Xylem water cryogenic vacuum extraction: Testing correction methods with CaviTron-based apple twig sampling

Cryogenic vacuum extraction (CVE), the most popular method for obtaining xylem water for isotopic analysis of ecosystems, is known to cause delta H-2 offset and may lead to biased inference on transpiration. While delta H-2 offset correction methods have been proposed recently, their accuracy in identifying plant water sources remains uncertain. In this study, capitalizing on the latest CaviTron centrifuge technique (CACT), we evaluated the accuracy of three current delta H-2 offset correction methods-CVE-bulk xylem water after the 8.1 parts per thousand correction method (8.1 parts per thousand Mean-Correc), CVE-bulk xylem water after the soil water isotope line correction method (Line-Correc), and CVE-bulk xylem water after the rehydration correction method (REHYDR)-for plant water source identification. We tested the three correction approaches on five apple (Malus pumila Mill) orchards. Our results revealed that although delta O-18 values of the CVE-bulk xylem water and the CACT-sap water matched, the delta H-2 values differed (P < 0.05), with a mean difference of -10.23 parts per thousand +/- 3.75 parts per thousand. By applying the available correction methods to the CVE-bulk xylem water, the delta H-2 offset was reduced to 2.13 +/- 3.75 parts per thousand (8.1 parts per thousand Mean-Correc), 4.43 +/- 4.48 parts per thousand (Line-Correc), and -0.44 +/- 3.69 parts per thousand (REHYDR). The corrections also improved the Z-score. However, substantial delta H-2 offset variations remained after the corrections. Moreover, when applying the corrected xylem water isotopic compositions in plant water apportioning analysis, the Line-Correc method resulted in significant differences (P < 0.05) in 1 out of 80 cases, while the 8.1 parts per thousand Mean-Correc (0/80) and REHYDR (0/80) methods did not. Our results suggest that for the available correction methods for cryogenic extraction to reduce delta H-2 offsets, the REHYDR and 8.1 parts per thousand Mean-Correc methods outperformed the Line-Correc method in plant water source identification. Thus, we recommended that the REHYDR method and 8.1 parts per thousand Mean-Correc method be used to identify plant water sources of CVE-extracted bulk xylem water.