Delayed and asymmetric responses of soil C pools and N fluxes to grassland/cropland conversions

Agricultural lands are subject to changes in use for a variety of environmental and economic reasons. Among these changes, shifts between crops and grasslands are particularly common. Understanding the effects of these changes on soil functioning is crucial to anticipating crop productivity and environmental impacts and yet has rarely been studied. The purpose of this study was to investigate the in situ temporal dynamics of soil N cycling and C pools after grassland/cropland conversions and risks of N losses from such systems. The experiment was conducted in western France at a long-term experimental site with a temperate oceanic climate. We followed soil N cycling and C pools for four experimental treatments over 36 months: (i) continuous grassland, (ii) continuous rotation of annual crops, (iii) conversion from grassland to a rotation of annual crops, and (iv) conversion from a rotation of annual crops to grassland. Gross N mineralization, potential N immobilization and potential nitrification were estimated using the 15 N dilution/enrichment technique and FLUAZ model, based on samples from the 0–10 cm soil layer (0, 0.75, 3, 12, 24 and 36 months after change in land use) and 10–30 cm soil layer (after 24 and 36 months) in all 4 treatments. Nitrogen pools (NH 4 + , NO 3 − , and organic N), total carbon (C), dissolved organic C, microbial biomass C, and C mineralization were also determined on all soil samples. The continuous grassland soil had rates of N mineralization and N immobilization twice as high as a rotation of annual crops, as well as a slightly lower rate of nitrification and a higher organic C content. Ploughing the grassland led to a rapid shift of soil organic matter pools, N fluxes and microbial activities towards characteristics of the cropland. These effects were attributed to cultivation mixing the soil layers. In contrast, the restoration of grassland on soil previously cropped with annual species did not significantly change soil N or C characteristics within 24 months although changes began to appear after 36 months. This time lag corresponded to the time needed for the complete establishment of root systems for the grasses. Therefore we found a strong asymmetry in the patterns of change in gross N fluxes and C pools between the two conversions. Overall, we found that responses in N fluxes were strongly correlated to soil C pools, except for nitrification. The ratio of potential nitrification to gross ammonium immobilization, which indicates the risk of N losses in agricultural systems, was greater in annual crop rotation than grassland. This indicates that adding grasslands into a rotation of annual crops does not necessarily reduce the risk of N loss within 2 years but does reduce the risk after 3 years.