Charles Keeling’s Mauna Loa carbon dioxide record is a tour de force, one of the great works of science of the 20th century. It shows the inexorably rising levels of carbon dioxide in our atmosphere. But its iconic status today derives from hypotheses confirmed and ideas well understood, things that were anything but certain when Keeling launched the effort. His son does a nice job with a 50th anniversary story in this week’s Science (sub. req.):

At the outset, the decision to place the instrument at Mauna Loa was a gamble. Existing measurements suggested that atmospheric CO₂ concentrations varied widely depending on the place and time. Given this variability, could a meaningful record be recovered from an instrument parked in one location?

Roger Revelle had an alternative idea: campaigns involving ships and airplanes, repeated every decade or so, to look for trends. But Keeling won out, and it is worth thinking about how our understanding today might be different if Revelle’s approach had been used:

[A] CO₂ record degraded to include only one point every decade or two loses its convincing message. Variations from survey to survey may be instrumental artifacts, or the apparent trend may be a random fluctuation.