Introduction

Understanding plant rooting depth and the sources of plant water is critical for effective water resource management (Fan et al., 2017; Miguez-Macho & Fan, 2021; Penna et al., 2018). For over four decades, scientists have used stable isotope analyses of plant water and potential source water (such as soil water from different depths and groundwater) to infer rooting depth and characterise plant water sources (e.g. Ehleringer & Dawson, 1992; Goldsmith et al., 2012; Thorburn et al., 1993; White et al., 1985; Zencich et al., 2002). Central to this isotope-enabled approach is the extraction of water from plant stems.
One widely adopted technique for this purpose is cryogenic vacuum distillation (CVD). In CVD, water is extracted through sublimation under vacuum conditions. While relatively easy to implement in the laboratory, a limitation of CVD is that it provides a bulk measurement of total stem water (hereafter ‘bulk stem water’), rather than a more targeted measurement of the water that transits from root uptake through xylem conduits and contributes to transpiration (hereafter ‘xylem water’). Less destructive and more selective in-situ alternatives have recently emerged (Kübert et al., 2023; Kühnhammer et al., 2022; Marshall et al., 2020; Volkmann et al., 2016), but CVD remains the most commonly used method among the scientific community (de la Casa et al., 2022; Millar et al., 2022; Millar et al., 2018).
Recent studies have raised concerns about the use of CVD for plant water sourcing investigations (Barbeta et al., 2022; Chen et al., 2020; Wen et al., 2022). These concerns have resulted from increasing evidence of systematic deuterium offsets in plant water, reaching up to \($-40\textperthousand$\) relative to source water (e.g. Barbeta et al., 2019; de la Casa et al., 2022; Duvert et al., 2022; Ellsworth & Williams, 2007; Lin & Sternberg, 1993; Poca et al., 2019; Tetzlaff et al., 2021). The underlying reasons for these CVD-induced deuterium offsets are now becoming clearer, and may be related to a combination of (1) hydrogen exchange with organic tissues during the CVD extraction process (Chen et al., 2020; Diao et al., 2022; Wen et al., 2022) and (2) isotopic heterogeneity between xylem water and the water stored in non-conductive tissues, as CVD extracts both water pools (Barbeta et al., 2022; Bowers & Williams, 2022; Wen et al., 2022).
Much of the evidence regarding these CVD-induced biases has been derived from controlled laboratory and potted experiments, and few studies have examined isotopic biases under field conditions. Several authors have conducted rehydration experiments with labelled water (Chen et al., 2020; Diao et al., 2022; Wen et al., 2022), but such experiments may lead to artificially high deuterium biases (Diao et al., 2022). Conversely, field studies often lack source water isotopic data, making bias inferences uncertain (Bowers & Williams, 2022; Zuecco et al., 2022). To our knowledge, Barbeta et al. (2022) and He et al. (2023) are the only authors who have investigated the differences between xylem water, bulk stem water and source water under natural field conditions. However, both studies focused on a limited number of tree species, warranting further investigation into how CVD-induced biases may vary across different species. Additionally, questions remain regarding the influence of environmental factors on deuterium biases, such as water availability and resulting tree water stress levels. A meta-analysis by de la Casa et al. (2022) suggests that environmental conditions may influence deuterium biases, but this hypothesis requires more detailed testing at specific locations.
As an alternative to CVD approaches, the Cavitron, a custom-made rotor fitted to a centrifuge (Cochard, 2002; Cochard et al., 2005), is now emerging as a powerful tool for non-destructive extraction of xylem water and subsequent evaluation of CVD-induced isotopic biases (Barbeta et al., 2022; He et al., 2023; Wen et al., 2022). In this study, we use Cavitron-extracted xylem water as a reference to assess the extent and magnitude of CVD-induced biases across several tree species under natural field conditions in the Australian tropics. We deliberately selected tree species spanning a range of water stress levels, as indicated by pre-dawn leaf water potentials. We also measured the isotopic composition of groundwater and soil water to characterise source water. Our work addresses the following questions:
Does the isotopic composition of Cavitron-extracted xylem water align with source water, and does CVD-extracted bulk stem water exhibit a deuterium offset?
Does the CVD-induced deuterium bias occur consistently across tree species, regardless of water stress conditions and xylem water isotopic composition?
We provide a detailed operating procedure for our Cavitron xylem water extraction (see Supplementary Information). Following the recommendations of Millar et al. (2022), we hope that this procedure can contribute to the establishment of a unified protocol for use by other research groups.