Performance test of the soil nitrogen balance estimation system using lysimeter experiments conducted in andosol-based cropping systems

Abstract

The purpose of this study was to investigate the performance of a simple empirical model that was previously developed to support rational fertilizer management in upland farms in Japan. Statistical criteria and graphical displays were used to evaluate the capability of the soil nitrogen (N) balance estimation system to estimate nitrate-N (NO-3-N) leaching. Longterm lysimeter experimental data of three Andosol-based cropping systems in Abashiri, Mito and Miyakonojo were used in this study. Input data included soil characteristics, crop properties, climatic data and evaluation period. Results showed good fitness (low relative root mean square error) and precision (R 2=0.57, 0.93) in Abashiri and Mito, respectively. Simulated NO- 3-N leaching agreed with observed trends for all N management schemes at the three locations. Overestimations followed by underestimations in Miyakonojo affected the system's precision (R2=0.30), thereby exposing its weakness. Simulated soil residual N strongly correlated to crop uptake and simulated and observed leached nitrate-nitrogen in Abashiri and Mito. Monthly leaching estimates showed peaks in March to April, June to July and September to October demonstrating the system's usefulness as a tool for rational fertilizer management. A sensitivity analysis demonstrated the system's ability to respond to N input and output data. RRMSE and R2 for high nutrient-input treatments indicated the system's ability to estimate NO- 3-N leaching.