Effect of hydrogen isotopic offset and input data on the isotope-based estimation of plant water sources

Abstract

The stable isotope of water has been widely used to investigate plant water resources with the assumption that root water uptake and plant water transport is a non-isotopic fractionation process. However, recent research revealed that significant isotopic offsets between xylem water and potential water sources, which may result in errors when estimating plant water sources. We test five different input data modes in Bayesian isotope mixing model [only δ18O (M−O), only δD (M−D), δD after subtracting the deviation with respect to the soil water line (SW-excess) from the δD of xylem samples (M−SD), δD and δ18O (M−OD), and δD and δ18O after subtracting the SW-excess from the δD of xylem samples (M−OSD)] to identify sources of water absorbed by Chinese wolfberries (Lycium chinense Miller). The results showed that there was significant difference (p < 0.05) between the M−O and M−D modes. The evaluation indexes (RESE, N, MaxE, MinE, MAPE) performed well in the M−SD, M−O, and M−OSD modes. Furthermore, the M−SD, M−OSD, and M−O had consistent contribution proportions for the three soil layers. In addition, the corrected δD values in the dual and single isotope modes performed better than the original δD in quantifying water source contribution. This study provides critical insights into choosing a suitable method for isotope-based identification of plant water sources.