Modelling long-term changes in forest productivity and soil nitrogen supply following conversion of pasture to Pinus radiata plantation

Recent experimental studies have revealed that soil carbon (C) and nitrogen (N) content and soil N availability often decrease following conversion of improved pasture to Pinus radiata plantations. These decreases are a concern partly because of possible negative consequences for future forest productivity. This issue is investigated by applying the G'DAY model of C and N cycling in pasture and forest ecosystems to simulate the replacement of improved, legume-rich pasture by P. radiata plantations that are grown over several harvest cycles at a site in New Zealand. We illustrate how process models can be used for analysing constraints on long-term productivity by performing a sensitivity analysis of G'DAY's response to various rates of N removal. We find that simulated productivity declines over successive forest rotations and that the rate of decline is sensitive to N losses through wood harvesting, slash removal, and N leakage (i.e. leaching and soil gaseous emission). However, even when these N losses are zero, forest productivity still declines because of a gradual depletion of labile soil N reserves. Simulations are used to evaluate the mean annual increment (MAI) in wood volume over forty 30-year forest rotations. With harvesting only, simulated MAI declines from 44 m(3) ha(-1)yr(-1) in the first rotation to 18 m; ha(-1)yr(-1) in the fortieth rotation. The simulated MAI in the fortieth rotation is 13 m(3) ha(-1)yr(-1) with harvesting and leakage, and 11 m(3) ha(-1)yr(-1) with harvesting, leakage and 50% slash removal. If the simulation with harvesting only is modified so N removed in harvests is replaced by an equivalent fertiliser addition, MAI is 28 m(3) ha(-1)yr(-1) in the fortieth rotation. Results presented are tentative and should be regarded with caution until the model is fully tested.