Differences between stem and branch xylem water isotope composition in four tropical tree species

Stable isotope studies (delta H-2 and delta O-18) of water within plants are providing new information on water sources, competitive interactions and water use patterns under natural conditions. This is based on the assumption that there is no fractionation at the time of water uptake or during its transport within trees. However, previous studies have found fractionation does occur in water taken up through the roots in halophytic and xerophytic plants. It is unclear how widespread such fractionation is in other species. In this study, we tested this fractionation by comparing stem and branch xylem water isotopes over the period of one day (from 8 am to 5 pm) in four wet tropical rainforest tree species (Dendrocnide photinophylla, Aphananthe philippinensis, Daphnandra repandula and Mallotus polyadenos). We found branch water isotope ratios (delta H-2 and delta O-18) were enriched compared with stem xylem water isotopes. D. photinophylla had a significantly different branch delta O-18 ratio than A. philippinensis, D. repandula and M. polyadenos. In contrast, there were no significant differences in stem xylem delta O-18 among the four observed tree species. We found clear differences in the stable isotope ratios of delta O-18 for stem and branch xylem water for D. photinophylla and D. repandula of upto -0.85% and 0.50%, respectively. Remarkable delta H-2 differences were also found between stem and branch xylem water isotope ratios for A. philippinensis, D. repandula and M. polyadenos, being upto -12.14%, -16.17% and -9.65%, respectively. A dual isotope (delta H-2-delta O-18) plot showed branch water values were more enriched than stem xylem water values for D. photinophylla and D. repandula, indicating a clear difference between stem and branch xylem water. This study suggests that delta H-2 and delta O-18 fractionation could be a species-specific phenomenon in tropical trees, which has important implications for plant water source identification and evapotranspiration partitioning.