HOW DO YOU KNOW?
Ask these questions of the science (for the scientists will not answer you):
How do you know? What instruments have you used to determine temperatures? What is their precision, their accuracy? (There is a difference.) Where and when have temperature readings been taken? What attention has been paid to the siting of recording instruments? What biasing affect has that siting had on readings? What length of timeline do you have? How accurate are the records (how diligent the record keeping, gathering, and collating)? What methods have been used to determine trends? What steps have been taken to deal with outliers — apparently aberrant data points, missing blocks of data, sudden changes in trends?
Follow the answers to those questions until the threads unravel in your hands. Do not stop until you get reliable and accurate answers.
Now. From your “primary school science,” recall what you know — or should know — about temperature and the atmosphere?
Ask yourself these questions:
In the day-night cycle, when do the highest and lowest temperatures occur? What bearing does a daily average have on those data points? What meaning can be derived from an average?
Over a metropolitan area, what is the delta (the value of change) between high and low minimums and maximums on a daily basis? How does this vary seasonally?
For example: the temperature at 10AM at the international airport is X. What would you expect the temperature to be in the city center? In a residential suburban neighborhood? On a college campus? In a rural setting — say a wheatfield? At a sewage treatment plant? Repeat for 5PM, 10PM, and 5AM. What variation do you see? Do the various temperature trends (per location) change at different rates? What do you think might be the probable cause?
How far apart would these various sites be?
Compare the resolution (expressed as, say, the number of datapoints per square mile) of your thought experiment to the actual recording network. Then ask yourself:
Is the resolution of the “network” you used for your local temperatures sufficient to give you a clear and accurate picture of “the” temperature of your town or city? How does it compare in density and coverage with the real-world climatological network?
Over the course of a year, what meaning could you derive from daily “averages” of your temperature observing and recording network?
Does this value scale to a global network?
What does a global average temperature mean?
What does a change of less than a degree (any scale) in global average temperature mean?