It is generally accepted that ice cores archive amount-weighted water stable isotope signals. In order to achieve an improved understanding of the nature of water stable isotope signals stored in ice cores annual δ18O and δ2H averages (i.e. amount-weighted) were calculated for two Antarctic meteorological stations, Vernadsky and Hal-ley Bay, using monthly precipitation amount and monthly net accumulation as weights, respectively. These were then compared with the annual mean δ18O δ2H and records of the nearest available ice cores. In addition, at the stations, both arithmetic means (i.e. time-weighted) and amount-weighted (precipitation amount and net accumulation used as weights) annual air temperature averages were calculated and then compared to amount weighted annual mean δ18O and δ2H using correlation- and regression analyses. The main hypothesis was that amount weighted annual mean water isotope and temperature records from the stations would be able to replicate the annual water isotope signal stored in ice cores to a higher degree. Results showed that (i) amount weighting is incapable of ameliorating the signal replication between the stations and the ice cores, while arithmetic means gave the stronger linear relationships; (ii) post depositional processes may have a more determining effect on the isotopic composition of the firn than expected; and (iii) mean annual air temperature provided the closest match to ice core derived annual water isotope records. This latter conveys a similar message to that of recent findings, in as much as ambient temperature, via equilibrium isotope fractionation, is imprinted into the uppermost snow layer by vapor exchange even between precipitation events. Together, these observations imply that ice core stable water isotope records can be a more continuous archive of near-surface temperature changes than hitherto believed.