Modelling long-term yield and soil organic matter dynamics in a maize cropping system

Long-term cropping system experiments are one of the most reliable sources of information for informing sustainable agriculture and predicting future trends. When combined with crop modelling, expansion of findings on optimised management approaches is possible. In this study, results from a South African semi-arid region long-term (66 years) maize (Zea mays L.) trial are presented and combined with crop modelling to identify the impacts of fertilisation and residue management on yield and soil organic matter (SOM) levels. Simulated and observed results generally agreed well in calibration and testing exercises with APSIM. For the fertilised treatment, residue retention led to a 41% increase in average yield over the long term, and for unfertilised treatment the average yield increase was even higher at 59%. The greatest SOM decline of 46% was observed for the unfertilised plus residue removal treatment (over 66 years and considering a 60 cm soil depth). Fertilising and retaining residue reduced the SOM decline to 18%. Using only fertiliser without residue retention did not lead to a declining yield trend over the long-term for this soil. The study indicated that the APSIM model can be used to explore the ecological intensification of maize production in sub-Saharan Africa. Further attention is recommended, however, on testing the simulation of subsoil SOM dynamics. The results of this study give insight into soil fertility in low-input maize production systems and quantify the benefits of N fertiliser and residue retention guided by long-term measured data.