Assessing the Effectiveness of Different Kinetic Models in Simulating the Soil Nitrification Process of Ammonium Fertilizers

Precise estimation of the soil nitrification process of ammonium fertilizers is crucial to improving soil fertility and reducing environmental pollution since nitrification is closely related to ammonia volatilization and nitrate leaching. However, the applicability and effectiveness of different kinetic models in simulating the soil nitrification process have not been systematically evaluated in previous studies. Here, we compared the effectiveness of one self-established model (T-function) and three commonly-used kinetic models (L-function, zero-order kinetic model, and first-order kinetic model) using data extracted from peer-reviewed publications. Results showed that the average determination coefficients (R²) of the T-function and L-function were 3% higher than that of the first-order kinetic model, while the average root mean square errors (RMSE) of the T-function and L-function were 30% lower than that of the first-order kinetic model. In addition, first-order kinetic model could not function properly when it was applied to simulate linear data. Zero-order kinetic model was not suitable for predicting soil nitrification process because its mathematical nature was inconsistent with the actual soil nitrification process. Further investigations of the parameters of the T-function and L-function revealed that p1 (parameter 1) represented the soil net nitrate production of the applied N fertilizer, while p1 × p3 represented the soil nitrification rate. The response of the parameters of the T-function was more sensitive to soil pH, TN, and C/N ratio changes than those of the L-function. Therefore, the T-function is slightly more accurate than the L-function, but it requires more data points to achieve the best performance. In general, T-function and L-function are the most suitable models for simulating the soil nitrification process of N fertilizers and have the potential to be incorporated into soil N cycling models.