Nitrogen management in crop rotations after the break-up of grassland: Insights from modelling

Evaluation data consisted of two rotation types over two years (maize-maize and barley-mustard-maize) with two nitrogen (N) fertilizer schemes (zero and standard fertilizer: 160 for maize and 120 kg N/ha for barley) after the break-up of grassland. Experiments were carried out at three different sites with contrasting soils in north-west Germany. Results showed that APSIM was capable of simulating the crop rotations and fertilizer applications satisfactorily: Total biomass (n = 21) was reproduced with a root mean square error (RMSE) of 1139 kg/ha against an observed mean of 9915 kg/ha across crops. Total N uptake (n = 21) was simulated well with a RMSE of 22 kg/ha (against observed mean 144 kg/ha). Simulated soil mineral N in the top 0-30 cm (n = 253) and 0-90 cm (n = 33) showed a high index of agreement (IA) of 0.90 and 0.86, respectively. Comparisons observed vs simulated over time confirmed that APSIM was able to capture the N dynamics in the soil. Extractable soil water was also modelled well. Leached nitrate (n = 16) was simulated with a RMSE of 50 kg N/ha, whereby APSIM captured the high nitrate losses of up to 240 kg N/ha/winter period caused by the high mineralization and the fertilization. In the long-term the simulation experiment showed that fertilization of maize did not result in additional biomass, but in higher leaching losses. Mustard was effective in reducing nitrate leaching but is difficult to implement in practice. Finally, the study demonstrated that crop modelling complements conventional analysis very well in identifying environmentally sound and profitable management practices for complex situations in soil-crop systems such as grassland break-up.