Improving continuous-flow analysis of triple oxygen isotopes in ice cores: insights from replicate measurements

Abstract

Stable water isotope measurements from polar ice cores provide high-resolution information about past hydrologic conditions and are therefore important for understanding earth's climate system. Routine high-resolution measurements of δ18O, δD, and deuterium excess are made by continuous-flow analysis (CFA) methods that include laser spectrometers. Cavity ring-down laser spectroscopy (CRDS) allows for simultaneous measurements of all stable water isotopes, including δ17O and 17O excess (Δ17O); however, the limitations of CFA methodologies for Δ17O are not well understood. Here, we describe a measurement methodology for all stable water isotopes that uses a CFA system coupled with a CRDS instrument. We make repeated measurements of an ice-core section using this method to explore the reproducibility of CFA–CRDS measurements for Δ17O. Our data demonstrate that the CFA–CRDS method can make high-precision measurements of Δ17O (< 5 per meg at averaging times > 3000 s). We show that the variations within our CFA ice-core measurements are well matched in magnitude and timing by the variations within the discrete CRDS measurements; we find that calibration offsets generate most of the variability among the replicate datasets. When these offsets are accounted for, the precision of CFA–CRDS ice-core data for Δ17O is as good as the precision of Δ17O for continuous reference water measurements. We demonstrate that this method can detect seasonal variability in Δ17O in Greenland ice, and our work suggests that the measurement resolution of CFA–CRDS is largely defined by the melt and measurement rate. We suggest that CFA–CRDS has the potential to increase measurement resolution of δ17O and Δ17O in ice cores, but also highlight the importance of developing calibration strategies with attention to Δ17O.