The functioning of human-managed grassland ecosystems strongly depends on how common management practices will affect grassland "belowground compartment" including soil biogeochemistry and plant roots. Key questions remain about how animal grazing, liming (e.g., the addition of CaCO to soils), and nutrient fertilization might affect, in the long-term, soil nutrient cycling and multiple root traits. Here we focus on a mesotrophic grassland located in Berkshire, UK, where contrasting levels of rabbit grazing, liming, and different inorganic fertilizers have been applied since 1991. We ask how (1) soil nitrogen (N) availability and cycling, (2) total root mass, (3) root mass decomposition, and (4) arbuscular mycorrhizal fungal (AMF) root colonization might respond to 22 years of very different management. We found that liming strongly affected total root mass, root decomposition, root AMF colonization as well as soil N availability and cycling and that these effects were mainly driven by liming-induced increases in soil pH. Increases in soil pH were associated with significant (1) decreases in root mass, (2) increases in root mass decomposability and in the mineralization of N in decomposing root detritus, and (3) increases in AMF infection. Soil pH was also significantly related to greater N availability (i.e., soil NO levels) and to lower δ N natural abundance, which suggests more efficient N uptake by plants in limed soils as we found in our study. The application of multiple nutrients (N, P, K, Mg) also reduced total root mass, while N-only fertilization was associated with greater AMF infection. Surprisingly the long-term impact of grazing was generally weak and not significant on most plant and soil parameters. Despite soil pH affecting most belowground variables, changes in soil pH were not associated with any change in soil C and N stocks. Because liming can improve nutrient cycling (and benefits soil pH and grass yields) without negatively affecting soil C sequestration, we suggest that regular liming applications may provide management solutions for increasing the long-term sustainability of permanent grassland.

Abstract. The functioning of human-managed grassland ecosystems strongly depends on how common management practices (e.g., animal grazing and the chronic addition of fertilizing materials to soils) interact to influence plant and soil element stoichiometry. Here we use data from a 22-yr-long grassland experiment to address whether and how plant element stoichiometry (i.e., carbon [C], nitrogen [N], phosphorus [P] ratios) might respond to (1) animal grazing, (2) agricultural liming (i.e., CaCO 3 ) applications, and (3) nutrient fertilization. We also ask whether plant C:N:P stoichiometry could predict changes in soil N and P availability and in soil C, N, and P stocks. We found that grassland management significantly affected plant C:N:P ratios as predicted by ecological stoichiometry theory. For example, plant aboveground and belowground C:N and C:P ratios decreased under chronic N and P fertilization, respectively. Plant C:N and C:P ratios were significantly greater in unfertilized (control) soils. Also plant C:N ratios were highest under P-only additions, whereas plant C:P ratios were highest under N-only additions. However, unpredictable changes in C:N:P ratios also occurred, suggesting that plant tissue chemistry may not be a simple reflection of soil nutrient availability. Changes in plant C:nutrient ratios well predicted variation in soil nutrient availability, but not in soil C, N, and P stocks. Contrary to expectations, soil C stocks significantly increased with decreasing plant C:N ratios in the nutrient-fertilized grasslands and not with increasing plant C:N ratios in the unfertilized grasslands. We suggest that a better mechanistic understanding of the negative relationship between plant C:N stoichiometry and soil C accrual will greatly help in improving the sustainability of human-managed grasslands.