Drought-induced tree mortality is occurring more frequently in the world, with both the direct impact of drought (i.e., heat and drought events) and the mortality predisposition (e.g., tree nutrition status) influencing death or survival of trees. The oxygen and hydrogen isotopic compositions (δ¹⁸O and δ²H) of plant water are widely used as hydrological indicators. Both elements in water are tightly correlated and are the key source for the isotopic composition of plant organic matter. Yet, recent studies show that the relationship between δ¹⁸O and δ²H values in organic matter is weaker and more divergent. This is probably caused by physiological and metabolic processes (i.e., assimilation, assimilate allocation, use of reserves) that are integrated into δ²H but not into δ¹⁸O. This let us hypothesize that δ²H can function as a useful tool in tree mortality research to study assimilate and storage related reason (i.e., carbon starvation) of tree death, but more knowledge is urgently needed.

To test our hypothesis, we studied the pre-disposal fertilization and drought effects on δ¹⁸O and δ²H values in plant water and organic matter in a greenhouse experiment. We planted three years old saplings of Abies alba, Acer pseudoplatanus, Picea abies, Pinus sylvestris and Quercus petraea, half of which were treated with a slow-release formula fertilizer (control (F0) and fertilization (F+)); one year later, half of the F0 and half of the F+ plants were selected for a lethal drought treatment (control (D0) and drought (D+)), i.e., D0 plants were watered to field capacity, while D+ plants received no more water until they died. After 6 weeks of drought, leaf and twig samples were collected for δ¹⁸O and δ²H analyses of plant water. After the D+ plants died (i.e., 9-15 weeks after start of drought), additional leaf and stem material were collected from a same number of D0 and D+ plants. Organic matter of leaf and tree-ring of the recent year were prepared for δ¹⁸O and δ²H analyses. Additional physiological and metabolic factors were measured to examine the treatment effects.

Across all species, we found that the pre-disposal fertilization had no significant effect on δ¹⁸O and δ²H values of plant water and organic matter. On the other hand, the drought treatment significantly increased both δ¹⁸O and δ²H values of leaf and twig water, while it only significantly increased δ²H values of leaf and tree-ring organic matter. These results indicate that δ²H in leaf and tree-ring organic matter in dying trees can capture drought-induced tree mortality signals. We propose that the ²H-enrichment in the drought-exposed trees might be related to (i) the imprint of ²H-enriched signal of plant water; (ii) drought-induced changes in the metabolic processes of sugar biosynthesis; (iii) drought-induced changes in the use of carbon reserves. In summary, our study supports the idea that hydrogen isotopes can function as a potential diagnostic tool in tree mortality studies.