Correcting for Biogenic Gas Matrix Effects on Laser-Based Pore Water-Vapor Stable Isotope Measurements

• Main Authors: Benjamin Gralher, Barbara Herbstritt, Markus Weiler, Leonard I. Wassenaar, Christine Stumpp
• Format: Article
• Language: English
• Published: Wiley 2018-03-01
• Series: Vadose Zone Journal
• Online Access: https://dl.sciencesocieties.org/publications/vzj/articles/17/1/170157

The isotopic composition (δH, δO) of pore water is an invaluable tracer for the minimally invasive study of subsurface water flow and transport processes. Here, we evaluated a method for pore water isotope analysis that combines laser-based isotope analyzers and water-vapor isotope equilibration using evaporation-proof metalized sample bags. We tested inflation atmospheres (dry air vs. pure N) and the impact of biogenic gas (CO, CH) accumulation for storage times of up to 4 wk. Samples were analyzed with a water isotope analyzer (Picarro L2120-) and a gas chromatograph. Air-inflated water vapor samples showed a greater range of gas matrix effects (δO: 9.63‰; δH: 21.7‰) than N–inflated samples (δO: 7.49‰; δH: 10.6‰) induced by nonuniform buildup of biogenic CO, starting immediately after sample preparation. However, only air-inflated samples could be reliably corrected using instrument-specific sensitivity factors that were empirically determined by interpretation of periodically repeated isotope measurements. Corrected water isotope data were confirmed by similarity with local precipitation and suction cup isotope data. Residual uncertainties were well below the natural variations of soil water isotope values and independent of storage time, thus allowing for consistently reliable interpretations of soil water isotope profiles. We conclude that, especially for pore water sampling that requires small sample volumes and/or long storage times, metalized sample bags should be used to prevent evaporation notwithstanding the enhanced buildup of biogenic gases. Further, if gas matrix effects cannot be excluded, air inflation is preferred over pure N, as only in that case can reliable postcorrections be performed by using internal data only.