Validating the RothC model with long-term experiments in dryland areas of China

The soil carbon cycle is complex and affected by soil properties, field management, and climate conditions. To simulate this cycle, the Rothamsted Carbon Model (RothC) has been widely used in Europe and around the world, but it is seldomly validated before using. This study validated the RothC model with seven long-term experiments in dryland areas of China. These experiments include 27 fertilizer treatments with various organic fertilizers, such as crop residues, animal manure, and commercial organic fertilizers. Overall, the model performed well in predicting soil organic carbon (SOC) changes, except for sites with monthly mean air temperature lower than -18.27 ‰, which is the lower bound used to calculate the rate modifying factor of temperature in this model. For specific treatments, only 2 out of the 27 were accurately modeled, while the SOC in 24 of the treatments was underestimated. The potential reason is that C inputs from crop rhizodeposition were not considered in our modelling. In addition, the input data for crop root-shoot ratio is highly uncertain according to our literature review, while this parameter saliently affected the model performance based on our sensitive analysis. In summary, the RothC model must be validated and calibrated before used to predict SOC in a specific site of China, although it may perform well in large-scale modelling.
An accurate input data for root shoot ratio is crucial to ensure the model accuracy.