Sensitivity analysis of C and N modules in biogeochemical crop and grassland models following manure addition to soil

Process-based crop and grassland models estimating carbon (C) and nitrogen (N) dynamics are widely used to investigate best management practices in agriculture. They integrate several processes in a complex structure, but studies where modules corresponding to specific processes extracted from the whole model structure are assessed independently are uncommon. With the support of documented aerobic incubation trials in manure-amended soils, a sensitivity analysis was performed on the C\u2013N cycling processes of four modules (MOD1\u20134), corresponding to the models APSIM, EPIC, FASSET and STICS. The results showed that the parameter \u2018substrate use efficiency\u2019 had the most effect on the predicted values of net CO2 emissions and net N mineralization, together with the C/N ratio of the soil microbial biomass. They explained 74\u201375% on average of both output variances, whereas parameters determining manure C and N partitioning and first-order decomposition constants of manure pools explained, on average, an additional 17\u201319%. Efforts should be focused on calibrating these parameters for more accurate simulations. The greater sensitivity of both outputs to parameters related to manure pools in more complex modules (MOD2\u20134) facilitates their adaptation to specific contexts, whereas MOD1 probably requires that parameters related to soil pools are also adapted to specific applications. Parameter interactions were limited, becoming noticeable only in situations of N-limited soil organic matter decomposition. Models MOD1 and MOD3 allowed the C/N ratio of the soil microbial biomass to vary temporarily; therefore, they were less sensitive to mineral N availability and more easily adapted to a wide range of situations. This study provides essential information to support the development of state-of-the-art biogeochemical models. Highlights: We compared four C\u2013N modules embedded in process-based biogeochemical models. We used sensitivity analysis to assess the simulation of manure decomposition in soil. We identified a few parameters that influenced CO2 emissions and N mineralization. We found that substrate use efficiency explained most of the output variance for all models