Other proposed proxy methods include analyzing the carbon isotope ratio of soil carbonate , B / Ca ratio of marine calcite , or the phases of evaporate minerals which are CO 2 dependent . This later method probably involves the fewest assumptions of them all but unfortunately can only be applied to very rare geological deposits ( so far just 3 in the last 50 million years ; Lowenstein and Demicco , 2006 ). It does , however , confirm the predictions from boron isotopes and alkenones that pCO 2 was much higher than modern in at least some portion of the Paleogene period .
v Cautionary remarks
From the above discussion it can be seen that all the proxy methods are somewhat experimental in their application , especially over millions of years . Various laboratories are working hard to develop them and better constrain the assumptions . This means that the methods themselves are changing as they evolve , often making old pCO 2 estimates non- comparable with newer ones because different assumptions are made and as new knowledge is acquired . All the methods produce ‘ noise ’ in the records which arises not just from analytical uncertainty but also from the secondary assumptions ( like undetected variations in temperature for example ). Potentially more worryingly , the proxies might involve systematic biases of one sort or another . Hence to understand the record in any given publication it is necessary to examine the assumptions carefully . There is no space to do that here - we shall just examine the pCO 2 reconstructions themselves .
The good news is that the three proxies listed all work under controlled experimental conditions and also work fairly well at reconstructing the glacial-interglacial cycles in pCO 2 seen in the ice core records . It so happens that for different reasons , all the methods work less well at higher pCO 2 . The stomatal method is the most problematic at levels of CO 2 higher than modern because the calibration data sets are small and variable and non-linear ( the effect seems to flatten out at high CO 2 ). Moreover the biological relations of fossil leaves to their modern counterparts become increasingly distant with geological time . For those reasons I will focus mainly on alkenone and boron estimates .
3 . The records i . Pearson and Palmer ( 2000 )