Fig . 1 . Cenozoic pCO 2 reconstruction of Pearson and Palmer ( 2000 ). Red line added at 400 ppm .
First up we will examine the 60 million year record of Pearson and Palmer ( 2000 ) which used boron isotopes on planktonic foraminifera from deep sea cores . Note that in this and all the other plots time runs from right to left . The youngest pCO 2 estimate that is higher than 400 ppm in this record is in the early Miocene about 23 Ma ( million years ago ). Unfortunately there is a data gap in the Oligocene and upper Eocene . All but one of the estimates for the period 40-60 Ma are higher than 400 ppm , many of them considerably so .
Ongoing research since 2000 has shown that this record suffers from two main problems , both of which make the quantitative estimates problematic . The first is that our understanding of the boron isotope fractionation between the two aqueous species ( boric acid and borate ) has been substantially revised ( Klochko et al ., 2006 ) which would change the calculations . In fact it is impossible to calculate pCO 2 from the raw isotope data if we use the modern fractionation factor ! The second is that the measurements were made using Negative Thermal Ionization Mass Spectrometry ( N-TIMS ) and it has become increasingly clear that this method produces data that is strongly offset from measurements made using Multicollector Inductively Coupled Plasma Mass Spectrometry ( MC-ICP- MS ). It so happens that these two serious biases work against each other so that if we were to apply a method offset ( for example based on Foster et al ., 2012 ), the general pattern and trends shown by the data can be calculated although the absolute values will be somewhat different . This may seem like a desperate argument to make from one of the authors of this old paper (!) but it is backed by the observation that using the same methods and assumptions we were able to produce reasonable and