Retaining Detail In Cenozoic CO2 Curves

Palaeo CO₂ concentration data has very variable density across the Cenozoic. Some events (such as the PETM) and intervals (such as glacial-interglacial cycles) are covered by high resolution datasets, whereas at other times the spacing between datapoints is much greater. Because of this variable data density, combining the datapoints into a line describing the evolution of palaeo CO₂ usually focusses on either a short interval or on the long term trend (by blurring short term detail). We present a new approach which uses basis splines to produce a Cenozoic CO₂ curve. The spline approach allows us to produce a curve which retains details in times where we have the requisite data density without introducing problematic artefacts at times with lower data density. The spline method makes it possible for the first time to produce a single curve which is sensible regardless of the timespan of interest.