Atmospheric water

In the context of a globally warming climate it is crucial to study the climate variability in the past and to understand the underlying mechanisms (IPCC 2007). Precipitation deposited on the polar ice caps provides a means to retrieve information on temperature changes and atmospheric composition on time scales from one to almost one million years, with sub-annual resolution in the most recent centuries.

It is now generally accepted that the water oxygen and hydrogen isotope signals (d¹⁸O and d²H) in ice cores act as proxies of paleo-temperatures, and numerous ice-cores, both Arctic (a.o., DYE3, GISP, (North)GRIP) and Antarctic (such as the Vostok and EPICA cores), have been drilled and analyzed for d¹⁸O and/or d²H in the last decades. For a correct interpretation of the resulting datasets, the temporal relationship between the isotope signal and the (local, paleo) temperature needs to be established over the entire time scale spanned by the age of the ice. However, this calibration of the paleo-thermometer remains problematic. The spatial relation between the isotopic ratios d¹⁸O and d²H of precipitation and local temperature was recognized in the early sixties (Dansgaard 1964). The transfer of the present day spatial relationship between isotope signal and temperature need not, however, (and in fact does not) correspond to the temporal isotope – temperature relation in history, especially not during glacial periods, with their quite different climatic conditions. It is now widely recognized that the isotope – temperature relations vary in time, as well as in space (see, e.g., IAEA 1992, Cuffey et al. 1994, Van Lipzig et al. 2002, Masson-Delmotte et al. 2008). For this reason attempts are ongoing to provide a more physical basis of the isotope – temperature relation.

To address this principal problem with the interpretation of ice core isotope signals in terms of paleoclimate, there are two different approaches. The first looks for independent archived temperature signals that can be used for the calibration of the ice core isotope signal over time. In Greenland, inversion of borehole temperature profiles, as well as a method based on the diffusion of air during abrupt climate changes, has effectively challenged the use of the present-day spatial isotope–temperature relation. Neither of these two methods can be applied to Antarctic deep ice cores, because of the low accumulation rate and slower climate changes, respectively.

The second approach tries to model the global water cycle to the extent that the isotope signals in precipitation can be adequately reconstructed. Our work (the PhD projects of Janek Landsberg and Mathieu Casado) will contribute to this approach by measuring the isotopic composition of moisture carried towards and deposited on Antarctica, thus constraining the numerical models.