The surface temperature records are made up of lots of speculation and estimation. Urban warming effect in land records, for instance, is often corrected for by linear regression referenced to area population - pure speculation. But acceptable by those who do their PC peer review - but clearly flawed. Consider the variables: increased energy use in the urban area, air conditioning, cars, heating, elevators, lighting, heat absorbing materials, suburban workforce, higher employee efficiency, etc., etc.
I've noticed that local weather reports often have my nearby city 5 degrees F warmer than my surroundings. Apply that to the temperature records, and you might find global cooling!
Most "official" weather stations are located at airports or in the center of downtowns in major cities... surrounded by lots of concrete, i.e. major heat absorbers and retainers... as well as heat engines in airconditioners. Discount those from any "warming" scenarios and look at data from only rural sources and the findings are very different: any "warming" trend disappears.