Crop production, soil carbon and nutrient balances as affected by fertilisation in a Mollisol agroecosystem

A 19-year field experiment on a Mollisol agroecosystem was carried out to study the productivity of a wheat-maize-soybean rotation and the changes in soil carbon and nutrient status in response to different fertiliser applications in Northeast China. The experiment consisted of seven fertiliser treatments: (1) unfertilised control, (2) annual application of P and K fertilisers, (3) N and K fertilisers, (4) N and P fertilisers, (5) N, P and K fertilisers, (6) N, K and second level P fertilisers, and (7) N, P and second level K fertilisers. Without fertiliser, the Mollisols could support an average yield of 1.88 t ha−1 for wheat, 3.89 t ha−1 for maize and 2.12 t ha−1 for soybean, compared to yields of 3.20, 9.30 and 2.45 t ha−1 respectively for wheat, maize and soybean if the crop nutrient demands were met. At the potential yield level, the N, P and K removal by wheat are 79 kg N ha−1, 15 kg P ha−1, and 53 kg K ha−1, by maize are 207 kg N ha−1, 47 kg P ha−1, and 180 kg K ha−1, by soybean are 174 kg N ha−1, 18 kg P ha−1, and 55 kg K ha−1. Crop yield, change in soil organic carbon (SOC), and the total and available nutrient status were used to evaluate the fertility of this soil over different time periods. This study showed that a fertiliser strategy that was able to maintain yields in the short term (19 years) would not maintain the long term fertility of these soils. Although organic carbon levels did not rise to the level of virgin soil in any treatment, a combination of N, P and K fertiliser that approximated crop export was required to stabilise SOC and prevent a decline in the total store of soil nutrients.