the scientific illiteracy of much of the public, the susceptibility of the public to the substitution of repetition for truth, and the exploitation of these weaknesses by politicians, environmental promoters, and, after 20 years of media drum beating, many others as well. Climate is always changing. We have had ice ages and warmer periods when alligators were found in Spitzbergen. Ice ages have occurred in a hundred thousand year cycle for the last 700 thousand years, and there have been previous periods that appear to have been warmer than the present despite CO2 levels being lower than they are now. More recently, we have had the medieval warm period and the little ice age. During the latter, alpine glaciers advanced to the chagrin of overrun villages. Since the beginning of the 19th Century these glaciers have been retreating. Frankly, we don’t fully understand either the advance or the retreat.
For small changes in climate associated with tenths of a degree, there is no need for any external cause. The earth is never exactly in equilibrium. The motions of the massive oceans where heat is moved between deep layers and the surface provides variability on time scales from years to centuries. Recent work (Tsonis et al, 2007), suggests that this variability is enough to account for all climate change since the 19th Century.
For warming since 1979, there is a further problem. The dominant role of cumulus convection in the tropics requires that temperature approximately follow what is called a moist adiabatic profile. This requires that warming in the tropical upper troposphere be 2-3 times greater than at the surface. Indeed, all models do show this, but the data doesn’t and this means that something is wrong with the data. It is well known that above about 2 km altitude, the tropical temperatures are pretty homogeneous in the horizontal so that sampling is not a problem. Below two km (roughly the height of what is referred to as the trade wind inversion), there is much more horizontal variability, and, therefore, there is a profound sampling problem. Under the circumstances, it is reasonable to conclude that the problem resides in the surface data, and that the actual trend at the surface is about 60% too large. Even the claimed trend is larger than what models would have projected but for the inclusion of an arbitrary fudge factor due to aerosol cooling. The discrepancy was reported by Lindzen (2007) and by Douglass et al (2007). Inevitably in climate science, when data conflicts with models, a small coterie of scientists can be counted upon to modify the data. Thus, Santer, et al (2008), argue that stretching uncertainties in observations and models might marginally eliminate the inconsistency. That the data should always need correcting to agree with models is totally implausible and indicative of a certain corruption within the climate science community.
The rest here.