Identifying the underpinnings of δ²H discrepancies between plant stem and soil water: extraction-induced methodological artifacts versus biological fractionation effects

Abstract. Recent studies have reported widespread presence of hydrogen isotope offset (HIO) between cryogenically-extracted plant stem and soil water, challenging the long-standing assumption that the isotopic composition of stem xylem water reliably represents that of its source water. Despite intensive researches on this topic over the past decade, it remains debated as to whether and/or to what extent HIO originates from extraction-related artifacts or from in situ isotope mixing/fractionation during water transport from soil to plants. Here, we used cryogenic vacuum distillation (CVD) to extract stem and soil water from eight species (trees, shrubs, and grasses) grown under two humidity regimes. We quantified species-specific HIO, tested its associations with ecophysiological and environmental variables, and conducted immersion-based rehydration experiments to assess CVD-induced biases. Across species, HIO ranged from −7.2‰ to 3.2‰: trees were consistently negative, whereas shrubs and grasses were near-zero to slightly positive. Rehydration experiments revealed CVD-induced δ²H biases in stem (−4.5‰) and soil water (−2.5‰). When these extraction-related biases in both stem and soil water were simultaneously corrected, species-level HIO (mean = 0.2‰) was no longer different from zero, and showed no significant correlations with ecophysiological or environmental variables. These results suggest that apparent HIO is largely driven by CVD-induced artifacts rather than ecophysiological/environmental processes that cause isotopic fractionation during water transport along the soil-xylem continuum. We conclude that simultaneously correcting CVD-induced biases in both stem and soil water is critical to avoid spurious HIO signals and to improve isotope-based estimation of plant water sources.