Stable isotope disequilibrium between soil bound water and soil bulk water – Implications for estimations of plant water sources

Determining the origin of plant water sources remains a persistent challenge due to the complex nature of eco-hydrological processes. Spatial and temporal fluctuations in plant water sources are frequently traced using the stable isotope compositions of soil water, which is commonly extracted using the cryogenic vacuum distillation technique (CVD). However, utilizing CVD for soil samples may not accurately reflect the true source of water for plant use, because the soil bulk water may not be fully mixed, and its water stable isotope composition could vary with different soil matric potentials. In this study, twig, soil, stream water, and event-based throughfall/precipitation samples were systematically collected during the pre-rainy season (May to July 2019) and the rainy season (August and September 2019) in an upper Heihe River catchment in northwest China. A suction lysimeter (SUC) method was used in the field to collect soil water at a matrix potential > -0.05 MPa, while the CVD and centrifugation techniques (CEN) were used in the laboratory to extract soil bulk water and water over a matrix potential of -2.5 MPa, respectively. The proportional contributions of soil water from various depths to plant water were estimated using the MixSIAR model for native Qinghai Spruce (Picea crassifolia) and Qilian Juniper (Juniperus przewalskii) trees. Through the pre-rainy and the rainy seasons, the shallow soil 0–15 cm contributed between 20 ± 9 % and 75 ± 22 % to the plant water use, while the 15–30 cm soil depth supplied between 25 ± 22 % and 80 ± 9 % of water at the Spruce site. The depth levels of 0–15 cm, 15–30 cm, 30–45 cm, and 45–60 cm contributed on average 27 ± 22 %, 29 ± 23 %, 24 ± 20 %, and 21 ± 20 % soil water to the plants at the Juniper site, respectively. In comparison to the CEN technique for soil samples, soil water showed significantly (p < 0.05) more negative δ¹⁸O (and δ²H) values up to -2.38 ± 1.23 ‰ (-19.0 ± 7.6 ‰) for the CVD method, and -0.43 ± 0.81 ‰ (-2.9 ± 5.8 ‰) for SUC methods. These results suggest an obvious stable isotope disequilibrium between soil bound water (a matrix potential lower than -2.5 MPa) and soil bulk water stable isotope compositions, which indicates that the calculation of soil water contributions to plants could be significantly biased if only the CVD method is used. Employing routinely both CEN and CVD methods for soil samples can lead to much more reliable results and improve our understanding of water source contributions at different hydraulic potentials to plant use.