Just as I thought; you have no quantitative analytic answer. You can't understand the example I gave. That's OK. I understand. Don't worry about it.
The basic reason for that example was to illustrate that the change in levels of a variable is all that is important in assessing change over time, and your example works to that very same end. I am satisfied that I proved my point using your example. QED.
You can always go to wikipedia and change their analysis.
If I want a handle on the base line temperature so I can measure change in my living conditions, I would put the thermometers in the areas I generally occupy. (I generally don't spend much time behind the refrigerator)
However if I were interested in the thermal energy rise in the house, then I must also include a representative sampling near all heat sources, or heat sinks (in the winter.) That's not important to a resident, but it is important for the average temperature of the earth
NASA also says the sea levels are rising even though the ocean cooled.
http://www.nasa.gov/home/hqnews/2006/sep/HQ_06318_Ocean_Cooling.html
One reason could be that glacial ice is raising the sea level. Also melting ice releases the heat of fusion which is 344 Joules/gram. There are .24 calories/Joule. That means each gram of ice will require 82 calories of heat from the surrounding water. That is significant cooling considering the fact that 1 calorie of energy will raise 1 gram of water by 1 deg. C.
You forgot to answer my question. You say that the spread in mean temperatures is at least 2 deg. C. The spread in several papers that I found is only 1 deg. C. Can you cite the sources you found that lead you to believe the temperature spread is "at least 2 deg. C"?
