You either didn't understand my explanation or didn't retain it.
In it, I gave the example of Wellington, and I quote:
" Temps before the mid-1920s were recorded at Thorndon, near sea level, but then the recording station moved to Kelburn at 125 m above sea level. It’s pretty basic meteorology that temperatures fall as you move above sea level, so the two stations are not directly comparable
This sort of correction is commonplace, and not remotely controversial amongst meteorologists and climatologists who are trying to build long term records from disparate data series. Apart from station moves, changes of thermometer and alterations in the environment around the recording station can all have an affect on temperature data, and are commonly corrected for. "
In the example I gave for Wellington they relied on calculating the statistical correlations between the two sites. When you have a robust correlation set, you can calculate how much your error confidence levels need to change - if at all. There's proven statistical methods for that.
Now it's true I've had a cup of wine with my dinner and so there is a chance I might wake up in the morning and think of some obscure counterexample, but right now I cannot think of any field of scientific endeavour dealing with large and disparate data sets which does not practice data homogenisation and cleansing techniques to some extent.
It is incredibly commonplace with robust and proven statistical techniques to verify.
Your hangup on this point just shows complete and utter scientific illiteracy.
