One of my biggest questions when trying to determine how much warming the globe has experienced due to the activities of man kind is what should be average? By default, 1951 to 1980 is considered "average" because it was selected as "average" and is used as the base line for determining the impact of the Greenhouse effect.
With the gold standard of global surface temperature reconstructions, using the 1950 to 1980 baseline everyone is familiar with.
This base line provides a better view of the internal oscillations between the hemispheres. The northern hemisphere with less ocean volume and more land area has less thermal mass so would be more sensitive to changes in forcing.
Another minor issue is that land means above sea level. As elevation increases the thermal mass of the air would decrease meaning the variation in temperature for a given unit of forcing would be greater at higher elevation than at sea level.
This may not be properly accounted for with the satellite measurements versus each other and the surface station data unless the weighting filters of each satellite temperature product perfectly match.
Working backwards from the more abundant and arguably more accurate data, the view of climate change, changes. This view should reduce some of the increasing uncertainty.