Why does rainfall affect the trend in soil carbon after converting pastures to forests? A possible explanation based on nitrogen dynamics

When trees are planted onto former pastures, soil carbon stocks typically either remain constant or decrease, with decreases more common in regions with higher rainfall. We conducted a modelling analysis to assess whether those changes in soil carbon, especially the interaction with rainfall, could be understood through consideration of nitrogen balances. The study was based on simulations with the whole-system ecophysiological model CenW which allowed explicit modelling of both carbon and nitrogen pools and their fluxes through plants and soil organic matter. We found that in a modelled coniferous forest without excess water input, total system nitrogen stocks remained similar to pre-forestation values because there were few pathways for nitrogen losses, and without biological nitrogen fixation or fertiliser inputs, gains were restricted to small inputs from atmospheric deposition. However, tree biomass and the litter layer accumulated considerable amounts of nitrogen. This accumulation of nitrogen came at the expense of depleting soil nitrogen stocks. With the change from input of grass litter that is low in lignin to forest litter with higher lignin concentration, organic-matter C:N ratios increased so that more carbon could be stored per unit of soil nitrogen which partly negated the effect of reduced nitrogen stocks. The increase in C:N ratios was initially confined to the surface litter layer because of slow transfer of material to the mineral soil. Over a period of decades, soil C:N ratios eventually increased in the soil as well. Simulations with different amounts of precipitation showed that greater amounts of nitrogen were leached from systems where water supply exceeded the plants' requirements. Reduced nitrogen stocks then caused a subsequent reduction in soil organic carbon stocks. These simulations thus provided a consistent explanation for the observation of greater losses of soil organic carbon in high-rainfall systems after converting pastures to forests. More generally, the simulations showed that explicit modelling of the nitrogen cycle can put important constraints on possible changes in soil-carbon stocks that may occur after land-use change