Re: Cut the sh*t!
We can use any reasonable bucket of how that is measured.
This is the key problem. There ain't one. The biggest problem with temperature data sets is that a significant number of ground stations which were originally located in the countryside upon establishment have ended with the grounds of the neigbouring metropolis over time.
As a result the time series are:
1. Not continuous while presented as such.
2. The set within Eu/Western Russia and USA which is representative is actually quite small and covers only a relatively recent period of time when having automated stations with radio connectivity to the central office became technically feasible.
3. I have yet to see a single study which instead of sucking numbers out of thin air (Berkley sets inclusive) uses strictly _ONLY_ stations which are outiside urban areas by more than 100km+. There are such stations (mostly coastal observations from stations associated with lighthouses), however there is no analysis which uses just them. Everyone finds it "essential" to stick into the equations Eu and USA data which is corrupted by _LOCAL_ industiral/urban heat.
Note - I am not saying anything about CO2, models, etc. I just want to see statistics done on the only 100% clean dataset which is readily available - just stick solely to observations taken at maritime navigational facilities (lighthouses, etc) and throw out all the ones within 100km of a large city. I have yet to see any and I could not care less about the "scientific value of the guesstimate used in the correction factors for urban area data".