Quantitative evidence of overestimated rhizodeposition using 15N leaf-labelling

Nitrogen (N) rhizodeposition is defined as the release of N from living plant roots into the soil and root turnover. The proportion of N in the soil derived from rhizodeposition (NdfR) is usually determined using ¹⁵N labelling of the plant. This isotope approach assumes that i) the enrichment of rhizodeposits is equal to the root enrichment and that ii) the root remains homogeneously enriched over space and iii) over time. The aim of this study was to quantify the bias resulting from a possible violation of the mentioned assumptions and to study the causative factors of bias.We conducted two experiments with single-pulse ¹⁵N-urea leaf-labelled red clover (Trifolium pratense L.). In the rhizodeposition experiment, we simultaneously observed the changes in substrate N concentration to obtain the mass-based rhizodeposits and determined the isotope-predicted rhizodeposits using the isotope approach. By comparing isotope-predicted to mass-based rhizodeposits we quantified the bias of the isotope approach for a period of 6 weeks. In the root distribution experiment, we observed the root ¹⁵N enrichment over space and time (4 weeks) by sampling roots grown within certain periods relative to the labelling. In both experiments we monitored the ¹⁵N distribution between shoots and roots.We observed violations of the three assumptions of the isotope approach. The average root enrichment increased over time in the root distribution experiment, but remained constant in the rhizodeposition experiment. Significant long-term translocation of ¹⁵N from shoot to root during the whole experiment (over)-compensated for growth dilution. Spatial root enrichment varied within a factor of 3, peaking in roots grown 2–8 days after labelling (d.a.l.). We observed a significant leakage of 0.5 ± 0.2% of the applied ¹⁵N within the first day after labelling corresponding to an overestimation of first-day rhizodeposits by 1100 ± 800% which translates to a calculated enrichment of 9 ± 6 atom% ¹⁵N excess for first-day rhizodeposits compared to 0.77 ± 0.09 atom% ¹⁵N excess of the root.The leaked ¹⁵N together with ordinary rhizodeposits (rhizodeposits released after the first day of labelling) led to an overestimation of N rhizodeposition by 70 ± 30% at the end of the six weeks lasting experiment using the isotope approach. The observed ¹⁵N distribution in roots and long-term ¹⁵N translocation from shoots to roots did not correspond to the expected distribution following classical pulse labelling. Thus, leaf-labelling with ¹⁵N-urea should not be considered a pure pulse-labelling method.