Application of the RothC model to the results of long-term experiments on typical upland soils in northern China

Arable land can be either a source or a sink for atmospheric carbon dioxide depending on its management. It is important to assess changes in soil organic carbon (SOC) under future climate change scenarios using models at regional or global scales. This paper aims to calibrate the RothC model on non-waterlogged soils in northern China to obtain the necessary model input parameters for later use in large-scale studies. Data sets from three long-term experiments in northern China were used to evaluate the performance of the RothC soil carbon turnover model. The plant carbon input rate, an important model input parameter, was calibrated using experimental data under typical rotation systems with different fertilization. The results showed that RothC accurately simulated the changes in SOC across a wide area of northern China (northeast, north, and northwest China. The modelling error expressed as root mean square error for four treatments (nil, manure, fertilizer, fertilizer + manure) at three sites were less than 20.2%, and less than 7.8% if occasional extreme measured values were omitted. The simulation biases expressed as M (i.e. relative error) for all treatments at the three sites were non-significant. Observed trends in SOC included a decrease for the nil (no fertilizer or manure) treatment and an increase for the treatments which received both manure and fertilizers. The experiments also indicated that manures applied at an appropriate rate were more effective in increasing or maintaining SOC than fertilizers which were more effective in increasing crop yields.