Speak loudly and carry a busted hockey stick
The so-called settled science of predicting future temperatures is rooted in absurd generalisations, unrepresentative sampling, trick graphs and statistical sleights of hand. The louder they shout, the more apparent their bogus discipline's flaws follies.
The average temperature for the Earth, or any region or even any specific place is very difficult to determine with any accuracy. At any given time surface air temperatures around the world range over about 100°C. Even in the same place they can vary by nearly that much seasonally and as much as 30°C or more in a day. Weather stations are relatively few and located very irregularly. Well maintained stations with good records going back a century or more can be counted on one’s fingers. Even then only maximum and minimum temperatures or ones at a few particular times of day are usually available. Maintenance, siting, and surrounding land use also all have influences on the temperatures recorded.
The purported 0.7°C of average global warming over the past century is highly uncertain. It is in fact less than the margin of error in our ability to determine the average temperature anywhere, much less globally. What portion of any such warming might be due to due to anthropogenic CO2 emissions is even less certain. There are, however, numerous phenomena which are affected by temperature and which can provide good evidence of relative warming or cooling and, in some cases, even actual temperatures. These include growth rings in trees, corals and stalactites, borehole temperature profiles and the isotopic and biologic signatures in core samples from sediments or glaciers. In addition, historical accounts of crops grown, harvest times, freezes, sea ice, river levels, glacial advances or retreats and other such records provide clear indication of warming and cooling.