Tree-ring isotope-based 20th century reconstructions of the seasonal origin of water sourced by trees: Advances and limitations

The hydrogen (δ²H) and oxygen (δ¹⁸O) isotopic signatures of tree rings depend on that of the environmental water sources, such as precipitation and soil water, taken up by trees (i.e., "source water"). Analyzing δ²H and δ¹⁸O of tree rings is thus a promising approach for reconstructing the spatio-temporal origins of tree water sources. However, such reconstructions remain rare, likely due to methodological challenges, including the analysis of hydrogen isotopes in tree rings and the availability of historical source water isotope data.

In this study, we present a first attempt to reconstruct the temporal origins of water used by trees during the 20th century (1901–1995) with annually resolved tree-ring δ¹⁸O time series. The reconstruction is based on a δ¹⁸O chronology of whole wood, sampled from the latewood of spruce (Picea abies) at Bettlachstock, Switzerland. Our choice of site and species reflects a conservative approach, as a transfer function linking δ¹⁸O of tree-ring cellulose to the δ¹⁸O of source waters (e.g., stem xylem water and soil solutions) was recently established over a 17-year period (2006–2022) at the same site. After accounting for the isotopic offset between whole wood and cellulose, we estimated δ¹⁸O values of soil solution (80 cm depth) and stem xylem water during the growing season (May–September) using a linear transfer function. Further, using modeled precipitation δ¹⁸O data and the estimated δ¹⁸O of soil solution and xylem water, we deduced interannual variations in the seasonal origin index (SOI) of soil solution and xylem water during the 20th century.

Our results show that the reconstructed δ¹⁸O values and SOI of xylem water were higher than those of soil solutions, suggesting a greater contribution of summer water to xylem water than to soil solutions. Interestingly, while conditions from 1900 to 1970 remained relatively stable, we observed abrupt increases in SOI for both soil solutions and stem xylem water between 1970 and 1995. These recent changes were not due to an increase in summer precipitation amount but may be linked to shifts in seasonal precipitation patterns, causing a relative increase in the contribution of summer precipitation in tree water sources.

Despite these findings, uncertainties in precipitation isotope data and transfer functions need further investigation to draw more definitive conclusions. We hope this study will stimulate discussion on the advances and limitations of using tree-ring isotopes to reconstruct historical water sources.