Calculating uncertainty for the RICE ice core continuous flow analysis water isotope record

Abstract

We describe a systematic approach to the calibration and uncertainty estimation of a high-resolution continuous flow analysis (CFA) water isotope (δ2H, δ18O) record from the Roosevelt Island Climate Evolution (RICE) Antarctic ice core. Our method establishes robust uncertainty estimates for CFA δ2H and δ18O measurements, comparable to those reported for discrete sample δ2H and δ18O analysis. Data were calibrated using a time-weighted two-point linear calibration with two standards measured both before and after continuously melting 3 or 4 m of ice core. The error at each data point was calculated as the quadrature sum of three factors: Allan variance error, scatter over our averaging interval (error of the variance) and calibration error (error of the mean). Final mean total uncertainty for the entire record is δ2H = 0.74 ‰ and δ18O = 0.21 ‰. Uncertainties vary through the data set and were exacerbated by a range of factors, which typically could not be isolated due to the requirements of the multi-instrument CFA campaign. These factors likely occurred in combination and included ice quality, ice breaks, upstream equipment failure, contamination with drill fluid and leaks or valve degradation. We demonstrate that our methodology for documenting uncertainty was effective across periods of uneven system performance and delivered a significant achievement in the precision of high-resolution CFA water isotope measurements.