Global sensitivity and uncertainty analysis of the dynamic simulation of crop N uptake by using various N dilution curve approaches

Crop nitrogen (N) uptake is a key process in soil-crop models. This process affects crop growth and soil N cycling and determines crop quality. However, crop N uptake modeling remains uncertain because of the various N dilution curve approaches adopted in soil-crop models. In this study, four different representative N dilution curve approaches (i.e., DAISY, M1; CERES and RZWQM, M2; EPIC, M3; and CROPSYST or STICS, M4) were incorporated into a soil-crop model platform (WHCNS), and their effects on crop N uptake and crop growth simulation under different water and N stresses were evaluated via global sensitivity analysis methods. The three-year field experiment data of winter wheat-summer maize rotation under different water and N management practices were used to test the model. Results showed that the WHCNS model performed well in modeling the supplies of soil water and mineral N. The values of statistical indices for crop N uptake, LAI, dry matter and yield simulation by the four methods were within the acceptable ranges, and had the relative mean square error (RRMSE) < 24 %, index of agreement (IA) > 0.81 and Nash and Sutcliffe index (NSE) > 0.37. However, the M2 method performed well using the minimum input parameters, and hence recommended to simulate crop N uptake in soil-crop models. In this study, the dataset of high water and N input treatment was more suitable for model parameter estimation to reduce uncertainty, and the datasets of middle and low water and N input treatments was appropriate to validate the model. These information were useful to guide selecting the modeling method and the model calibration dataset.