Quantifying nitrogen leaching in wheat (Triticum aestivum L.) using lysimeter stable isotope conservative tracer and modelling techniques

Abstract

Nitrogen (N) leaching is one of the important pathways that leads to water pollution, and previous studies have highlighted the difficulty in measuring it. The purpose of this study was to evaluate different techniques used to quantify nitrate-N (NO3-N) leaching load and determine fertiliser N use efficiency (FNUE). Lysimeter and field trial sites were planted with wheat (Triticum aestivum) (PAN3400 cultivar) at the University of Pretoria Experimental Farm, Hatfield, Pretoria. Two weighing lysimeters and a drain gauge were installed at the lysimeter trial site. Water content sensors and suction cups (SCs) were installed at 0.15, 0.3, 0.5 and 0.7 m depths in the weighing lysimeters and close to the drain gauge, while SCs, wetting front detectors (WFDs) and water content sensors were installed at 0.25 and 0.5 m in the field trial site. The crop was fertilised with 200 kg N ha-1 at both sites, but no fertiliser was applied on unfertilised control plots at the field trial site. High density drip irrigation was used at both sites, and bromide (Br-) was applied to all field plots at 0.020 kg m-2. Water was sampled from the SCs, WFDs and the bottom of weighing lysimeters and drain gauge to determine soil water NO3-N and Br- concentrations. Soil samples collected before and after the trial, and plant samples taken at tillering, flowering and physiological maturity were analysed for plant N% and stable 15N natural abundance using a mass spectrometer. Phenological and growth data from the lysimeter trial were used to calibrate Agricultural Production Systems sIMulator (APSIM) model for the first time (according to our knowledge) on wheat in South Africa. The model was then validated using data from the field trial. The drain gauge drained more frequently and in greater amounts than the weighing lysimeters, and NO3- was observed in drainage water from the drain gauge, but was undetectable from the weighing lysimeters drainage possibly because of saturated bottom layer that promoted denitrification. Based on stable 15N natural abundance, the FNUE was 68%, so the fertilised crop did not use 32% of the applied fertiliser. Good correlation was noted between the flag leaf and total plant N% at physiological maturity, indicating that flag leaf can be used to determine the FNUE without requiring whole plant analysis. The potential NO3-N leaching determined using a Br- conservative tracer was 51.5 kg ha-1 season-1. In fertilised plots, the calibrated model predicted NO3-N leaching of 22.7 kg ha-1 season-1, which was slightly lower than the drain gauge measured NO3-N leaching 24.9 kg ha-1 season. Therefore, the drain gauge shows excellent promise in quantifying N leaching but will require further testing under a range of cropping systems. Since the measured drain gauge and simulated NO3-N leaching agreed, and variables such as grain yield, total above dry matter, leaf area index and soil water content were reasonably simulated, the APSIM model can be applied to wheat cropping systems to improve N management decisions. The model confirmed that proper timing of N applications can reduce leaching losses, but further tests are required in several wheat growing agro-ecological zones to explore N management options that minimise N leaching losses. Even without measurements and/or modelling of N losses and crop uptake, results of this study for wheat indicate that the 15N stable isotopes can be used on its own to estimate FNUE, but more studies from different soil types and on wheat varieties are required to verify the trends observed in this study.