It appears to me that such a minor change might be well within the normal measuring error. Too many variables to control to get a repeatable reading?
What do you think?
Perhaps. I'm not sure how they measured this but I suspect methods superior to those in the earlier 20th century. Thus if we can't trust these new methods for a .25 degree change, we probably can't trust older methods calibrated against newer ones with a .5 degree change. This would leave us uncertain if the planet has warmed this century at all.
>> It appears to me that such a minor change might be well within the normal measuring error. Too many variables to control to get a repeatable reading? What do you think? <<
Yes and no: You can see if the temperature has gone up or down, even if you can’t calibrate very precisely. And the goal of the NCDC is to get enough readings to statistically smooth over any chance invalidity. But such statistical smoothing presumes that the invalidities have no specific trend. It certainly has been proven that the data collection over the US mainland has been horribly invalid, with a known upward trend. The thought is that ocean temperature readings may be somewhat better, but that problem still exists.