Compilation of stable isotope data from East Antarctic snow pits and firn cores

Stable water isotopes in polar ice provide a wealth of information about past climate evolution. Snow-pit studies allow us to relate observed weather and climate conditions to the measured isotope variations in the snow. They therefore offer the possibility to test our understanding of how isotope signals are formed and stored in firn and ice. As stable water isotopes in the snowfall are strongly correlated to air temperature, isotopes in the near-surface snow are thought to record the seasonal cycle at a given site. Accordingly, the number of seasonal cycles observed over a given depth should depend on the accumulation rate of snow. However, snow-pit studies from different accumulation conditions in East Antarctica reported similar isotopic variability and comparable apparent cycles in the d18 O and dD profiles with typical wavelengths of ~ 20cm. These observations are unexpected as the accumulation rates strongly differ between the sites, ranging from 20 to 80mm w.e. yr -1 (~ 6-21cm of snow per year). Various mechanism have been proposed to explain the isotopic variations individually at each site; however, none of these is consistent with the similarity of the different profiles independent of the local accumulation conditions. Here, we systematically analyse the properties and origins of isotopic variations in high-resolution firn profiles from eight East Antarctic sites. First, we confirm the suggested cycle length (mean distance between peaks) of ~ 20cm by counting the isotopic maxima. Spectral analysis further shows a strong similarity between the sites but indicates no dominant periodic features. Furthermore, the apparent cycle length increases with depth for most East Antarctic sites, which is inconsistent with burial and compression of a regular seasonal cycle. We show that these results can be explained by isotopic diffusion acting on a noise-dominated isotope signal. The firn diffusion length is rather stable across the Antarctic Plateau and thus leads to similar power spectral densities of the isotopic variations. This in turn implies a similar distance between isotopic maxima in the firn profiles. Our results explain a large set of observations discussed in the literature, providing a simple explanation for the interpretation of apparent cycles in shallow isotope records, without invoking complex mechanisms. Finally, the results underline previous suggestions that isotope signals in single ice cores from low-accumulation regions have a small signal-to-noise ratio and thus likely do not allow the reconstruction of interannual to decadal climate variations.