The ability of crop models to predict soil organic carbon changes in a maize cropping system under contrasting fertilization and residues management: Evidence from a long-term experiment

This study assesses the ability of an ensemble of crop models (MME) to predict the impacts of fertilization and crop residue management on soil organic carbon (SOC) and aboveground biomass (AGB) in a long-term experiment (LTE) based on continuous maize cropping systems. Data from a LTE in Northern Italy were used. Treatments included continuous grain (MG) or silage (MS) maize, fertilized with mineral, cattle slurry, and farmyard manure. The MME median resulted the best predictor of the observed values. Models performance was better when simulating MG than MS, and for crops treated with mineral compared to organic fertilizers. The ability to predict the dynamics of SOC was affected by the model used and by the year × residues management and year × fertilizer interactions. The model and the residue × fertilizer interaction affected the ability to simulate AGB dynamics. Results showed that a MME can effectively predict the long-term dynamics of SOC and maize crop production under contrasting fertilization and crop residue management, and thus their potential for climate change mitigation. The uncertainty in the simulation of SOC is related to the model routines simulating SOC partitioning and to the complexity of the interactions between management factors over time.