Prediction of nitrogen leaching loss from summer maize production in North China

North China has seen intensive flood irrigation and excessive nitrogen (N) fertilization over the past four decades as a main cereal crop-producing region in China. N leaching from farmland in this region has rapidly increased with agricultural intensification, and the non-point source pollution has become increasingly prominent. It is necessary to quantify the amount of N leaching during crop production systematically. Literature on N leaching loss from summer maize production in North China published from 1980–2021 was screened, and soil properties and agricultural management practices were chosen as independent variables to predict N leaching loss based on linear, exponential, polynomial, and multiple regression models. Soil properties included soil organic matter, total N, clay content, sand content, pH, and depth, and agricultural management practices included straw incorporation, N application, and soil water. The results showed that soil water and N fertilizer input significantly influenced N leaching loss. Soil organic matter, soil total N, and clay content positively correlated with the total N leaching amount, whereas straw incorporation, soil depth, pH, and sand content negatively correlated with the total N leaching amount. For the single-factor simulation model, the exponential equation was more appropriate for quantifying total N leaching loss with fertilizer N input than the linear equation, indicating the importance of optimizing fertilizer N in summer maize production in North China. It also indicated that the risk of excess N leaching from summer maize production in North China was relatively high after a certain threshold of fertilizer N input, and optimization of N fertilization should be adopted as an important practice. Unlike many previous studies that directly selected fertilizer N input for predicting N leaching loss, this study combined N (total N rate, N surplus) and water (water input, water balance, water percolation) in various combinations to obtain an optimal prediction combination. The combination of the total N rate and water percolation had the highest R² (0.3413). The stepwise regression equation of Y_{total N leaching loss}=−23.07+1.14X_{soil organic matter}+0.34X_{clay content}−0.13X_{sand content}+0.06X_{total N rate}+0.18X_{water percolation} (R²=0.414) was better than the prediction effects of exponential, linear, and polynomial models. The standardized regression coefficients of the predictive variables were 0.18, 0.11, 0.07, 0.23, and 0.31 for soil organic matter, clay content, sand content, total N rate, and water percolation, respectively, which showed that water percolation was the most important, followed by total N rate and soil organic matter. Considering the complexity of the water percolation calculation process, the water input can be used to replace water percolation in the equation, that is, Y_{total N leaching loss}=−18.60+0.64X_{soil organic matter}−10.27X_{straw incorporation}−0.30X_{sand content}+0.13X_{total N rate}+0.04X_{water input}; however, the prediction accuracy of the regression equation was affected. Future research on predicting N leaching loss in North China should focus on accurately quantifying water percolation. The quantitative model obtained in this study provides technical support for precise N management and effective pollution prevention in North China.