Fate of the nitrogen from the beech litter to the tree on the long term in beech forests. Modelling and use of labelled litter

Forest soils are a physical media and main source of plant nutrients. Increasing anthropogenic stresses may lead to a decreasing fertility and thus a decrease in biomass production. Improved understanding of biogeochemical cycles in forest ecosystems allows a better management of soil resources to sustain biomass production and satisfy the human needs. With phosphorus, nitrogen (N) is an essential nutrient for plant growth. Although present in large quantities in nature, N remains a limiting factor for many ecosystems. Moreover, it is the only nutrient that is not released by mineral weathering. The main source of N for plants comes from the decomposition and mineralization of the litter. The main objective of this PhD was to study the fate of litter-derived N on the long term (one decade) in the soil-plant system in temperate forest ecosystems. The factors controlling the processes of stabilization, immobilisation, tree uptake and N loss leading to the formation of soil N stocks and contributing to tree nutrition were also highlighted. For this, a robust and simple model consisting of two modules - soil and tree - was built and validated from a nitrogen-15 (¹⁵N) labelling experiment. The labelling experiment took place ten years ago in ten European beech forests. The ¹⁵N enriched litter was produced on a dedicated site and was set on the soil around selected trees in the ten sites. The ¹⁵N enrichment was measured regularly for one decade in the decomposing litter, the forest floor (0 - 30 cm) and the leaves. Monitoring of the tracer ¹⁵N in the soil profile highlighted a low transfer of N into mineral soil horizons over a decade. There was no isotopic enrichment of the soil layers deeper than 10 cm depth and the maximum of ¹⁵N recovered in the soil was measured in the first 5 cm of the soil (Ah horizon). The ¹⁵N peak of the litter-derived N recovered in the soil was measured between 2 and 4 years after the labelling in the ten sites and in the 0 - 5 cm layer. In the leaves of the trees that received the labelled litter the kinetic of the amount of ¹⁵N over time showed a first phase of increasing leaf ¹⁵N illustrating an increasing N availability due to preferential mineralization of litter N followed by a second phase of decreasing leaf ¹⁵N due to stabilization and vertical distribution of litter N into the soil profile. Modeling the fate of litter ¹⁵N in the soil and the tree resulted in a model with three compartments in the soil. The first compartment is dedicated to litter decomposition products and the two others soil compartments correspond to mineral horizons with different N turnover rates. The analysis of the kinetic of the leaf ¹⁵N over a decade as well as the model parameters allows linking the form of the humus and the dynamics of the litter-derived N uptake by trees. Mull humus sites showed that the homogenization of N in the soil profile and the sequestration in earthworm faeces may slow down the accessibility of plants to the source of N but extended it in time by progressive release of N. At moder humus sites, trees quickly and directly mine for litter N from the organic horizons by ectomycorrhizal associations. The modeling study also shows that the litter contributes to the tree N nutrition over a decade and at a maximum rate 4 to 5 years after the litterfall. Neglecting atmospheric N inputs, about 25 % of the total N needs of the tree (leaves, trunk, branches, and roots) are covered by the reallocation of N from the tree, and 75% are from the soil N uptake. The use of litter derived N by trees is also function of tree growth, site productivity and tree age, young trees do not have the same N requirements than older trees. (author)