Fundamental controls on triple oxygen-isotope ratios in Antarctic precipitation and ice cores

Stable isotope ratios of water (deltaD and delta 18O) in polar precipitation and ice cores have long been used to study past climate variations and the hydrological cycle. Recently-developed methods permit the precise measurement of delta17O and the 17O-excess, relative to the delta17O vs. delta 18O meteoric water line. The novel isotope parameter "17O excess" provides an additional tool for investigating the global hydrological cycle.;Early experimental and modeling studies showed that 17O excess in atmospheric water vapor is sensitive to relative humidity during evaporation from the ocean surface, and suggested that there was little fractionation during condensation. It was therefore expected that 17Oexcess in polar snow could be used as an indicator for humidity in the ocean source regions where polar moisture originates. Later work shows that the magnitude of 17Oexcess change between the last glacial period and the Holocene warm period, measured in Antarctic ice cores, increases from the Antarctic coast towards the interior, suggested significant fractionation during transport. Full interpretation of these conflicting results has been challenging, hindered in part by the labor-intensive nature of making 17Oexcess measurements and by the lack of an accepted standard for reporting 17O excess values.;This thesis provides a new, comprehensive assessment of the 17 Oexcess of Antarctic precipitation and ice core data. The contributions from this work also include improvements to 17O excess measurement techniques, using both isotope-ratio mass spectrometry and collaborative developments in laser spectroscopy, and a formal calibration of international water standards for 17Oexcess. It further addresses both the spatial and temporal variations observed in Antarctic 17Oexcess values, providing a coherent explanation for both.;New Antarctic 17Oexcess measurements from this work show that there is a strong negative spatial gradient of 17O excess in snowfall towards the interior of Antarctica, a similar spatial pattern to the glacial-interglacial change in 17Oexcess , and a smaller-amplitude seasonal cycle in West Antarctica than in the interior of East Antarctica. These measurements, when combined with earlier published work, provide the most complete view of the spatial distribution and temporal variability of 17Oexcess to date.;Model studies, using both an intermediate complexity isotope model (ICM) and an isotope-enabled general circulation model (GCM), have permitted a thorough investigation of the most relevant and important processes affecting 17Oexcess in Antarctica. The model simulations show that changes in source relative humidity have only a modest effect on 17 Oexcess in polar precipitation, and can not account for the full seasonal cycle amplitude, nor the large glacial-interglacial 17Oexcess changes observed in Antarctic ice cores. The spatial gradient of 17Oexcess in modern precipitation, along with the large amplitude seasonal cycle in East Antarctica and the greater magnitude of 17Oexcess change for interior sites between glacial and interglacial periods, can be explained by kinetic isotope fractionation during snow formation under supersaturated conditions. The model experiments further show that the influence of moisture recharge is important to the evolution of 17Oexcess in poleward-moving air masses. The seasonal presence of sea ice is also a significant factor affecting 17Oexcess. Greater sea ice concentration or extent reduces evaporative recharge and increases the spatial area over which kinetic fractionation processes are important; both these factors tend to lower 17Oexcess.