Modeling nitrogen dynamics across contrasting yield zones to close yield gaps in a commercial maize system

Abstract

CONTEXT

Improving nitrogen use efficiency (NUE) in maize systems demands a mechanistic understanding of how nitrogen (N) dynamics interact with climate variability, irrigation regimes, and within-field spatial heterogeneity. Yet, N fate pathways such as leaching, volatilization, and denitrification remain difficult to quantify across diverse zones of productivity under real-farm conditions.

OBJECTIVE

This study aimed to simulate the fate and flow of applied nitrogen across management zones with contrasting yield stability and water-holding capacity, using a process-based model under long-term weather and irrigation scenarios.

METHODS

The DSSAT CSM-CERES-Maize model was calibrated and evaluated using independent field datasets from two growing seasons across two management zones (MZs) delineated by yield stability. The calibrated model was assessed for LAI, soil moisture, yield, soil N, and N uptake. After model evaluation, simulations were conducted over 40 years (1984–2022) of weather data, categorized into drought, normal, and wet years using the Standardized Precipitation Index (SPI), and under three irrigation scenarios: rainfed, 50 %, and 30 % management allowable depletions (MAD).

RESULTS AND CONCLUSIONS

Upon calibration, DSSAT CSM CERES Maize showed good agreement between observed and simulated values for crop growth, development, and soil water and N dynamics (NRMSE <15 %), supporting the use of the model for management zone-based N accounting. In 2022 high yield management zone (MZ1) exhibited higher N uptake (305 kg N ha⁻¹) and greater N losses via leaching (16.1 kg N ha⁻¹) and volatilization (33 kg N ha⁻¹), while low yield management zone (MZ2) showed elevated residual nitrate at harvest (60 kg N ha⁻¹) and higher denitrification (6.2 kg N ha⁻¹), indicating lower NUE. The long-term simulations indicated that leaching peaked under wet years (up to 33.8 kg N ha⁻¹), whereas drought years induced maximum volatilization (65.5 kg ha⁻¹). MZ2 required 45 % more irrigation to match MZ1 yields, reflecting soil-based water limitations. Moderate MAD irrigation regimes (50 %) reduced total N losses without yield penalties.

SIGNIFICANCE

This study demonstrates that integrating spatial field variability with long-term climate and irrigation scenarios is critical for diagnosing N inefficiencies and optimizing N use. Mechanistic modeling provides a scalable decision-support tool to design adaptive nutrients and water management strategies tailored to heterogeneous production systems.