Modelling soil carbon and nitrogen dynamics using measurable and conceptual soil organic matter pools in APSIM

Measurable soil organic carbon (SOC) fractions are recommended to be used in soil organic matter (SOM) models to replace the conceptual pools in order to reduce uncertainty related to model initialization. However, changes in SOM pools require derivation of the decomposition rate for the new pools and also impact on simulation of nutrient cycling in soil. Here, we used the Agricultural Production Systems sIMulator (APSIM) to explore the impact of adopting measurable SOM pools on simulated SOC and nitrogen dynamics under three agro-ecosystems at Brigalow, Tarlee and Wagga Wagga in Australia. The decomposition rates for the measurable pools were derived based on the data in the top 30 cm soil at Brigalow. With those decomposition rates, the model could predict the variation in observed total SOC, and also captured the dynamics of the measured SOC fractions at the three sites. However, the re-parameterized model overestimated soil nitrogen. Extending the model simulations to centurial scales led to the distinct difference in simulated soil carbon and nitrogen dynamics, due to the difference in pool sizes, decomposition rates and C:N ratios between the two approaches. Sensitivity analysis based on measurable pool approach indicated that simulated soil carbon dynamics was most sensitive to the partitioning of carbon in the sub-pools of new plant material, decomposition rates of particulate and humic organic matter. The simulated soil nitrogen dynamics was most sensitive to the C:N ratio of humic organic matter. In order to accurately simulate the long-term carbon and nitrogen dynamics, more detailed information on the composition of the measurable SOM pools, particularly their C:N ratios and the underlying mechanisms controlling their decomposition and transformation among different pools across different environments are needed.