How Hot Does It Have to Be to Break a Record?

So who hasn’t seen one of the bajillion recent stories saying 2014 is going to set the instrumental record for the highest average global surface temperature? May we throw a teense of cold water on that hot news?

Annual temperatures are calculated by averaging up monthly readings, so the last data point that we have is October. The National Climatic Data Center, a part of the Department of Commerce, estimates that global average temperature was a record high of 58.46°F. The previous record was 58.45°.

The key word is “estimates.” When a scientist measures something—with a ruler, a scale, or a thermometer, for example—there’s always a measurement error owing to properties of the measuring device or even the skill of the scientist. When it comes to global temperature, scientists are averaging data from over a thousand thermometers scattered about the planet. Some are well-taken care of, and some are not. Some may have traces of urban warming in them. Nor is the number of readings exactly the same from year to year, or even from month to month.

The result is that there is a central estimate (58.46°) and a 95% confidence range as to where the “true” value lies. 

The most recent and most transparent error analysis of global temperatures has been done by a group called Berkeley Earth. For October, they find that the 95% confidence range is 0.10°F, or +/- 0.05°.

So, using the normal rules of science, is 58.46° then distinguishable from 58.45°? In a word, “NO.”