How does plant leaf senescence of grasland species influence decomposition kinetics and litter compounds dynamics?

National audience

anglais. The influence of litter quality on plant litter decomposition rates is a crucial aspect of the soils C cycle. In grassland ecosystems, leaf litter, which is not removed either by herbivores or by mowing, returns to soil after the senescence process (brown litter). In grassland managed by mowing, another significant proportion of litter returns to the soil before senescence through harvesting losses (green litter). We hypothesized that changes in leaf tissue quality due to the senescence process would lead to contrasting decomposition dynamics of brown litter compared to green litter. Our conceptual approach included the monitoring of decomposition of green (fresh leaves) and brown litter (dead leaves, still attached to the plant) of three different grassland species (Lolium perenne, Festuca arundinacea and Dactylis glomerata) during a 1 year field incubation. After 0, 2, 4, 20 and 44 weeks, we retrieved the litterbags and analysed the remaining material for carbon and nitrogen content and stable isotope composition. Additionally, we determined the lignin content and composition by CuO oxidation and the non-cellulosic neutral carbohydrate content and composition after TFA hydrolysis. As expected, green litter, being higher in N and soluble compounds, while showing a lower C:N ratio and lower lignin contents compared to brown litter, was degraded at a higher rate. Carbon decomposition kinetics suggests that both leaf litter types consist of two pools with contrasting turnover times. The size of the active pool was related to the initial content of soluble plant litter compounds and the size of the recalcitrant pool was related to the lignin to N ratio of initial plant material. More lignin was lost from green litter compared to brown litter. P-coumaryl-type lignin units were decomposed at a higher rate than vanillyl and syringyl units. Total non cellulosic polysaccharide content showed little changes for both litter types. In contrast, the ratios of hexoses/pentoses (C6/C5) and desoxy sugars/pentoses (desoxy/C5) increased during decomposition of green litter only. This is an indication for an increasing contribution of microbial derived compounds being consistant with the higher decomposition rate of this material. Our results showed that grassland management (grazing versus mowing) could influence soil carbon sequestration through different proportions of green and brown litter returned to soil.