Abstract The stable isotopes of hydrogen and oxygen in xylem water are often used to investigate tree water sources. But this traditional approach does not acknowledge the contribution of water stored in the phloem to transpiration and how this may affect xylem water and source water interpretations. Additionally, there is a prevailing assumption that there is no isotope fractionation during tree water transport. Here, we systematically sampled xylem and phloem water at daily and subdaily resolutions in a large lysimeter planted with Salix viminalis . Stem diurnal change in phloem water storage and transpiration rates were also measured. Our results show that phloem water is significantly less enriched in heavy isotopes than xylem water. At subdaily resolution, we observed a larger isotopic difference between xylem and phloem during phloem water refilling and under periods of tree water deficit. These findings contrast with the expectation of heavy‐isotope enriched water in phloem due to downward transport of enriched leaf water isotopic signatures. Because of previous evidence of aquaporin mediated phloem and xylem water transport and higher osmotic permeability of lighter hydrogen isotopologues across aquaporins, we propose that radial water transport across the xylem–phloem boundary may drive the relative depletion of heavy isotopes in phloem and their relative enrichment in xylem.

Closure of the soil water balance is fundamental to ecohydrology. But closing the soil water balance with hydrometric information offers no insight into the age distribution of water transiting the soil column via deep drainage or the combination of soil evaporation and transpiration. This is a major challenge in our discipline currently; tracing the water balance is the needed next step. Here we report results from a controlled tracer experiment aimed at both closing the soil water balance and tracing its individual components. This was carried out on a 2.5 m3 lysimeter planted with a willow tree. We applied 25 mm of isotopically enriched water on top of the lysimeter and tracked it for 43 days through the soil water, the bottom drainage, and the plant xylem. We then destructively sampled the system to quantify the remaining isotope mass. More than 900 water samples were collected for stable isotope analysis to trace the labeled irrigation. We then used these data to quantify when and where the labeled irrigation became the source of plant uptake or deep percolation. Evapotranspiration dominated the water balance outflow (88%). Tracing the transpiration flux showed further that transpiration was soil water that had fallen as precipitation 1–2 months prior. The tracer breakthrough in transpiration was complex and different from the breakthrough curves observed within the soil or in the bottom drainage. Given the lack of direct experimental data on travel time to transpiration, these results provide a first balance closure where all the relevant outflows are traced.

The stable isotopes of hydrogen and oxygen (δ2H and δ18O, respectively) have been widely used to investigate tree water source partitioning. These tracers have shed new light on patterns of tree water use in time and space. However, there are several limiting factors to this methodology (e.g., the difficult assessment of isotope fractionation in trees, and the labor‐intensity associated with the collection of significant sample sizes) and the use of isotopes alone has not been enough to provide a mechanistic understanding of source water partitioning. Here, we combine isotope data in xylem and soil water with measurements of tree's physiological information including tree water deficit (TWD), fine root distribution, and soil matric potential, to investigate the mechanism driving tree water source partitioning. We used a 2 m3 lysimeter with willow trees (Salix viminalis) planted within, to conduct a high spatial–temporal resolution experiment. TWD provided an integrated response of plant water status to water supply and demand. The combined isotopic and TWD measurement showed that short‐term variation (within days) in source water partitioning is determined mainly by plant hydraulic response to changes in soil matric potential. We observed changes in the relationship between soil matric potential and TWD that are matched by shifts in source water partitioning. Our results show that tree water use is a dynamic process on the time scale of days. These findings demonstrate tree's plasticity to water supply over days can be identified with high‐resolution measurements of plant water status. Our results further support that root distribution alone is not an indicator of water uptake dynamics. Overall, we show that combining physiological measurements with traditional isotope tracing can reveal mechanistic insights into plant responses to changing environmental conditions.